rpc/grpcclientidentity: remove hard-coded deadline in listener adatper causing crash

Verified once again that grpc.DialContext is indeed non-blocking. However, it checks in a defer stmt that the passed dial is not ctx.Done(). That is highly unusual if the dial is non-blocking. But it might still happen, maybe because of machine suspend during the function call and before the defer stmt is executed. panic: context deadline exceeded goroutine 49 [running]: github.com/zrepl/zrepl/rpc/grpcclientidentity/grpchelper.ClientConn(0x1906ea0, 0xc0003ea1e0, 0x1921620, 0xc0002da660, 0x0) /gopath/src/github.com/zrepl/zrepl/rpc/grpcclientidentity/grpchelper/authlistener_grpc_adaptor_wrapper.go:49 +0x38c github.com/zrepl/zrepl/rpc.NewClient(0x1906f00, 0xc0002d60f0, 0x1921620, 0xc0002da640, 0x1921620, 0xc0002da660, 0x1921620, 0xc0002da6a0, 0x1921620) /gopath/src/github.com/zrepl/zrepl/rpc/rpc_client.go:53 +0x199 github.com/zrepl/zrepl/daemon/job.(*modePush).ConnectEndpoints(0xc0000d1e90, 0x1921620, 0xc0002da640, 0x1921620, 0xc0002da660, 0x1921620, 0xc0002da6a0, 0x1906f00, 0xc0002d60f0) /gopath/src/github.com/zrepl/zrepl/daemon/job/active.go:105 +0x15d github.com/zrepl/zrepl/daemon/job.(*ActiveSide).do(0xc0000d6120, 0x1918720, 0xc00020f170) /gopath/src/github.com/zrepl/zrepl/daemon/job/active.go:356 +0x236 github.com/zrepl/zrepl/daemon/job.(*ActiveSide).Run(0xc0000d6120, 0x1918720, 0xc00009c660) /gopath/src/github.com/zrepl/zrepl/daemon/job/active.go:347 +0x289 github.com/zrepl/zrepl/daemon.(*jobs).start.func1(0xc0000fc880, 0x1921620, 0xc0002da120, 0x191a320, 0xc0000d6120, 0x1918720, 0xc0002d6a80)
docs: include Arch Linux installation
2019-10-10 14:02:12 +02:00 · 2019-10-06 20:38:00 +02:00 · 2019-10-06 16:23:20 +02:00 · 2019-10-05 14:59:51 +02:00 · 2019-10-05 14:58:49 +02:00 · 2019-10-03 11:57:19 +02:00
270 changed files with 18654 additions and 5479 deletions
@@ -0,0 +1,110 @@
+version: 2.0
+workflows:
+  version: 2
+  build:
+    jobs:
+      - build-1.11
+      - build-1.12
+      - build-1.13
+      - build-latest
+jobs:
+  # build-latest serves as the template
+  # we use YAML anchors & aliases to exchange the docker image (and hence Go version used for the build)
+  build-latest: &build-latest
+    description: Builds zrepl
+    parameters:
+      image:
+        description: "the docker image that the job should use"
+        type: string
+    docker:
+      - image: circleci/golang:latest
+    environment:
+      # required by lazy.sh
+      TERM: xterm
+    working_directory: /go/src/github.com/zrepl/zrepl
+    steps:
+      - run:
+          name: Setup environment variables
+          command: |
+            # used by pip (for docs)
+            echo 'export PATH="$HOME/.local/bin:$PATH"' >> $BASH_ENV
+            # we use modules
+            echo 'export GO111MODULE=on' >> $BASH_ENV
+
+      - restore_cache:
+          keys:
+            - source
+            - protobuf
+
+      - checkout
+
+      - save_cache:
+          key: source
+          paths:
+            - ".git"
+      
+      # install deps
+      - run: wget https://github.com/protocolbuffers/protobuf/releases/download/v3.6.1/protoc-3.6.1-linux-x86_64.zip
+      - run: echo "6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip" | sha256sum -c
+      - run: sudo unzip -d /usr protoc-3.6.1-linux-x86_64.zip
+      - save_cache:
+          key: protobuf
+          paths:
+            - "/usr/include/google/protobuf"
+            
+      - run: sudo apt update && sudo apt install python3 python3-pip libgirepository1.0-dev
+      - run: ./lazy.sh devsetup
+
+      - run: make
+      - run: make vet
+      - run: make test
+      - run: make lint
+      - run: make release
+
+
+      - store_artifacts:
+          path: ./artifacts/release
+          when: always
+
+      - run:
+          shell: /bin/bash -eo pipefail
+          when: always
+          command: |
+            if [ -n "$CIRCLE_PR_NUMBER" ]; then # CIRCLE_PR_NUMBER is guaranteed to be only present in forked PRs (external)
+              echo "Forked PR detected. Sry, can't trust you with credentials to external artifact store, use CircleCI's instead."
+              exit 0
+            fi
+            set -u # from now on
+            
+            # Download and install minio
+            curl -sSL https://dl.minio.io/client/mc/release/linux-amd64/mc -o ${GOPATH}/bin/mc
+            chmod +x ${GOPATH}/bin/mc
+            mc config host add --api s3v4 zrepl-minio https://minio.cschwarz.com ${MINIO_ACCESS_KEY} ${MINIO_SECRET_KEY}
+            
+            # Upload artifacts
+            echo "$CIRCLE_BUILD_URL" > ./artifacts/release/cirlceci_build_url
+            mc cp -r artifacts/release "zrepl-minio/zrepl-ci-artifacts/${CIRCLE_SHA1}/${CIRCLE_JOB}/"
+
+            # Push Artifact Link to GitHub
+            REPO="zrepl/zrepl"
+            COMMIT="${CIRCLE_SHA1}"
+            JOB_NAME="${CIRCLE_JOB}"
+            curl "https://api.github.com/repos/$REPO/statuses/$COMMIT?access_token=$GITHUB_COMMIT_STATUS_TOKEN" \
+              -H "Content-Type: application/json" \
+              -X POST \
+              -d '{"context":"zrepl/publish-ci-artifacts", "state": "success", "description":"CI Build Artifacts for '"$JOB_NAME"'", "target_url":"https://minio.cschwarz.com/minio/zrepl-ci-artifacts/'"$COMMIT"'/"}'
+
+  build-1.11:
+    <<: *build-latest
+    docker:
+    - image: circleci/golang:1.11
+
+  build-1.12:
+    <<: *build-latest
+    docker:
+    - image: circleci/golang:1.12
+
+  build-1.13:
+    <<: *build-latest
+    docker:
+    - image: circleci/golang:1.13
@@ -0,0 +1,4 @@
+patreon: zrepl
+liberapay: zrepl
+custom: https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96
+ko_fi: problame
@@ -0,0 +1,10 @@
+linters:
+  enable:
+  - goimports
+
+issues:
+  exclude-rules:
+    - path: _test\.go
+      linters:
+        - errcheck
+
@@ -0,0 +1,70 @@
+dist: xenial
+services:
+- docker
+
+env: # for allow_failures: https://docs.travis-ci.com/user/customizing-the-build/
+matrix:
+  include:
+
+  - language: go
+    name: "Build in Docker (docs/installation.rst)"
+    script:
+    - sudo docker build -t zrepl_build -f build.Dockerfile .
+    - |
+      sudo docker run -it --rm \
+        -v "${PWD}:/go/src/github.com/zrepl/zrepl" \
+        --user "$(id -u):$(id -g)" \
+        zrepl_build make vendordeps release
+
+  - &zrepl_build_template
+    language: go
+    go_import_path: github.com/zrepl/zrepl
+    before_install:
+        - wget https://github.com/protocolbuffers/protobuf/releases/download/v3.6.1/protoc-3.6.1-linux-x86_64.zip
+        - echo "6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip" | sha256sum -c
+        - sudo unzip -d /usr protoc-3.6.1-linux-x86_64.zip
+        - ./lazy.sh godep
+        - make vendordeps
+    script:
+      - make
+      - make vet
+      - make test
+      - make lint
+      - make artifacts/zrepl-freebsd-amd64
+      - make artifacts/zrepl-linux-amd64
+      - make artifacts/zrepl-darwin-amd64
+    go:
+    - "1.11"
+
+  - <<: *zrepl_build_template
+    go:
+    - "1.12"
+
+  - <<: *zrepl_build_template
+    go:
+    - "master"
+
+  - &zrepl_docs_template 
+    language: python
+    python:
+    - "3.4"
+    install:
+    - sudo apt-get install libgirepository1.0-dev
+    - pip install -r docs/requirements.txt
+    script:
+    - make docs
+  - <<: *zrepl_docs_template
+    python:
+    - "3.5"
+  - <<: *zrepl_docs_template
+    python:
+    - "3.6"
+  - <<: *zrepl_docs_template
+    python:
+    - "3.7"
+  
+
+  allow_failures:
+  - <<: *zrepl_build_template
+    go:
+    - "master"
@@ -1,313 +0,0 @@
-# This file is autogenerated, do not edit; changes may be undone by the next 'dep ensure'.
-
-
-[[projects]]
-  branch = "master"
-  digest = "1:c0bec5f9b98d0bc872ff5e834fac186b807b656683bd29cb82fb207a1513fabb"
-  name = "github.com/beorn7/perks"
-  packages = ["quantile"]
-  pruneopts = ""
-  revision = "3a771d992973f24aa725d07868b467d1ddfceafb"
-
-[[projects]]
-  digest = "1:56c130d885a4aacae1dd9c7b71cfe39912c7ebc1ff7d2b46083c8812996dc43b"
-  name = "github.com/davecgh/go-spew"
-  packages = ["spew"]
-  pruneopts = ""
-  revision = "346938d642f2ec3594ed81d874461961cd0faa76"
-  version = "v1.1.0"
-
-[[projects]]
-  digest = "1:e988ed0ca0d81f4d28772760c02ee95084961311291bdfefc1b04617c178b722"
-  name = "github.com/fatih/color"
-  packages = ["."]
-  pruneopts = ""
-  revision = "5b77d2a35fb0ede96d138fc9a99f5c9b6aef11b4"
-  version = "v1.7.0"
-
-[[projects]]
-  branch = "master"
-  digest = "1:5d0a2385edf4ba44f3b7b76bc0436ceb8f62bf55aa5d540a9eb9ec6c58d86809"
-  name = "github.com/ftrvxmtrx/fd"
-  packages = ["."]
-  pruneopts = ""
-  revision = "c6d800382fff6dc1412f34269f71b7f83bd059ad"
-
-[[projects]]
-  branch = "master"
-  digest = "1:4416ad04caf950d004921268f777e418f8c328080ccf6a91e799f977dc128808"
-  name = "github.com/gdamore/encoding"
-  packages = ["."]
-  pruneopts = ""
-  revision = "b23993cbb6353f0e6aa98d0ee318a34728f628b9"
-
-[[projects]]
-  digest = "1:d8088c614a5bffaf38a40c116226e7d230750bafb77942dca25eab779ceeaf1f"
-  name = "github.com/gdamore/tcell"
-  packages = [
-    ".",
-    "termbox",
-    "terminfo",
-  ]
-  pruneopts = ""
-  revision = "de7e78efa4a71b3f36c7154989c529dbdf9ae623"
-  version = "v1.1.0"
-
-[[projects]]
-  digest = "1:6a4a01d58b227c4b6b11111b9f172ec5c17682b82724e58e6daf3f19f4faccd8"
-  name = "github.com/go-logfmt/logfmt"
-  packages = ["."]
-  pruneopts = ""
-  revision = "390ab7935ee28ec6b286364bba9b4dd6410cb3d5"
-  version = "v0.3.0"
-
-[[projects]]
-  digest = "1:3dd078fda7500c341bc26cfbc6c6a34614f295a2457149fc1045cab767cbcf18"
-  name = "github.com/golang/protobuf"
-  packages = ["proto"]
-  pruneopts = ""
-  revision = "aa810b61a9c79d51363740d207bb46cf8e620ed5"
-  version = "v1.2.0"
-
-[[projects]]
-  branch = "master"
-  digest = "1:cb09475f771b9167fb9333629f5d6a7161572602ea040f1094602b0dc8709878"
-  name = "github.com/jinzhu/copier"
-  packages = ["."]
-  pruneopts = ""
-  revision = "db4671f3a9b8df855e993f7c94ec5ef1ffb0a23b"
-
-[[projects]]
-  branch = "master"
-  digest = "1:1ed9eeebdf24aadfbca57eb50e6455bd1d2474525e0f0d4454de8c8e9bc7ee9a"
-  name = "github.com/kr/logfmt"
-  packages = ["."]
-  pruneopts = ""
-  revision = "b84e30acd515aadc4b783ad4ff83aff3299bdfe0"
-
-[[projects]]
-  branch = "master"
-  digest = "1:591a2778aa6e896980757ea87e659b3aa13d8c0e790310614028463a31c0998b"
-  name = "github.com/kr/pretty"
-  packages = ["."]
-  pruneopts = ""
-  revision = "cfb55aafdaf3ec08f0db22699ab822c50091b1c4"
-
-[[projects]]
-  branch = "master"
-  digest = "1:109e0a9b9e74c5c8adf94a2dc4dd4a9ca2a183d4d87ffecd4e62db69a5ede55a"
-  name = "github.com/kr/text"
-  packages = ["."]
-  pruneopts = ""
-  revision = "7cafcd837844e784b526369c9bce262804aebc60"
-
-[[projects]]
-  digest = "1:863ba095ae0954100963826333bd1503c169424bb798d9ce800d9f6651f4c234"
-  name = "github.com/lucasb-eyer/go-colorful"
-  packages = ["."]
-  pruneopts = ""
-  revision = "345fbb3dbcdb252d9985ee899a84963c0fa24c82"
-  version = "v1.0"
-
-[[projects]]
-  digest = "1:9ea83adf8e96d6304f394d40436f2eb44c1dc3250d223b74088cc253a6cd0a1c"
-  name = "github.com/mattn/go-colorable"
-  packages = ["."]
-  pruneopts = ""
-  revision = "167de6bfdfba052fa6b2d3664c8f5272e23c9072"
-  version = "v0.0.9"
-
-[[projects]]
-  digest = "1:78229b46ddb7434f881390029bd1af7661294af31f6802e0e1bedaad4ab0af3c"
-  name = "github.com/mattn/go-isatty"
-  packages = ["."]
-  pruneopts = ""
-  revision = "0360b2af4f38e8d38c7fce2a9f4e702702d73a39"
-  version = "v0.0.3"
-
-[[projects]]
-  digest = "1:82b912465c1da0668582a7d1117339c278e786c2536b3c3623029a0c7141c2d0"
-  name = "github.com/mattn/go-runewidth"
-  packages = ["."]
-  pruneopts = ""
-  revision = "ce7b0b5c7b45a81508558cd1dba6bb1e4ddb51bb"
-  version = "v0.0.3"
-
-[[projects]]
-  digest = "1:4c23ced97a470b17d9ffd788310502a077b9c1f60221a85563e49696276b4147"
-  name = "github.com/matttproud/golang_protobuf_extensions"
-  packages = ["pbutil"]
-  pruneopts = ""
-  revision = "3247c84500bff8d9fb6d579d800f20b3e091582c"
-  version = "v1.0.0"
-
-[[projects]]
-  digest = "1:7365acd48986e205ccb8652cc746f09c8b7876030d53710ea6ef7d0bd0dcd7ca"
-  name = "github.com/pkg/errors"
-  packages = ["."]
-  pruneopts = ""
-  revision = "645ef00459ed84a119197bfb8d8205042c6df63d"
-  version = "v0.8.0"
-
-[[projects]]
-  digest = "1:256484dbbcd271f9ecebc6795b2df8cad4c458dd0f5fd82a8c2fa0c29f233411"
-  name = "github.com/pmezard/go-difflib"
-  packages = ["difflib"]
-  pruneopts = ""
-  revision = "792786c7400a136282c1664665ae0a8db921c6c2"
-  version = "v1.0.0"
-
-[[projects]]
-  branch = "master"
-  digest = "1:1392748e290ca66ac8447ef24961f8ae9e1d846a53af0f58a5a0256982ce0577"
-  name = "github.com/problame/go-netssh"
-  packages = ["."]
-  pruneopts = ""
-  revision = "c56ad38d2c91397ad3c8dd9443d7448e328a9e9e"
-
-[[projects]]
-  branch = "master"
-  digest = "1:8c63c44f018bd52b03ebad65c9df26aabbc6793138e421df1c8c84c285a45bc6"
-  name = "github.com/problame/go-rwccmd"
-  packages = ["."]
-  pruneopts = ""
-  revision = "391d2c78c8404a9683d79f75dd24ab53040f89f7"
-
-[[projects]]
-  digest = "1:c2ba1c9dc003c15856e4529dac028cacba08ee8924300f058b3467cde9acf7a9"
-  name = "github.com/problame/go-streamrpc"
-  packages = [
-    ".",
-    "internal/pdu",
-  ]
-  pruneopts = ""
-  revision = "de6f6a4041c77f700f02d8fe749e54efa50811f7"
-  version = "v0.4"
-
-[[projects]]
-  branch = "master"
-  digest = "1:ebf8ffdde9bdbf9c83e22121875c68c01d821776523546554b2a3ff6f72773ab"
-  name = "github.com/prometheus/client_golang"
-  packages = [
-    "prometheus",
-    "prometheus/promhttp",
-  ]
-  pruneopts = ""
-  revision = "e11c6ff8170beca9d5fd8b938e71165eeec53ac6"
-
-[[projects]]
-  branch = "master"
-  digest = "1:60aca47f4eeeb972f1b9da7e7db51dee15ff6c59f7b401c1588b8e6771ba15ef"
-  name = "github.com/prometheus/client_model"
-  packages = ["go"]
-  pruneopts = ""
-  revision = "99fa1f4be8e564e8a6b613da7fa6f46c9edafc6c"
-
-[[projects]]
-  branch = "master"
-  digest = "1:af21ee3e0a8212f17bb317cd7237f9920bcb2641a291ac111f30f63b3cab817f"
-  name = "github.com/prometheus/common"
-  packages = [
-    "expfmt",
-    "internal/bitbucket.org/ww/goautoneg",
-    "model",
-  ]
-  pruneopts = ""
-  revision = "d0f7cd64bda49e08b22ae8a730aa57aa0db125d6"
-
-[[projects]]
-  branch = "master"
-  digest = "1:61df0898746840afc7be5dc2c3eeec83022fab70df11ecee5b16c85e912cf5ed"
-  name = "github.com/prometheus/procfs"
-  packages = [
-    ".",
-    "internal/util",
-    "nfs",
-    "xfs",
-  ]
-  pruneopts = ""
-  revision = "8b1c2da0d56deffdbb9e48d4414b4e674bd8083e"
-
-[[projects]]
-  branch = "master"
-  digest = "1:146327ce93be37e68bd3ff8541090d96da8cb3adc9e35d57570e9170a29f6bf6"
-  name = "github.com/spf13/cobra"
-  packages = ["."]
-  pruneopts = ""
-  revision = "b78744579491c1ceeaaa3b40205e56b0591b93a3"
-
-[[projects]]
-  digest = "1:261bc565833ef4f02121450d74eb88d5ae4bd74bfe5d0e862cddb8550ec35000"
-  name = "github.com/spf13/pflag"
-  packages = ["."]
-  pruneopts = ""
-  revision = "e57e3eeb33f795204c1ca35f56c44f83227c6e66"
-  version = "v1.0.0"
-
-[[projects]]
-  digest = "1:3926a4ec9a4ff1a072458451aa2d9b98acd059a45b38f7335d31e06c3d6a0159"
-  name = "github.com/stretchr/testify"
-  packages = [
-    "assert",
-    "require",
-  ]
-  pruneopts = ""
-  revision = "69483b4bd14f5845b5a1e55bca19e954e827f1d0"
-  version = "v1.1.4"
-
-[[projects]]
-  branch = "v2"
-  digest = "1:9d92186f609a73744232323416ddafd56fae67cb552162cc190ab903e36900dd"
-  name = "github.com/zrepl/yaml-config"
-  packages = ["."]
-  pruneopts = ""
-  revision = "af27d27978ad95808723a62d87557d63c3ff0605"
-
-[[projects]]
-  branch = "master"
-  digest = "1:9c286cf11d0ca56368185bada5dd6d97b6be4648fc26c354fcba8df7293718f7"
-  name = "golang.org/x/sys"
-  packages = ["unix"]
-  pruneopts = ""
-  revision = "bf42f188b9bc6f2cf5b8ee5a912ef1aedd0eba4c"
-
-[[projects]]
-  digest = "1:5acd3512b047305d49e8763eef7ba423901e85d5dd2fd1e71778a0ea8de10bd4"
-  name = "golang.org/x/text"
-  packages = [
-    "encoding",
-    "encoding/internal/identifier",
-    "internal/gen",
-    "transform",
-    "unicode/cldr",
-  ]
-  pruneopts = ""
-  revision = "f21a4dfb5e38f5895301dc265a8def02365cc3d0"
-  version = "v0.3.0"
-
-[solve-meta]
-  analyzer-name = "dep"
-  analyzer-version = 1
-  input-imports = [
-    "github.com/fatih/color",
-    "github.com/gdamore/tcell/termbox",
-    "github.com/go-logfmt/logfmt",
-    "github.com/golang/protobuf/proto",
-    "github.com/jinzhu/copier",
-    "github.com/kr/pretty",
-    "github.com/mattn/go-isatty",
-    "github.com/pkg/errors",
-    "github.com/problame/go-netssh",
-    "github.com/problame/go-rwccmd",
-    "github.com/problame/go-streamrpc",
-    "github.com/prometheus/client_golang/prometheus",
-    "github.com/prometheus/client_golang/prometheus/promhttp",
-    "github.com/spf13/cobra",
-    "github.com/spf13/pflag",
-    "github.com/stretchr/testify/assert",
-    "github.com/stretchr/testify/require",
-    "github.com/zrepl/yaml-config",
-    "golang.org/x/sys/unix",
-  ]
-  solver-name = "gps-cdcl"
-  solver-version = 1
@@ -1,75 +0,0 @@
-ignored = [ "github.com/inconshreveable/mousetrap" ]
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/ftrvxmtrx/fd"
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/jinzhu/copier"
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/kr/pretty"
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/mitchellh/go-homedir"
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/mitchellh/mapstructure"
-
-[[constraint]]
-  name = "github.com/pkg/errors"
-  version = "0.8.0"
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/spf13/cobra"
-
-[[constraint]]
-  name = "github.com/spf13/viper"
-  version = "1.0.0"
-
-[[constraint]]
-  name = "github.com/stretchr/testify"
-  version = "1.1.4"
-
-[[constraint]]
-  branch = "v2"
-  name = "github.com/zrepl/yaml-config"
-
-[[constraint]]
-  name = "github.com/go-logfmt/logfmt"
-  version = "*"
-
-[[constraint]]
-  name = "github.com/problame/go-rwccmd"
-  branch = "master"
-
-[[constraint]]
-  name = "github.com/problame/go-netssh"
-  branch = "master"
-
-[[constraint]]
-  name = "github.com/prometheus/client_golang"
-  branch = "master"
-
-[[constraint]]
-  name = "github.com/golang/protobuf"
-  version = "1.2.0"
-
-[[constraint]]
-  name = "github.com/fatih/color"
-  version = "1.7.0"
-
-[[constraint]]
-  name = "github.com/problame/go-streamrpc"
-  version = "0.4.0"
-
-
-
-[[constraint]]
-  name = "github.com/gdamore/tcell"
-  version = "1.0.0"
@@ -1,94 +1,65 @@
-.PHONY: generate build test vet cover release docs docs-clean clean vendordeps
+.PHONY: generate build test vet cover release docs docs-clean clean format lint platformtest
 .DEFAULT_GOAL := build

-ROOT := github.com/zrepl/zrepl
-SUBPKGS += client
-SUBPKGS += config
-SUBPKGS += daemon
-SUBPKGS += daemon/filters
-SUBPKGS += daemon/job
-SUBPKGS += daemon/logging
-SUBPKGS += daemon/nethelpers
-SUBPKGS += daemon/pruner
-SUBPKGS += daemon/snapper
-SUBPKGS += daemon/streamrpcconfig
-SUBPKGS += daemon/transport
-SUBPKGS += daemon/transport/connecter
-SUBPKGS += daemon/transport/serve
-SUBPKGS += endpoint
-SUBPKGS += logger
-SUBPKGS += pruning
-SUBPKGS += pruning/retentiongrid
-SUBPKGS += replication
-SUBPKGS += replication/fsrep
-SUBPKGS += replication/pdu
-SUBPKGS += replication/internal/queue
-SUBPKGS += replication/internal/diff
-SUBPKGS += tlsconf
-SUBPKGS += util
-SUBPKGS += util/socketpair
-SUBPKGS += util/watchdog
-SUBPKGS += util/envconst
-SUBPKGS += version
-SUBPKGS += zfs
-
-_TESTPKGS := $(ROOT) $(foreach p,$(SUBPKGS),$(ROOT)/$(p))
-
 ARTIFACTDIR := artifacts

 ifdef ZREPL_VERSION
    _ZREPL_VERSION := $(ZREPL_VERSION)
 endif
 ifndef _ZREPL_VERSION
-    _ZREPL_VERSION := $(shell git describe --dirty 2>/dev/null || echo "ZREPL_BUILD_INVALID_VERSION" )
+    _ZREPL_VERSION := $(shell git describe --always --dirty 2>/dev/null || echo "ZREPL_BUILD_INVALID_VERSION" )
    ifeq ($(_ZREPL_VERSION),ZREPL_BUILD_INVALID_VERSION) # can't use .SHELLSTATUS because Debian Stretch is still on gmake 4.1
        $(error cannot infer variable ZREPL_VERSION using git and variable is not overriden by make invocation)
    endif
 endif
+GO := go
 GO_LDFLAGS := "-X github.com/zrepl/zrepl/version.zreplVersion=$(_ZREPL_VERSION)"
+GO_MOD_READONLY := -mod=readonly
+GO_BUILDFLAGS := $(GO_MOD_READONLY)
+GO_BUILD := GO111MODULE=on $(GO) build $(GO_BUILDFLAGS) -v -ldflags $(GO_LDFLAGS)

-GO_BUILD := go build -ldflags $(GO_LDFLAGS)
-
+# keep in sync with vet target
 RELEASE_BINS := $(ARTIFACTDIR)/zrepl-freebsd-amd64
 RELEASE_BINS += $(ARTIFACTDIR)/zrepl-linux-amd64
+RELEASE_BINS += $(ARTIFACTDIR)/zrepl-linux-arm64
 RELEASE_BINS += $(ARTIFACTDIR)/zrepl-darwin-amd64

 RELEASE_NOARCH := $(ARTIFACTDIR)/zrepl-noarch.tar
-THIS_PLATFORM_RELEASE_BIN := $(shell bash -c 'source <(go env) && echo "zrepl-$${GOOS}-$${GOARCH}"' )
-
-vendordeps:
-	dep ensure -v -vendor-only
+THIS_PLATFORM_RELEASE_BIN := $(shell bash -c 'source <($(GO) env) && echo "zrepl-$${GOOS}-$${GOARCH}"' )

 generate: #not part of the build, must do that manually
-	protoc -I=replication/pdu --go_out=replication/pdu replication/pdu/pdu.proto
-	@for pkg in $(_TESTPKGS); do\
-		go generate "$$pkg" || exit 1; \
-	done;
+	protoc -I=replication/logic/pdu --go_out=plugins=grpc:replication/logic/pdu replication/logic/pdu/pdu.proto
+	$(GO) generate $(GO_BUILDFLAGS) -x ./...
+
+format:
+	goimports -srcdir . -local 'github.com/zrepl/zrepl' -w $(shell find . -type f -name '*.go' -not -path "./vendor/*" -not -name '*.pb.go' -not -name '*_enumer.go')
+
+lint:
+	golangci-lint run ./...

 build:
-	@echo "INFO: In case of missing dependencies, run 'make vendordeps'"
 	$(GO_BUILD) -o "$(ARTIFACTDIR)/zrepl"

 test:
-	@for pkg in $(_TESTPKGS); do \
-		echo "Testing $$pkg"; \
-		go test "$$pkg" || exit 1; \
-	done;
+	$(GO) test $(GO_BUILDFLAGS) ./...
+	# TODO compile the tests for each supported platform
+	# but `go test -c ./...` is not supported

 vet:
-	@for pkg in $(_TESTPKGS); do \
-		echo "Vetting $$pkg"; \
-		go vet "$$pkg" || exit 1; \
-	done;
+	$(GO) vet $(GO_BUILDFLAGS) ./...
+	# for each supported platform to cover conditional compilation
+	# (keep in sync with RELEASE_BINS)
+	GOOS=freebsd	GOARCH=amd64 	$(GO) vet $(GO_BUILDFLAGS) ./...
+	GOOS=linux		GOARCH=amd64 	$(GO) vet $(GO_BUILDFLAGS) ./...
+	GOOS=linux		GOARCH=arm64 	$(GO) vet $(GO_BUILDFLAGS) ./...
+	GOOS=darwin		GOARCH=amd64 	$(GO) vet $(GO_BUILDFLAGS) ./...

-cover: artifacts
-	@for pkg in $(_TESTPKGS); do \
-		profile="$(ARTIFACTDIR)/cover-$$(basename $$pkg).out"; \
-		go test -coverprofile "$$profile" $$pkg || exit 1; \
-		if [ -f "$$profile" ]; then \
-   			go tool cover -html="$$profile" -o "$${profile}.html" || exit 2; \
-		fi; \
-	done;
+ZREPL_PLATFORMTEST_POOLNAME := zreplplatformtest
+ZREPL_PLATFORMTEST_IMAGEPATH := /tmp/zreplplatformtest.pool.img
+$(ARTIFACTDIR)/zrepl_platformtest:
+	$(GO_BUILD) -o "$(ARTIFACTDIR)/zrepl_platformtest" ./platformtest/harness
+platformtest: $(ARTIFACTDIR)/zrepl_platformtest
+	"$(ARTIFACTDIR)/zrepl_platformtest" -poolname "$(ZREPL_PLATFORMTEST_POOLNAME)" -imagepath "$(ZREPL_PLATFORMTEST_IMAGEPATH)"

 $(ARTIFACTDIR):
 	mkdir -p "$@"
@@ -99,8 +70,9 @@ $(ARTIFACTDIR)/docs: $(ARTIFACTDIR)
 $(ARTIFACTDIR)/bash_completion: $(RELEASE_BINS)
 	artifacts/$(THIS_PLATFORM_RELEASE_BIN) bashcomp "$@"

+.PHONY: $(ARTIFACTDIR)/go_version.txt
 $(ARTIFACTDIR)/go_version.txt:
-	go version > $@
+	$(GO) version > $@

 docs: $(ARTIFACTDIR)/docs
 	make -C docs \
@@ -112,12 +84,11 @@ docs-clean:
 		clean \
 		BUILDDIR=../artifacts/docs

-
 .PHONY: $(RELEASE_BINS)
 # TODO: two wildcards possible
-$(RELEASE_BINS): $(ARTIFACTDIR)/zrepl-%-amd64: generate $(ARTIFACTDIR) vet test
-	@echo "INFO: In case of missing dependencies, run 'make vendordeps'"
-	GOOS=$* GOARCH=amd64   $(GO_BUILD) -o "$(ARTIFACTDIR)/zrepl-$*-amd64"
+$(RELEASE_BINS): $(ARTIFACTDIR)/zrepl-%: generate $(ARTIFACTDIR) vet test lint
+	STEM=$*; GOOS="$${STEM%%-*}"; GOARCH="$${STEM##*-}"; export GOOS GOARCH; \
+		$(GO_BUILD) -o "$(ARTIFACTDIR)/zrepl-$$GOOS-$$GOARCH"

 $(RELEASE_NOARCH): docs $(ARTIFACTDIR)/bash_completion $(ARTIFACTDIR)/go_version.txt
 	tar --mtime='1970-01-01' --sort=name \
@@ -133,9 +104,13 @@ release: $(RELEASE_BINS) $(RELEASE_NOARCH)
 	cp $^ "$(ARTIFACTDIR)/release"
 	cd "$(ARTIFACTDIR)/release" && sha512sum $$(ls | sort) > sha512sum.txt
 	@# note that we use ZREPL_VERSION and not _ZREPL_VERSION because we want to detect the override
-	@if git describe --dirty 2>/dev/null | grep dirty >/dev/null; then \
-		if [ "$(ZREPL_VERSION)" == "" ]; then \
-			echo "[WARN] git checkout is dirty and make variable ZREPL_VERSION was not used to override"; \
+	@if git describe --always --dirty 2>/dev/null | grep dirty >/dev/null; then \
+        echo '[INFO] either git reports checkout is dirty or git is not installed or this is not a git checkout'; \
+		if [ "$(ZREPL_VERSION)" = "" ]; then \
+			echo '[WARN] git checkout is dirty and make variable ZREPL_VERSION was not used to override'; \
+			git status; \
+			echo "git diff:";  \
+			git diff | cat; \
 			exit 1; \
 		fi; \
 	fi;
@@ -1,3 +1,10 @@
+[![GitHub license](https://img.shields.io/github/license/zrepl/zrepl.svg)](https://github.com/zrepl/zrepl/blob/master/LICENSE)
+[![Language: Go](https://img.shields.io/badge/language-Go-6ad7e5.svg)](https://golang.org/)
+[![User Docs](https://img.shields.io/badge/docs-web-blue.svg)](https://zrepl.github.io)
+[![Donate via PayPal](https://img.shields.io/badge/donate-paypal-yellow.svg)](https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96)
+[![Donate via Liberapay](https://img.shields.io/liberapay/receives/zrepl.svg?logo=liberapay)](https://liberapay.com/zrepl/donate)
+[![Twitter](https://img.shields.io/twitter/url/https/github.com/zrepl/zrepl.svg?style=social)](https://twitter.com/intent/tweet?text=Wow:&url=https%3A%2F%2Fgithub.com%2Fzrepl%2Fzrepl)
+
 # zrepl
 zrepl is a one-stop ZFS backup & replication solution.

@@ -26,13 +33,16 @@ Check out the *Coding Workflow* section below for details.
 * Make sure your distro is compatible with the paths in `docs/installation.rst`.
 * Ship a default config that adheres to your distro's `hier` and logging system.
 * Ship a service manager file and _please_ try to upstream it to this repository.
+  * `dist/systemd` contains a Systemd unit template.
+* Ship other material provided in `./dist`, e.g. in `/usr/share/zrepl/`.
 * Use `make release ZREPL_VERSION='mydistro-1.2.3_1'`
    * Your distro's name and any versioning supplemental to zrepl's (e.g. package revision) should be in this string
+* Use `make platformtest` **on a test system** to validate that zrepl's abstractions on top of ZFS work with the system ZFS.
 * Make sure you are informed about new zrepl versions, e.g. by subscribing to GitHub's release RSS feed.

 ## Developer Documentation

-zrepl is written in [Go](https://golang.org) and uses [`dep`](https://github.com/golang/dep) to manage dependencies.
+zrepl is written in [Go](https://golang.org) and uses [Go modules](https://github.com/golang/go/wiki/Modules) to manage dependencies.
 The documentation is written in [ReStructured Text](http://docutils.sourceforge.net/rst.html) using the [Sphinx](https://www.sphinx-doc.org) framework.

 To get started, run `./lazy.sh devsetup` to easily install build dependencies and read `docs/installation.rst -> Compiling from Source`.
@@ -45,21 +55,53 @@ Make sure to develop an understanding how zrepl is typically used by studying th
 ### Project Structure

 ```
-├── cmd
-│   ├── endpoint            # implementations of endpoints for package replication
-│   ├── sampleconf          # example configuration
+├── artifacts               # build artifcats generate by make
+├── cli                     # wrapper around CLI package cobra
+├── client                  # all subcommands that are not `daemon`
+├── config                  # config data types (=> package yaml-config)
+│   └── samples
+├── daemon                  # the implementation of `zrepl daemon` subcommand
+│   ├── filters
+│   ├── job                 # job implementations
+│   ├── logging             # logging outlets + formatters
+│   ├── nethelpers
+│   ├── prometheus
+│   ├── pruner              # pruner implementation
+│   ├── snapper             # snapshotter implementation
 ├── docs                    # sphinx-based documentation
+├── dist                    # supplemental material for users & package maintainers
 │   ├── **/*.rst            # documentation in reStructuredText
 │   ├── sphinxconf
 │   │   └── conf.py         # sphinx config (see commit 445a280 why its not in docs/)
 │   ├── requirements.txt    # pip3 requirements to build documentation
 │   ├── publish.sh          # shell script for automated rendering & deploy to zrepl.github.io repo
-│   ├── public_git          # checkout of zrepl.github.io managed by above shell script
-├── logger                  # logger package used by zrepl
-├── replication             # replication functionality
-├── rpc                     # rpc protocol implementation
+│   └── public_git          # checkout of zrepl.github.io managed by above shell script
+├── endpoint                # implementation of replication endpoints (=> package replication)
+├── logger                  # our own logger package
+├── pruning                 # pruning rules (the logic, not the actual execution)
+│   └── retentiongrid
+├── replication
+│   ├── driver              # the driver of the replication logic (status reporting, error handling)
+│   ├── logic               # planning & executing replication steps via rpc
+|   |   └── pdu             # the generated gRPC & protobuf code used in replication (and endpoints)
+│   └── report              # the JSON-serializable report datastructures exposed to the client
+├── rpc                     # the hybrid gRPC + ./dataconn RPC client: connects to a remote replication.Endpoint
+│   ├── dataconn            # Bulk data-transfer RPC protocol
+│   ├── grpcclientidentity  # adaptor to inject package transport's 'client identity' concept into gRPC contexts
+│   ├── netadaptor          # adaptor to convert a package transport's Connecter and Listener into net.* primitives
+│   ├── transportmux        # TCP connecter and listener used to split control & data traffic
+│   └── versionhandshake    # replication protocol version handshake perfomed on newly established connections
+├── tlsconf                 # abstraction for Go TLS server + client config
+├── transport               # transports implementation
+│   ├── fromconfig
+│   ├── local
+│   ├── ssh
+│   ├── tcp
+│   └── tls
 ├── util
-└── zfs                     # ZFS wrappers, filesystemm diffing
+├── vendor                  # managed by dep
+├── version                 # abstraction for versions (filled during build by Makefile)
+└── zfs                     # zfs(8) wrappers
 ```

 ### Coding Workflow
@@ -4,19 +4,26 @@ RUN apt-get update && apt-get install -y \
    python3-pip \
    unzip

-RUN wget https://github.com/protocolbuffers/protobuf/releases/download/v3.6.1/protoc-3.6.1-linux-x86_64.zip
-RUN echo "6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip" | sha256sum -c
-RUN unzip -d /usr protoc-3.6.1-linux-x86_64.zip
+ADD build.installprotoc.bash ./
+RUN bash build.installprotoc.bash

 ADD lazy.sh /tmp/lazy.sh
 ADD docs/requirements.txt /tmp/requirements.txt
 ENV ZREPL_LAZY_DOCS_REQPATH=/tmp/requirements.txt
-RUN /tmp/lazy.sh devsetup
+RUN /tmp/lazy.sh docdep

 # prepare volume mount of git checkout to /zrepl
-RUN mkdir -p /go/src/github.com/zrepl/zrepl
-RUN chmod -R 0777 /go
+RUN mkdir -p /src/github.com/zrepl/zrepl
 RUN mkdir -p /.cache && chmod -R 0777 /.cache
-WORKDIR /go/src/github.com/zrepl/zrepl

+# $GOPATH is /go
+# Go 1.12 doesn't use modules within GOPATH, but 1.13 and later do
+# => store source outside of GOPATH
+WORKDIR /src
+
+# Install build tools (e.g. protobuf generator, stringer) into $GOPATH/bin
+ADD go.mod go.sum ./
+RUN /tmp/lazy.sh godep
+
+RUN chmod -R 0777 /go

@@ -0,0 +1,25 @@
+#!/usr/bin/env bash
+set -euo pipefail
+set -x
+
+MACH=$(uname -m)
+MACH="${MACH/aarch64/aarch_64}"
+
+VERSION=3.6.1
+FILENAME=protoc-"$VERSION"-linux-"$MACH".zip
+
+if [ -e "$FILENAME" ]; then
+    echo "$FILENAME" already exists 1>&2
+    exit 1
+fi
+
+wget https://github.com/protocolbuffers/protobuf/releases/download/v"$VERSION"/"$FILENAME"
+
+stat "$FILENAME"
+
+sha256sum -c --ignore-missing <<EOF
+6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip
+af8e5aaaf39ddec62ec8dd2be1b8d9602c6da66564883a16393ade5f71170922  protoc-3.6.1-linux-aarch_64.zip
+EOF
+
+unzip -d /usr "$FILENAME"
@@ -0,0 +1,23 @@
+//+build windows
+
+// This package cannot actually be built, and since the Travis
+// tests run go test ./..., we have to avoid a build attempt.
+// Windows is not supported atm, so this works ¯\_(ツ)_/¯
+
+// This package is a pseudo-user of build-time dependencies for zrepl,
+// mostly for various code-generation tools.
+//
+// The imports are necessary to satisfy go dep
+package main
+
+import (
+	"fmt"
+
+	_ "github.com/alvaroloes/enumer"
+	_ "github.com/golang/protobuf/protoc-gen-go"
+	_ "golang.org/x/tools/cmd/stringer"
+)
+
+func main() {
+	fmt.Println("just a placeholder")
+}
@@ -2,10 +2,12 @@ package cli

 import (
 	"fmt"
+	"os"
+
 	"github.com/spf13/cobra"
 	"github.com/spf13/pflag"
+
 	"github.com/zrepl/zrepl/config"
-	"os"
 )

 var rootArgs struct {
@@ -23,7 +25,10 @@ var bashcompCmd = &cobra.Command{
 	Run: func(cmd *cobra.Command, args []string) {
 		if len(args) != 1 {
 			fmt.Fprintf(os.Stderr, "specify exactly one positional agument\n")
-			cmd.Usage()
+			err := cmd.Usage()
+			if err != nil {
+				panic(err)
+			}
 			os.Exit(1)
 		}
 		if err := rootCmd.GenBashCompletionFile(args[0]); err != nil {
@@ -42,6 +47,7 @@ func init() {
 type Subcommand struct {
 	Use              string
 	Short            string
+	Example          string
 	NoRequireConfig  bool
 	Run              func(subcommand *Subcommand, args []string) error
 	SetupFlags       func(f *pflag.FlagSet)
@@ -94,6 +100,7 @@ func addSubcommandToCobraCmd(c *cobra.Command, s *Subcommand) {
 	cmd := cobra.Command{
 		Use:     s.Use,
 		Short:   s.Short,
+		Example: s.Example,
 	}
 	if s.SetupSubcommands == nil {
 		cmd.Run = s.run
@@ -108,7 +115,6 @@ func addSubcommandToCobraCmd(c *cobra.Command, s *Subcommand) {
 	c.AddCommand(&cmd)
 }

-
 func Run() {
 	if err := rootCmd.Execute(); err != nil {
 		os.Exit(1)
@@ -3,16 +3,18 @@ package client
 import (
 	"encoding/json"
 	"fmt"
+	"os"
+
 	"github.com/kr/pretty"
 	"github.com/pkg/errors"
 	"github.com/spf13/pflag"
 	"github.com/zrepl/yaml-config"
+
 	"github.com/zrepl/zrepl/cli"
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/job"
 	"github.com/zrepl/zrepl/daemon/logging"
 	"github.com/zrepl/zrepl/logger"
-	"os"
 )

 var configcheckArgs struct {
@@ -28,14 +30,22 @@ var ConfigcheckCmd = &cli.Subcommand{
 		f.StringVar(&configcheckArgs.what, "what", "all", "what to print [all|config|jobs|logging]")
 	},
 	Run: func(subcommand *cli.Subcommand, args []string) error {
-		formatMap := map[string]func(interface{}) {
+		formatMap := map[string]func(interface{}){
 			"": func(i interface{}) {},
-			"pretty": func(i interface{}) { pretty.Println(i) },
+			"pretty": func(i interface{}) {
+				if _, err := pretty.Println(i); err != nil {
+					panic(err)
+				}
+			},
 			"json": func(i interface{}) {
-				json.NewEncoder(os.Stdout).Encode(subcommand.Config())
+				if err := json.NewEncoder(os.Stdout).Encode(subcommand.Config()); err != nil {
+					panic(err)
+				}
 			},
 			"yaml": func(i interface{}) {
-				yaml.NewEncoder(os.Stdout).Encode(subcommand.Config())
+				if err := yaml.NewEncoder(os.Stdout).Encode(subcommand.Config()); err != nil {
+					panic(err)
+				}
 			},
 		}

@@ -71,8 +81,7 @@ var ConfigcheckCmd = &cli.Subcommand{
 			}
 		}

-
-		whatMap := map[string]func() {
+		whatMap := map[string]func(){
 			"all": func() {
 				o := struct {
 					config  *config.Config
@@ -109,4 +118,3 @@ var ConfigcheckCmd = &cli.Subcommand{
 		}
 	},
 }
-
@@ -4,10 +4,11 @@ import (
 	"bytes"
 	"context"
 	"encoding/json"
-	"github.com/pkg/errors"
 	"io"
 	"net"
 	"net/http"
+
+	"github.com/pkg/errors"
 )

 func controlHttpClient(sockpath string) (client http.Client, err error) {
@@ -35,7 +36,7 @@ func jsonRequestResponse(c http.Client, endpoint string, req interface{}, res in

 	if resp.StatusCode != http.StatusOK {
 		var msg bytes.Buffer
-		io.CopyN(&msg, resp.Body, 4096)
+		_, _ = io.CopyN(&msg, resp.Body, 4096) // ignore error, just display what we got
 		return errors.Errorf("%s", msg.String())
 	}

@@ -0,0 +1,104 @@
+package client
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/pkg/errors"
+	"github.com/spf13/pflag"
+
+	"github.com/zrepl/zrepl/zfs"
+
+	"github.com/zrepl/zrepl/cli"
+	"github.com/zrepl/zrepl/config"
+)
+
+var (
+	MigrateCmd = &cli.Subcommand{
+		Use:   "migrate",
+		Short: "perform migration of the on-disk / zfs properties",
+		SetupSubcommands: func() []*cli.Subcommand {
+			return migrations
+		},
+	}
+)
+
+var migrations = []*cli.Subcommand{
+	&cli.Subcommand{
+		Use: "0.0.X:0.1:placeholder",
+		Run: doMigratePlaceholder0_1,
+		SetupFlags: func(f *pflag.FlagSet) {
+			f.BoolVar(&migratePlaceholder0_1Args.dryRun, "dry-run", false, "dry run")
+		},
+	},
+}
+
+var migratePlaceholder0_1Args struct {
+	dryRun bool
+}
+
+func doMigratePlaceholder0_1(sc *cli.Subcommand, args []string) error {
+	if len(args) != 0 {
+		return fmt.Errorf("migration does not take arguments, got %v", args)
+	}
+
+	cfg := sc.Config()
+
+	ctx := context.Background()
+	allFSS, err := zfs.ZFSListMapping(ctx, zfs.NoFilter())
+	if err != nil {
+		return errors.Wrap(err, "cannot list filesystems")
+	}
+
+	type workItem struct {
+		jobName string
+		rootFS  *zfs.DatasetPath
+		fss     []*zfs.DatasetPath
+	}
+	var wis []workItem
+	for i, j := range cfg.Jobs {
+		var rfsS string
+		switch job := j.Ret.(type) {
+		case *config.SinkJob:
+			rfsS = job.RootFS
+		case *config.PullJob:
+			rfsS = job.RootFS
+		default:
+			fmt.Printf("ignoring job %q (%d/%d, type %T)\n", j.Name(), i, len(cfg.Jobs), j.Ret)
+			continue
+		}
+		rfs, err := zfs.NewDatasetPath(rfsS)
+		if err != nil {
+			return errors.Wrapf(err, "root fs for job %q is not a valid dataset path", j.Name())
+		}
+		var fss []*zfs.DatasetPath
+		for _, fs := range allFSS {
+			if fs.HasPrefix(rfs) {
+				fss = append(fss, fs)
+			}
+		}
+		wis = append(wis, workItem{j.Name(), rfs, fss})
+	}
+
+	for _, wi := range wis {
+		fmt.Printf("job %q => migrate filesystems below root_fs %q\n", wi.jobName, wi.rootFS.ToString())
+		if len(wi.fss) == 0 {
+			fmt.Printf("\tno filesystems\n")
+			continue
+		}
+		for _, fs := range wi.fss {
+			fmt.Printf("\t%q ... ", fs.ToString())
+			r, err := zfs.ZFSMigrateHashBasedPlaceholderToCurrent(fs, migratePlaceholder0_1Args.dryRun)
+			if err != nil {
+				fmt.Printf("error: %s\n", err)
+			} else if !r.NeedsModification {
+				fmt.Printf("unchanged (placeholder=%v)\n", r.OriginalState.IsPlaceholder)
+			} else {
+				fmt.Printf("migrate (placeholder=%v) (old value = %q)\n",
+					r.OriginalState.IsPlaceholder, r.OriginalState.RawLocalPropertyValue)
+			}
+		}
+	}
+
+	return nil
+}
@@ -0,0 +1,16 @@
+package client
+
+import "testing"
+
+func TestMigrationsUnambiguousNames(t *testing.T) {
+	names := make(map[string]bool)
+	for _, mig := range migrations {
+		if _, ok := names[mig.Use]; ok {
+			t.Errorf("duplicate migration name %q", mig.Use)
+			t.FailNow()
+			return
+		} else {
+			names[mig.Use] = true
+		}
+	}
+}
@@ -2,11 +2,12 @@ package client

 import (
 	"errors"
+	"log"
+	"os"
+
 	"github.com/zrepl/zrepl/cli"
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon"
-	"log"
-	"os"
 )

 var pprofArgs struct {
@@ -2,6 +2,7 @@ package client

 import (
 	"github.com/pkg/errors"
+
 	"github.com/zrepl/zrepl/cli"
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon"
@@ -2,16 +2,6 @@ package client

 import (
 	"fmt"
-	"github.com/gdamore/tcell/termbox"
-	"github.com/pkg/errors"
-	"github.com/spf13/pflag"
-	"github.com/zrepl/yaml-config"
-	"github.com/zrepl/zrepl/cli"
-	"github.com/zrepl/zrepl/daemon"
-	"github.com/zrepl/zrepl/daemon/job"
-	"github.com/zrepl/zrepl/daemon/pruner"
-	"github.com/zrepl/zrepl/replication"
-	"github.com/zrepl/zrepl/replication/fsrep"
 	"io"
 	"math"
 	"net/http"
@@ -20,8 +10,72 @@ import (
 	"strings"
 	"sync"
 	"time"
+
+	// tcell is the termbox-compatbile library for abstracting away escape sequences, etc.
+	// as of tcell#252, the number of default distributed terminals is relatively limited
+	// additional terminal definitions can be included via side-effect import
+	// See https://github.com/gdamore/tcell/blob/master/terminfo/base/base.go
+	// See https://github.com/gdamore/tcell/issues/252#issuecomment-533836078
+	"github.com/gdamore/tcell/termbox"
+	_ "github.com/gdamore/tcell/terminfo/s/screen" // tmux on FreeBSD 11 & 12 without ncurses
+
+	"github.com/pkg/errors"
+	"github.com/spf13/pflag"
+	"github.com/zrepl/yaml-config"
+
+	"github.com/zrepl/zrepl/cli"
+	"github.com/zrepl/zrepl/daemon"
+	"github.com/zrepl/zrepl/daemon/job"
+	"github.com/zrepl/zrepl/daemon/pruner"
+	"github.com/zrepl/zrepl/daemon/snapper"
+	"github.com/zrepl/zrepl/replication/report"
 )

+type byteProgressMeasurement struct {
+	time time.Time
+	val  int64
+}
+
+type bytesProgressHistory struct {
+	last        *byteProgressMeasurement // pointer as poor man's optional
+	changeCount int
+	lastChange  time.Time
+	bpsAvg      float64
+}
+
+func (p *bytesProgressHistory) Update(currentVal int64) (bytesPerSecondAvg int64, changeCount int) {
+
+	if p.last == nil {
+		p.last = &byteProgressMeasurement{
+			time: time.Now(),
+			val:  currentVal,
+		}
+		return 0, 0
+	}
+
+	if p.last.val != currentVal {
+		p.changeCount++
+		p.lastChange = time.Now()
+	}
+
+	if time.Since(p.lastChange) > 3*time.Second {
+		p.last = nil
+		return 0, 0
+	}
+
+	deltaV := currentVal - p.last.val
+	deltaT := time.Since(p.last.time)
+	rate := float64(deltaV) / deltaT.Seconds()
+
+	factor := 0.3
+	p.bpsAvg = (1-factor)*p.bpsAvg + factor*rate
+
+	p.last.time = time.Now()
+	p.last.val = currentVal
+
+	return int64(p.bpsAvg), p.changeCount
+}
+
 type tui struct {
 	x, y   int
 	indent int
@@ -29,20 +83,23 @@ type tui struct {
 	lock   sync.Mutex //For report and error
 	report map[string]job.Status
 	err    error
+
+	jobFilter string
+
+	replicationProgress map[string]*bytesProgressHistory // by job name
 }

 func newTui() tui {
-	return tui{}
+	return tui{
+		replicationProgress: make(map[string]*bytesProgressHistory),
+	}
 }

-func (t *tui) moveCursor(x, y int) {
-	t.x += x
-	t.y += y
-}
+const INDENT_MULTIPLIER = 4

 func (t *tui) moveLine(dl int, col int) {
 	t.y += dl
-	t.x = t.indent*4 + col
+	t.x = t.indent*INDENT_MULTIPLIER + col
 }

 func (t *tui) write(text string) {
@@ -60,6 +117,40 @@ func (t *tui) printf(text string, a ...interface{}) {
 	t.write(fmt.Sprintf(text, a...))
 }

+func wrap(s string, width int) string {
+	var b strings.Builder
+	for len(s) > 0 {
+		rem := width
+		if rem > len(s) {
+			rem = len(s)
+		}
+		if idx := strings.IndexAny(s, "\n\r"); idx != -1 && idx < rem {
+			rem = idx + 1
+		}
+		untilNewline := strings.TrimRight(s[:rem], "\n\r")
+		s = s[rem:]
+		if len(untilNewline) == 0 {
+			continue
+		}
+		b.WriteString(untilNewline)
+		b.WriteString("\n")
+	}
+	return strings.TrimRight(b.String(), "\n\r")
+}
+
+func (t *tui) printfDrawIndentedAndWrappedIfMultiline(format string, a ...interface{}) {
+	whole := fmt.Sprintf(format, a...)
+	width, _ := termbox.Size()
+	if !strings.ContainsAny(whole, "\n\r") && t.x+len(whole) <= width {
+		t.printf(format, a...)
+	} else {
+		t.addIndent(1)
+		t.newline()
+		t.write(wrap(whole, width-INDENT_MULTIPLIER*t.indent))
+		t.addIndent(-1)
+	}
+}
+
 func (t *tui) newline() {
 	t.moveLine(1, 0)
 }
@@ -76,6 +167,7 @@ func (t *tui) addIndent(indent int) {

 var statusFlags struct {
 	Raw bool
+	Job string
 }

 var StatusCmd = &cli.Subcommand{
@@ -83,6 +175,7 @@ var StatusCmd = &cli.Subcommand{
 	Short: "show job activity or dump as JSON for monitoring",
 	SetupFlags: func(f *pflag.FlagSet) {
 		f.BoolVar(&statusFlags.Raw, "raw", false, "dump raw status description from zrepl daemon")
+		f.StringVar(&statusFlags.Job, "job", "", "only dump specified job")
 	},
 	Run: runStatus,
 }
@@ -94,14 +187,17 @@ func runStatus(s *cli.Subcommand, args []string) error {
 	}

 	if statusFlags.Raw {
-		resp, err := httpc.Get("http://unix"+daemon.ControlJobEndpointStatus)
+		resp, err := httpc.Get("http://unix" + daemon.ControlJobEndpointStatus)
 		if err != nil {
 			return err
 		}
 		defer resp.Body.Close()
 		if resp.StatusCode != http.StatusOK {
 			fmt.Fprintf(os.Stderr, "Received error response:\n")
-			io.CopyN(os.Stderr, resp.Body, 4096)
+			_, err := io.CopyN(os.Stderr, resp.Body, 4096)
+			if err != nil {
+				return err
+			}
 			return errors.Errorf("exit")
 		}
 		if _, err := io.Copy(os.Stdout, resp.Body); err != nil {
@@ -114,6 +210,7 @@ func runStatus(s *cli.Subcommand, args []string) error {
 	t.lock.Lock()
 	t.err = errors.New("Got no report yet")
 	t.lock.Unlock()
+	t.jobFilter = statusFlags.Job

 	err = termbox.Init()
 	if err != nil {
@@ -140,7 +237,7 @@ func runStatus(s *cli.Subcommand, args []string) error {
 	ticker := time.NewTicker(500 * time.Millisecond)
 	defer ticker.Stop()
 	go func() {
-		for _ = range ticker.C {
+		for range ticker.C {
 			update()
 		}
 	}()
@@ -167,6 +264,15 @@ loop:

 }

+func (t *tui) getReplicationProgresHistory(jobName string) *bytesProgressHistory {
+	p, ok := t.replicationProgress[jobName]
+	if !ok {
+		p = &bytesProgressHistory{}
+		t.replicationProgress[jobName] = p
+	}
+	return p
+}
+
 func (t *tui) draw() {
 	t.lock.Lock()
 	defer t.lock.Unlock()
@@ -180,19 +286,29 @@ func (t *tui) draw() {
 		t.write(t.err.Error())
 	} else {
 		//Iterate over map in alphabetical order
-		keys := make([]string, len(t.report))
-		i := 0
-		for k, _ := range t.report {
-			keys[i] = k
-			i++
-		}
-		sort.Strings(keys)
-
-		for _, k := range keys {
-			v := t.report[k]
+		keys := make([]string, 0, len(t.report))
+		for k := range t.report {
 			if len(k) == 0 || daemon.IsInternalJobName(k) { //Internal job
 				continue
 			}
+			if t.jobFilter != "" && k != t.jobFilter {
+				continue
+			}
+			keys = append(keys, k)
+		}
+		sort.Strings(keys)
+
+		if len(keys) == 0 {
+			t.setIndent(0)
+			t.printf("no jobs to display")
+			t.newline()
+			termbox.Flush()
+			return
+		}
+
+		for _, k := range keys {
+			v := t.report[k]
+
 			t.setIndent(0)

 			t.printf("Job: %s", k)
@@ -202,7 +318,66 @@ func (t *tui) draw() {
 			t.setIndent(1)
 			t.newline()

-			if v.Type != job.TypePush && v.Type != job.TypePull {
+			if v.Type == job.TypePush || v.Type == job.TypePull {
+				activeStatus, ok := v.JobSpecific.(*job.ActiveSideStatus)
+				if !ok || activeStatus == nil {
+					t.printf("ActiveSideStatus is null")
+					t.newline()
+					continue
+				}
+
+				t.printf("Replication:")
+				t.newline()
+				t.addIndent(1)
+				t.renderReplicationReport(activeStatus.Replication, t.getReplicationProgresHistory(k))
+				t.addIndent(-1)
+
+				t.printf("Pruning Sender:")
+				t.newline()
+				t.addIndent(1)
+				t.renderPrunerReport(activeStatus.PruningSender)
+				t.addIndent(-1)
+
+				t.printf("Pruning Receiver:")
+				t.newline()
+				t.addIndent(1)
+				t.renderPrunerReport(activeStatus.PruningReceiver)
+				t.addIndent(-1)
+
+				if v.Type == job.TypePush {
+					t.printf("Snapshotting:")
+					t.newline()
+					t.addIndent(1)
+					t.renderSnapperReport(activeStatus.Snapshotting)
+					t.addIndent(-1)
+				}
+
+			} else if v.Type == job.TypeSnap {
+				snapStatus, ok := v.JobSpecific.(*job.SnapJobStatus)
+				if !ok || snapStatus == nil {
+					t.printf("SnapJobStatus is null")
+					t.newline()
+					continue
+				}
+				t.printf("Pruning snapshots:")
+				t.newline()
+				t.addIndent(1)
+				t.renderPrunerReport(snapStatus.Pruning)
+				t.addIndent(-1)
+				t.printf("Snapshotting:")
+				t.newline()
+				t.addIndent(1)
+				t.renderSnapperReport(snapStatus.Snapshotting)
+				t.addIndent(-1)
+			} else if v.Type == job.TypeSource {
+
+				st := v.JobSpecific.(*job.PassiveStatus)
+				t.printf("Snapshotting:\n")
+				t.addIndent(1)
+				t.renderSnapperReport(st.Snapper)
+				t.addIndent(-1)
+
+			} else {
 				t.printf("No status representation for job type '%s', dumping as YAML", v.Type)
 				t.newline()
 				asYaml, err := yaml.Marshal(v.JobSpecific)
@@ -215,103 +390,96 @@ func (t *tui) draw() {
 				t.newline()
 				continue
 			}
-
-			pushStatus, ok := v.JobSpecific.(*job.ActiveSideStatus)
-			if !ok || pushStatus == nil {
-				t.printf("ActiveSideStatus is null")
-				t.newline()
-				continue
-			}
-
-			t.printf("Replication:")
-			t.newline()
-			t.addIndent(1)
-			t.renderReplicationReport(pushStatus.Replication)
-			t.addIndent(-1)
-
-			t.printf("Pruning Sender:")
-			t.newline()
-			t.addIndent(1)
-			t.renderPrunerReport(pushStatus.PruningSender)
-			t.addIndent(-1)
-
-			t.printf("Pruning Receiver:")
-			t.newline()
-			t.addIndent(1)
-			t.renderPrunerReport(pushStatus.PruningReceiver)
-			t.addIndent(-1)
-
 		}
 	}
 	termbox.Flush()
 }

-func (t *tui) renderReplicationReport(rep *replication.Report) {
+func (t *tui) renderReplicationReport(rep *report.Report, history *bytesProgressHistory) {
 	if rep == nil {
 		t.printf("...\n")
 		return
 	}

-	all := make([]*fsrep.Report, 0, len(rep.Completed)+len(rep.Pending) + 1)
-	all = append(all, rep.Completed...)
-	all = append(all, rep.Pending...)
-	if rep.Active != nil {
-		all = append(all, rep.Active)
+	if rep.WaitReconnectError != nil {
+		t.printfDrawIndentedAndWrappedIfMultiline("Connectivity: %s", rep.WaitReconnectError)
+		t.newline()
 	}
-	sort.Slice(all, func(i, j int) bool {
-		return all[i].Filesystem < all[j].Filesystem
+	if !rep.WaitReconnectSince.IsZero() {
+		delta := time.Until(rep.WaitReconnectUntil).Round(time.Second)
+		if rep.WaitReconnectUntil.IsZero() || delta > 0 {
+			var until string
+			if rep.WaitReconnectUntil.IsZero() {
+				until = "waiting indefinitely"
+			} else {
+				until = fmt.Sprintf("hard fail in %s @ %s", delta, rep.WaitReconnectUntil)
+			}
+			t.printfDrawIndentedAndWrappedIfMultiline("Connectivity: reconnecting with exponential backoff (since %s) (%s)",
+				rep.WaitReconnectSince, until)
+		} else {
+			t.printfDrawIndentedAndWrappedIfMultiline("Connectivity: reconnects reached hard-fail timeout @ %s", rep.WaitReconnectUntil)
+		}
+		t.newline()
+	}
+
+	// TODO visualize more than the latest attempt by folding all attempts into one
+	if len(rep.Attempts) == 0 {
+		t.printf("no attempts made yet")
+		return
+	} else {
+		t.printf("Attempt #%d", len(rep.Attempts))
+		if len(rep.Attempts) > 1 {
+			t.printf(". Previous attempts failed with the following statuses:")
+			t.newline()
+			t.addIndent(1)
+			for i, a := range rep.Attempts[:len(rep.Attempts)-1] {
+				t.printfDrawIndentedAndWrappedIfMultiline("#%d: %s (failed at %s) (ran %s)", i+1, a.State, a.FinishAt, a.FinishAt.Sub(a.StartAt))
+				t.newline()
+			}
+			t.addIndent(-1)
+		} else {
+			t.newline()
+		}
+	}
+
+	latest := rep.Attempts[len(rep.Attempts)-1]
+	sort.Slice(latest.Filesystems, func(i, j int) bool {
+		return latest.Filesystems[i].Info.Name < latest.Filesystems[j].Info.Name
 	})

-	state, err := replication.StateString(rep.Status)
-	if err != nil {
-		t.printf("Status: %q (parse error: %q)\n", rep.Status, err)
-		return
-	}
-
-	t.printf("Status: %s", state)
+	t.printf("Status: %s", latest.State)
 	t.newline()
-	if rep.Problem != "" {
-		t.printf("Problem: %s", rep.Problem)
+	if latest.State == report.AttemptPlanningError {
+		t.printf("Problem: ")
+		t.printfDrawIndentedAndWrappedIfMultiline("%s", latest.PlanError)
+		t.newline()
+	} else if latest.State == report.AttemptFanOutError {
+		t.printf("Problem: one or more of the filesystems encountered errors")
 		t.newline()
 	}
-	if rep.SleepUntil.After(time.Now()) && !state.IsTerminal() {
-		t.printf("Sleeping until %s (%s left)\n", rep.SleepUntil, rep.SleepUntil.Sub(time.Now()))
-	}

-	if state != replication.Planning && state != replication.PlanningError {
+	if latest.State != report.AttemptPlanning && latest.State != report.AttemptPlanningError {
+		// Draw global progress bar
 		// Progress: [---------------]
-		sumUpFSRep := func(rep *fsrep.Report) (transferred, total int64) {
-			for _, s := range rep.Pending {
-				transferred += s.Bytes
-				total  += s.ExpectedBytes
-			}
-			for _, s := range rep.Completed {
-				transferred += s.Bytes
-				total += s.ExpectedBytes
-			}
-			return
-		}
-		var transferred, total int64
-		for _, fs := range all {
-			fstx, fstotal := sumUpFSRep(fs)
-			transferred += fstx
-			total += fstotal
-		}
+		expected, replicated := latest.BytesSum()
+		rate, changeCount := history.Update(replicated)
 		t.write("Progress: ")
-		t.drawBar(80, transferred, total)
-		t.write(fmt.Sprintf(" %s / %s", ByteCountBinary(transferred), ByteCountBinary(total)))
+		t.drawBar(50, replicated, expected, changeCount)
+		t.write(fmt.Sprintf(" %s / %s @ %s/s", ByteCountBinary(replicated), ByteCountBinary(expected), ByteCountBinary(rate)))
 		t.newline()
-	}

 		var maxFSLen int
-	for _, fs := range all {
-		if len(fs.Filesystem) > maxFSLen {
-			maxFSLen = len(fs.Filesystem)
+		for _, fs := range latest.Filesystems {
+			if len(fs.Info.Name) > maxFSLen {
+				maxFSLen = len(fs.Info.Name)
 			}
 		}
-	for _, fs := range all {
-		printFilesystemStatus(fs, t, fs == rep.Active, maxFSLen)
+		for _, fs := range latest.Filesystems {
+			t.printFilesystemStatus(fs, false, maxFSLen) // FIXME bring 'active' flag back
 		}
+
+	}
+
 }

 func (t *tui) renderPrunerReport(r *pruner.Report) {
@@ -332,15 +500,12 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 	if r.Error != "" {
 		t.printf("Error: %s\n", r.Error)
 	}
-	if r.SleepUntil.After(time.Now()) {
-		t.printf("Sleeping until %s (%s left)\n", r.SleepUntil, r.SleepUntil.Sub(time.Now()))
-	}

 	type commonFS struct {
 		*pruner.FSReport
 		completed bool
 	}
-	all := make([]commonFS, 0, len(r.Pending) + len(r.Completed))
+	all := make([]commonFS, 0, len(r.Pending)+len(r.Completed))
 	for i := range r.Pending {
 		all = append(all, commonFS{&r.Pending[i], false})
 	}
@@ -349,9 +514,9 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 	}

 	switch state {
-	case pruner.Plan: fallthrough
-	case pruner.PlanWait: fallthrough
-	case pruner.ErrPerm:
+	case pruner.Plan:
+		fallthrough
+	case pruner.PlanErr:
 		return
 	}

@@ -378,7 +543,7 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 	t.write("[")
 	t.write(times("=", progress))
 	t.write(">")
-	t.write(times("-", 80 - progress))
+	t.write(times("-", 80-progress))
 	t.write("]")
 	t.printf(" %d/%d snapshots", completedDestroyCount, totalDestroyCount)
 	t.newline()
@@ -391,8 +556,18 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 	for _, fs := range all {
 		t.write(rightPad(fs.Filesystem, maxFSname, " "))
 		t.write(" ")
-		if fs.Error != "" {
-			t.printf("ERROR:    %s\n", fs.Error) // whitespace is padding
+		if !fs.SkipReason.NotSkipped() {
+			t.printf("skipped: %s\n", fs.SkipReason)
+			continue
+		}
+		if fs.LastError != "" {
+			if strings.ContainsAny(fs.LastError, "\r\n") {
+				t.printf("ERROR:")
+				t.printfDrawIndentedAndWrappedIfMultiline("%s\n", fs.LastError)
+			} else {
+				t.printfDrawIndentedAndWrappedIfMultiline("ERROR: %s\n", fs.LastError)
+			}
+			t.newline()
 			continue
 		}

@@ -400,7 +575,7 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 			len(fs.DestroyList), len(fs.SnapshotList))

 		if fs.completed {
-			t.printf( "Completed %s\n", pruneRuleActionStr)
+			t.printf("Completed  %s\n", pruneRuleActionStr)
 			continue
 		}

@@ -415,23 +590,84 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {

 }

-const snapshotIndent = 1
-func calculateMaxFSLength(all []*fsrep.Report) (maxFS, maxStatus int) {
-	for _, e := range all {
-		if len(e.Filesystem) > maxFS {
-			maxFS = len(e.Filesystem)
-		}
-		all2 := make([]*fsrep.StepReport, 0, len(e.Pending) + len(e.Completed))
-		all2 = append(all2, e.Pending...)
-		all2 = append(all2, e.Completed...)
-		for _, e2 := range all2 {
-			elen := len(e2.Problem) + len(e2.From) + len(e2.To) + 60 // random spacing, units, labels, etc
-			if elen > maxStatus {
-				maxStatus = elen
-			}
-		}
-	}
+func (t *tui) renderSnapperReport(r *snapper.Report) {
+	if r == nil {
+		t.printf("<snapshot type does not have a report>\n")
 		return
+	}
+
+	t.printf("Status: %s", r.State)
+	t.newline()
+
+	if r.Error != "" {
+		t.printf("Error: %s\n", r.Error)
+	}
+	if !r.SleepUntil.IsZero() {
+		t.printf("Sleep until: %s\n", r.SleepUntil)
+	}
+
+	sort.Slice(r.Progress, func(i, j int) bool {
+		return strings.Compare(r.Progress[i].Path, r.Progress[j].Path) == -1
+	})
+
+	t.addIndent(1)
+	defer t.addIndent(-1)
+	dur := func(d time.Duration) string {
+		return d.Round(100 * time.Millisecond).String()
+	}
+
+	type row struct {
+		path, state, duration, remainder, hookReport string
+	}
+	var widths struct {
+		path, state, duration int
+	}
+	rows := make([]*row, len(r.Progress))
+	for i, fs := range r.Progress {
+		r := &row{
+			path:  fs.Path,
+			state: fs.State.String(),
+		}
+		if fs.HooksHadError {
+			r.hookReport = fs.Hooks // FIXME render here, not in daemon
+		}
+		switch fs.State {
+		case snapper.SnapPending:
+			r.duration = "..."
+			r.remainder = ""
+		case snapper.SnapStarted:
+			r.duration = dur(time.Since(fs.StartAt))
+			r.remainder = fmt.Sprintf("snap name: %q", fs.SnapName)
+		case snapper.SnapDone:
+			fallthrough
+		case snapper.SnapError:
+			r.duration = dur(fs.DoneAt.Sub(fs.StartAt))
+			r.remainder = fmt.Sprintf("snap name: %q", fs.SnapName)
+		}
+		rows[i] = r
+		if len(r.path) > widths.path {
+			widths.path = len(r.path)
+		}
+		if len(r.state) > widths.state {
+			widths.state = len(r.state)
+		}
+		if len(r.duration) > widths.duration {
+			widths.duration = len(r.duration)
+		}
+	}
+
+	for _, r := range rows {
+		path := rightPad(r.path, widths.path, " ")
+		state := rightPad(r.state, widths.state, " ")
+		duration := rightPad(r.duration, widths.duration, " ")
+		t.printf("%s %s %s", path, state, duration)
+		t.printfDrawIndentedAndWrappedIfMultiline(" %s", r.remainder)
+		if r.hookReport != "" {
+			t.printfDrawIndentedAndWrappedIfMultiline("%s", r.hookReport)
+		}
+		t.newline()
+	}
+
 }

 func times(str string, n int) (out string) {
@@ -445,18 +681,13 @@ func rightPad(str string, length int, pad string) string {
 	if len(str) > length {
 		return str[:length]
 	}
-	return str + times(pad, length-len(str))
+	return str + strings.Repeat(pad, length-len(str))
 }

+var arrowPositions = `>\|/`

-func leftPad(str string, length int, pad string) string {
-	if len(str) > length {
-		return str[len(str)-length:]
-	}
-	return times(pad, length-len(str)) + str
-}
-
-func (t *tui) drawBar(length int, bytes, totalBytes int64) {
+// changeCount = 0 indicates stall / no progresss
+func (t *tui) drawBar(length int, bytes, totalBytes int64, changeCount int) {
 	var completedLength int
 	if totalBytes > 0 {
 		completedLength = int(int64(length) * bytes / totalBytes)
@@ -469,66 +700,46 @@ func (t *tui) drawBar(length int, bytes, totalBytes int64) {

 	t.write("[")
 	t.write(times("=", completedLength))
-	t.write(">")
+	t.write(string(arrowPositions[changeCount%len(arrowPositions)]))
 	t.write(times("-", length-completedLength))
 	t.write("]")
 }

-func StringStepState(s fsrep.StepState) string {
-	switch s {
-	case fsrep.StepReplicationReady: return "Ready"
-	case fsrep.StepReplicationRetry: return "Retry"
-	case fsrep.StepMarkReplicatedReady: return "MarkReady"
-	case fsrep.StepMarkReplicatedRetry: return "MarkRetry"
-	case fsrep.StepPermanentError: return "PermanentError"
-	case fsrep.StepCompleted: return "Completed"
-	default:
-		return fmt.Sprintf("UNKNOWN %d", s)
-	}
-}
+func (t *tui) printFilesystemStatus(rep *report.FilesystemReport, active bool, maxFS int) {

-func filesystemStatusString(rep *fsrep.Report, active bool, fsWidth int) (line string, bytes, totalBytes int64) {
-	bytes = int64(0)
-	totalBytes = int64(0)
-	for _, s := range rep.Pending {
-		bytes += s.Bytes
-		totalBytes += s.ExpectedBytes
-	}
-	for _, s := range rep.Completed {
-		bytes += s.Bytes
-		totalBytes += s.ExpectedBytes
-	}
-
-	next := ""
-	if rep.Problem != "" {
-		next = " problem: " + rep.Problem
-	} else if len(rep.Pending) > 0 {
-		if rep.Pending[0].From != "" {
-			next = fmt.Sprintf(" next: %s => %s", rep.Pending[0].From, rep.Pending[0].To)
-		} else {
-			next = fmt.Sprintf(" next: %s (full)", rep.Pending[0].To)
-		}
-	}
-	status := fmt.Sprintf("%s (step %d/%d, %s/%s)%s",
-		rep.Status,
-		len(rep.Completed), len(rep.Pending) + len(rep.Completed),
-		ByteCountBinary(bytes), ByteCountBinary(totalBytes),
-		next,
+	expected, replicated := rep.BytesSum()
+	status := fmt.Sprintf("%s (step %d/%d, %s/%s)",
+		strings.ToUpper(string(rep.State)),
+		rep.CurrentStep, len(rep.Steps),
+		ByteCountBinary(replicated), ByteCountBinary(expected),
 	)
+
 	activeIndicator := " "
 	if active {
 		activeIndicator = "*"
 	}
-	line = fmt.Sprintf("%s %s %s",
+	t.printf("%s %s %s ",
 		activeIndicator,
-		rightPad(rep.Filesystem, fsWidth, " "),
+		rightPad(rep.Info.Name, maxFS, " "),
 		status)
-	return line, bytes, totalBytes
-}

-func printFilesystemStatus(rep *fsrep.Report, t *tui, active bool, maxFS int) {
-	totalStatus, _, _ := filesystemStatusString(rep, active, maxFS)
-	t.write(totalStatus)
+	next := ""
+	if err := rep.Error(); err != nil {
+		next = err.Err
+	} else if rep.State != report.FilesystemDone {
+		if nextStep := rep.NextStep(); nextStep != nil {
+			if nextStep.IsIncremental() {
+				next = fmt.Sprintf("next: %s => %s", nextStep.Info.From, nextStep.Info.To)
+			} else {
+				next = fmt.Sprintf("next: %s (full)", nextStep.Info.To)
+			}
+		} else {
+			next = "" // individual FSes may still be in planning state
+		}
+
+	}
+	t.printfDrawIndentedAndWrappedIfMultiline("%s", next)
+
 	t.newline()
 }

@@ -1,13 +1,15 @@
 package client

 import (
-	"github.com/zrepl/zrepl/cli"
 	"os"

+	"github.com/problame/go-netssh"
+
+	"github.com/zrepl/zrepl/cli"
+	"github.com/zrepl/zrepl/config"
+
 	"context"
 	"errors"
-	"github.com/problame/go-netssh"
-	"github.com/zrepl/zrepl/config"
 	"log"
 	"path"
 )
@@ -2,17 +2,20 @@ package client

 import (
 	"fmt"
+
+	"github.com/pkg/errors"
 	"github.com/spf13/pflag"
+
 	"github.com/zrepl/zrepl/cli"
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/filters"
 	"github.com/zrepl/zrepl/zfs"
 )

-var TestCmd = &cli.Subcommand {
+var TestCmd = &cli.Subcommand{
 	Use: "test",
 	SetupSubcommands: func() []*cli.Subcommand {
-		return []*cli.Subcommand{testFilter}
+		return []*cli.Subcommand{testFilter, testPlaceholder}
 	},
 }

@@ -50,8 +53,10 @@ func runTestFilterCmd(subcommand *cli.Subcommand, args []string) error {
 		return err
 	}
 	switch j := job.Ret.(type) {
-	case *config.SourceJob: confFilter = j.Filesystems
-	case *config.PushJob: confFilter = j.Filesystems
+	case *config.SourceJob:
+		confFilter = j.Filesystems
+	case *config.PushJob:
+		confFilter = j.Filesystems
 	default:
 		return fmt.Errorf("job type %T does not have filesystems filter", j)
 	}
@@ -106,3 +111,68 @@ func runTestFilterCmd(subcommand *cli.Subcommand, args []string) error {
 	}
 	return nil
 }
+
+var testPlaceholderArgs struct {
+	ds  string
+	all bool
+}
+
+var testPlaceholder = &cli.Subcommand{
+	Use:   "placeholder [--all | --dataset DATASET]",
+	Short: fmt.Sprintf("list received placeholder filesystems (zfs property %q)", zfs.PlaceholderPropertyName),
+	Example: `
+	placeholder --all
+	placeholder --dataset path/to/sink/clientident/fs`,
+	NoRequireConfig: true,
+	SetupFlags: func(f *pflag.FlagSet) {
+		f.StringVar(&testPlaceholderArgs.ds, "dataset", "", "dataset path (not required to exist)")
+		f.BoolVar(&testPlaceholderArgs.all, "all", false, "list tab-separated placeholder status of all filesystems")
+	},
+	Run: runTestPlaceholder,
+}
+
+func runTestPlaceholder(subcommand *cli.Subcommand, args []string) error {
+
+	var checkDPs []*zfs.DatasetPath
+
+	// all actions first
+	if testPlaceholderArgs.all {
+		out, err := zfs.ZFSList([]string{"name"})
+		if err != nil {
+			return errors.Wrap(err, "could not list ZFS filesystems")
+		}
+		for _, row := range out {
+			dp, err := zfs.NewDatasetPath(row[0])
+			if err != nil {
+				panic(err)
+			}
+			checkDPs = append(checkDPs, dp)
+		}
+	} else {
+		dp, err := zfs.NewDatasetPath(testPlaceholderArgs.ds)
+		if err != nil {
+			return err
+		}
+		if dp.Empty() {
+			return fmt.Errorf("must specify --dataset DATASET or --all")
+		}
+		checkDPs = append(checkDPs, dp)
+	}
+
+	fmt.Printf("IS_PLACEHOLDER\tDATASET\tzrepl:placeholder\n")
+	for _, dp := range checkDPs {
+		ph, err := zfs.ZFSGetFilesystemPlaceholderState(dp)
+		if err != nil {
+			return errors.Wrap(err, "cannot get placeholder state")
+		}
+		if !ph.FSExists {
+			panic("placeholder state inconsistent: filesystem " + ph.FS + " must exist in this context")
+		}
+		is := "yes"
+		if !ph.IsPlaceholder {
+			is = "no"
+		}
+		fmt.Printf("%s\t%s\t%s\n", is, dp.ToString(), ph.RawLocalPropertyValue)
+	}
+	return nil
+}
@@ -2,12 +2,14 @@ package client

 import (
 	"fmt"
+	"os"
+
 	"github.com/spf13/pflag"
+
 	"github.com/zrepl/zrepl/cli"
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon"
 	"github.com/zrepl/zrepl/version"
-	"os"
 )

 var versionArgs struct {
@@ -2,14 +2,16 @@ package config

 import (
 	"fmt"
-	"github.com/pkg/errors"
-	"github.com/zrepl/yaml-config"
 	"io/ioutil"
+	"log/syslog"
 	"os"
 	"reflect"
 	"regexp"
 	"strconv"
 	"time"
+
+	"github.com/pkg/errors"
+	"github.com/zrepl/yaml-config"
 )

 type Config struct {
@@ -33,12 +35,18 @@ type JobEnum struct {
 func (j JobEnum) Name() string {
 	var name string
 	switch v := j.Ret.(type) {
-	case *PushJob: name = v.Name
-	case *SinkJob: name = v.Name
-	case *PullJob: name = v.Name
-	case *SourceJob: name = v.Name
+	case *SnapJob:
+		name = v.Name
+	case *PushJob:
+		name = v.Name
+	case *SinkJob:
+		name = v.Name
+	case *PullJob:
+		name = v.Name
+	case *SourceJob:
+		name = v.Name
 	default:
-		panic(fmt.Sprintf("unknownn job type %T", v))
+		panic(fmt.Sprintf("unknown job type %T", v))
 	}
 	return name
 }
@@ -58,6 +66,15 @@ type PassiveJob struct {
 	Debug JobDebugSettings `yaml:"debug,optional"`
 }

+type SnapJob struct {
+	Type         string            `yaml:"type"`
+	Name         string            `yaml:"name"`
+	Pruning      PruningLocal      `yaml:"pruning"`
+	Debug        JobDebugSettings  `yaml:"debug,optional"`
+	Snapshotting SnapshottingEnum  `yaml:"snapshotting"`
+	Filesystems  FilesystemsFilter `yaml:"filesystems"`
+}
+
 type PushJob struct {
 	ActiveJob    `yaml:",inline"`
 	Snapshotting SnapshottingEnum  `yaml:"snapshotting"`
@@ -67,7 +84,38 @@ type PushJob struct {
 type PullJob struct {
 	ActiveJob `yaml:",inline"`
 	RootFS    string                   `yaml:"root_fs"`
-	Interval  time.Duration `yaml:"interval,positive"`
+	Interval  PositiveDurationOrManual `yaml:"interval"`
+}
+
+type PositiveDurationOrManual struct {
+	Interval time.Duration
+	Manual   bool
+}
+
+var _ yaml.Unmarshaler = (*PositiveDurationOrManual)(nil)
+
+func (i *PositiveDurationOrManual) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
+	var s string
+	if err := u(&s, true); err != nil {
+		return err
+	}
+	switch s {
+	case "manual":
+		i.Manual = true
+		i.Interval = 0
+	case "":
+		return fmt.Errorf("value must not be empty")
+	default:
+		i.Manual = false
+		i.Interval, err = time.ParseDuration(s)
+		if err != nil {
+			return err
+		}
+		if i.Interval <= 0 {
+			return fmt.Errorf("value must be a positive duration, got %q", s)
+		}
+	}
+	return nil
 }

 type SinkJob struct {
@@ -91,6 +139,7 @@ type SnapshottingPeriodic struct {
 	Type     string        `yaml:"type"`
 	Prefix   string        `yaml:"prefix"`
 	Interval time.Duration `yaml:"interval,positive"`
+	Hooks    HookList      `yaml:"hooks,optional"`
 }

 type SnapshottingManual struct {
@@ -130,7 +179,6 @@ type Global struct {
 	Monitoring []MonitoringEnum       `yaml:"monitoring,optional"`
 	Control    *GlobalControl         `yaml:"control,optional,fromdefaults"`
 	Serve      *GlobalServe           `yaml:"serve,optional,fromdefaults"`
-	RPC        *RPCConfig             `yaml:"rpc,optional,fromdefaults"`
 }

 func Default(i interface{}) {
@@ -145,39 +193,28 @@ func Default(i interface{}) {
 	}
 }

-type RPCConfig struct {
-	Timeout             time.Duration `yaml:"timeout,optional,positive,default=10s"`
-	TxChunkSize         uint32        `yaml:"tx_chunk_size,optional,default=32768"`
-	RxStructuredMaxLen  uint32        `yaml:"rx_structured_max,optional,default=16777216"`
-	RxStreamChunkMaxLen uint32        `yaml:"rx_stream_chunk_max,optional,default=16777216"`
-	RxHeaderMaxLen      uint32        `yaml:"rx_header_max,optional,default=40960"`
-	SendHeartbeatInterval             time.Duration `yaml:"send_heartbeat_interval,optional,positive,default=5s"`
-
-}
-
 type ConnectEnum struct {
 	Ret interface{}
 }

 type ConnectCommon struct {
 	Type string `yaml:"type"`
-	RPC  *RPCConfig `yaml:"rpc,optional"`
 }

 type TCPConnect struct {
 	ConnectCommon `yaml:",inline"`
-	Address       string        `yaml:"address"`
-	DialTimeout   time.Duration `yaml:"dial_timeout,positive,default=10s"`
+	Address       string        `yaml:"address,hostport"`
+	DialTimeout   time.Duration `yaml:"dial_timeout,zeropositive,default=10s"`
 }

 type TLSConnect struct {
 	ConnectCommon `yaml:",inline"`
-	Address       string        `yaml:"address"`
+	Address       string        `yaml:"address,hostport"`
 	Ca            string        `yaml:"ca"`
 	Cert          string        `yaml:"cert"`
 	Key           string        `yaml:"key"`
 	ServerCN      string        `yaml:"server_cn"`
-	DialTimeout   time.Duration `yaml:"dial_timeout,positive,default=10s"`
+	DialTimeout   time.Duration `yaml:"dial_timeout,zeropositive,default=10s"`
 }

 type SSHStdinserverConnect struct {
@@ -189,13 +226,14 @@ type SSHStdinserverConnect struct {
 	TransportOpenCommand []string      `yaml:"transport_open_command,optional"` //TODO unused
 	SSHCommand           string        `yaml:"ssh_command,optional"`            //TODO unused
 	Options              []string      `yaml:"options,optional"`
-	DialTimeout          time.Duration `yaml:"dial_timeout,positive,default=10s"`
+	DialTimeout          time.Duration `yaml:"dial_timeout,zeropositive,default=10s"`
 }

 type LocalConnect struct {
 	ConnectCommon  `yaml:",inline"`
 	ListenerName   string        `yaml:"listener_name"`
 	ClientIdentity string        `yaml:"client_identity"`
+	DialTimeout    time.Duration `yaml:"dial_timeout,zeropositive,default=2s"`
 }

 type ServeEnum struct {
@@ -204,23 +242,22 @@ type ServeEnum struct {

 type ServeCommon struct {
 	Type string `yaml:"type"`
-	RPC  *RPCConfig `yaml:"rpc,optional"`
 }

 type TCPServe struct {
 	ServeCommon `yaml:",inline"`
-	Listen      string            `yaml:"listen"`
+	Listen      string            `yaml:"listen,hostport"`
 	Clients     map[string]string `yaml:"clients"`
 }

 type TLSServe struct {
 	ServeCommon      `yaml:",inline"`
-	Listen           string        `yaml:"listen"`
+	Listen           string        `yaml:"listen,hostport"`
 	Ca               string        `yaml:"ca"`
 	Cert             string        `yaml:"cert"`
 	Key              string        `yaml:"key"`
 	ClientCNs        []string      `yaml:"client_cns"`
-	HandshakeTimeout time.Duration `yaml:"handshake_timeout,positive,default=10s"`
+	HandshakeTimeout time.Duration `yaml:"handshake_timeout,zeropositive,default=10s"`
 }

 type StdinserverServer struct {
@@ -250,6 +287,7 @@ type PruneKeepLastN struct {
 type PruneKeepRegex struct { // FIXME rename to KeepRegex
 	Type   string `yaml:"type"`
 	Regex  string `yaml:"regex"`
+	Negate bool   `yaml:"negate,optional,default=false"`
 }

 type LoggingOutletEnum struct {
@@ -270,12 +308,13 @@ type StdoutLoggingOutlet struct {

 type SyslogLoggingOutlet struct {
 	LoggingOutletCommon `yaml:",inline"`
+	Facility            *SyslogFacility `yaml:"facility,optional,fromdefaults"`
 	RetryInterval       time.Duration   `yaml:"retry_interval,positive,default=10s"`
 }

 type TCPLoggingOutlet struct {
 	LoggingOutletCommon `yaml:",inline"`
-	Address             string               `yaml:"address"`
+	Address             string               `yaml:"address,hostport"`
 	Net                 string               `yaml:"net,default=tcp"`
 	RetryInterval       time.Duration        `yaml:"retry_interval,positive,default=10s"`
 	TLS                 *TCPLoggingOutletTLS `yaml:"tls,optional"`
@@ -293,9 +332,17 @@ type MonitoringEnum struct {

 type PrometheusMonitoring struct {
 	Type   string `yaml:"type"`
-	Listen string `yaml:"listen"`
+	Listen string `yaml:"listen,hostport"`
 }

+type SyslogFacility syslog.Priority
+
+func (f *SyslogFacility) SetDefault() {
+	*f = SyslogFacility(syslog.LOG_LOCAL0)
+}
+
+var _ yaml.Defaulter = (*SyslogFacility)(nil)
+
 type GlobalControl struct {
 	SockPath string `yaml:"sockpath,default=/var/run/zrepl/control"`
 }
@@ -316,6 +363,38 @@ type JobDebugSettings struct {
 	RPCLog bool `yaml:"rpc_log,optional,default=false"`
 }

+type HookList []HookEnum
+
+type HookEnum struct {
+	Ret interface{}
+}
+
+type HookCommand struct {
+	Path               string            `yaml:"path"`
+	Timeout            time.Duration     `yaml:"timeout,optional,positive,default=30s"`
+	Filesystems        FilesystemsFilter `yaml:"filesystems,optional,default={'<': true}"`
+	HookSettingsCommon `yaml:",inline"`
+}
+
+type HookPostgresCheckpoint struct {
+	HookSettingsCommon `yaml:",inline"`
+	DSN                string            `yaml:"dsn"`
+	Timeout            time.Duration     `yaml:"timeout,optional,positive,default=30s"`
+	Filesystems        FilesystemsFilter `yaml:"filesystems"` // required, user should not CHECKPOINT for every FS
+}
+
+type HookMySQLLockTables struct {
+	HookSettingsCommon `yaml:",inline"`
+	DSN                string            `yaml:"dsn"`
+	Timeout            time.Duration     `yaml:"timeout,optional,positive,default=30s"`
+	Filesystems        FilesystemsFilter `yaml:"filesystems"`
+}
+
+type HookSettingsCommon struct {
+	Type       string `yaml:"type"`
+	ErrIsFatal bool   `yaml:"err_is_fatal,optional,default=false"`
+}
+
 func enumUnmarshal(u func(interface{}, bool) error, types map[string]interface{}) (interface{}, error) {
 	var in struct {
 		Type string
@@ -339,6 +418,7 @@ func enumUnmarshal(u func(interface{}, bool) error, types map[string]interface{}

 func (t *JobEnum) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
 	t.Ret, err = enumUnmarshal(u, map[string]interface{}{
+		"snap":   &SnapJob{},
 		"push":   &PushJob{},
 		"sink":   &SinkJob{},
 		"pull":   &PullJob{},
@@ -362,7 +442,7 @@ func (t *ServeEnum) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
 		"tcp":         &TCPServe{},
 		"tls":         &TLSServe{},
 		"stdinserver": &StdinserverServer{},
-		"local"      : &LocalServe{},
+		"local":       &LocalServe{},
 	})
 	return
 }
@@ -401,6 +481,69 @@ func (t *MonitoringEnum) UnmarshalYAML(u func(interface{}, bool) error) (err err
 	return
 }

+func (t *SyslogFacility) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
+	var s string
+	if err := u(&s, true); err != nil {
+		return err
+	}
+	var level syslog.Priority
+	switch s {
+	case "kern":
+		level = syslog.LOG_KERN
+	case "user":
+		level = syslog.LOG_USER
+	case "mail":
+		level = syslog.LOG_MAIL
+	case "daemon":
+		level = syslog.LOG_DAEMON
+	case "auth":
+		level = syslog.LOG_AUTH
+	case "syslog":
+		level = syslog.LOG_SYSLOG
+	case "lpr":
+		level = syslog.LOG_LPR
+	case "news":
+		level = syslog.LOG_NEWS
+	case "uucp":
+		level = syslog.LOG_UUCP
+	case "cron":
+		level = syslog.LOG_CRON
+	case "authpriv":
+		level = syslog.LOG_AUTHPRIV
+	case "ftp":
+		level = syslog.LOG_FTP
+	case "local0":
+		level = syslog.LOG_LOCAL0
+	case "local1":
+		level = syslog.LOG_LOCAL1
+	case "local2":
+		level = syslog.LOG_LOCAL2
+	case "local3":
+		level = syslog.LOG_LOCAL3
+	case "local4":
+		level = syslog.LOG_LOCAL4
+	case "local5":
+		level = syslog.LOG_LOCAL5
+	case "local6":
+		level = syslog.LOG_LOCAL6
+	case "local7":
+		level = syslog.LOG_LOCAL7
+	default:
+		return fmt.Errorf("invalid syslog level: %q", s)
+	}
+	*t = SyslogFacility(level)
+	return nil
+}
+
+func (t *HookEnum) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
+	t.Ret, err = enumUnmarshal(u, map[string]interface{}{
+		"command":             &HookCommand{},
+		"postgres-checkpoint": &HookPostgresCheckpoint{},
+		"mysql-lock-tables":   &HookMySQLLockTables{},
+	})
+	return
+}
+
 var ConfigFileDefaultLocations = []string{
 	"/etc/zrepl/zrepl.yml",
 	"/usr/local/etc/zrepl/zrepl.yml",
@@ -1,10 +1,13 @@
 package config

 import (
+	"fmt"
+	"log/syslog"
+	"testing"
+
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"github.com/zrepl/yaml-config"
-	"testing"
 )

 func testValidGlobalSection(t *testing.T, s string) *Config {
@@ -22,7 +25,7 @@ jobs:
 `
 	_, err := ParseConfigBytes([]byte(jobdef))
 	require.NoError(t, err)
-	return testValidConfig(t, s + jobdef)
+	return testValidConfig(t, s+jobdef)
 }

 func TestOutletTypes(t *testing.T) {
@@ -72,6 +75,36 @@ global:
 	assert.Equal(t, ":9091", conf.Global.Monitoring[0].Ret.(*PrometheusMonitoring).Listen)
 }

+func TestSyslogLoggingOutletFacility(t *testing.T) {
+	type SyslogFacilityPriority struct {
+		Facility string
+		Priority syslog.Priority
+	}
+	syslogFacilitiesPriorities := []SyslogFacilityPriority{
+		{"", syslog.LOG_LOCAL0}, // default
+		{"kern", syslog.LOG_KERN}, {"daemon", syslog.LOG_DAEMON}, {"auth", syslog.LOG_AUTH},
+		{"syslog", syslog.LOG_SYSLOG}, {"lpr", syslog.LOG_LPR}, {"news", syslog.LOG_NEWS},
+		{"uucp", syslog.LOG_UUCP}, {"cron", syslog.LOG_CRON}, {"authpriv", syslog.LOG_AUTHPRIV},
+		{"ftp", syslog.LOG_FTP}, {"local0", syslog.LOG_LOCAL0}, {"local1", syslog.LOG_LOCAL1},
+		{"local2", syslog.LOG_LOCAL2}, {"local3", syslog.LOG_LOCAL3}, {"local4", syslog.LOG_LOCAL4},
+		{"local5", syslog.LOG_LOCAL5}, {"local6", syslog.LOG_LOCAL6}, {"local7", syslog.LOG_LOCAL7},
+	}
+
+	for _, sFP := range syslogFacilitiesPriorities {
+		logcfg := fmt.Sprintf(`
+global:
+  logging:
+  - type: syslog
+    level: info
+    format: human
+    facility: %s
+`, sFP.Facility)
+		conf := testValidGlobalSection(t, logcfg)
+		assert.Equal(t, 1, len(*conf.Global.Logging))
+		assert.True(t, SyslogFacility(sFP.Priority) == *(*conf.Global.Logging)[0].Ret.(*SyslogLoggingOutlet).Facility)
+	}
+}
+
 func TestLoggingOutletEnumList_SetDefaults(t *testing.T) {
 	e := &LoggingOutletEnumList{}
 	var i yaml.Defaulter = e
@@ -2,6 +2,7 @@ package config

 import (
 	"testing"
+
 	"github.com/stretchr/testify/assert"
 )

@@ -0,0 +1,41 @@
+package config
+
+import (
+	"fmt"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/zrepl/yaml-config"
+)
+
+func TestPositiveDurationOrManual(t *testing.T) {
+	cases := []struct {
+		Comment, Input string
+		Result         *PositiveDurationOrManual
+	}{
+		{"empty is error", "", nil},
+		{"negative is error", "-1s", nil},
+		{"zero seconds is error", "0s", nil},
+		{"zero is error", "0", nil},
+		{"non-manual is error", "something", nil},
+		{"positive seconds works", "1s", &PositiveDurationOrManual{Manual: false, Interval: 1 * time.Second}},
+		{"manual works", "manual", &PositiveDurationOrManual{Manual: true, Interval: 0}},
+	}
+	for _, tc := range cases {
+		t.Run(tc.Comment, func(t *testing.T) {
+			var out struct {
+				FieldName PositiveDurationOrManual `yaml:"fieldname"`
+			}
+			input := fmt.Sprintf("\nfieldname: %s\n", tc.Input)
+			err := yaml.UnmarshalStrict([]byte(input), &out)
+			if tc.Result == nil {
+				assert.Error(t, err)
+				t.Logf("%#v", out)
+			} else {
+				assert.Equal(t, *tc.Result, out.FieldName)
+			}
+		})
+	}
+
+}
@@ -1,86 +0,0 @@
-package config
-
-import (
-	"github.com/stretchr/testify/assert"
-	"testing"
-	"time"
-)
-
-func TestRPC(t *testing.T) {
-	conf := testValidConfig(t, `
-jobs:
- name: pull_servers
-  type: pull
-  connect:
-    type: tcp
-    address: "server1.foo.bar:8888"
-    rpc:
-      timeout: 20s # different form default, should merge
-  root_fs: "pool2/backup_servers"
-  interval: 10m
-  pruning:
-    keep_sender:
-      - type: not_replicated
-    keep_receiver:
-      - type: last_n
-        count: 100
-
- name: pull_servers2
-  type: pull
-  connect:
-    type: tcp
-    address: "server1.foo.bar:8888"
-    rpc:
-      tx_chunk_size: 0xabcd # different from default, should merge
-  root_fs: "pool2/backup_servers"
-  interval: 10m
-  pruning:
-    keep_sender:
-    - type: not_replicated
-    keep_receiver:
-    - type: last_n
-      count: 100
-
- type: sink
-  name: "laptop_sink"
-  root_fs: "pool2/backup_laptops"
-  serve:
-    type: tcp
-    listen: "192.168.122.189:8888"
-    clients: {
-	"10.23.42.23":"client1"
-    }
-    rpc:
-      rx_structured_max: 0x2342
-
- type: sink
-  name: "other_sink"
-  root_fs: "pool2/backup_laptops"
-  serve:
-    type: tcp
-    listen: "192.168.122.189:8888"
-    clients: {
-	"10.23.42.23":"client1"
-    }
-    rpc:
-      send_heartbeat_interval: 10s
-
-`)
-
-	assert.Equal(t, 20*time.Second, conf.Jobs[0].Ret.(*PullJob).Connect.Ret.(*TCPConnect).RPC.Timeout)
-	assert.Equal(t, uint32(0xabcd), conf.Jobs[1].Ret.(*PullJob).Connect.Ret.(*TCPConnect).RPC.TxChunkSize)
-	assert.Equal(t, uint32(0x2342), conf.Jobs[2].Ret.(*SinkJob).Serve.Ret.(*TCPServe).RPC.RxStructuredMaxLen)
-	assert.Equal(t, 10*time.Second, conf.Jobs[3].Ret.(*SinkJob).Serve.Ret.(*TCPServe).RPC.SendHeartbeatInterval)
-	defConf := RPCConfig{}
-	Default(&defConf)
-	assert.Equal(t, defConf.Timeout, conf.Global.RPC.Timeout)
-}
-
-func TestGlobal_DefaultRPCConfig(t *testing.T) {
-	assert.NotPanics(t, func() {
-		var c RPCConfig
-		Default(&c)
-		assert.NotNil(t, c)
-		assert.Equal(t, c.TxChunkSize, uint32(1)<<15)
-	})
-}
@@ -2,9 +2,10 @@ package config

 import (
 	"fmt"
-	"github.com/stretchr/testify/assert"
 	"testing"
 	"time"
+
+	"github.com/stretchr/testify/assert"
 )

 func TestSnapshotting(t *testing.T) {
@@ -37,7 +38,30 @@ jobs:
    interval: 10m
 `

-	fillSnapshotting := func(s string) string {return fmt.Sprintf(tmpl, s)}
+	hooks := `
+  snapshotting:
+    type: periodic
+    prefix: zrepl_
+    interval: 10m
+    hooks:
+    - type: command
+      path: /tmp/path/to/command
+    - type: command
+      path: /tmp/path/to/command
+      filesystems: { "zroot<": true, "<": false }
+    - type: postgres-checkpoint
+      dsn: "host=localhost port=5432 user=postgres sslmode=disable"
+      filesystems: {
+          "tank/postgres/data11": true
+      }
+    - type: mysql-lock-tables
+      dsn: "root@tcp(localhost)/"
+      filesystems: {
+        "tank/mysql": true
+      }
+`
+
+	fillSnapshotting := func(s string) string { return fmt.Sprintf(tmpl, s) }
 	var c *Config

 	t.Run("manual", func(t *testing.T) {
@@ -51,7 +75,16 @@ jobs:
 		snp := c.Jobs[0].Ret.(*PushJob).Snapshotting.Ret.(*SnapshottingPeriodic)
 		assert.Equal(t, "periodic", snp.Type)
 		assert.Equal(t, 10*time.Minute, snp.Interval)
-		assert.Equal(t, "zrepl_" , snp.Prefix)
+		assert.Equal(t, "zrepl_", snp.Prefix)
+	})
+
+	t.Run("hooks", func(t *testing.T) {
+		c = testValidConfig(t, fillSnapshotting(hooks))
+		hs := c.Jobs[0].Ret.(*PushJob).Snapshotting.Ret.(*SnapshottingPeriodic).Hooks
+		assert.Equal(t, hs[0].Ret.(*HookCommand).Filesystems["<"], true)
+		assert.Equal(t, hs[1].Ret.(*HookCommand).Filesystems["zroot<"], true)
+		assert.Equal(t, hs[2].Ret.(*HookPostgresCheckpoint).Filesystems["tank/postgres/data11"], true)
+		assert.Equal(t, hs[3].Ret.(*HookMySQLLockTables).Filesystems["tank/mysql"], true)
 	})

 }
@@ -1,11 +1,18 @@
 package config

 import (
-	"github.com/kr/pretty"
-	"github.com/stretchr/testify/require"
+	"bufio"
+	"bytes"
+	"fmt"
 	"path"
 	"path/filepath"
+	"strings"
 	"testing"
+	"text/template"
+
+	"github.com/kr/pretty"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 )

 func TestSampleConfigsAreParsedWithoutErrors(t *testing.T) {
@@ -35,8 +42,22 @@ func TestSampleConfigsAreParsedWithoutErrors(t *testing.T) {

 }

+// template must be a template/text template with a single '{{ . }}' as placehodler for val
+//nolint[:deadcode,unused]
+func testValidConfigTemplate(t *testing.T, tmpl string, val string) *Config {
+	tmp, err := template.New("master").Parse(tmpl)
+	if err != nil {
+		panic(err)
+	}
+	var buf bytes.Buffer
+	err = tmp.Execute(&buf, val)
+	if err != nil {
+		panic(err)
+	}
+	return testValidConfig(t, buf.String())
+}

-func testValidConfig(t *testing.T, input string) (*Config) {
+func testValidConfig(t *testing.T, input string) *Config {
 	t.Helper()
 	conf, err := testConfig(t, input)
 	require.NoError(t, err)
@@ -48,3 +69,20 @@ func testConfig(t *testing.T, input string) (*Config, error) {
 	t.Helper()
 	return ParseConfigBytes([]byte(input))
 }
+
+func trimSpaceEachLineAndPad(s, pad string) string {
+	var out strings.Builder
+	scan := bufio.NewScanner(strings.NewReader(s))
+	for scan.Scan() {
+		fmt.Fprintf(&out, "%s%s\n", pad, bytes.TrimSpace(scan.Bytes()))
+	}
+	return out.String()
+}
+
+func TestTrimSpaceEachLineAndPad(t *testing.T) {
+	foo := `
+	foo
+	bar baz 
+	`
+	assert.Equal(t, "  \n  foo\n  bar baz\n  \n", trimSpaceEachLineAndPad(foo, "  "))
+}
@@ -0,0 +1,99 @@
+package config
+
+import (
+	"fmt"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func TestTransportConnect(t *testing.T) {
+	tmpl := `
+jobs:
+- name: foo
+  type: push
+  connect:
+%s
+  filesystems: {"<": true}
+  snapshotting:
+    type: manual
+  pruning:
+    keep_sender:
+    - type: last_n
+      count: 10
+    keep_receiver:
+    - type: last_n
+      count: 10
+`
+
+	mconf := func(s string) string { return fmt.Sprintf(tmpl, s) }
+
+	type test struct {
+		Name        string
+		ExpectError bool
+		Connect     string
+	}
+
+	testTable := []test{
+		{
+			Name:        "tcp_with_address_and_port",
+			ExpectError: false,
+			Connect: `
+			type: tcp
+			address: 10.0.0.23:42
+			`,
+		},
+		{
+			Name:        "tls_with_host_and_port",
+			ExpectError: false,
+			Connect: `
+			type: tls
+			address: "server1.foo.bar:8888"
+			ca:   /etc/zrepl/ca.crt
+			cert: /etc/zrepl/backupserver.fullchain
+			key:  /etc/zrepl/backupserver.key
+			server_cn: "server1"
+			`,
+		},
+		{
+			Name:        "tcp_without_port",
+			ExpectError: true,
+			Connect: `
+			type: tcp
+			address: 10.0.0.23
+			`,
+		},
+		{
+			Name:        "tls_without_port",
+			ExpectError: true,
+			Connect: `
+			type: tls
+			address: 10.0.0.23
+			ca:   /etc/zrepl/ca.crt
+			cert: /etc/zrepl/backupserver.fullchain
+			key:  /etc/zrepl/backupserver.key
+			server_cn: "server1"
+			`,
+		},
+	}
+
+	for _, tc := range testTable {
+		t.Run(tc.Name, func(t *testing.T) {
+			require.NotEmpty(t, tc.Connect)
+			connect := trimSpaceEachLineAndPad(tc.Connect, "    ")
+			conf := mconf(connect)
+			config, err := testConfig(t, conf)
+			if tc.ExpectError && err == nil {
+				t.Errorf("expected test failure, but got valid config %v", config)
+				return
+			}
+			if !tc.ExpectError && err != nil {
+				t.Errorf("not expecint test failure but got error: %s", err)
+				return
+			}
+			t.Logf("error=%v", err)
+			t.Logf("config=%v", config)
+		})
+	}
+
+}
@@ -31,10 +31,6 @@ func (i *RetentionInterval) KeepCount() int {

 const RetentionGridKeepCountAll int = -1

-type RetentionGrid struct {
-	intervals []RetentionInterval
-}
-
 func (t *RetentionIntervalList) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
 	var in string
 	if err := u(&in, true); err != nil {
@@ -0,0 +1,22 @@
+#!/usr/bin/env bash
+set -e
+
+[ "$ZREPL_DRYRUN" = "true" ] && DRYRUN="echo DRYRUN: "
+
+pre_snapshot() {
+    $DRYRUN date
+}
+
+post_snapshot() {
+    $DRYRUN date
+}
+
+case "$ZREPL_HOOKTYPE" in
+    pre_snapshot|post_snapshot)
+        "$ZREPL_HOOKTYPE"
+        ;;
+    *)
+        printf 'Unrecognized hook type: %s\n' "$ZREPL_HOOKTYPE"
+        exit 255
+        ;;
+esac
@@ -0,0 +1,14 @@
+jobs:
+- name: snapjob
+  type: snap
+  filesystems: {
+    "tank/frequently_changed<": true,
+  }
+  snapshotting:
+    type: periodic
+    interval: 2m
+    prefix: zrepl_snapjob_
+  pruning:
+    keep:
+      - type: last_n
+        count: 60
@@ -5,16 +5,20 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
-	"github.com/pkg/errors"
-	"github.com/prometheus/client_golang/prometheus"
-	"github.com/zrepl/zrepl/daemon/job"
-	"github.com/zrepl/zrepl/daemon/nethelpers"
-	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/version"
 	"io"
 	"net"
 	"net/http"
 	"time"
+
+	"github.com/pkg/errors"
+	"github.com/prometheus/client_golang/prometheus"
+
+	"github.com/zrepl/zrepl/daemon/job"
+	"github.com/zrepl/zrepl/daemon/nethelpers"
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/util/envconst"
+	"github.com/zrepl/zrepl/version"
+	"github.com/zrepl/zrepl/zfs"
 )

 type controlJob struct {
@@ -38,6 +42,8 @@ func (j *controlJob) Name() string { return jobNameControl }

 func (j *controlJob) Status() *job.Status { return &job.Status{Type: job.TypeInternal} }

+func (j *controlJob) OwnedDatasetSubtreeRoot() (p *zfs.DatasetPath, ok bool) { return nil, false }
+
 var promControl struct {
 	requestBegin    *prometheus.CounterVec
 	requestFinished *prometheus.HistogramVec
@@ -56,7 +62,7 @@ func (j *controlJob) RegisterMetrics(registerer prometheus.Registerer) {
 		Subsystem: "control",
 		Name:      "request_finished",
 		Help:      "time it took a request to finih",
-		Buckets: []float64{1e-6, 10e-6, 100e-6, 500e-6, 1e-3,10e-3, 100e-3, 200e-3,400e-3,800e-3, 1, 10, 20},
+		Buckets:   []float64{1e-6, 10e-6, 100e-6, 500e-6, 1e-3, 10e-3, 100e-3, 200e-3, 400e-3, 800e-3, 1, 10, 20},
 	}, []string{"endpoint"})
 	registerer.MustRegister(promControl.requestBegin)
 	registerer.MustRegister(promControl.requestFinished)
@@ -81,10 +87,16 @@ func (j *controlJob) Run(ctx context.Context) {
 	}

 	pprofServer := NewPProfServer(ctx)
+	if listen := envconst.String("ZREPL_DAEMON_AUTOSTART_PPROF_SERVER", ""); listen != "" {
+		pprofServer.Control(PprofServerControlMsg{
+			Run:               true,
+			HttpListenAddress: listen,
+		})
+	}

 	mux := http.NewServeMux()
 	mux.Handle(ControlJobEndpointPProf,
-		requestLogger{log: log, handler: jsonRequestResponder{func(decoder jsonDecoder) (interface{}, error) {
+		requestLogger{log: log, handler: jsonRequestResponder{log, func(decoder jsonDecoder) (interface{}, error) {
 			var msg PprofServerControlMsg
 			err := decoder(&msg)
 			if err != nil {
@@ -95,18 +107,19 @@ func (j *controlJob) Run(ctx context.Context) {
 		}}})

 	mux.Handle(ControlJobEndpointVersion,
-		requestLogger{log: log, handler: jsonResponder{func() (interface{}, error) {
+		requestLogger{log: log, handler: jsonResponder{log, func() (interface{}, error) {
 			return version.NewZreplVersionInformation(), nil
 		}}})

 	mux.Handle(ControlJobEndpointStatus,
-		requestLogger{log: log, handler: jsonResponder{func() (interface{}, error) {
+		// don't log requests to status endpoint, too spammy
+		jsonResponder{log, func() (interface{}, error) {
 			s := j.jobs.status()
 			return s, nil
-		}}})
+		}})

 	mux.Handle(ControlJobEndpointSignal,
-		requestLogger{log: log, handler: jsonRequestResponder{func(decoder jsonDecoder) (interface{}, error) {
+		requestLogger{log: log, handler: jsonRequestResponder{log, func(decoder jsonDecoder) (interface{}, error) {
 			type reqT struct {
 				Name string
 				Op   string
@@ -131,8 +144,8 @@ func (j *controlJob) Run(ctx context.Context) {
 	server := http.Server{
 		Handler: mux,
 		// control socket is local, 1s timeout should be more than sufficient, even on a loaded system
-		WriteTimeout: 1*time.Second,
-		ReadTimeout: 1*time.Second,
+		WriteTimeout: 1 * time.Second,
+		ReadTimeout:  1 * time.Second,
 	}

 outer:
@@ -147,7 +160,10 @@ outer:
 		select {
 		case <-ctx.Done():
 			log.WithError(ctx.Err()).Info("context done")
-			server.Shutdown(context.Background())
+			err := server.Shutdown(context.Background())
+			if err != nil {
+				log.WithError(err).Error("cannot shutdown server")
+			}
 			break outer
 		case err = <-served:
 			if err != nil {
@@ -161,33 +177,50 @@ outer:
 }

 type jsonResponder struct {
+	log      Logger
 	producer func() (interface{}, error)
 }

 func (j jsonResponder) ServeHTTP(w http.ResponseWriter, r *http.Request) {
+	logIoErr := func(err error) {
+		if err != nil {
+			j.log.WithError(err).Error("control handler io error")
+		}
+	}
 	res, err := j.producer()
 	if err != nil {
+		j.log.WithError(err).Error("control handler error")
 		w.WriteHeader(http.StatusInternalServerError)
-		io.WriteString(w, err.Error())
+		_, err = io.WriteString(w, err.Error())
+		logIoErr(err)
 		return
 	}
 	var buf bytes.Buffer
 	err = json.NewEncoder(&buf).Encode(res)
 	if err != nil {
+		j.log.WithError(err).Error("control handler json marshal error")
 		w.WriteHeader(http.StatusInternalServerError)
-		io.WriteString(w, err.Error())
+		_, err = io.WriteString(w, err.Error())
 	} else {
-		io.Copy(w, &buf)
+		_, err = io.Copy(w, &buf)
 	}
+	logIoErr(err)
 }

 type jsonDecoder = func(interface{}) error

 type jsonRequestResponder struct {
+	log      Logger
 	producer func(decoder jsonDecoder) (interface{}, error)
 }

 func (j jsonRequestResponder) ServeHTTP(w http.ResponseWriter, r *http.Request) {
+	logIoErr := func(err error) {
+		if err != nil {
+			j.log.WithError(err).Error("control handler io error")
+		}
+	}
+
 	var decodeError error
 	decoder := func(i interface{}) error {
 		err := json.NewDecoder(r.Body).Decode(&i)
@@ -199,22 +232,28 @@ func (j jsonRequestResponder) ServeHTTP(w http.ResponseWriter, r *http.Request)
 	//If we had a decode error ignore output of producer and return error
 	if decodeError != nil {
 		w.WriteHeader(http.StatusBadRequest)
-		io.WriteString(w, decodeError.Error())
+		_, err := io.WriteString(w, decodeError.Error())
+		logIoErr(err)
 		return
 	}
 	if producerErr != nil {
+		j.log.WithError(producerErr).Error("control handler error")
 		w.WriteHeader(http.StatusInternalServerError)
-		io.WriteString(w, producerErr.Error())
+		_, err := io.WriteString(w, producerErr.Error())
+		logIoErr(err)
 		return
 	}

 	var buf bytes.Buffer
 	encodeErr := json.NewEncoder(&buf).Encode(res)
 	if encodeErr != nil {
+		j.log.WithError(producerErr).Error("control handler json marhsal error")
 		w.WriteHeader(http.StatusInternalServerError)
-		io.WriteString(w, encodeErr.Error())
+		_, err := io.WriteString(w, encodeErr.Error())
+		logIoErr(err)
 	} else {
-		io.Copy(w, &buf)
+		_, err := io.Copy(w, &buf)
+		logIoErr(err)
 	}
 }

@@ -3,8 +3,16 @@ package daemon
 import (
 	"context"
 	"fmt"
+	"os"
+	"os/signal"
+	"strings"
+	"sync"
+	"syscall"
+	"time"
+
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
+
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/job"
 	"github.com/zrepl/zrepl/daemon/job/reset"
@@ -12,12 +20,6 @@ import (
 	"github.com/zrepl/zrepl/daemon/logging"
 	"github.com/zrepl/zrepl/logger"
 	"github.com/zrepl/zrepl/version"
-	"os"
-	"os/signal"
-	"strings"
-	"sync"
-	"syscall"
-	"time"
 )

 func Run(conf *config.Config) error {
@@ -74,12 +76,11 @@ func Run(conf *config.Config) error {
 			return errors.Errorf("unknown monitoring job #%d (type %T)", i, v)
 		}
 		if err != nil {
-			return errors.Wrapf(err,"cannot build monitorin gjob #%d", i)
+			return errors.Wrapf(err, "cannot build monitorin gjob #%d", i)
 		}
 		jobs.start(ctx, job, true)
 	}

-
 	log.Info("starting daemon")

 	// start regular jobs
@@ -117,8 +118,6 @@ func newJobs() *jobs {

 const (
 	logJobField string = "job"
-	logTaskField   string = "task"
-	logSubsysField string = "subsystem"
 )

 func (s *jobs) wait() <-chan struct{} {
@@ -2,10 +2,12 @@ package filters

 import (
 	"fmt"
+	"strings"
+
 	"github.com/pkg/errors"
+
 	"github.com/zrepl/zrepl/endpoint"
 	"github.com/zrepl/zrepl/zfs"
-	"strings"
 )

 type DatasetMapFilter struct {
@@ -0,0 +1,89 @@
+package filters
+
+import (
+	"github.com/zrepl/zrepl/zfs"
+)
+
+import "testing"
+
+func TestDatasetMapFilter(t *testing.T) {
+
+	type testCase struct {
+		name      string
+		filter    map[string]string
+		checkPass map[string]bool // each entry is checked to match the filter's `pass` return value
+	}
+
+	tcs := []testCase{
+		{
+			"default_no_match",
+			map[string]string{},
+			map[string]bool{
+				"":      false,
+				"foo":   false,
+				"zroot": false,
+			},
+		},
+		{
+			"more_specific_path_has_precedence",
+			map[string]string{
+				"tank<":         "ok",
+				"tank/tmp<":     "!",
+				"tank/home/x<":  "!",
+				"tank/home/x/1": "ok",
+			},
+			map[string]bool{
+				"zroot":         false,
+				"tank":          true,
+				"tank/tmp":      false,
+				"tank/tmp/foo":  false,
+				"tank/home/x":   false,
+				"tank/home/y":   true,
+				"tank/home/x/1": true,
+				"tank/home/x/2": false,
+			},
+		},
+		{
+			"precedence_of_specific_over_subtree_wildcard_on_same_path",
+			map[string]string{
+				"tank/home/bob":  "ok",
+				"tank/home/bob<": "!",
+			},
+			map[string]bool{
+				"tank/home/bob":           true,
+				"tank/home/bob/downloads": false,
+			},
+		},
+	}
+
+	for tc := range tcs {
+		t.Run(tcs[tc].name, func(t *testing.T) {
+			c := tcs[tc]
+			f := NewDatasetMapFilter(len(c.filter), true)
+			for p, a := range c.filter {
+				err := f.Add(p, a)
+				if err != nil {
+					t.Fatalf("incorrect filter spec: %s", err)
+				}
+			}
+			for p, checkPass := range c.checkPass {
+				zp, err := zfs.NewDatasetPath(p)
+				if err != nil {
+					t.Fatalf("incorrect path spec: %s", err)
+				}
+				pass, err := f.Filter(zp)
+				if err != nil {
+					t.Fatalf("unexpected filter error: %s", err)
+				}
+				ok := pass == checkPass
+				failstr := "OK"
+				if !ok {
+					failstr = "FAIL"
+					t.Fail()
+				}
+				t.Logf("%-40q  %5v  (exp=%v act=%v)", p, failstr, checkPass, pass)
+			}
+		})
+	}
+
+}
@@ -1,8 +1,9 @@
 package filters

 import (
-	"github.com/zrepl/zrepl/zfs"
 	"strings"
+
+	"github.com/zrepl/zrepl/zfs"
 )

 type AnyFSVFilter struct{}
@@ -17,7 +18,6 @@ func (AnyFSVFilter) Filter(t zfs.VersionType, name string) (accept bool, err err
 	return true, nil
 }

-
 type PrefixFilter struct {
 	prefix    string
 	fstype    zfs.VersionType
@@ -0,0 +1,35 @@
+// Code generated by "stringer -type=Edge"; DO NOT EDIT.
+
+package hooks
+
+import "strconv"
+
+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[Pre-1]
+	_ = x[Callback-2]
+	_ = x[Post-4]
+}
+
+const (
+	_Edge_name_0 = "PreCallback"
+	_Edge_name_1 = "Post"
+)
+
+var (
+	_Edge_index_0 = [...]uint8{0, 3, 11}
+)
+
+func (i Edge) String() string {
+	switch {
+	case 1 <= i && i <= 2:
+		i -= 1
+		return _Edge_name_0[_Edge_index_0[i]:_Edge_index_0[i+1]]
+	case i == 4:
+		return _Edge_name_1
+	default:
+		return "Edge(" + strconv.FormatInt(int64(i), 10) + ")"
+	}
+}
@@ -0,0 +1,52 @@
+package hooks
+
+import (
+	"fmt"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type List []Hook
+
+func HookFromConfig(in config.HookEnum) (Hook, error) {
+	switch v := in.Ret.(type) {
+	case *config.HookCommand:
+		return NewCommandHook(v)
+	case *config.HookPostgresCheckpoint:
+		return PgChkptHookFromConfig(v)
+	case *config.HookMySQLLockTables:
+		return MyLockTablesFromConfig(v)
+	default:
+		return nil, fmt.Errorf("unknown hook type %T", v)
+	}
+}
+
+func ListFromConfig(in *config.HookList) (r *List, err error) {
+	hl := make(List, len(*in))
+
+	for i, h := range *in {
+		hl[i], err = HookFromConfig(h)
+		if err != nil {
+			return nil, fmt.Errorf("create hook #%d: %s", i+1, err)
+		}
+	}
+
+	return &hl, nil
+}
+
+func (l List) CopyFilteredForFilesystem(fs *zfs.DatasetPath) (ret List, err error) {
+	ret = make(List, 0, len(l))
+
+	for _, h := range l {
+		var passFilesystem bool
+		if passFilesystem, err = h.Filesystems().Filter(fs); err != nil {
+			return nil, err
+		}
+		if passFilesystem {
+			ret = append(ret, h)
+		}
+	}
+
+	return ret, nil
+}
@@ -0,0 +1,34 @@
+// Package hooks implements pre- and post snapshot hooks.
+//
+// Plan is a generic executor for ExpectStepReports before and after an activity specified in a callback.
+// It provides a reporting facility that can be polled while the plan is executing to gather progress information.
+//
+// This package also provides all supported hook type implementations and abstractions around them.
+//
+// Use For Other Kinds Of ExpectStepReports
+//
+// This package REQUIRES REFACTORING before it can be used for other activities than snapshots, e.g. pre- and post-replication:
+//
+// The Hook interface requires a hook to provide a Filesystems() filter, which doesn't make sense for
+// all kinds of activities.
+//
+// The hook implementations should move out of this package.
+// However, there is a lot of tight coupling which to untangle isn't worth it ATM.
+//
+// How This Package Is Used By Package Snapper
+//
+// Deserialize a config.List using ListFromConfig().
+// Then it MUST filter the list to only contain hooks for a particular filesystem using
+// hooksList.CopyFilteredForFilesystem(fs).
+//
+// Then create a CallbackHook using NewCallbackHookForFilesystem().
+//
+// Pass all of the above to NewPlan() which provides a Report() and Run() method:
+//
+// Plan.Run(ctx context.Context,dryRun bool) executes the plan and take a context as argument that should contain a logger added using hooks.WithLogger()).
+// The value of dryRun is passed through to the hooks' Run() method.
+// Command hooks make it available in the environment variable ZREPL_DRYRUN.
+//
+// Plan.Report() can be called while Plan.Run() is executing to give an overview of plan execution progress (future use in "zrepl status").
+//
+package hooks
@@ -0,0 +1,292 @@
+package hooks
+
+import (
+	"context"
+	"fmt"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/zrepl/zrepl/zfs"
+)
+
+// Re-export type here so that
+// every file in package hooks doesn't
+// have to import github.com/zrepl/zrepl/zfs
+type Filter zfs.DatasetFilter
+
+type Hook interface {
+	Filesystems() Filter
+
+	// If true and the Pre edge invocation of Run fails, Post edge will not run and other Pre edges will not run.
+	ErrIsFatal() bool
+
+	// Run is invoked by HookPlan for a Pre edge.
+	// If HookReport.HadError() == false, the Post edge will be invoked, too.
+	Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport
+
+	String() string
+}
+
+type Phase string
+
+const (
+	PhaseSnapshot = Phase("snapshot")
+	PhaseTesting  = Phase("testing")
+)
+
+func (p Phase) String() string {
+	return string(p)
+}
+
+//go:generate stringer -type=Edge
+type Edge uint
+
+const (
+	Pre = Edge(1 << iota)
+	Callback
+	Post
+)
+
+func (e Edge) StringForPhase(phase Phase) string {
+	return fmt.Sprintf("%s_%s", e.String(), phase.String())
+}
+
+//go:generate enumer -type=StepStatus -trimprefix=Step
+type StepStatus int
+
+const (
+	StepPending StepStatus = 1 << iota
+	StepExec
+	StepOk
+	StepErr
+	StepSkippedDueToFatalErr
+	StepSkippedDueToPreErr
+)
+
+type HookReport interface {
+	String() string
+	HadError() bool
+	Error() string
+}
+
+type Step struct {
+	Hook       Hook
+	Edge       Edge
+	Status     StepStatus
+	Begin, End time.Time
+	// Report may be nil
+	// FIXME cannot serialize this for client status, but contains interesting info (like what error happened)
+	Report HookReport
+	state  map[interface{}]interface{}
+}
+
+func (s Step) String() (out string) {
+	fatal := "~"
+	if s.Hook.ErrIsFatal() && s.Edge == Pre {
+		fatal = "!"
+	}
+	runTime := "..."
+	if s.Status != StepPending {
+		t := s.End.Sub(s.Begin)
+		runTime = t.Round(time.Millisecond).String()
+	}
+	return fmt.Sprintf("[%s] [%5s] %s [%s]  %s", s.Status, runTime, fatal, s.Edge, s.Hook)
+}
+
+type Plan struct {
+	mtx sync.RWMutex
+
+	steps []*Step
+	pre   []*Step // protected by mtx
+	cb    *Step
+	post  []*Step // not reversed, i.e. entry at index i corresponds to pre-edge in pre[i]
+
+	phase Phase
+	env   Env
+}
+
+func NewPlan(hooks *List, phase Phase, cb *CallbackHook, extra Env) (*Plan, error) {
+
+	var pre, post []*Step
+	// TODO sanity check unique name of hook?
+	for _, hook := range *hooks {
+		state := make(map[interface{}]interface{})
+		preE := &Step{
+			Hook:   hook,
+			Edge:   Pre,
+			Status: StepPending,
+			state:  state,
+		}
+		pre = append(pre, preE)
+		postE := &Step{
+			Hook:   hook,
+			Edge:   Post,
+			Status: StepPending,
+			state:  state,
+		}
+		post = append(post, postE)
+	}
+
+	cbE := &Step{
+		Hook:   cb,
+		Edge:   Callback,
+		Status: StepPending,
+	}
+
+	steps := make([]*Step, 0, len(pre)+len(post)+1)
+	steps = append(steps, pre...)
+	steps = append(steps, cbE)
+	for i := len(post) - 1; i >= 0; i-- {
+		steps = append(steps, post[i])
+	}
+
+	plan := &Plan{
+		phase: phase,
+		env:   extra,
+		steps: steps,
+		pre:   pre,
+		post:  post,
+		cb:    cbE,
+	}
+
+	return plan, nil
+}
+
+type PlanReport []Step
+
+func (p *Plan) Report() PlanReport {
+	p.mtx.RLock()
+	defer p.mtx.RUnlock()
+	rep := make([]Step, len(p.steps))
+	for i := range rep {
+		rep[i] = *p.steps[i]
+	}
+	return rep
+}
+
+func (r PlanReport) HadError() bool {
+	for _, e := range r {
+		if e.Status == StepErr {
+			return true
+		}
+	}
+	return false
+}
+
+func (r PlanReport) HadFatalError() bool {
+	for _, e := range r {
+		if e.Status == StepSkippedDueToFatalErr {
+			return true
+		}
+	}
+	return false
+}
+
+func (r PlanReport) String() string {
+	stepStrings := make([]string, len(r))
+	for i, e := range r {
+		stepStrings[i] = fmt.Sprintf("%02d %s", i+1, e)
+	}
+	return strings.Join(stepStrings, "\n")
+}
+
+func (p *Plan) Run(ctx context.Context, dryRun bool) {
+	p.mtx.RLock()
+	defer p.mtx.RUnlock()
+	w := func(f func()) {
+		p.mtx.RUnlock()
+		defer p.mtx.RLock()
+		p.mtx.Lock()
+		defer p.mtx.Unlock()
+		f()
+	}
+	runHook := func(s *Step, ctx context.Context, edge Edge) HookReport {
+		w(func() { s.Status = StepExec })
+		begin := time.Now()
+		r := s.Hook.Run(ctx, edge, p.phase, dryRun, p.env, s.state)
+		end := time.Now()
+		w(func() {
+			s.Report = r
+			s.Status = StepOk
+			if r.HadError() {
+				s.Status = StepErr
+			}
+			s.Begin, s.End = begin, end
+		})
+		return r
+	}
+
+	l := getLogger(ctx)
+
+	// it's a stack, execute until we reach the end of the list (last item in)
+	// or fail inbetween
+	l.Info("run pre-edges in configuration order")
+	next := 0
+	for ; next < len(p.pre); next++ {
+		e := p.pre[next]
+		l := l.WithField("hook", e.Hook)
+		r := runHook(e, ctx, Pre)
+		if r.HadError() {
+			l.WithError(r).Error("hook invocation failed for pre-edge")
+			if e.Hook.ErrIsFatal() {
+				l.Error("the hook run was aborted due to a fatal error in this hook")
+				break
+			}
+		}
+	}
+
+	hadFatalErr := next != len(p.pre)
+	if hadFatalErr {
+		l.Error("fatal error in a pre-snapshot hook invocation")
+		l.Error("no snapshot will be taken")
+		l.Error("only running post-edges for successful pre-edges")
+		w(func() {
+			p.post[next].Status = StepSkippedDueToFatalErr
+			for i := next + 1; i < len(p.pre); i++ {
+				p.pre[i].Status = StepSkippedDueToFatalErr
+				p.post[i].Status = StepSkippedDueToFatalErr
+			}
+			p.cb.Status = StepSkippedDueToFatalErr
+		})
+		return
+	}
+
+	l.Info("running callback")
+	cbR := runHook(p.cb, ctx, Callback)
+	if cbR.HadError() {
+		l.WithError(cbR).Error("callback failed")
+	}
+
+	l.Info("run post-edges for successful pre-edges in reverse configuration order")
+
+	// the constructor produces pre and post entries
+	// post is NOT reversed
+	next-- // now at index of last executed pre-edge
+	for ; next >= 0; next-- {
+		e := p.post[next]
+		l := l.WithField("hook", e.Hook)
+
+		if p.pre[next].Status != StepOk {
+			if p.pre[next].Status != StepErr {
+				panic(fmt.Sprintf("expecting a pre-edge hook report to be either Ok or Err, got %s", p.pre[next].Status))
+			}
+			l.Info("skip post-edge because pre-edge failed")
+			w(func() {
+				e.Status = StepSkippedDueToPreErr
+			})
+			continue
+		}
+
+		report := runHook(e, ctx, Post)
+
+		if report.HadError() {
+			l.WithError(report).Error("hook invocation failed for post-edge")
+			l.Error("subsequent post-edges run regardless of this post-edge failure")
+		}
+
+		// ErrIsFatal is only relevant for Pre
+
+	}
+
+}
@@ -0,0 +1,103 @@
+package hooks
+
+import (
+	"bufio"
+	"bytes"
+	"context"
+	"sync"
+
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/util/envconst"
+)
+
+type contextKey int
+
+const (
+	contextKeyLog contextKey = 0
+)
+
+type Logger = logger.Logger
+
+func WithLogger(ctx context.Context, log Logger) context.Context {
+	return context.WithValue(ctx, contextKeyLog, log)
+}
+
+func GetLogger(ctx context.Context) Logger { return getLogger(ctx) }
+
+func getLogger(ctx context.Context) Logger {
+	if log, ok := ctx.Value(contextKeyLog).(Logger); ok {
+		return log
+	}
+	return logger.NewNullLogger()
+}
+
+const MAX_HOOK_LOG_SIZE_DEFAULT int = 1 << 20
+
+type logWriter struct {
+	/*
+		Mutex prevents:
+			concurrent writes to buf, scanner in Write([]byte)
+			data race on scanner vs Write([]byte)
+				and concurrent write to buf (call to buf.Reset())
+				in Close()
+
+		(Also, Close() should generally block until any Write() call completes.)
+	*/
+	mtx     *sync.Mutex
+	buf     bytes.Buffer
+	scanner *bufio.Scanner
+	logger  Logger
+	level   logger.Level
+	field   string
+}
+
+func NewLogWriter(mtx *sync.Mutex, logger Logger, level logger.Level, field string) *logWriter {
+	w := new(logWriter)
+	w.mtx = mtx
+	w.scanner = bufio.NewScanner(&w.buf)
+	w.logger = logger
+	w.level = level
+	w.field = field
+	return w
+}
+
+func (w *logWriter) log(line string) {
+	w.logger.WithField(w.field, line).Log(w.level, "hook output")
+}
+
+func (w *logWriter) logUnreadBytes() error {
+	for w.scanner.Scan() {
+		w.log(w.scanner.Text())
+	}
+	if w.buf.Cap() > envconst.Int("ZREPL_MAX_HOOK_LOG_SIZE", MAX_HOOK_LOG_SIZE_DEFAULT) {
+		w.buf.Reset()
+	}
+
+	return nil
+}
+
+func (w *logWriter) Write(in []byte) (int, error) {
+	w.mtx.Lock()
+	defer w.mtx.Unlock()
+
+	n, bufErr := w.buf.Write(in)
+	if bufErr != nil {
+		return n, bufErr
+	}
+
+	err := w.logUnreadBytes()
+	if err != nil {
+		return n, err
+	}
+	// Always reset the scanner for the next Write
+	w.scanner = bufio.NewScanner(&w.buf)
+
+	return n, nil
+}
+
+func (w *logWriter) Close() (err error) {
+	w.mtx.Lock()
+	defer w.mtx.Unlock()
+
+	return w.logUnreadBytes()
+}
@@ -0,0 +1,65 @@
+package hooks
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type HookJobCallback func(ctx context.Context) error
+
+type CallbackHook struct {
+	cb            HookJobCallback
+	filter        Filter
+	displayString string
+}
+
+func NewCallbackHookForFilesystem(displayString string, fs *zfs.DatasetPath, cb HookJobCallback) *CallbackHook {
+	filter, _ := filters.DatasetMapFilterFromConfig(map[string]bool{fs.ToString(): true})
+	return NewCallbackHook(displayString, cb, filter)
+}
+
+func NewCallbackHook(displayString string, cb HookJobCallback, filter Filter) *CallbackHook {
+	return &CallbackHook{
+		cb:            cb,
+		filter:        filter,
+		displayString: displayString,
+	}
+}
+
+func (h *CallbackHook) Filesystems() Filter {
+	return h.filter
+}
+
+func (h *CallbackHook) ErrIsFatal() bool {
+	return false // callback is by definition
+}
+
+func (h *CallbackHook) String() string {
+	return h.displayString
+}
+
+type CallbackHookReport struct {
+	Name string
+	Err  error
+}
+
+func (r *CallbackHookReport) String() string {
+	if r.HadError() {
+		return r.Error()
+	}
+	return r.Name
+}
+
+func (r *CallbackHookReport) HadError() bool { return r.Err != nil }
+
+func (r *CallbackHookReport) Error() string {
+	return fmt.Sprintf("%s error: %s", r.Name, r.Err)
+}
+
+func (h *CallbackHook) Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport {
+	err := h.cb(ctx)
+	return &CallbackHookReport{h.displayString, err}
+}
@@ -0,0 +1,192 @@
+package hooks
+
+import (
+	"context"
+	"fmt"
+	"io"
+	"math"
+	"os"
+	"os/exec"
+	"sort"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/util/circlog"
+	"github.com/zrepl/zrepl/util/envconst"
+)
+
+type HookEnvVar string
+
+const (
+	EnvType     HookEnvVar = "ZREPL_HOOKTYPE"
+	EnvDryRun   HookEnvVar = "ZREPL_DRYRUN"
+	EnvFS       HookEnvVar = "ZREPL_FS"
+	EnvSnapshot HookEnvVar = "ZREPL_SNAPNAME"
+	EnvTimeout  HookEnvVar = "ZREPL_TIMEOUT"
+)
+
+type Env map[HookEnvVar]string
+
+func NewHookEnv(edge Edge, phase Phase, dryRun bool, timeout time.Duration, extra Env) Env {
+	r := Env{
+		EnvTimeout: fmt.Sprintf("%.f", math.Floor(timeout.Seconds())),
+	}
+
+	edgeString := edge.StringForPhase(phase)
+	r[EnvType] = strings.ToLower(edgeString)
+
+	var dryRunString string
+	if dryRun {
+		dryRunString = "true"
+	} else {
+		dryRunString = ""
+	}
+	r[EnvDryRun] = dryRunString
+
+	for k, v := range extra {
+		r[k] = v
+	}
+
+	return r
+}
+
+type CommandHook struct {
+	edge       Edge
+	filter     Filter
+	errIsFatal bool
+	command    string
+	timeout    time.Duration
+}
+
+type CommandHookReport struct {
+	Command                      string
+	Args                         []string // currently always empty
+	Env                          Env
+	Err                          error
+	CapturedStdoutStderrCombined []byte
+}
+
+func (r *CommandHookReport) String() string {
+	// Reproduces a POSIX shell-compatible command line
+	var cmdLine strings.Builder
+	sep := ""
+
+	// Make sure environment variables are always
+	// printed in the same order
+	var hookEnvKeys []HookEnvVar
+	for k := range r.Env {
+		hookEnvKeys = append(hookEnvKeys, k)
+	}
+	sort.Slice(hookEnvKeys, func(i, j int) bool { return string(hookEnvKeys[i]) < string(hookEnvKeys[j]) })
+
+	for _, k := range hookEnvKeys {
+		cmdLine.WriteString(fmt.Sprintf("%s%s='%s'", sep, k, r.Env[k]))
+		sep = " "
+	}
+
+	cmdLine.WriteString(fmt.Sprintf("%s%s", sep, r.Command))
+	for _, a := range r.Args {
+		cmdLine.WriteString(fmt.Sprintf("%s'%s'", sep, a))
+	}
+
+	return fmt.Sprintf("command hook invocation: \"%s\"", cmdLine.String()) // no %q to make copy-pastable
+}
+func (r *CommandHookReport) Error() string {
+	if r.Err == nil {
+		return ""
+	}
+	return fmt.Sprintf("%s FAILED with error: %s", r.String(), r.Err)
+}
+
+func (r *CommandHookReport) HadError() bool {
+	return r.Err != nil
+}
+
+func NewCommandHook(in *config.HookCommand) (r *CommandHook, err error) {
+	r = &CommandHook{
+		errIsFatal: in.ErrIsFatal,
+		command:    in.Path,
+		timeout:    in.Timeout,
+	}
+
+	r.filter, err = filters.DatasetMapFilterFromConfig(in.Filesystems)
+	if err != nil {
+		return nil, fmt.Errorf("cannot parse filesystem filter: %s", err)
+	}
+
+	r.edge = Pre | Post
+
+	return r, nil
+}
+
+func (h *CommandHook) Filesystems() Filter {
+	return h.filter
+}
+
+func (h *CommandHook) ErrIsFatal() bool {
+	return h.errIsFatal
+}
+
+func (h *CommandHook) String() string {
+	return h.command
+}
+
+func (h *CommandHook) Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport {
+	l := getLogger(ctx).WithField("command", h.command)
+
+	cmdCtx, cancel := context.WithTimeout(ctx, h.timeout)
+	defer cancel()
+
+	cmdExec := exec.CommandContext(cmdCtx, h.command)
+
+	hookEnv := NewHookEnv(edge, phase, dryRun, h.timeout, extra)
+	cmdEnv := os.Environ()
+	for k, v := range hookEnv {
+		cmdEnv = append(cmdEnv, fmt.Sprintf("%s=%s", k, v))
+	}
+	cmdExec.Env = cmdEnv
+
+	var scanMutex sync.Mutex
+	combinedOutput, err := circlog.NewCircularLog(envconst.Int("ZREPL_MAX_HOOK_LOG_SIZE", MAX_HOOK_LOG_SIZE_DEFAULT))
+	if err != nil {
+		return &CommandHookReport{Err: err}
+	}
+	logErrWriter := NewLogWriter(&scanMutex, l, logger.Warn, "stderr")
+	logOutWriter := NewLogWriter(&scanMutex, l, logger.Info, "stdout")
+	defer logErrWriter.Close()
+	defer logOutWriter.Close()
+
+	cmdExec.Stderr = io.MultiWriter(logErrWriter, combinedOutput)
+	cmdExec.Stdout = io.MultiWriter(logOutWriter, combinedOutput)
+
+	report := &CommandHookReport{
+		Command: h.command,
+		Env:     hookEnv,
+		// no report.Args
+	}
+
+	err = cmdExec.Start()
+	if err != nil {
+		report.Err = err
+		return report
+	}
+
+	err = cmdExec.Wait()
+	combinedOutputBytes := combinedOutput.Bytes()
+	report.CapturedStdoutStderrCombined = make([]byte, len(combinedOutputBytes))
+	copy(report.CapturedStdoutStderrCombined, combinedOutputBytes)
+	if err != nil {
+		if cmdCtx.Err() == context.DeadlineExceeded {
+			report.Err = fmt.Errorf("timed out after %s: %s", h.timeout, err)
+			return report
+		}
+		report.Err = err
+		return report
+	}
+
+	return report
+}
@@ -0,0 +1,145 @@
+package hooks
+
+import (
+	"context"
+	"database/sql"
+	"fmt"
+	"strings"
+
+	sqldriver "database/sql/driver"
+
+	"github.com/go-sql-driver/mysql"
+
+	"github.com/pkg/errors"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+// https://dev.mysql.com/doc/mysql-backup-excerpt/5.7/en/backup-methods.html
+//
+// Making Backups Using a File System Snapshot:
+//
+//  If you are using a Veritas file system, you can make a backup like this:
+//
+//    	From a client program, execute FLUSH TABLES WITH READ LOCK.
+//    	From another shell, execute mount vxfs snapshot.
+//    	From the first client, execute UNLOCK TABLES.
+//    	Copy files from the snapshot.
+//    	Unmount the snapshot.
+//
+//	Similar snapshot capabilities may be available in other file systems, such as LVM or ZFS.
+type MySQLLockTables struct {
+	errIsFatal  bool
+	connector   sqldriver.Connector
+	filesystems Filter
+}
+
+type myLockTablesStateKey int
+
+const (
+	myLockTablesConnection myLockTablesStateKey = 1 + iota
+)
+
+func MyLockTablesFromConfig(in *config.HookMySQLLockTables) (*MySQLLockTables, error) {
+	conf, err := mysql.ParseDSN(in.DSN)
+	if err != nil {
+		return nil, errors.Wrap(err, "`dsn` invalid")
+	}
+	cn, err := mysql.NewConnector(conf)
+	if err != nil {
+		return nil, errors.Wrap(err, "`connect` invalid")
+	}
+
+	filesystems, err := filters.DatasetMapFilterFromConfig(in.Filesystems)
+	if err != nil {
+		return nil, errors.Wrap(err, "`filesystems` invalid")
+	}
+
+	return &MySQLLockTables{
+		in.ErrIsFatal,
+		cn,
+		filesystems,
+	}, nil
+}
+
+func (h *MySQLLockTables) ErrIsFatal() bool    { return h.errIsFatal }
+func (h *MySQLLockTables) Filesystems() Filter { return h.filesystems }
+func (h *MySQLLockTables) String() string      { return "MySQL FLUSH TABLES WITH READ LOCK" }
+
+type MyLockTablesReport struct {
+	What string
+	Err  error
+}
+
+func (r *MyLockTablesReport) HadError() bool { return r.Err != nil }
+func (r *MyLockTablesReport) Error() string  { return r.String() }
+func (r *MyLockTablesReport) String() string {
+	var s strings.Builder
+	s.WriteString(r.What)
+	if r.Err != nil {
+		fmt.Fprintf(&s, ": %s", r.Err)
+	}
+	return s.String()
+}
+
+func (h *MySQLLockTables) Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport {
+	fs, ok := extra[EnvFS]
+	if !ok {
+		panic(extra)
+	}
+	dp, err := zfs.NewDatasetPath(fs)
+	if err != nil {
+		panic(err)
+	}
+	if pass, err := h.filesystems.Filter(dp); err != nil {
+		return &MyLockTablesReport{What: "filesystem filter", Err: err}
+	} else if !pass {
+		getLogger(ctx).Debug("filesystem does not match filter, skipping")
+		return &MyLockTablesReport{What: "filesystem filter skipped this filesystem", Err: nil}
+	}
+
+	switch edge {
+	case Pre:
+		err := h.doRunPre(ctx, dp, dryRun, state)
+		return &MyLockTablesReport{"FLUSH TABLES WITH READ LOCK", err}
+	case Post:
+		err := h.doRunPost(ctx, dp, dryRun, state)
+		return &MyLockTablesReport{"UNLOCK TABLES", err}
+	}
+	return &MyLockTablesReport{What: "skipped this edge", Err: nil}
+}
+
+func (h *MySQLLockTables) doRunPre(ctx context.Context, fs *zfs.DatasetPath, dry bool, state map[interface{}]interface{}) (err error) {
+	db := sql.OpenDB(h.connector)
+	defer func(err *error) {
+		if *err != nil {
+			db.Close()
+		}
+	}(&err)
+
+	getLogger(ctx).Debug("do FLUSH TABLES WITH READ LOCK")
+	_, err = db.ExecContext(ctx, "FLUSH TABLES WITH READ LOCK")
+	if err != nil {
+		return
+	}
+
+	state[myLockTablesConnection] = db
+
+	return nil
+}
+
+func (h *MySQLLockTables) doRunPost(ctx context.Context, fs *zfs.DatasetPath, dry bool, state map[interface{}]interface{}) error {
+
+	db := state[myLockTablesConnection].(*sql.DB)
+	defer db.Close()
+
+	getLogger(ctx).Debug("do UNLOCK TABLES")
+	_, err := db.ExecContext(ctx, "UNLOCK TABLES")
+	if err != nil {
+		return err
+	}
+
+	return nil
+}
@@ -0,0 +1,98 @@
+package hooks
+
+import (
+	"context"
+	"database/sql"
+	"fmt"
+	"math"
+	"time"
+
+	"github.com/lib/pq"
+	"github.com/pkg/errors"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type PgChkptHook struct {
+	errIsFatal  bool
+	connector   *pq.Connector
+	filesystems Filter
+}
+
+func PgChkptHookFromConfig(in *config.HookPostgresCheckpoint) (*PgChkptHook, error) {
+	filesystems, err := filters.DatasetMapFilterFromConfig(in.Filesystems)
+	if err != nil {
+		return nil, errors.Wrap(err, "`filesystems` invalid")
+	}
+	cn, err := pq.NewConnector(in.DSN)
+	if err != nil {
+		return nil, errors.Wrap(err, "`dsn` invalid")
+	}
+
+	return &PgChkptHook{
+		in.ErrIsFatal,
+		cn,
+		filesystems,
+	}, nil
+}
+
+func (h *PgChkptHook) ErrIsFatal() bool    { return h.errIsFatal }
+func (h *PgChkptHook) Filesystems() Filter { return h.filesystems }
+func (h *PgChkptHook) String() string      { return "postgres checkpoint" }
+
+type PgChkptHookReport struct{ Err error }
+
+func (r *PgChkptHookReport) HadError() bool { return r.Err != nil }
+func (r *PgChkptHookReport) Error() string  { return r.Err.Error() }
+func (r *PgChkptHookReport) String() string {
+	if r.Err != nil {
+		return fmt.Sprintf("postgres CHECKPOINT failed: %s", r.Err)
+	} else {
+		return "postgres CHECKPOINT completed"
+	}
+}
+
+func (h *PgChkptHook) Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport {
+	if edge != Pre {
+		return &PgChkptHookReport{nil}
+	}
+	fs, ok := extra[EnvFS]
+	if !ok {
+		panic(extra)
+	}
+	dp, err := zfs.NewDatasetPath(fs)
+	if err != nil {
+		panic(err)
+	}
+	err = h.doRunPre(ctx, dp, dryRun)
+	return &PgChkptHookReport{err}
+}
+
+func (h *PgChkptHook) doRunPre(ctx context.Context, fs *zfs.DatasetPath, dry bool) error {
+
+	if pass, err := h.filesystems.Filter(fs); err != nil || !pass {
+		getLogger(ctx).Debug("filesystem does not match filter, skipping")
+		return err
+	}
+
+	db := sql.OpenDB(h.connector)
+	defer db.Close()
+	dl, ok := ctx.Deadline()
+	if ok {
+		timeout := uint64(math.Floor(time.Until(dl).Seconds() * 1000)) // TODO go1.13 milliseconds
+		getLogger(ctx).WithField("statement_timeout", timeout).Debug("setting statement timeout for CHECKPOINT")
+		_, err := db.ExecContext(ctx, "SET statement_timeout TO ?", timeout)
+		if err != nil {
+			return err
+		}
+	}
+	if dry {
+		getLogger(ctx).Debug("dry-run - use ping instead of CHECKPOINT")
+		return db.PingContext(ctx)
+	}
+	getLogger(ctx).Info("execute CHECKPOINT command")
+	_, err := db.ExecContext(ctx, "CHECKPOINT")
+	return err
+}
@@ -0,0 +1,484 @@
+package hooks_test
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"os"
+	"regexp"
+	"testing"
+	"text/template"
+
+	"github.com/stretchr/testify/require"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/hooks"
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type comparisonAssertionFunc func(require.TestingT, interface{}, interface{}, ...interface{})
+type valueAssertionFunc func(require.TestingT, interface{}, ...interface{})
+
+type expectStep struct {
+	ExpectedEdge hooks.Edge
+	ExpectStatus hooks.StepStatus
+	OutputTest   valueAssertionFunc
+	ErrorTest    valueAssertionFunc
+}
+
+type testCase struct {
+	Name           string
+	Config         []string
+	IsSlow         bool
+	SuppressOutput bool
+
+	ExpectCallbackSkipped bool
+	ExpectHadFatalErr     bool
+	ExpectHadError        bool
+	ExpectStepReports     []expectStep
+}
+
+func curryLeft(f comparisonAssertionFunc, expected interface{}) valueAssertionFunc {
+	return curry(f, expected, false)
+}
+
+func curryRight(f comparisonAssertionFunc, expected interface{}) valueAssertionFunc {
+	return curry(f, expected, true)
+}
+
+func curry(f comparisonAssertionFunc, expected interface{}, right bool) (ret valueAssertionFunc) {
+	ret = func(t require.TestingT, s interface{}, v ...interface{}) {
+		var x interface{}
+		var y interface{}
+		if right {
+			x = s
+			y = expected
+		} else {
+			x = expected
+			y = s
+		}
+
+		if len(v) > 0 {
+			f(t, x, y, v)
+		} else {
+			f(t, x, y)
+		}
+	}
+	return
+}
+
+func TestHooks(t *testing.T) {
+	testFSName := "testpool/testdataset"
+	testSnapshotName := "testsnap"
+
+	tmpl, err := template.New("TestHooks").Parse(`
+jobs:
+- name: TestHooks
+  type: snap
+  filesystems: {"<": true}
+  snapshotting:
+    type: periodic
+    interval: 1m
+    prefix: zrepl_snapjob_
+    hooks:
+    {{- template "List" . }}
+  pruning:
+    keep:
+      - type: last_n
+        count: 10
+`)
+	if err != nil {
+		panic(err)
+	}
+
+	regexpTest := func(s string) valueAssertionFunc {
+		return curryLeft(require.Regexp, regexp.MustCompile(s))
+	}
+
+	containsTest := func(s string) valueAssertionFunc {
+		return curryRight(require.Contains, s)
+	}
+
+	testTable := []testCase{
+		testCase{
+			Name: "no_hooks",
+			ExpectStepReports: []expectStep{
+				{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+			},
+		},
+		testCase{
+			Name:           "timeout",
+			IsSlow:         true,
+			ExpectHadError: true,
+			Config:         []string{`{type: command, path: {{.WorkDir}}/test/test-timeout.sh, timeout: 2s}`},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s ZREPL_TIMEOUT=2", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest(`timed out after 2(.\d+)?s`),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToPreErr,
+				},
+			},
+		},
+		testCase{
+			Name:           "check_env",
+			Config:         []string{`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`},
+			ExpectHadError: false,
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST post_testing %s@%s", testFSName, testSnapshotName)),
+				},
+			},
+		},
+
+		testCase{
+			Name:                  "nonfatal_pre_error_continues",
+			ExpectCallbackSkipped: false,
+			ExpectHadError:        true,
+			Config:                []string{`{type: command, path: {{.WorkDir}}/test/test-error.sh}`},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR pre_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToPreErr, // post-edge is not executed for failing pre-edge
+				},
+			},
+		},
+
+		testCase{
+			Name:                  "pre_error_fatal_skips_subsequent_pre_edges_and_callback_and_its_post_edge_and_post_edges",
+			ExpectCallbackSkipped: true,
+			ExpectHadFatalErr:     true,
+			ExpectHadError:        true,
+			Config: []string{
+				`{type: command, path: {{.WorkDir}}/test/test-error.sh, err_is_fatal: true}`,
+				`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`,
+			},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR pre_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepSkippedDueToFatalErr,
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepSkippedDueToFatalErr},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToFatalErr,
+				},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToFatalErr,
+				},
+			},
+		},
+
+		testCase{
+			Name:              "post_error_fails_are_ignored_even_if_fatal",
+			ExpectHadFatalErr: false, // only occurs during Post, so it's not a fatal error
+			ExpectHadError:    true,
+			Config: []string{
+				`{type: command, path: {{.WorkDir}}/test/test-post-error.sh, err_is_fatal: true}`,
+				`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`,
+			},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					// No-action run of test-post-error.sh
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   require.Empty,
+				},
+				expectStep{
+					// Pre run of test-report-env.sh
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST post_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR post_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+			},
+		},
+
+		testCase{
+			Name:   "cleanup_check_env",
+			Config: []string{`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST post_testing %s@%s", testFSName, testSnapshotName)),
+				},
+			},
+		},
+
+		testCase{
+			Name:              "pre_error_cancels_post",
+			Config:            []string{`{type: command, path: {{.WorkDir}}/test/test-pre-error-post-ok.sh}`},
+			ExpectHadError:    true,
+			ExpectHadFatalErr: false,
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR pre_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToPreErr,
+				},
+			},
+		},
+
+		testCase{
+			Name: "pre_error_does_not_cancel_other_posts_but_itself",
+			Config: []string{
+				`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`,
+				`{type: command, path: {{.WorkDir}}/test/test-pre-error-post-ok.sh}`,
+			},
+			ExpectHadError:    true,
+			ExpectHadFatalErr: false,
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR pre_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToPreErr,
+				},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST post_testing %s@%s", testFSName, testSnapshotName)),
+				},
+			},
+		},
+
+		testCase{
+			Name:              "exceed_buffer_limit",
+			SuppressOutput:    true,
+			Config:            []string{`{type: command, path: {{.WorkDir}}/test/test-large-stdout.sh}`},
+			ExpectHadError:    false,
+			ExpectHadFatalErr: false,
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest: func(t require.TestingT, s interface{}, v ...interface{}) {
+						require.Len(t, s, 1<<20)
+					},
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					// No-action run of above hook
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   require.Empty,
+				},
+			},
+		},
+
+		/*
+			Following not intended to test functionality of
+			filter package. Only to demonstrate that hook
+			filters are being applied. The following should
+			result in NO hooks running. If it does run, a
+			fatal hooks.RunReport will be returned.
+		*/
+		testCase{
+			Name: "exclude_all_filesystems",
+			Config: []string{
+				`{type: command, path: {{.WorkDir}}/test/test-error.sh, err_is_fatal: true, filesystems: {"<": false}}`,
+			},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Callback,
+					ExpectStatus: hooks.StepOk,
+				},
+			},
+		},
+	}
+
+	parseHookConfig := func(t *testing.T, in string) *config.Config {
+		t.Helper()
+		conf, err := config.ParseConfigBytes([]byte(in))
+		require.NoError(t, err)
+		require.NotNil(t, conf)
+		return conf
+	}
+
+	fillHooks := func(tt *testCase) string {
+		// make hook path absolute
+		cwd, err := os.Getwd()
+		if err != nil {
+			panic("os.Getwd() failed")
+		}
+		var hooksTmpl string = "\n"
+		for _, l := range tt.Config {
+			hooksTmpl += fmt.Sprintf("    - %s\n", l)
+		}
+		tmpl.New("List").Parse(hooksTmpl)
+
+		var outBytes bytes.Buffer
+		data := struct {
+			WorkDir string
+		}{
+			WorkDir: cwd,
+		}
+		if err := tmpl.Execute(&outBytes, data); err != nil {
+			panic(err)
+		}
+
+		return outBytes.String()
+	}
+
+	var c *config.Config
+	fs, err := zfs.NewDatasetPath(testFSName)
+	require.NoError(t, err)
+
+	log := logger.NewTestLogger(t)
+
+	var cbReached bool
+	cb := hooks.NewCallbackHookForFilesystem("testcallback", fs, func(_ context.Context) error {
+		cbReached = true
+		return nil
+	})
+
+	hookEnvExtra := hooks.Env{
+		hooks.EnvFS:       fs.ToString(),
+		hooks.EnvSnapshot: testSnapshotName,
+	}
+
+	for _, tt := range testTable {
+		if testing.Short() && tt.IsSlow {
+			continue
+		}
+
+		t.Run(tt.Name, func(t *testing.T) {
+			c = parseHookConfig(t, fillHooks(&tt))
+			snp := c.Jobs[0].Ret.(*config.SnapJob).Snapshotting.Ret.(*config.SnapshottingPeriodic)
+			hookList, err := hooks.ListFromConfig(&snp.Hooks)
+			require.NoError(t, err)
+
+			filteredHooks, err := hookList.CopyFilteredForFilesystem(fs)
+			require.NoError(t, err)
+			plan, err := hooks.NewPlan(&filteredHooks, hooks.PhaseTesting, cb, hookEnvExtra)
+			require.NoError(t, err)
+			t.Logf("REPORT PRE EXECUTION:\n%s", plan.Report())
+
+			cbReached = false
+
+			ctx := context.Background()
+			if testing.Verbose() && !tt.SuppressOutput {
+				ctx = hooks.WithLogger(ctx, log)
+			}
+			plan.Run(ctx, false)
+			report := plan.Report()
+
+			t.Logf("REPORT POST EXECUTION:\n%s", report)
+
+			/*
+			 * TEST ASSERTIONS
+			 */
+
+			t.Logf("len(runReports)=%v", len(report))
+			t.Logf("len(tt.ExpectStepReports)=%v", len(tt.ExpectStepReports))
+			require.Equal(t, len(tt.ExpectStepReports), len(report), "ExpectStepReports must be same length as expected number of hook runs, excluding possible Callback")
+
+			// Check if callback ran, when required
+			if tt.ExpectCallbackSkipped {
+				require.False(t, cbReached, "callback ran but should not have run")
+			} else {
+				require.True(t, cbReached, "callback should have run but did not")
+			}
+
+			// Check if a fatal run error occurred and was expected
+			require.Equal(t, tt.ExpectHadFatalErr, report.HadFatalError(), "non-matching HadFatalError")
+			require.Equal(t, tt.ExpectHadError, report.HadError(), "non-matching HadError")
+
+			if tt.ExpectHadFatalErr {
+				require.True(t, tt.ExpectHadError, "ExpectHadFatalErr implies ExpectHadError")
+			}
+			if !tt.ExpectHadError {
+				require.False(t, tt.ExpectHadFatalErr, "!ExpectHadError implies !ExpectHadFatalErr")
+			}
+
+			// Iterate through each expected hook run
+			for i, hook := range tt.ExpectStepReports {
+				t.Logf("expecting report conforming to %v", hook)
+
+				exp, act := hook.ExpectStatus, report[i].Status
+				require.Equal(t, exp, act, "%s != %s", exp, act)
+
+				// Check for required ExpectedEdge
+				require.NotZero(t, hook.ExpectedEdge, "each hook must have an ExpectedEdge")
+				require.Equal(t, hook.ExpectedEdge, report[i].Edge,
+					"incorrect edge: expected %q, actual %q", hook.ExpectedEdge.String(), report[i].Edge.String(),
+				)
+
+				// Check for expected output
+				if hook.OutputTest != nil {
+					require.IsType(t, (*hooks.CommandHookReport)(nil), report[i].Report)
+					chr := report[i].Report.(*hooks.CommandHookReport)
+					hook.OutputTest(t, string(chr.CapturedStdoutStderrCombined))
+				}
+
+				// Check for expected errors
+				if hook.ErrorTest != nil {
+					hook.ErrorTest(t, string(report[i].Report.Error()))
+				}
+			}
+
+		})
+	}
+}
@@ -0,0 +1,77 @@
+// Code generated by "enumer -type=StepStatus -trimprefix=Step"; DO NOT EDIT.
+
+//
+package hooks
+
+import (
+	"fmt"
+)
+
+const (
+	_StepStatusName_0 = "PendingExec"
+	_StepStatusName_1 = "Ok"
+	_StepStatusName_2 = "Err"
+	_StepStatusName_3 = "SkippedDueToFatalErr"
+	_StepStatusName_4 = "SkippedDueToPreErr"
+)
+
+var (
+	_StepStatusIndex_0 = [...]uint8{0, 7, 11}
+	_StepStatusIndex_1 = [...]uint8{0, 2}
+	_StepStatusIndex_2 = [...]uint8{0, 3}
+	_StepStatusIndex_3 = [...]uint8{0, 20}
+	_StepStatusIndex_4 = [...]uint8{0, 18}
+)
+
+func (i StepStatus) String() string {
+	switch {
+	case 1 <= i && i <= 2:
+		i -= 1
+		return _StepStatusName_0[_StepStatusIndex_0[i]:_StepStatusIndex_0[i+1]]
+	case i == 4:
+		return _StepStatusName_1
+	case i == 8:
+		return _StepStatusName_2
+	case i == 16:
+		return _StepStatusName_3
+	case i == 32:
+		return _StepStatusName_4
+	default:
+		return fmt.Sprintf("StepStatus(%d)", i)
+	}
+}
+
+var _StepStatusValues = []StepStatus{1, 2, 4, 8, 16, 32}
+
+var _StepStatusNameToValueMap = map[string]StepStatus{
+	_StepStatusName_0[0:7]:  1,
+	_StepStatusName_0[7:11]: 2,
+	_StepStatusName_1[0:2]:  4,
+	_StepStatusName_2[0:3]:  8,
+	_StepStatusName_3[0:20]: 16,
+	_StepStatusName_4[0:18]: 32,
+}
+
+// StepStatusString retrieves an enum value from the enum constants string name.
+// Throws an error if the param is not part of the enum.
+func StepStatusString(s string) (StepStatus, error) {
+	if val, ok := _StepStatusNameToValueMap[s]; ok {
+		return val, nil
+	}
+	return 0, fmt.Errorf("%s does not belong to StepStatus values", s)
+}
+
+// StepStatusValues returns all values of the enum
+func StepStatusValues() []StepStatus {
+	return _StepStatusValues
+}
+
+// IsAStepStatus returns "true" if the value is listed in the enum definition. "false" otherwise
+func (i StepStatus) IsAStepStatus() bool {
+	for _, v := range _StepStatusValues {
+		if i == v {
+			return true
+		}
+	}
+	return false
+}
@@ -0,0 +1,5 @@
+#!/bin/sh -eu
+
+>&2 echo "TEST ERROR $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME"
+
+exit 1
@@ -0,0 +1,18 @@
+#!/bin/sh -eu
+
+# Exceed default hook log size of 1<<20 bytes by double
+# The Go test should fail if buffer size exceeds 1MB
+
+pre_testing() {
+	ln=0
+	while :; do printf '%06d: 012345678901234567890123456789012345678901234567890123456789012345678901\n' "$ln"; ln="$(($ln + 1))"; done \
+	| head -c$(( 2 << 20 ))
+}
+
+case "$ZREPL_HOOKTYPE" in
+	pre_testing)
+		"$ZREPL_HOOKTYPE";;
+	*)
+		# Not handled by this script
+		exit 0;;
+esac
@@ -0,0 +1,15 @@
+#!/bin/sh -eu
+
+post_testing() {
+	>&2 echo "TEST ERROR $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME"
+
+	exit 1
+}
+
+case "$ZREPL_HOOKTYPE" in
+	post_testing)
+		"$ZREPL_HOOKTYPE";;
+	*)
+		# Not handled by this script
+		exit 0;;
+esac
@@ -0,0 +1,11 @@
+#!/bin/sh -eu
+
+if [ "$ZREPL_HOOKTYPE" = "pre_testing" ]; then
+	>&2 echo "TEST ERROR $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME"
+	exit 1
+elif [ "$ZREPL_HOOKTYPE" = "post_testing" ]; then
+	echo "TEST $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME"
+else
+	printf "Unknown hook type: %s" "$ZREPL_HOOKTYPE" 
+	exit 255
+fi
@@ -0,0 +1,3 @@
+#!/bin/sh -eu
+
+echo "TEST $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME ${ZREPL_DRYRUN:+(dry run)}"
@@ -0,0 +1,5 @@
+#!/bin/sh -eu
+
+echo "TEST $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME ZREPL_TIMEOUT=$ZREPL_TIMEOUT"
+
+exec sleep $(($ZREPL_TIMEOUT + 1))
@@ -2,34 +2,37 @@ package job

 import (
 	"context"
+	"sync"
+	"time"
+
 	"github.com/pkg/errors"
-	"github.com/problame/go-streamrpc"
 	"github.com/prometheus/client_golang/prometheus"
+
 	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
 	"github.com/zrepl/zrepl/daemon/job/reset"
 	"github.com/zrepl/zrepl/daemon/job/wakeup"
-	"github.com/zrepl/zrepl/daemon/transport/connecter"
-	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/daemon/logging"
 	"github.com/zrepl/zrepl/daemon/pruner"
+	"github.com/zrepl/zrepl/daemon/snapper"
 	"github.com/zrepl/zrepl/endpoint"
 	"github.com/zrepl/zrepl/replication"
-	"github.com/zrepl/zrepl/util/envconst"
-	"github.com/zrepl/zrepl/util/watchdog"
+	"github.com/zrepl/zrepl/replication/driver"
+	"github.com/zrepl/zrepl/replication/logic"
+	"github.com/zrepl/zrepl/replication/report"
+	"github.com/zrepl/zrepl/rpc"
+	"github.com/zrepl/zrepl/transport"
+	"github.com/zrepl/zrepl/transport/fromconfig"
 	"github.com/zrepl/zrepl/zfs"
-	"sync"
-	"github.com/zrepl/zrepl/daemon/logging"
-	"github.com/zrepl/zrepl/daemon/snapper"
-	"time"
 )

 type ActiveSide struct {
 	mode      activeMode
 	name      string
-	clientFactory *connecter.ClientFactory
+	connecter transport.Connecter

 	prunerFactory *pruner.PrunerFactory

-
 	promRepStateSecs    *prometheus.HistogramVec // labels: state
 	promPruneSecs       *prometheus.HistogramVec // labels: prune_side
 	promBytesReplicated *prometheus.CounterVec   // labels: filesystem
@@ -38,18 +41,34 @@ type ActiveSide struct {
 	tasks    activeSideTasks
 }

+//go:generate enumer -type=ActiveSideState
+type ActiveSideState int
+
+const (
+	ActiveSideReplicating ActiveSideState = 1 << iota
+	ActiveSidePruneSender
+	ActiveSidePruneReceiver
+	ActiveSideDone // also errors
+)
+
 type activeSideTasks struct {
-	replication *replication.Replication
+	state ActiveSideState
+
+	// valid for state ActiveSideReplicating, ActiveSidePruneSender, ActiveSidePruneReceiver, ActiveSideDone
+	replicationReport driver.ReportFunc
 	replicationCancel context.CancelFunc
+
+	// valid for state ActiveSidePruneSender, ActiveSidePruneReceiver, ActiveSideDone
 	prunerSender, prunerReceiver *pruner.Pruner
+
+	// valid for state ActiveSidePruneReceiver, ActiveSideDone
 	prunerSenderCancel, prunerReceiverCancel context.CancelFunc
 }

 func (a *ActiveSide) updateTasks(u func(*activeSideTasks)) activeSideTasks {
 	a.tasksMtx.Lock()
 	defer a.tasksMtx.Unlock()
-	var copy activeSideTasks
-	copy = a.tasks
+	copy := a.tasks
 	if u == nil {
 		return copy
 	}
@@ -59,28 +78,64 @@ func (a *ActiveSide) updateTasks(u func(*activeSideTasks)) activeSideTasks {
 }

 type activeMode interface {
-	SenderReceiver(client *streamrpc.Client) (replication.Sender, replication.Receiver, error)
+	ConnectEndpoints(rpcLoggers rpc.Loggers, connecter transport.Connecter)
+	DisconnectEndpoints()
+	SenderReceiver() (logic.Sender, logic.Receiver)
 	Type() Type
 	RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{})
+	SnapperReport() *snapper.Report
+	ResetConnectBackoff()
 }

 type modePush struct {
+	setupMtx sync.Mutex
+	sender   *endpoint.Sender
+	receiver *rpc.Client
 	fsfilter endpoint.FSFilter
 	snapper  *snapper.PeriodicOrManual
 }

-func (m *modePush) SenderReceiver(client *streamrpc.Client) (replication.Sender, replication.Receiver, error) {
-	sender := endpoint.NewSender(m.fsfilter)
-	receiver := endpoint.NewRemote(client)
-	return sender, receiver, nil
+func (m *modePush) ConnectEndpoints(loggers rpc.Loggers, connecter transport.Connecter) {
+	m.setupMtx.Lock()
+	defer m.setupMtx.Unlock()
+	if m.receiver != nil || m.sender != nil {
+		panic("inconsistent use of ConnectEndpoints and DisconnectEndpoints")
+	}
+	m.sender = endpoint.NewSender(m.fsfilter)
+	m.receiver = rpc.NewClient(connecter, loggers)
+}
+
+func (m *modePush) DisconnectEndpoints() {
+	m.setupMtx.Lock()
+	defer m.setupMtx.Unlock()
+	m.receiver.Close()
+	m.sender = nil
+	m.receiver = nil
+}
+
+func (m *modePush) SenderReceiver() (logic.Sender, logic.Receiver) {
+	m.setupMtx.Lock()
+	defer m.setupMtx.Unlock()
+	return m.sender, m.receiver
 }

 func (m *modePush) Type() Type { return TypePush }

-func (m *modePush) RunPeriodic(ctx context.Context, wakeUpCommon chan <- struct{}) {
+func (m *modePush) RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{}) {
 	m.snapper.Run(ctx, wakeUpCommon)
 }

+func (m *modePush) SnapperReport() *snapper.Report {
+	return m.snapper.Report()
+}
+
+func (m *modePush) ResetConnectBackoff() {
+	m.setupMtx.Lock()
+	defer m.setupMtx.Unlock()
+	if m.receiver != nil {
+		m.receiver.ResetConnectBackoff()
+	}
+}

 func modePushFromConfig(g *config.Global, in *config.PushJob) (*modePush, error) {
 	m := &modePush{}
@@ -98,20 +153,46 @@ func modePushFromConfig(g *config.Global, in *config.PushJob) (*modePush, error)
 }

 type modePull struct {
+	setupMtx sync.Mutex
+	receiver *endpoint.Receiver
+	sender   *rpc.Client
 	rootFS   *zfs.DatasetPath
-	interval time.Duration
+	interval config.PositiveDurationOrManual
 }

-func (m *modePull) SenderReceiver(client *streamrpc.Client) (replication.Sender, replication.Receiver, error) {
-	sender := endpoint.NewRemote(client)
-	receiver, err := endpoint.NewReceiver(m.rootFS)
-	return sender, receiver, err
+func (m *modePull) ConnectEndpoints(loggers rpc.Loggers, connecter transport.Connecter) {
+	m.setupMtx.Lock()
+	defer m.setupMtx.Unlock()
+	if m.receiver != nil || m.sender != nil {
+		panic("inconsistent use of ConnectEndpoints and DisconnectEndpoints")
+	}
+	m.receiver = endpoint.NewReceiver(m.rootFS, false)
+	m.sender = rpc.NewClient(connecter, loggers)
+}
+
+func (m *modePull) DisconnectEndpoints() {
+	m.setupMtx.Lock()
+	defer m.setupMtx.Unlock()
+	m.sender.Close()
+	m.sender = nil
+	m.receiver = nil
+}
+
+func (m *modePull) SenderReceiver() (logic.Sender, logic.Receiver) {
+	m.setupMtx.Lock()
+	defer m.setupMtx.Unlock()
+	return m.sender, m.receiver
 }

 func (*modePull) Type() Type { return TypePull }

 func (m *modePull) RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{}) {
-	t := time.NewTicker(m.interval)
+	if m.interval.Manual {
+		GetLogger(ctx).Info("manual pull configured, periodic pull disabled")
+		// "waiting for wakeups" is printed in common ActiveSide.do
+		return
+	}
+	t := time.NewTicker(m.interval.Interval)
 	defer t.Stop()
 	for {
 		select {
@@ -130,11 +211,20 @@ func (m *modePull) RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{}
 	}
 }

+func (m *modePull) SnapperReport() *snapper.Report {
+	return nil
+}
+
+func (m *modePull) ResetConnectBackoff() {
+	m.setupMtx.Lock()
+	defer m.setupMtx.Unlock()
+	if m.sender != nil {
+		m.sender.ResetConnectBackoff()
+	}
+}
+
 func modePullFromConfig(g *config.Global, in *config.PullJob) (m *modePull, err error) {
 	m = &modePull{}
-	if in.Interval <= 0 {
-		return nil, errors.New("interval must be positive")
-	}
 	m.interval = in.Interval

 	m.rootFS, err = zfs.NewDatasetPath(in.RootFS)
@@ -157,17 +247,17 @@ func activeSide(g *config.Global, in *config.ActiveJob, mode activeMode) (j *Act
 		Subsystem:   "replication",
 		Name:        "state_time",
 		Help:        "seconds spent during replication",
-		ConstLabels: prometheus.Labels{"zrepl_job":j.name},
+		ConstLabels: prometheus.Labels{"zrepl_job": j.name},
 	}, []string{"state"})
 	j.promBytesReplicated = prometheus.NewCounterVec(prometheus.CounterOpts{
 		Namespace:   "zrepl",
 		Subsystem:   "replication",
 		Name:        "bytes_replicated",
 		Help:        "number of bytes replicated from sender to receiver per filesystem",
-		ConstLabels: prometheus.Labels{"zrepl_job":j.name},
+		ConstLabels: prometheus.Labels{"zrepl_job": j.name},
 	}, []string{"filesystem"})

-	j.clientFactory, err = connecter.FromConfig(g, in.Connect)
+	j.connecter, err = fromconfig.ConnecterFromConfig(g, in.Connect)
 	if err != nil {
 		return nil, errors.Wrap(err, "cannot build client")
 	}
@@ -177,7 +267,7 @@ func activeSide(g *config.Global, in *config.ActiveJob, mode activeMode) (j *Act
 		Subsystem:   "pruning",
 		Name:        "time",
 		Help:        "seconds spent in pruner",
-		ConstLabels: prometheus.Labels{"zrepl_job":j.name},
+		ConstLabels: prometheus.Labels{"zrepl_job": j.name},
 	}, []string{"prune_side"})
 	j.prunerFactory, err = pruner.NewPrunerFactory(in.Pruning, j.promPruneSecs)
 	if err != nil {
@@ -196,8 +286,9 @@ func (j *ActiveSide) RegisterMetrics(registerer prometheus.Registerer) {
 func (j *ActiveSide) Name() string { return j.name }

 type ActiveSideStatus struct {
-	Replication *replication.Report
+	Replication                    *report.Report
 	PruningSender, PruningReceiver *pruner.Report
+	Snapshotting                   *snapper.Report
 }

 func (j *ActiveSide) Status() *Status {
@@ -205,8 +296,8 @@ func (j *ActiveSide) Status() *Status {

 	s := &ActiveSideStatus{}
 	t := j.mode.Type()
-	if tasks.replication != nil {
-		s.Replication = tasks.replication.Report()
+	if tasks.replicationReport != nil {
+		s.Replication = tasks.replicationReport()
 	}
 	if tasks.prunerSender != nil {
 		s.PruningSender = tasks.prunerSender.Report()
@@ -214,9 +305,19 @@ func (j *ActiveSide) Status() *Status {
 	if tasks.prunerReceiver != nil {
 		s.PruningReceiver = tasks.prunerReceiver.Report()
 	}
+	s.Snapshotting = j.mode.SnapperReport()
 	return &Status{Type: t, JobSpecific: s}
 }

+func (j *ActiveSide) OwnedDatasetSubtreeRoot() (rfs *zfs.DatasetPath, ok bool) {
+	pull, ok := j.mode.(*modePull)
+	if !ok {
+		_ = j.mode.(*modePush) // make sure we didn't introduce a new job type
+		return nil, false
+	}
+	return pull.rootFS.Copy(), true
+}
+
 func (j *ActiveSide) Run(ctx context.Context) {
 	log := GetLogger(ctx)
 	ctx = logging.WithSubsystemLoggers(ctx, log)
@@ -238,6 +339,7 @@ outer:
 			break outer

 		case <-wakeup.Wait(ctx):
+			j.mode.ResetConnectBackoff()
 		case <-periodicDone:
 		}
 		invocationCount++
@@ -250,15 +352,15 @@ func (j *ActiveSide) do(ctx context.Context) {

 	log := GetLogger(ctx)
 	ctx = logging.WithSubsystemLoggers(ctx, log)
+	loggers := rpc.GetLoggersOrPanic(ctx) // filled by WithSubsystemLoggers
+	j.mode.ConnectEndpoints(loggers, j.connecter)
+	defer j.mode.DisconnectEndpoints()

 	// allow cancellation of an invocation (this function)
 	ctx, cancelThisRun := context.WithCancel(ctx)
 	defer cancelThisRun()
-	runDone := make(chan struct{})
-	defer close(runDone)
 	go func() {
 		select {
-		case <-runDone:
 		case <-reset.Wait(ctx):
 			log.Info("reset received, cancelling current invocation")
 			cancelThisRun()
@@ -266,64 +368,7 @@ func (j *ActiveSide) do(ctx context.Context) {
 		}
 	}()

-	// watchdog
-	go func() {
-		// if no progress after 1 minute, kill the task
-		wdto := envconst.Duration("ZREPL_JOB_WATCHDOG_TIMEOUT", 1*time.Minute)
-		log.WithField("watchdog_timeout", wdto.String()).Debug("starting watchdog")
-
-		t := time.NewTicker(wdto)
-		defer t.Stop()
-
-		var (
-			rep, prunerSender, prunerReceiver watchdog.Progress
-		)
-		for {
-			select {
-			case <-runDone:
-				return
-			case <-ctx.Done():
-				return
-			case <-t.C: // fall
-			}
-
-			log := log.WithField("watchdog_timeout", wdto.String()) // shadowing!
-
-			j.updateTasks(func(tasks *activeSideTasks) {
-				if tasks.replication != nil &&
-					!tasks.replication.Progress.ExpectProgress(&rep) &&
-					!tasks.replication.State().IsTerminal() {
-					log.Error("replication did not make progress, cancelling")
-					tasks.replicationCancel()
-				}
-				if tasks.prunerSender != nil &&
-					!tasks.prunerSender.Progress.ExpectProgress(&prunerSender) &&
-					!tasks.prunerSender.State().IsTerminal() {
-					log.Error("pruner:sender did not make progress, cancelling")
-					tasks.prunerSenderCancel()
-				}
-				if tasks.prunerReceiver != nil &&
-					!tasks.prunerReceiver.Progress.ExpectProgress(&prunerReceiver) &&
-					!tasks.prunerReceiver.State().IsTerminal() {
-					log.Error("pruner:receiver did not make progress, cancelling")
-					tasks.prunerReceiverCancel()
-				}
-			})
-			log.WithField("replication_progress", rep.String()).
-				WithField("pruner_sender_progress", prunerSender.String()).
-				WithField("pruner_receiver_progress", prunerReceiver.String()).
-				Debug("watchdog did run")
-
-		}
-	}()
-
-	client, err := j.clientFactory.NewClient()
-	if err != nil {
-		log.WithError(err).Error("factory cannot instantiate streamrpc client")
-	}
-	defer client.Close(ctx)
-
-	sender, receiver, err := j.mode.SenderReceiver(client)
+	sender, receiver := j.mode.SenderReceiver()

 	{
 		select {
@@ -332,14 +377,18 @@ func (j *ActiveSide) do(ctx context.Context) {
 		default:
 		}
 		ctx, repCancel := context.WithCancel(ctx)
-		tasks := j.updateTasks(func(tasks *activeSideTasks) {
+		var repWait driver.WaitFunc
+		j.updateTasks(func(tasks *activeSideTasks) {
 			// reset it
 			*tasks = activeSideTasks{}
 			tasks.replicationCancel = repCancel
-			tasks.replication = replication.NewReplication(j.promRepStateSecs, j.promBytesReplicated)
+			tasks.replicationReport, repWait = replication.Do(
+				ctx, logic.NewPlanner(j.promRepStateSecs, j.promBytesReplicated, sender, receiver),
+			)
+			tasks.state = ActiveSideReplicating
 		})
 		log.Info("start replication")
-		tasks.replication.Drive(ctx, sender, receiver)
+		repWait(true) // wait blocking
 		repCancel()   // always cancel to free up context resources
 	}

@@ -353,6 +402,7 @@ func (j *ActiveSide) do(ctx context.Context) {
 		tasks := j.updateTasks(func(tasks *activeSideTasks) {
 			tasks.prunerSender = j.prunerFactory.BuildSenderPruner(ctx, sender, sender)
 			tasks.prunerSenderCancel = senderCancel
+			tasks.state = ActiveSidePruneSender
 		})
 		log.Info("start pruning sender")
 		tasks.prunerSender.Prune()
@@ -369,10 +419,16 @@ func (j *ActiveSide) do(ctx context.Context) {
 		tasks := j.updateTasks(func(tasks *activeSideTasks) {
 			tasks.prunerReceiver = j.prunerFactory.BuildReceiverPruner(ctx, receiver, sender)
 			tasks.prunerReceiverCancel = receiverCancel
+			tasks.state = ActiveSidePruneReceiver
 		})
 		log.Info("start pruning receiver")
 		tasks.prunerReceiver.Prune()
 		log.Info("finished pruning receiver")
 		receiverCancel()
 	}
+
+	j.updateTasks(func(tasks *activeSideTasks) {
+		tasks.state = ActiveSideDone
+	})
+
 }
@@ -0,0 +1,67 @@
+// Code generated by "enumer -type=ActiveSideState"; DO NOT EDIT.
+
+//
+package job
+
+import (
+	"fmt"
+)
+
+const (
+	_ActiveSideStateName_0 = "ActiveSideReplicatingActiveSidePruneSender"
+	_ActiveSideStateName_1 = "ActiveSidePruneReceiver"
+	_ActiveSideStateName_2 = "ActiveSideDone"
+)
+
+var (
+	_ActiveSideStateIndex_0 = [...]uint8{0, 21, 42}
+	_ActiveSideStateIndex_1 = [...]uint8{0, 23}
+	_ActiveSideStateIndex_2 = [...]uint8{0, 14}
+)
+
+func (i ActiveSideState) String() string {
+	switch {
+	case 1 <= i && i <= 2:
+		i -= 1
+		return _ActiveSideStateName_0[_ActiveSideStateIndex_0[i]:_ActiveSideStateIndex_0[i+1]]
+	case i == 4:
+		return _ActiveSideStateName_1
+	case i == 8:
+		return _ActiveSideStateName_2
+	default:
+		return fmt.Sprintf("ActiveSideState(%d)", i)
+	}
+}
+
+var _ActiveSideStateValues = []ActiveSideState{1, 2, 4, 8}
+
+var _ActiveSideStateNameToValueMap = map[string]ActiveSideState{
+	_ActiveSideStateName_0[0:21]:  1,
+	_ActiveSideStateName_0[21:42]: 2,
+	_ActiveSideStateName_1[0:23]:  4,
+	_ActiveSideStateName_2[0:14]:  8,
+}
+
+// ActiveSideStateString retrieves an enum value from the enum constants string name.
+// Throws an error if the param is not part of the enum.
+func ActiveSideStateString(s string) (ActiveSideState, error) {
+	if val, ok := _ActiveSideStateNameToValueMap[s]; ok {
+		return val, nil
+	}
+	return 0, fmt.Errorf("%s does not belong to ActiveSideState values", s)
+}
+
+// ActiveSideStateValues returns all values of the enum
+func ActiveSideStateValues() []ActiveSideState {
+	return _ActiveSideStateValues
+}
+
+// IsAActiveSideState returns "true" if the value is listed in the enum definition. "false" otherwise
+func (i ActiveSideState) IsAActiveSideState() bool {
+	for _, v := range _ActiveSideStateValues {
+		if i == v {
+			return true
+		}
+	}
+	return false
+}
@@ -2,7 +2,11 @@ package job

 import (
 	"fmt"
+	"sort"
+	"strings"
+
 	"github.com/pkg/errors"
+
 	"github.com/zrepl/zrepl/config"
 )

@@ -18,6 +22,22 @@ func JobsFromConfig(c *config.Config) ([]Job, error) {
 		}
 		js[i] = j
 	}
+
+	// receiving-side root filesystems must not overlap
+	{
+		rfss := make([]string, 0, len(js))
+		for _, j := range js {
+			jrfs, ok := j.OwnedDatasetSubtreeRoot()
+			if !ok {
+				continue
+			}
+			rfss = append(rfss, jrfs.ToString())
+		}
+		if err := validateReceivingSidesDoNotOverlap(rfss); err != nil {
+			return nil, err
+		}
+	}
+
 	return js, nil
 }

@@ -45,6 +65,11 @@ func buildJob(c *config.Global, in config.JobEnum) (j Job, err error) {
 		if err != nil {
 			return cannotBuildJob(err, v.Name)
 		}
+	case *config.SnapJob:
+		j, err = snapJobFromConfig(c, v)
+		if err != nil {
+			return cannotBuildJob(err, v.Name)
+		}
 	case *config.PushJob:
 		m, err := modePushFromConfig(c, v)
 		if err != nil {
@@ -69,3 +94,27 @@ func buildJob(c *config.Global, in config.JobEnum) (j Job, err error) {
 	return j, nil

 }
+
+func validateReceivingSidesDoNotOverlap(receivingRootFSs []string) error {
+	if len(receivingRootFSs) == 0 {
+		return nil
+	}
+	rfss := make([]string, len(receivingRootFSs))
+	copy(rfss, receivingRootFSs)
+	sort.Slice(rfss, func(i, j int) bool {
+		return strings.Compare(rfss[i], rfss[j]) == -1
+	})
+	// idea:
+	//   no path in rfss must be prefix of another
+	//
+	// rfss is now lexicographically sorted, which means that
+	// if i is prefix of j, i < j (in lexicographical order)
+	// thus,
+	// if any i is prefix of i+n (n >= 1), there is overlap
+	for i := 0; i < len(rfss)-1; i++ {
+		if strings.HasPrefix(rfss[i+1], rfss[i]) {
+			return fmt.Errorf("receiving jobs with overlapping root filesystems are forbidden")
+		}
+	}
+	return nil
+}
@@ -0,0 +1,42 @@
+package job
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestValidateReceivingSidesDoNotOverlap(t *testing.T) {
+	type testCase struct {
+		err   bool
+		input []string
+	}
+	tcs := []testCase{
+		{false, nil},
+		{false, []string{}},
+		{false, []string{""}}, // not our job to determine valid paths
+		{false, []string{"a"}},
+		{false, []string{"some/path"}},
+		{false, []string{"zroot/sink1", "zroot/sink2", "zroot/sink3"}},
+		{true, []string{"zroot/b", "zroot/b"}},
+		{true, []string{"zroot/foo", "zroot/foo/bar", "zroot/baz"}},
+		{false, []string{"a/x", "b/x"}},
+		{false, []string{"a", "b"}},
+		{true, []string{"a", "a"}},
+		{true, []string{"a/x/y", "a/x"}},
+		{true, []string{"a/x", "a/x/y"}},
+		{true, []string{"a/x", "b/x", "a/x/y"}},
+		{true, []string{"a", "a/b", "a/c", "a/b"}},
+		{true, []string{"a/b", "a/c", "a/b", "a/d", "a/c"}},
+	}
+
+	for _, tc := range tcs {
+		t.Logf("input: %v", tc.input)
+		err := validateReceivingSidesDoNotOverlap(tc.input)
+		if tc.err {
+			assert.Error(t, err)
+		} else {
+			assert.NoError(t, err)
+		}
+	}
+}
@@ -4,8 +4,11 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
+
 	"github.com/prometheus/client_golang/prometheus"
+
 	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/zfs"
 )

 type Logger = logger.Logger
@@ -27,18 +30,21 @@ func WithLogger(ctx context.Context, l Logger) context.Context {
 	return context.WithValue(ctx, contextKeyLog, l)
 }

-
 type Job interface {
 	Name() string
 	Run(ctx context.Context)
 	Status() *Status
 	RegisterMetrics(registerer prometheus.Registerer)
+	// Jobs that return a subtree of the dataset hierarchy
+	// must return the root of that subtree as rfs and ok = true
+	OwnedDatasetSubtreeRoot() (rfs *zfs.DatasetPath, ok bool)
 }

 type Type string

 const (
 	TypeInternal Type = "internal"
+	TypeSnap     Type = "snap"
 	TypePush     Type = "push"
 	TypeSink     Type = "sink"
 	TypePull     Type = "pull"
@@ -59,7 +65,7 @@ func (s *Status) MarshalJSON() ([]byte, error) {
 	if err != nil {
 		return nil, err
 	}
-	m := map[string]json.RawMessage {
+	m := map[string]json.RawMessage{
 		"type":         typeJson,
 		string(s.Type): jobJSON,
 	}
@@ -84,16 +90,25 @@ func (s *Status) UnmarshalJSON(in []byte) (err error) {
 		return fmt.Errorf("field '%s', not found", key)
 	}
 	switch s.Type {
-	case TypePull: fallthrough
+	case TypeSnap:
+		var st SnapJobStatus
+		err = json.Unmarshal(jobJSON, &st)
+		s.JobSpecific = &st
+
+	case TypePull:
+		fallthrough
 	case TypePush:
 		var st ActiveSideStatus
 		err = json.Unmarshal(jobJSON, &st)
 		s.JobSpecific = &st
-	case TypeSource: fallthrough
+
+	case TypeSource:
+		fallthrough
 	case TypeSink:
 		var st PassiveStatus
 		err = json.Unmarshal(jobJSON, &st)
 		s.JobSpecific = &st
+
 	case TypeInternal:
 		// internal jobs do not report specifics
 	default:
@@ -2,29 +2,32 @@ package job

 import (
 	"context"
+	"fmt"
+
 	"github.com/pkg/errors"
-	"github.com/problame/go-streamrpc"
 	"github.com/prometheus/client_golang/prometheus"
+
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/filters"
 	"github.com/zrepl/zrepl/daemon/logging"
-	"github.com/zrepl/zrepl/daemon/transport/serve"
 	"github.com/zrepl/zrepl/daemon/snapper"
 	"github.com/zrepl/zrepl/endpoint"
+	"github.com/zrepl/zrepl/rpc"
+	"github.com/zrepl/zrepl/transport"
+	"github.com/zrepl/zrepl/transport/fromconfig"
 	"github.com/zrepl/zrepl/zfs"
-	"path"
 )

 type PassiveSide struct {
 	mode   passiveMode
 	name   string
-	l        serve.ListenerFactory
-	rpcConf  *streamrpc.ConnConfig
+	listen transport.AuthenticatedListenerFactory
 }

 type passiveMode interface {
-	ConnHandleFunc(ctx context.Context, conn serve.AuthenticatedConn) streamrpc.HandlerFunc
+	Handler() rpc.Handler
 	RunPeriodic(ctx context.Context)
+	SnapperReport() *snapper.Report // may be nil
 	Type() Type
 }

@@ -34,29 +37,12 @@ type modeSink struct {

 func (m *modeSink) Type() Type { return TypeSink }

-func (m *modeSink) ConnHandleFunc(ctx context.Context, conn serve.AuthenticatedConn) streamrpc.HandlerFunc {
-	log := GetLogger(ctx)
-
-	clientRootStr := path.Join(m.rootDataset.ToString(), conn.ClientIdentity())
-	clientRoot, err := zfs.NewDatasetPath(clientRootStr)
-	if err != nil {
-		log.WithError(err).
-			WithField("client_identity", conn.ClientIdentity()).
-			Error("cannot build client filesystem map (client identity must be a valid ZFS FS name")
-	}
-	log.WithField("client_root", clientRoot).Debug("client root")
-
-	local, err := endpoint.NewReceiver(clientRoot)
-	if err != nil {
-		log.WithError(err).Error("unexpected error: cannot convert mapping to filter")
-		return nil
-	}
-
-	h := endpoint.NewHandler(local)
-	return h.Handle
+func (m *modeSink) Handler() rpc.Handler {
+	return endpoint.NewReceiver(m.rootDataset, true)
 }

 func (m *modeSink) RunPeriodic(_ context.Context)  {}
+func (m *modeSink) SnapperReport() *snapper.Report { return nil }

 func modeSinkFromConfig(g *config.Global, in *config.SinkJob) (m *modeSink, err error) {
 	m = &modeSink{}
@@ -93,21 +79,23 @@ func modeSourceFromConfig(g *config.Global, in *config.SourceJob) (m *modeSource

 func (m *modeSource) Type() Type { return TypeSource }

-func (m *modeSource) ConnHandleFunc(ctx context.Context, conn serve.AuthenticatedConn) streamrpc.HandlerFunc {
-	sender := endpoint.NewSender(m.fsfilter)
-	h := endpoint.NewHandler(sender)
-	return h.Handle
+func (m *modeSource) Handler() rpc.Handler {
+	return endpoint.NewSender(m.fsfilter)
 }

 func (m *modeSource) RunPeriodic(ctx context.Context) {
 	m.snapper.Run(ctx, nil)
 }

+func (m *modeSource) SnapperReport() *snapper.Report {
+	return m.snapper.Report()
+}
+
 func passiveSideFromConfig(g *config.Global, in *config.PassiveJob, mode passiveMode) (s *PassiveSide, err error) {

 	s = &PassiveSide{mode: mode, name: in.Name}
-	if s.l, s.rpcConf, err = serve.FromConfig(g, in.Serve); err != nil {
-		return nil, errors.Wrap(err, "cannot build server")
+	if s.listen, err = fromconfig.ListenerFactoryFromConfig(g, in.Serve); err != nil {
+		return nil, errors.Wrap(err, "cannot build listener factory")
 	}

 	return s, nil
@@ -115,10 +103,24 @@ func passiveSideFromConfig(g *config.Global, in *config.PassiveJob, mode passive

 func (j *PassiveSide) Name() string { return j.name }

-type PassiveStatus struct {}
+type PassiveStatus struct {
+	Snapper *snapper.Report
+}

 func (s *PassiveSide) Status() *Status {
-	return &Status{Type: s.mode.Type()} // FIXME PassiveStatus
+	st := &PassiveStatus{
+		Snapper: s.mode.SnapperReport(),
+	}
+	return &Status{Type: s.mode.Type(), JobSpecific: st}
+}
+
+func (j *PassiveSide) OwnedDatasetSubtreeRoot() (rfs *zfs.DatasetPath, ok bool) {
+	sink, ok := j.mode.(*modeSink)
+	if !ok {
+		_ = j.mode.(*modeSource) // make sure we didn't introduce a new job type
+		return nil, false
+	}
+	return sink.rootDataset.Copy(), true
 }

 func (*PassiveSide) RegisterMetrics(registerer prometheus.Registerer) {}
@@ -127,70 +129,30 @@ func (j *PassiveSide) Run(ctx context.Context) {

 	log := GetLogger(ctx)
 	defer log.Info("job exiting")
-
-	l, err := j.l.Listen()
-	if err != nil {
-		log.WithError(err).Error("cannot listen")
-		return
-	}
-	defer l.Close()
-
+	ctx = logging.WithSubsystemLoggers(ctx, log)
 	{
-		ctx, cancel := context.WithCancel(logging.WithSubsystemLoggers(ctx, log)) // shadowing
+		ctx, cancel := context.WithCancel(ctx) // shadowing
 		defer cancel()
 		go j.mode.RunPeriodic(ctx)
 	}

-	log.WithField("addr", l.Addr()).Debug("accepting connections")
-	var connId int
-outer:
-	for {
-
-		select {
-		case res := <-accept(ctx, l):
-			if res.err != nil {
-				log.WithError(res.err).Info("accept error")
-				continue
+	handler := j.mode.Handler()
+	if handler == nil {
+		panic(fmt.Sprintf("implementation error: j.mode.Handler() returned nil: %#v", j))
 	}
-			conn := res.conn
-			connId++
-			connLog := log.
-				WithField("connID", connId)
-			connLog.
-				WithField("addr", conn.RemoteAddr()).
-				WithField("client_identity", conn.ClientIdentity()).
-				Info("handling connection")
-			go func() {
-				defer connLog.Info("finished handling connection")
-				defer conn.Close()
-				ctx := logging.WithSubsystemLoggers(ctx, connLog)
-				handleFunc := j.mode.ConnHandleFunc(ctx, conn)
-				if handleFunc == nil {
+
+	ctxInterceptor := func(handlerCtx context.Context) context.Context {
+		return logging.WithSubsystemLoggers(handlerCtx, log)
+	}
+
+	rpcLoggers := rpc.GetLoggersOrPanic(ctx) // WithSubsystemLoggers above
+	server := rpc.NewServer(handler, rpcLoggers, ctxInterceptor)
+
+	listener, err := j.listen()
+	if err != nil {
+		log.WithError(err).Error("cannot listen")
 		return
 	}
-				if err := streamrpc.ServeConn(ctx, conn, j.rpcConf, handleFunc); err != nil {
-					log.WithError(err).Error("error serving client")
-				}
-			}()
-
-		case <-ctx.Done():
-			break outer
-		}
-
-	}

-}
-
-type acceptResult struct {
-	conn serve.AuthenticatedConn
-	err  error
-}
-
-func accept(ctx context.Context, listener serve.AuthenticatedListener) <-chan acceptResult {
-	c := make(chan acceptResult, 1)
-	go func() {
-		conn, err := listener.Accept(ctx)
-		c <- acceptResult{conn, err}
-	}()
-	return c
+	server.Serve(ctx, listener)
 }
@@ -0,0 +1,170 @@
+package job
+
+import (
+	"context"
+	"fmt"
+	"sort"
+
+	"github.com/pkg/errors"
+	"github.com/prometheus/client_golang/prometheus"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/daemon/job/wakeup"
+	"github.com/zrepl/zrepl/daemon/logging"
+	"github.com/zrepl/zrepl/daemon/pruner"
+	"github.com/zrepl/zrepl/daemon/snapper"
+	"github.com/zrepl/zrepl/endpoint"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type SnapJob struct {
+	name     string
+	fsfilter zfs.DatasetFilter
+	snapper  *snapper.PeriodicOrManual
+
+	prunerFactory *pruner.LocalPrunerFactory
+
+	promPruneSecs *prometheus.HistogramVec // labels: prune_side
+
+	pruner *pruner.Pruner
+}
+
+func (j *SnapJob) Name() string { return j.name }
+
+func (j *SnapJob) Type() Type { return TypeSnap }
+
+func snapJobFromConfig(g *config.Global, in *config.SnapJob) (j *SnapJob, err error) {
+	j = &SnapJob{}
+	fsf, err := filters.DatasetMapFilterFromConfig(in.Filesystems)
+	if err != nil {
+		return nil, errors.Wrap(err, "cannnot build filesystem filter")
+	}
+	j.fsfilter = fsf
+
+	if j.snapper, err = snapper.FromConfig(g, fsf, in.Snapshotting); err != nil {
+		return nil, errors.Wrap(err, "cannot build snapper")
+	}
+	j.name = in.Name
+	j.promPruneSecs = prometheus.NewHistogramVec(prometheus.HistogramOpts{
+		Namespace:   "zrepl",
+		Subsystem:   "pruning",
+		Name:        "time",
+		Help:        "seconds spent in pruner",
+		ConstLabels: prometheus.Labels{"zrepl_job": j.name},
+	}, []string{"prune_side"})
+	j.prunerFactory, err = pruner.NewLocalPrunerFactory(in.Pruning, j.promPruneSecs)
+	if err != nil {
+		return nil, errors.Wrap(err, "cannot build snapjob pruning rules")
+	}
+	return j, nil
+}
+
+func (j *SnapJob) RegisterMetrics(registerer prometheus.Registerer) {
+	registerer.MustRegister(j.promPruneSecs)
+}
+
+type SnapJobStatus struct {
+	Pruning      *pruner.Report
+	Snapshotting *snapper.Report // may be nil
+}
+
+func (j *SnapJob) Status() *Status {
+	s := &SnapJobStatus{}
+	t := j.Type()
+	if j.pruner != nil {
+		s.Pruning = j.pruner.Report()
+	}
+	s.Snapshotting = j.snapper.Report()
+	return &Status{Type: t, JobSpecific: s}
+}
+
+func (j *SnapJob) OwnedDatasetSubtreeRoot() (rfs *zfs.DatasetPath, ok bool) {
+	return nil, false
+}
+
+func (j *SnapJob) Run(ctx context.Context) {
+	log := GetLogger(ctx)
+	ctx = logging.WithSubsystemLoggers(ctx, log)
+
+	defer log.Info("job exiting")
+
+	periodicDone := make(chan struct{})
+	ctx, cancel := context.WithCancel(ctx)
+	defer cancel()
+	go j.snapper.Run(ctx, periodicDone)
+
+	invocationCount := 0
+outer:
+	for {
+		log.Info("wait for wakeups")
+		select {
+		case <-ctx.Done():
+			log.WithError(ctx.Err()).Info("context")
+			break outer
+
+		case <-wakeup.Wait(ctx):
+		case <-periodicDone:
+		}
+		invocationCount++
+		invLog := log.WithField("invocation", invocationCount)
+		j.doPrune(WithLogger(ctx, invLog))
+	}
+}
+
+// Adaptor that implements pruner.History around a pruner.Target.
+// The ReplicationCursor method is Get-op only and always returns
+// the filesystem's most recent version's GUID.
+//
+// TODO:
+// This is a work-around for the current package daemon/pruner
+// and package pruning.Snapshot limitation: they require the
+//  `Replicated` getter method be present, but obviously,
+// a local job like SnapJob can't deliver on that.
+// But the pruner.Pruner gives up on an FS if no replication
+// cursor is present, which is why this pruner returns the
+// most recent filesystem version.
+type alwaysUpToDateReplicationCursorHistory struct {
+	// the Target passed as Target to BuildLocalPruner
+	target pruner.Target
+}
+
+var _ pruner.History = (*alwaysUpToDateReplicationCursorHistory)(nil)
+
+func (h alwaysUpToDateReplicationCursorHistory) ReplicationCursor(ctx context.Context, req *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error) {
+	if req.GetGet() == nil {
+		return nil, fmt.Errorf("unsupported ReplicationCursor request: SnapJob only supports GETting a (faked) cursor")
+	}
+	fsvReq := &pdu.ListFilesystemVersionsReq{
+		Filesystem: req.GetFilesystem(),
+	}
+	res, err := h.target.ListFilesystemVersions(ctx, fsvReq)
+	if err != nil {
+		return nil, err
+	}
+	fsvs := res.GetVersions()
+	if len(fsvs) <= 0 {
+		return &pdu.ReplicationCursorRes{Result: &pdu.ReplicationCursorRes_Notexist{Notexist: true}}, nil
+	}
+	// always return must recent version
+	sort.Slice(fsvs, func(i, j int) bool {
+		return fsvs[i].CreateTXG < fsvs[j].CreateTXG
+	})
+	mostRecent := fsvs[len(fsvs)-1]
+	return &pdu.ReplicationCursorRes{Result: &pdu.ReplicationCursorRes_Guid{Guid: mostRecent.GetGuid()}}, nil
+}
+
+func (h alwaysUpToDateReplicationCursorHistory) ListFilesystems(ctx context.Context, req *pdu.ListFilesystemReq) (*pdu.ListFilesystemRes, error) {
+	return h.target.ListFilesystems(ctx, req)
+}
+
+func (j *SnapJob) doPrune(ctx context.Context) {
+	log := GetLogger(ctx)
+	ctx = logging.WithSubsystemLoggers(ctx, log)
+	sender := endpoint.NewSender(j.fsfilter)
+	j.pruner = j.prunerFactory.BuildLocalPruner(ctx, sender, alwaysUpToDateReplicationCursorHistory{sender})
+	log.Info("start pruning")
+	j.pruner.Prune()
+	log.Info("finished pruning")
+}
@@ -1,32 +0,0 @@
-package logging
-
-import (
-	"fmt"
-	"github.com/problame/go-streamrpc"
-	"github.com/zrepl/zrepl/logger"
-	"strings"
-)
-
-type streamrpcLogAdaptor = twoClassLogAdaptor
-
-type twoClassLogAdaptor struct {
-	logger.Logger
-}
-
-var _ streamrpc.Logger = twoClassLogAdaptor{}
-
-func (a twoClassLogAdaptor) Errorf(fmtStr string, args ...interface{}) {
-	const errorSuffix = ": %s"
-	if len(args) == 1 {
-		if err, ok := args[0].(error); ok && strings.HasSuffix(fmtStr, errorSuffix) {
-			msg := strings.TrimSuffix(fmtStr, errorSuffix)
-			a.WithError(err).Error(msg)
-			return
-		}
-	}
-	a.Logger.Error(fmt.Sprintf(fmtStr, args...))
-}
-
-func (a twoClassLogAdaptor) Infof(fmtStr string, args ...interface{}) {
-	a.Logger.Debug(fmt.Sprintf(fmtStr, args...))
-}
@@ -4,18 +4,24 @@ import (
 	"context"
 	"crypto/tls"
 	"crypto/x509"
+	"log/syslog"
+	"os"
+
 	"github.com/mattn/go-isatty"
 	"github.com/pkg/errors"
-	"github.com/problame/go-streamrpc"
+
 	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/hooks"
 	"github.com/zrepl/zrepl/daemon/pruner"
+	"github.com/zrepl/zrepl/daemon/snapper"
 	"github.com/zrepl/zrepl/endpoint"
 	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/replication"
+	"github.com/zrepl/zrepl/replication/driver"
+	"github.com/zrepl/zrepl/replication/logic"
+	"github.com/zrepl/zrepl/rpc"
+	"github.com/zrepl/zrepl/rpc/transportmux"
 	"github.com/zrepl/zrepl/tlsconf"
-	"os"
-	"github.com/zrepl/zrepl/daemon/snapper"
-	"github.com/zrepl/zrepl/daemon/transport/serve"
+	"github.com/zrepl/zrepl/transport"
 )

 func OutletsFromConfig(in config.LoggingOutletEnumList) (*logger.Outlets, error) {
@@ -32,7 +38,7 @@ func OutletsFromConfig(in config.LoggingOutletEnumList) (*logger.Outlets, error)
 	var syslogOutlets, stdoutOutlets int
 	for lei, le := range in {

-		outlet, minLevel, err := parseOutlet(le)
+		outlet, minLevel, err := ParseOutlet(le)
 		if err != nil {
 			return nil, errors.Wrapf(err, "cannot parse outlet #%d", lei)
 		}
@@ -60,22 +66,44 @@ func OutletsFromConfig(in config.LoggingOutletEnumList) (*logger.Outlets, error)

 }

+type Subsystem string
+
 const (
-	SubsysReplication = "repl"
-	SubsysStreamrpc   = "rpc"
-	SubsyEndpoint     = "endpoint"
+	SubsysReplication  Subsystem = "repl"
+	SubsyEndpoint      Subsystem = "endpoint"
+	SubsysPruning      Subsystem = "pruning"
+	SubsysSnapshot     Subsystem = "snapshot"
+	SubsysHooks        Subsystem = "hook"
+	SubsysTransport    Subsystem = "transport"
+	SubsysTransportMux Subsystem = "transportmux"
+	SubsysRPC          Subsystem = "rpc"
+	SubsysRPCControl   Subsystem = "rpc.ctrl"
+	SubsysRPCData      Subsystem = "rpc.data"
 )

 func WithSubsystemLoggers(ctx context.Context, log logger.Logger) context.Context {
-	ctx = replication.WithLogger(ctx, log.WithField(SubsysField, "repl"))
-	ctx = streamrpc.ContextWithLogger(ctx, streamrpcLogAdaptor{log.WithField(SubsysField, "rpc")})
-	ctx = endpoint.WithLogger(ctx, log.WithField(SubsysField, "endpoint"))
-	ctx = pruner.WithLogger(ctx, log.WithField(SubsysField, "pruning"))
-	ctx = snapper.WithLogger(ctx, log.WithField(SubsysField, "snapshot"))
-	ctx = serve.WithLogger(ctx, log.WithField(SubsysField, "serve"))
+	ctx = logic.WithLogger(ctx, log.WithField(SubsysField, SubsysReplication))
+	ctx = driver.WithLogger(ctx, log.WithField(SubsysField, SubsysReplication))
+	ctx = endpoint.WithLogger(ctx, log.WithField(SubsysField, SubsyEndpoint))
+	ctx = pruner.WithLogger(ctx, log.WithField(SubsysField, SubsysPruning))
+	ctx = snapper.WithLogger(ctx, log.WithField(SubsysField, SubsysSnapshot))
+	ctx = hooks.WithLogger(ctx, log.WithField(SubsysField, SubsysHooks))
+	ctx = transport.WithLogger(ctx, log.WithField(SubsysField, SubsysTransport))
+	ctx = transportmux.WithLogger(ctx, log.WithField(SubsysField, SubsysTransportMux))
+	ctx = rpc.WithLoggers(ctx,
+		rpc.Loggers{
+			General: log.WithField(SubsysField, SubsysRPC),
+			Control: log.WithField(SubsysField, SubsysRPCControl),
+			Data:    log.WithField(SubsysField, SubsysRPCData),
+		},
+	)
 	return ctx
 }

+func LogSubsystem(log logger.Logger, subsys Subsystem) logger.Logger {
+	return log.ReplaceField(SubsysField, subsys)
+}
+
 func parseLogFormat(i interface{}) (f EntryFormatter, err error) {
 	var is string
 	switch j := i.(type) {
@@ -98,7 +126,7 @@ func parseLogFormat(i interface{}) (f EntryFormatter, err error) {

 }

-func parseOutlet(in config.LoggingOutletEnum) (o logger.Outlet, level logger.Level, err error) {
+func ParseOutlet(in config.LoggingOutletEnum) (o logger.Outlet, level logger.Level, err error) {

 	parseCommon := func(common config.LoggingOutletCommon) (logger.Level, EntryFormatter, error) {
 		if common.Level == "" || common.Format == "" {
@@ -200,6 +228,7 @@ func parseSyslogOutlet(in *config.SyslogLoggingOutlet, formatter EntryFormatter)
 	out = &SyslogOutlet{}
 	out.Formatter = formatter
 	out.Formatter.SetMetadataFlags(MetadataNone)
+	out.Facility = syslog.Priority(*in.Facility)
 	out.RetryInterval = in.RetryInterval
 	return out, nil
 }
@@ -4,11 +4,13 @@ import (
 	"bytes"
 	"encoding/json"
 	"fmt"
+	"time"
+
 	"github.com/fatih/color"
 	"github.com/go-logfmt/logfmt"
 	"github.com/pkg/errors"
+
 	"github.com/zrepl/zrepl/logger"
-	"time"
 )

 const (
@@ -86,7 +88,7 @@ func (f *HumanFormatter) Format(e *logger.Entry) (out []byte, err error) {
 	prefixFields := []string{JobField, SubsysField}
 	prefixed := make(map[string]bool, len(prefixFields)+2)
 	for _, field := range prefixFields {
-		val, ok := e.Fields[field].(string)
+		val, ok := e.Fields[field]
 		if !ok {
 			continue
 		}
@@ -159,10 +161,16 @@ func (f *LogfmtFormatter) Format(e *logger.Entry) ([]byte, error) {
 	enc := logfmt.NewEncoder(&buf)

 	if f.metadataFlags&MetadataTime != 0 {
-		enc.EncodeKeyval(FieldTime, e.Time)
+		err := enc.EncodeKeyval(FieldTime, e.Time)
+		if err != nil {
+			return nil, errors.Wrap(err, "logfmt: encode time")
+		}
 	}
 	if f.metadataFlags&MetadataLevel != 0 {
-		enc.EncodeKeyval(FieldLevel, e.Level)
+		err := enc.EncodeKeyval(FieldLevel, e.Level)
+		if err != nil {
+			return nil, errors.Wrap(err, "logfmt: encode level")
+		}
 	}

 	// at least try and put job and task in front
@@ -179,8 +187,10 @@ func (f *LogfmtFormatter) Format(e *logger.Entry) ([]byte, error) {
 		prefixed[pf] = true
 	}

-	enc.EncodeKeyval(FieldMessage, e.Message)
-
+	err := enc.EncodeKeyval(FieldMessage, e.Message)
+	if err != nil {
+		return nil, errors.Wrap(err, "logfmt: encode message")
+	}
 	for k, v := range e.Fields {
 		if !prefixed[k] {
 			if err := logfmtTryEncodeKeyval(enc, k, v); err != nil {
@@ -199,7 +209,10 @@ func logfmtTryEncodeKeyval(enc *logfmt.Encoder, field, value interface{}) error
 	case nil: // ok
 		return nil
 	case logfmt.ErrUnsupportedValueType:
-		enc.EncodeKeyval(field, fmt.Sprintf("<%T>", value))
+		err := enc.EncodeKeyval(field, fmt.Sprintf("<%T>", value))
+		if err != nil {
+			return errors.Wrap(err, "cannot encode unsuuported value type Go type")
+		}
 		return nil
 	}
 	return errors.Wrapf(err, "cannot encode field '%s'", field)
@@ -4,12 +4,14 @@ import (
 	"bytes"
 	"context"
 	"crypto/tls"
-	"github.com/pkg/errors"
-	"github.com/zrepl/zrepl/logger"
 	"io"
 	"log/syslog"
 	"net"
 	"time"
+
+	"github.com/pkg/errors"
+
+	"github.com/zrepl/zrepl/logger"
 )

 type EntryFormatter interface {
@@ -28,7 +30,10 @@ func (h WriterOutlet) WriteEntry(entry logger.Entry) error {
 		return err
 	}
 	_, err = h.writer.Write(bytes)
-	h.writer.Write([]byte("\n"))
+	if err != nil {
+		return err
+	}
+	_, err = h.writer.Write([]byte("\n"))
 	return err
 }

@@ -92,8 +97,10 @@ func (h *TCPOutlet) outLoop(retryInterval time.Duration) {
 				conn = nil
 			}
 		}
-		conn.SetWriteDeadline(time.Now().Add(retryInterval))
+		err = conn.SetWriteDeadline(time.Now().Add(retryInterval))
+		if err == nil {
 			_, err = io.Copy(conn, msg)
+		}
 		if err != nil {
 			retry = time.Now().Add(retryInterval)
 			conn.Close()
@@ -124,6 +131,7 @@ func (h *TCPOutlet) WriteEntry(e logger.Entry) error {
 type SyslogOutlet struct {
 	Formatter          EntryFormatter
 	RetryInterval      time.Duration
+	Facility           syslog.Priority
 	writer             *syslog.Writer
 	lastConnectAttempt time.Time
 }
@@ -142,7 +150,7 @@ func (o *SyslogOutlet) WriteEntry(entry logger.Entry) error {
 		if now.Sub(o.lastConnectAttempt) < o.RetryInterval {
 			return nil // not an error toward logger
 		}
-		o.writer, err = syslog.New(syslog.LOG_LOCAL0, "zrepl")
+		o.writer, err = syslog.New(o.Facility, "zrepl")
 		o.lastConnectAttempt = time.Now()
 		if err != nil {
 			o.writer = nil
@@ -7,7 +7,7 @@ import (

 type Logger = logger.Logger

-var DaemonCmd = &cli.Subcommand {
+var DaemonCmd = &cli.Subcommand{
 	Use:   "daemon",
 	Short: "run the zrepl daemon",
 	Run: func(subcommand *cli.Subcommand, args []string) error {
@@ -1,10 +1,11 @@
 package nethelpers

 import (
-	"github.com/pkg/errors"
 	"net"
 	"os"
 	"path/filepath"
+
+	"github.com/pkg/errors"
 )

 func PreparePrivateSockpath(sockpath string) error {
@@ -7,11 +7,12 @@ import (
 	"context"
 	"net"
 	"net/http/pprof"
+
+	"github.com/zrepl/zrepl/daemon/job"
 )

 type pprofServer struct {
 	cc       chan PprofServerControlMsg
-	state    PprofServerControlMsg
 	listener net.Listener
 }

@@ -63,7 +64,14 @@ outer:
 			mux.Handle("/debug/pprof/profile", http.HandlerFunc(pprof.Profile))
 			mux.Handle("/debug/pprof/symbol", http.HandlerFunc(pprof.Symbol))
 			mux.Handle("/debug/pprof/trace", http.HandlerFunc(pprof.Trace))
-			go http.Serve(s.listener, mux)
+			go func() {
+				err := http.Serve(s.listener, mux)
+				if ctx.Err() != nil {
+					return
+				} else if err != nil {
+					job.GetLogger(ctx).WithError(err).Error("pprof server serve error")
+				}
+			}()
 			continue
 		}

@@ -2,14 +2,17 @@ package daemon

 import (
 	"context"
+	"net"
+	"net/http"
+
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
+
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/job"
 	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/rpc/dataconn/frameconn"
 	"github.com/zrepl/zrepl/zfs"
-	"net"
-	"net/http"
 )

 type prometheusJob struct {
@@ -41,6 +44,8 @@ func (j *prometheusJob) Name() string { return jobNamePrometheus }

 func (j *prometheusJob) Status() *job.Status { return &job.Status{Type: job.TypeInternal} }

+func (j *prometheusJob) OwnedDatasetSubtreeRoot() (p *zfs.DatasetPath, ok bool) { return nil, false }
+
 func (j *prometheusJob) RegisterMetrics(registerer prometheus.Registerer) {}

 func (j *prometheusJob) Run(ctx context.Context) {
@@ -49,6 +54,10 @@ func (j *prometheusJob) Run(ctx context.Context) {
 		panic(err)
 	}

+	if err := frameconn.PrometheusRegister(prometheus.DefaultRegisterer); err != nil {
+		panic(err)
+	}
+
 	log := job.GetLogger(ctx)

 	l, err := net.Listen("tcp", j.listen)
@@ -56,17 +65,15 @@ func (j *prometheusJob) Run(ctx context.Context) {
 		log.WithError(err).Error("cannot listen")
 	}
 	go func() {
-		select {
-		case <-ctx.Done():
+		<-ctx.Done()
 		l.Close()
-		}
 	}()

 	mux := http.NewServeMux()
 	mux.Handle("/metrics", promhttp.Handler())

 	err = http.Serve(l, mux)
-	if err != nil {
+	if err != nil && ctx.Err() == nil {
 		log.WithError(err).Error("error while serving")
 	}

@@ -86,4 +93,3 @@ func (o prometheusJobOutlet) WriteEntry(entry logger.Entry) error {
 	prom.taskLogEntries.WithLabelValues(o.jobName, entry.Level.String()).Inc()
 	return nil
 }
-
@@ -3,28 +3,31 @@ package pruner
 import (
 	"context"
 	"fmt"
+	"sort"
+	"strings"
+	"sync"
+	"time"
+
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
+
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/logger"
 	"github.com/zrepl/zrepl/pruning"
-	"github.com/zrepl/zrepl/replication/pdu"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
 	"github.com/zrepl/zrepl/util/envconst"
-	"github.com/zrepl/zrepl/util/watchdog"
-	"net"
-	"sort"
-	"sync"
-	"time"
 )

-// Try to keep it compatible with gitub.com/zrepl/zrepl/replication.Endpoint
+// Try to keep it compatible with gitub.com/zrepl/zrepl/endpoint.Endpoint
 type History interface {
 	ReplicationCursor(ctx context.Context, req *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error)
+	ListFilesystems(ctx context.Context, req *pdu.ListFilesystemReq) (*pdu.ListFilesystemRes, error)
 }

+// Try to keep it compatible with gitub.com/zrepl/zrepl/endpoint.Endpoint
 type Target interface {
-	ListFilesystems(ctx context.Context) ([]*pdu.Filesystem, error)
-	ListFilesystemVersions(ctx context.Context, fs string) ([]*pdu.FilesystemVersion, error) // fix depS
+	ListFilesystems(ctx context.Context, req *pdu.ListFilesystemReq) (*pdu.ListFilesystemRes, error)
+	ListFilesystemVersions(ctx context.Context, req *pdu.ListFilesystemVersionsReq) (*pdu.ListFilesystemVersionsRes, error)
 	DestroySnapshots(ctx context.Context, req *pdu.DestroySnapshotsReq) (*pdu.DestroySnapshotsRes, error)
 }

@@ -58,19 +61,15 @@ type args struct {
 type Pruner struct {
 	args args

-	Progress watchdog.KeepAlive
-
 	mtx sync.RWMutex

 	state State

-	// State ErrWait|ErrPerm
-	sleepUntil time.Time
+	// State PlanErr
 	err error

 	// State Exec
-	prunePending   []*fs
-	pruneCompleted []*fs
+	execQueue *execQueue
 }

 type PrunerFactory struct {
@@ -81,17 +80,30 @@ type PrunerFactory struct {
 	promPruneSecs                  *prometheus.HistogramVec
 }

-func checkContainsKeep1(rules []pruning.KeepRule) error {
-	if len(rules) == 0 {
-		return nil //No keep rules means keep all - ok
+type LocalPrunerFactory struct {
+	keepRules     []pruning.KeepRule
+	retryWait     time.Duration
+	promPruneSecs *prometheus.HistogramVec
+}
+
+func NewLocalPrunerFactory(in config.PruningLocal, promPruneSecs *prometheus.HistogramVec) (*LocalPrunerFactory, error) {
+	rules, err := pruning.RulesFromConfig(in.Keep)
+	if err != nil {
+		return nil, errors.Wrap(err, "cannot build pruning rules")
 	}
-	for _, e := range rules {
-		switch e.(type) {
-		case *pruning.KeepLastN:
-			return nil
+	for _, r := range in.Keep {
+		if _, ok := r.Ret.(*config.PruneKeepNotReplicated); ok {
+			// rule NotReplicated  for a local pruner doesn't make sense
+			// because no replication happens with that job type
+			return nil, fmt.Errorf("single-site pruner cannot support `not_replicated` keep rule")
 		}
 	}
-	return errors.New("sender keep rules must contain last_n or be empty so that the last snapshot is definitely kept")
+	f := &LocalPrunerFactory{
+		keepRules:     rules,
+		retryWait:     envconst.Duration("ZREPL_PRUNER_RETRY_INTERVAL", 10*time.Second),
+		promPruneSecs: promPruneSecs,
+	}
+	return f, nil
 }

 func NewPrunerFactory(in config.PruningSenderReceiver, promPruneSecs *prometheus.HistogramVec) (*PrunerFactory, error) {
@@ -116,7 +128,7 @@ func NewPrunerFactory(in config.PruningSenderReceiver, promPruneSecs *prometheus
 	f := &PrunerFactory{
 		senderRules:                    keepRulesSender,
 		receiverRules:                  keepRulesReceiver,
-		retryWait: envconst.Duration("ZREPL_PRUNER_RETRY_INTERVAL", 4 * time.Second),
+		retryWait:                      envconst.Duration("ZREPL_PRUNER_RETRY_INTERVAL", 10*time.Second),
 		considerSnapAtCursorReplicated: considerSnapAtCursorReplicated,
 		promPruneSecs:                  promPruneSecs,
 	}
@@ -155,67 +167,55 @@ func (f *PrunerFactory) BuildReceiverPruner(ctx context.Context, target Target,
 	return p
 }

+func (f *LocalPrunerFactory) BuildLocalPruner(ctx context.Context, target Target, receiver History) *Pruner {
+	p := &Pruner{
+		args: args{
+			ctx,
+			target,
+			receiver,
+			f.keepRules,
+			f.retryWait,
+			false, // considerSnapAtCursorReplicated is not relevant for local pruning
+			f.promPruneSecs.WithLabelValues("local"),
+		},
+		state: Plan,
+	}
+	return p
+}
+
 //go:generate enumer -type=State
 type State int

 const (
 	Plan State = 1 << iota
-	PlanWait
+	PlanErr
 	Exec
-	ExecWait
-	ErrPerm
+	ExecErr
 	Done
 )

-func (s State) statefunc() state {
-	var statemap = map[State]state{
-		Plan:     statePlan,
-		PlanWait: statePlanWait,
-		Exec:     stateExec,
-		ExecWait: stateExecWait,
-		ErrPerm:  nil,
-		Done:     nil,
-	}
-	return statemap[s]
-}
-
-func (s State) IsTerminal() bool {
-	return s.statefunc() == nil
-}
-
-type updater func(func(*Pruner)) State
-type state func(args *args, u updater) state
+type updater func(func(*Pruner))

 func (p *Pruner) Prune() {
 	p.prune(p.args)
 }

 func (p *Pruner) prune(args args) {
-	s := p.state.statefunc()
-	for s != nil {
-		pre := p.state
-		s = s(&args, func(f func(*Pruner)) State {
+	u := func(f func(*Pruner)) {
 		p.mtx.Lock()
 		defer p.mtx.Unlock()
 		f(p)
-			return p.state
-		})
-		post := p.state
-		GetLogger(args.ctx).
-			WithField("transition", fmt.Sprintf("%s=>%s", pre, post)).
-			Debug("state transition")
-		if err := p.Error(); err != nil {
-			GetLogger(args.ctx).
-				WithError(p.err).
-				WithField("state", post.String()).
-				Error("entering error state after error")
-		}
 	}
+	// TODO support automatic retries
+	// It is advisable to merge this code with package replication/driver before
+	// That will likely require re-modelling struct fs like replication/driver.attempt,
+	// including figuring out how to resume a plan after being interrupted by network errors
+	// The non-retrying code in this package should move straight to replication/logic.
+	doOneAttempt(&args, u)
 }

 type Report struct {
 	State              string
-	SleepUntil time.Time
 	Error              string
 	Pending, Completed []FSReport
 }
@@ -223,7 +223,8 @@ type Report struct {
 type FSReport struct {
 	Filesystem                string
 	SnapshotList, DestroyList []SnapshotReport
-	Error string
+	SkipReason                FSSkipReason
+	LastError                 string
 }

 type SnapshotReport struct {
@@ -238,26 +239,12 @@ func (p *Pruner) Report() *Report {

 	r := Report{State: p.state.String()}

-	if p.state & PlanWait|ExecWait != 0 {
-		r.SleepUntil = p.sleepUntil
-	}
-	if p.state & PlanWait|ExecWait|ErrPerm != 0 {
 	if p.err != nil {
 		r.Error = p.err.Error()
 	}
-	}

-	if p.state & Plan|PlanWait == 0 {
-		return &r
-	}
-
-	r.Pending = make([]FSReport, len(p.prunePending))
-	for i, fs := range p.prunePending{
-		r.Pending[i] = fs.Report()
-	}
-	r.Completed = make([]FSReport, len(p.pruneCompleted))
-	for i, fs := range p.pruneCompleted{
-		r.Completed[i] = fs.Report()
+	if p.execQueue != nil {
+		r.Pending, r.Completed = p.execQueue.Report()
 	}

 	return &r
@@ -269,18 +256,17 @@ func (p *Pruner) State() State {
 	return p.state
 }

-func (p *Pruner) Error() error {
-	p.mtx.Lock()
-	defer p.mtx.Unlock()
-	if p.state & (PlanWait|ExecWait|ErrPerm) != 0 {
-		return p.err
-	}
-	return nil
-}
-
 type fs struct {
 	path string

+	// permanent error during planning
+	planErr        error
+	planErrContext string
+
+	// if != "", the fs was skipped for planning and the field
+	// contains the reason
+	skipReason FSSkipReason
+
 	// snapshots presented by target
 	// (type snapshot)
 	snaps []pruning.Snapshot
@@ -289,14 +275,21 @@ type fs struct {
 	destroyList []pruning.Snapshot

 	mtx sync.RWMutex
-	// for Plan
-	err error
+
+	// only during Exec state, also used by execQueue
+	execErrLast error
 }

-func (f *fs) Update(err error) {
-	f.mtx.Lock()
-	defer f.mtx.Unlock()
-	f.err = err
+type FSSkipReason string
+
+const (
+	NotSkipped                   = ""
+	SkipPlaceholder              = "filesystem is placeholder"
+	SkipNoCorrespondenceOnSender = "filesystem has no correspondence on sender"
+)
+
+func (r FSSkipReason) NotSkipped() bool {
+	return r == NotSkipped
 }

 func (f *fs) Report() FSReport {
@@ -305,8 +298,15 @@ func (f *fs) Report() FSReport {

 	r := FSReport{}
 	r.Filesystem = f.path
-	if f.err != nil {
-		r.Error = f.err.Error()
+	r.SkipReason = f.skipReason
+	if !r.SkipReason.NotSkipped() {
+		return r
+	}
+
+	if f.planErr != nil {
+		r.LastError = f.planErr.Error()
+	} else if f.execErrLast != nil {
+		r.LastError = f.execErrLast.Error()
 	}

 	r.SnapshotList = make([]SnapshotReport, len(f.snaps))
@@ -315,7 +315,7 @@ func (f *fs) Report() FSReport {
 	}

 	r.DestroyList = make([]SnapshotReport, len(f.destroyList))
-	for i, snap := range f.destroyList{
+	for i, snap := range f.destroyList {
 		r.DestroyList[i] = snap.(snapshot).Report()
 	}

@@ -344,62 +344,70 @@ func (s snapshot) Replicated() bool { return s.replicated }

 func (s snapshot) Date() time.Time { return s.date }

-func shouldRetry(e error) bool {
-	if neterr, ok := e.(net.Error); ok {
-		return neterr.Temporary()
-	}
-	return false
-}
-
-func onErr(u updater, e error) state {
-	return u(func(p *Pruner) {
-		p.err = e
-		if !shouldRetry(e) {
-			p.state = ErrPerm
-			return
-		}
-		switch p.state {
-		case Plan:
-			p.state = PlanWait
-		case Exec:
-			p.state = ExecWait
-		default:
-			panic(p.state)
-		}
-	}).statefunc()
-}
-
-func statePlan(a *args, u updater) state {
+func doOneAttempt(a *args, u updater) {

 	ctx, target, receiver := a.ctx, a.target, a.receiver

-	tfss, err := target.ListFilesystems(ctx)
+	sfssres, err := receiver.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
 	if err != nil {
-		return onErr(u, err)
+		u(func(p *Pruner) {
+			p.state = PlanErr
+			p.err = err
+		})
+		return
+	}
+	sfss := make(map[string]*pdu.Filesystem)
+	for _, sfs := range sfssres.GetFilesystems() {
+		sfss[sfs.GetPath()] = sfs
 	}

+	tfssres, err := target.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
+	if err != nil {
+		u(func(p *Pruner) {
+			p.state = PlanErr
+			p.err = err
+		})
+		return
+	}
+	tfss := tfssres.GetFilesystems()
+
 	pfss := make([]*fs, len(tfss))
-fsloop:
+tfss_loop:
 	for i, tfs := range tfss {

 		l := GetLogger(ctx).WithField("fs", tfs.Path)
 		l.Debug("plan filesystem")

-
 		pfs := &fs{
 			path: tfs.Path,
 		}
 		pfss[i] = pfs

-		tfsvs, err := target.ListFilesystemVersions(ctx, tfs.Path)
+		if tfs.GetIsPlaceholder() {
+			pfs.skipReason = SkipPlaceholder
+			l.WithField("skip_reason", pfs.skipReason).Debug("skipping filesystem")
+			continue
+		} else if sfs := sfss[tfs.GetPath()]; sfs == nil {
+			pfs.skipReason = SkipNoCorrespondenceOnSender
+			l.WithField("skip_reason", pfs.skipReason).WithField("sfs", sfs.GetPath()).Debug("skipping filesystem")
+			continue
+		}
+
+		pfsPlanErrAndLog := func(err error, message string) {
+			t := fmt.Sprintf("%T", err)
+			pfs.planErr = err
+			pfs.planErrContext = message
+			l.WithField("orig_err_type", t).WithError(err).Error(fmt.Sprintf("%s: plan error, skipping filesystem", message))
+		}
+
+		tfsvsres, err := target.ListFilesystemVersions(ctx, &pdu.ListFilesystemVersionsReq{Filesystem: tfs.Path})
 		if err != nil {
-			l.WithError(err).Error("cannot list filesystem versions")
-			if shouldRetry(err) {
-				return onErr(u, err)
-			}
-			pfs.err = err
-			continue fsloop
+			pfsPlanErrAndLog(err, "cannot list filesystem versions")
+			continue tfss_loop
 		}
+		tfsvs := tfsvsres.GetVersions()
+		// no progress here since we could run in a live-lock (must have used target AND receiver before progress)
+
 		pfs.snaps = make([]pruning.Snapshot, 0, len(tfsvs))

 		rcReq := &pdu.ReplicationCursorReq{
@@ -410,19 +418,14 @@ fsloop:
 		}
 		rc, err := receiver.ReplicationCursor(ctx, rcReq)
 		if err != nil {
-			l.WithError(err).Error("cannot get replication cursor")
-			if shouldRetry(err) {
-				return onErr(u, err)
+			pfsPlanErrAndLog(err, "cannot get replication cursor bookmark")
+			continue tfss_loop
 		}
-			pfs.err = err
-			continue fsloop
+		if rc.GetNotexist() {
+			err := errors.New("replication cursor bookmark does not exist (one successful replication is required before pruning works)")
+			pfsPlanErrAndLog(err, "")
+			continue tfss_loop
 		}
-		if rc.GetError() != "" {
-			l.WithField("reqErr", rc.GetError()).Error("cannot get replication cursor")
-			pfs.err = fmt.Errorf("%s", rc.GetError())
-			continue fsloop
-		}
-

 		// scan from older to newer, all snapshots older than cursor are interpreted as replicated
 		sort.Slice(tfsvs, func(i, j int) bool {
@@ -445,9 +448,9 @@ fsloop:
 			}
 			creation, err := tfsv.CreationAsTime()
 			if err != nil {
-				pfs.err = fmt.Errorf("%s%s has invalid creation date: %s", tfs, tfsv.RelName(), err)
-				l.WithError(pfs.err).Error("")
-				continue fsloop
+				err := fmt.Errorf("%s: %s", tfsv.RelName(), err)
+				pfsPlanErrAndLog(err, "fs version with invalid creation date")
+				continue tfss_loop
 			}
 			// note that we cannot use CreateTXG because target and receiver could be on different pools
 			atCursor := tfsv.Guid == rc.GetGuid()
@@ -459,47 +462,61 @@ fsloop:
 			})
 		}
 		if preCursor {
-			pfs.err = fmt.Errorf("replication cursor not found in prune target filesystem versions")
-			l.WithError(pfs.err).Error("")
-			continue fsloop
+			pfsPlanErrAndLog(fmt.Errorf("replication cursor not found in prune target filesystem versions"), "")
+			continue tfss_loop
 		}

 		// Apply prune rules
 		pfs.destroyList = pruning.PruneSnapshots(pfs.snaps, a.rules)
-
 	}

-	return u(func(pruner *Pruner) {
+	u(func(pruner *Pruner) {
+		pruner.execQueue = newExecQueue(len(pfss))
 		for _, pfs := range pfss {
-			if pfs.err != nil {
-				pruner.pruneCompleted = append(pruner.pruneCompleted, pfs)
-			} else {
-				pruner.prunePending = append(pruner.prunePending, pfs)
-			}
+			pruner.execQueue.Put(pfs, nil, false)
 		}
 		pruner.state = Exec
-	}).statefunc()
+	})
+
+	for {
+		var pfs *fs
+		u(func(pruner *Pruner) {
+			pfs = pruner.execQueue.Pop()
+		})
+		if pfs == nil {
+			break
+		}
+		doOneAttemptExec(a, u, pfs)
+	}
+
+	var rep *Report
+	{
+		// must not hold lock for report
+		var pruner *Pruner
+		u(func(p *Pruner) {
+			pruner = p
+		})
+		rep = pruner.Report()
+	}
+	u(func(p *Pruner) {
+		if len(rep.Pending) > 0 {
+			panic("queue should not have pending items at this point")
+		}
+		hadErr := false
+		for _, fsr := range rep.Completed {
+			hadErr = hadErr || fsr.SkipReason.NotSkipped() && fsr.LastError != ""
+		}
+		if hadErr {
+			p.state = ExecErr
+		} else {
+			p.state = Done
+		}
+	})
+
 }

-func stateExec(a *args, u updater) state {
-
-	var pfs *fs
-	state := u(func(pruner *Pruner) {
-		if len(pruner.prunePending) == 0 {
-			nextState := Done
-			for _, pfs := range pruner.pruneCompleted {
-				if pfs.err != nil {
-					nextState = ErrPerm
-				}
-			}
-			pruner.state = nextState
-			return
-		}
-		pfs = pruner.prunePending[0]
-	})
-	if state != Exec {
-		return state.statefunc()
-	}
+// attempts to exec pfs, puts it back into the queue with the result
+func doOneAttemptExec(a *args, u updater, pfs *fs) {

 	destroyList := make([]*pdu.FilesystemVersion, len(pfs.destroyList))
 	for i := range destroyList {
@@ -509,41 +526,57 @@ func stateExec(a *args, u updater) state {
 			WithField("destroy_snap", destroyList[i].Name).
 			Debug("policy destroys snapshot")
 	}
-	pfs.Update(nil)
 	req := pdu.DestroySnapshotsReq{
 		Filesystem: pfs.path,
 		Snapshots:  destroyList,
 	}
-	_, err := a.target.DestroySnapshots(a.ctx, &req)
-	pfs.Update(err)
-	if err != nil && shouldRetry(err) {
-		return onErr(u, err)
+	GetLogger(a.ctx).WithField("fs", pfs.path).Debug("destroying snapshots")
+	res, err := a.target.DestroySnapshots(a.ctx, &req)
+	if err != nil {
+		u(func(pruner *Pruner) {
+			pruner.execQueue.Put(pfs, err, false)
+		})
+		return
 	}
-	// if it's not retryable, treat is like as being done
-
-	return u(func(pruner *Pruner) {
-		pruner.pruneCompleted = append(pruner.pruneCompleted, pfs)
-		pruner.prunePending = pruner.prunePending[1:]
-	}).statefunc()
-}
-
-func stateExecWait(a *args, u updater) state {
-	return doWait(Exec, a, u)
-}
-
-func statePlanWait(a *args, u updater) state {
-	return doWait(Plan, a, u)
-}
-
-func doWait(goback State, a *args, u updater) state {
-	timer := time.NewTimer(a.retryWait)
-	defer timer.Stop()
-	select {
-	case <-timer.C:
-		return u(func(pruner *Pruner) {
-			pruner.state = goback
-		}).statefunc()
-	case <-a.ctx.Done():
-		return onErr(u, a.ctx.Err())
+	// check if all snapshots were destroyed
+	destroyResults := make(map[string]*pdu.DestroySnapshotRes)
+	for _, fsres := range res.Results {
+		destroyResults[fsres.Snapshot.Name] = fsres
+	}
+	err = nil
+	destroyFails := make([]*pdu.DestroySnapshotRes, 0)
+	for _, reqDestroy := range destroyList {
+		res, ok := destroyResults[reqDestroy.Name]
+		if !ok {
+			err = fmt.Errorf("missing destroy-result for %s", reqDestroy.RelName())
+			break
+		} else if res.Error != "" {
+			destroyFails = append(destroyFails, res)
+		}
+	}
+	if err == nil && len(destroyFails) > 0 {
+		names := make([]string, len(destroyFails))
+		pairs := make([]string, len(destroyFails))
+		allSame := true
+		lastMsg := destroyFails[0].Error
+		for i := 0; i < len(destroyFails); i++ {
+			allSame = allSame && destroyFails[i].Error == lastMsg
+			relname := destroyFails[i].Snapshot.RelName()
+			names[i] = relname
+			pairs[i] = fmt.Sprintf("(%s: %s)", relname, destroyFails[i].Error)
+		}
+		if allSame {
+			err = fmt.Errorf("destroys failed %s: %s",
+				strings.Join(names, ", "), lastMsg)
+		} else {
+			err = fmt.Errorf("destroys failed: %s", strings.Join(pairs, ", "))
+		}
+	}
+	u(func(pruner *Pruner) {
+		pruner.execQueue.Put(pfs, err, err == nil)
+	})
+	if err != nil {
+		GetLogger(a.ctx).WithError(err).Error("target could not destroy snapshots")
+		return
 	}
 }
@@ -0,0 +1,82 @@
+package pruner
+
+import (
+	"sort"
+	"strings"
+	"sync"
+)
+
+type execQueue struct {
+	mtx                sync.Mutex
+	pending, completed []*fs
+}
+
+func newExecQueue(cap int) *execQueue {
+	q := execQueue{
+		pending:   make([]*fs, 0, cap),
+		completed: make([]*fs, 0, cap),
+	}
+	return &q
+}
+
+func (q *execQueue) Report() (pending, completed []FSReport) {
+	q.mtx.Lock()
+	defer q.mtx.Unlock()
+
+	pending = make([]FSReport, len(q.pending))
+	for i, fs := range q.pending {
+		pending[i] = fs.Report()
+	}
+	completed = make([]FSReport, len(q.completed))
+	for i, fs := range q.completed {
+		completed[i] = fs.Report()
+	}
+
+	return pending, completed
+}
+
+func (q *execQueue) HasCompletedFSWithErrors() bool {
+	q.mtx.Lock()
+	defer q.mtx.Unlock()
+	for _, fs := range q.completed {
+		if fs.execErrLast != nil {
+			return true
+		}
+	}
+	return false
+}
+
+func (q *execQueue) Pop() *fs {
+	if len(q.pending) == 0 {
+		return nil
+	}
+	fs := q.pending[0]
+	q.pending = q.pending[1:]
+	return fs
+}
+
+func (q *execQueue) Put(fs *fs, err error, done bool) {
+	fs.mtx.Lock()
+	fs.execErrLast = err
+	if done || err != nil {
+		fs.mtx.Unlock()
+		q.mtx.Lock()
+		q.completed = append(q.completed, fs)
+		q.mtx.Unlock()
+		return
+	}
+	fs.mtx.Unlock()
+
+	q.mtx.Lock()
+	// inefficient priority q
+	q.pending = append(q.pending, fs)
+	sort.SliceStable(q.pending, func(i, j int) bool {
+		q.pending[i].mtx.Lock()
+		defer q.pending[i].mtx.Unlock()
+		q.pending[j].mtx.Lock()
+		defer q.pending[j].mtx.Unlock()
+		return strings.Compare(q.pending[i].path, q.pending[j].path) == -1
+	})
+	q.mtx.Unlock()
+
+}
@@ -1,211 +0,0 @@
-package pruner
-
-import (
-	"context"
-	"fmt"
-	"github.com/stretchr/testify/assert"
-	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/pruning"
-	"github.com/zrepl/zrepl/replication/pdu"
-	"net"
-	"testing"
-	"time"
-)
-
-type mockFS struct {
-	path  string
-	snaps []string
-}
-
-func (m *mockFS) Filesystem() *pdu.Filesystem {
-	return &pdu.Filesystem{
-		Path: m.path,
-	}
-}
-
-func (m *mockFS) FilesystemVersions() []*pdu.FilesystemVersion {
-	versions := make([]*pdu.FilesystemVersion, len(m.snaps))
-	for i, v := range m.snaps {
-		versions[i] = &pdu.FilesystemVersion{
-			Type:     pdu.FilesystemVersion_Snapshot,
-			Name:     v,
-			Creation: pdu.FilesystemVersionCreation(time.Unix(0, 0)),
-			Guid: uint64(i),
-		}
-	}
-	return versions
-}
-
-type mockTarget struct {
-	fss                []mockFS
-	destroyed          map[string][]string
-	listVersionsErrs   map[string][]error
-	listFilesystemsErr []error
-	destroyErrs        map[string][]error
-}
-
-func (t *mockTarget) ListFilesystems(ctx context.Context) ([]*pdu.Filesystem, error) {
-	if len(t.listFilesystemsErr) > 0 {
-		e := t.listFilesystemsErr[0]
-		t.listFilesystemsErr = t.listFilesystemsErr[1:]
-		return nil, e
-	}
-	fss := make([]*pdu.Filesystem, len(t.fss))
-	for i := range fss {
-		fss[i] = t.fss[i].Filesystem()
-	}
-	return fss, nil
-}
-
-func (t *mockTarget) ListFilesystemVersions(ctx context.Context, fs string) ([]*pdu.FilesystemVersion, error) {
-	if len(t.listVersionsErrs[fs]) != 0 {
-		e := t.listVersionsErrs[fs][0]
-		t.listVersionsErrs[fs] = t.listVersionsErrs[fs][1:]
-		return nil, e
-	}
-
-	for _, mfs := range t.fss {
-		if mfs.path != fs {
-			continue
-		}
-		return mfs.FilesystemVersions(), nil
-	}
-	return nil, fmt.Errorf("filesystem %s does not exist", fs)
-}
-
-func (t *mockTarget) DestroySnapshots(ctx context.Context, req *pdu.DestroySnapshotsReq) (*pdu.DestroySnapshotsRes, error) {
-	fs, snaps := req.Filesystem, req.Snapshots
-	if len(t.destroyErrs[fs]) != 0 {
-		e := t.destroyErrs[fs][0]
-		t.destroyErrs[fs] = t.destroyErrs[fs][1:]
-		return nil, e
-	}
-	destroyed := t.destroyed[fs]
-	res := make([]*pdu.DestroySnapshotRes, len(snaps))
-	for i, s := range snaps {
-		destroyed = append(destroyed, s.Name)
-		res[i] = &pdu.DestroySnapshotRes{Error: "", Snapshot: s}
-	}
-	t.destroyed[fs] = destroyed
-	return &pdu.DestroySnapshotsRes{Results: res}, nil
-}
-
-type mockCursor struct {
-	snapname string
-	guid uint64
-}
-type mockHistory struct {
-	errs map[string][]error
-	cursors map[string]*mockCursor
-}
-
-func (r *mockHistory) ReplicationCursor(ctx context.Context, req *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error) {
-	fs := req.Filesystem
-	if len(r.errs[fs]) > 0 {
-		e := r.errs[fs][0]
-		r.errs[fs] = r.errs[fs][1:]
-		return nil, e
-	}
-	return &pdu.ReplicationCursorRes{Result: &pdu.ReplicationCursorRes_Guid{Guid: 0}}, nil
-}
-
-type stubNetErr struct {
-	msg string
-	temporary, timeout bool
-}
-
-var _ net.Error = stubNetErr{}
-
-func (e stubNetErr) Error() string {
-	return e.msg
-}
-
-func (e stubNetErr) Temporary() bool { return e.temporary }
-
-func (e stubNetErr) Timeout() bool { return e.timeout }
-
-func TestPruner_Prune(t *testing.T) {
-
-	var _ net.Error = &net.OpError{} // we use it below
-	target := &mockTarget{
-		listFilesystemsErr: []error{
-			stubNetErr{msg: "fakerror0", temporary: true},
-		},
-		listVersionsErrs: map[string][]error{
-			"zroot/foo": {
-				stubNetErr{msg: "fakeerror1", temporary: true}, // should be classified as temporaty
-				stubNetErr{msg: "fakeerror2", temporary: true,},
-			},
-		},
-		destroyErrs: map[string][]error{
-			"zroot/foo": {
-				fmt.Errorf("permanent error"),
-			},
-			"zroot/bar": {
-				stubNetErr{msg: "fakeerror3", temporary: true},
-			},
-		},
-		destroyed: make(map[string][]string),
-		fss: []mockFS{
-			{
-				path: "zroot/foo",
-				snaps: []string{
-					"keep_a",
-					"keep_b",
-					"drop_c",
-					"keep_d",
-				},
-			},
-			{
-				path: "zroot/bar",
-				snaps: []string{
-					"keep_e",
-					"keep_f",
-					"drop_g",
-				},
-			},
-			{
-				path: "zroot/baz",
-				snaps: []string{
-					"keep_h",
-					"drop_i",
-				},
-			},
-		},
-	}
-	history := &mockHistory{
-		errs: map[string][]error{
-			"zroot/foo": {
-				stubNetErr{msg: "fakeerror4", temporary: true},
-			},
-			"zroot/baz": {
-				fmt.Errorf("permanent error2"),
-			},
-		},
-	}
-
-	keepRules := []pruning.KeepRule{pruning.MustKeepRegex("^keep")}
-
-	p := Pruner{
-		args: args{
-			ctx: WithLogger(context.Background(), logger.NewTestLogger(t)),
-			target: target,
-			receiver: history,
-			rules: keepRules,
-			retryWait: 10*time.Millisecond,
-		},
-		state: Plan,
-	}
-	p.Prune()
-
-	exp := map[string][]string{
-		"zroot/bar": {"drop_g"},
-		// drop_c is prohibited by failing destroy
-		// drop_i is prohibiteed by failing ReplicationCursor call
-	}
-
-	assert.Equal(t, exp, target.destroyed)
-
-	//assert.Equal(t, map[string][]error{}, target.listVersionsErrs, "retried")
-
-}
@@ -1,5 +1,6 @@
 // Code generated by "enumer -type=State"; DO NOT EDIT.

+//
 package pruner

 import (
@@ -7,19 +8,17 @@ import (
 )

 const (
-	_StateName_0 = "PlanPlanWait"
+	_StateName_0 = "PlanPlanErr"
 	_StateName_1 = "Exec"
-	_StateName_2 = "ExecWait"
-	_StateName_3 = "ErrPerm"
-	_StateName_4 = "Done"
+	_StateName_2 = "ExecErr"
+	_StateName_3 = "Done"
 )

 var (
-	_StateIndex_0 = [...]uint8{0, 4, 12}
+	_StateIndex_0 = [...]uint8{0, 4, 11}
 	_StateIndex_1 = [...]uint8{0, 4}
-	_StateIndex_2 = [...]uint8{0, 8}
-	_StateIndex_3 = [...]uint8{0, 7}
-	_StateIndex_4 = [...]uint8{0, 4}
+	_StateIndex_2 = [...]uint8{0, 7}
+	_StateIndex_3 = [...]uint8{0, 4}
 )

 func (i State) String() string {
@@ -33,22 +32,19 @@ func (i State) String() string {
 		return _StateName_2
 	case i == 16:
 		return _StateName_3
-	case i == 32:
-		return _StateName_4
 	default:
 		return fmt.Sprintf("State(%d)", i)
 	}
 }

-var _StateValues = []State{1, 2, 4, 8, 16, 32}
+var _StateValues = []State{1, 2, 4, 8, 16}

 var _StateNameToValueMap = map[string]State{
 	_StateName_0[0:4]:  1,
-	_StateName_0[4:12]: 2,
+	_StateName_0[4:11]: 2,
 	_StateName_1[0:4]:  4,
-	_StateName_2[0:8]:  8,
-	_StateName_3[0:7]:  16,
-	_StateName_4[0:4]:  32,
+	_StateName_2[0:7]:  8,
+	_StateName_3[0:4]:  16,
 }

 // StateString retrieves an enum value from the enum constants string name.
@@ -1,18 +1,20 @@
 package snapper

 import (
-	"github.com/zrepl/zrepl/config"
-	"github.com/pkg/errors"
-	"time"
 	"context"
-	"github.com/zrepl/zrepl/daemon/filters"
 	"fmt"
-	"github.com/zrepl/zrepl/zfs"
 	"sort"
-	"github.com/zrepl/zrepl/logger"
 	"sync"
-)
+	"time"

+	"github.com/pkg/errors"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/daemon/hooks"
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/zfs"
+)

 //go:generate stringer -type=SnapState
 type SnapState uint
@@ -24,18 +26,20 @@ const (
 	SnapError
 )

+// All fields protected by Snapper.mtx
 type snapProgress struct {
 	state SnapState

 	// SnapStarted, SnapDone, SnapError
 	name     string
 	startAt  time.Time
+	hookPlan *hooks.Plan

 	// SnapDone
 	doneAt time.Time

-	// SnapErr
-	err error
+	// SnapErr TODO disambiguate state
+	runResults hooks.PlanReport
 }

 type args struct {
@@ -44,7 +48,9 @@ type args struct {
 	prefix         string
 	interval       time.Duration
 	fsf            *filters.DatasetMapFilter
-	snapshotsTaken chan<-struct{}
+	snapshotsTaken chan<- struct{}
+	hooks          *hooks.List
+	dryRun         bool
 }

 type Snapper struct {
@@ -57,7 +63,7 @@ type Snapper struct {
 	lastInvocation time.Time

 	// valid for state Snapshotting
-	plan map[*zfs.DatasetPath]snapProgress
+	plan map[*zfs.DatasetPath]*snapProgress

 	// valid for state SyncUp and Waiting
 	sleepUntil time.Time
@@ -70,7 +76,8 @@ type Snapper struct {
 type State uint

 const (
-	SyncUp State = 1<<iota
+	SyncUp State = 1 << iota
+	SyncUpErrWait
 	Planning
 	Snapshotting
 	Waiting
@@ -81,6 +88,7 @@ const (
 func (s State) sf() state {
 	m := map[State]state{
 		SyncUp:        syncUp,
+		SyncUpErrWait: wait,
 		Planning:      plan,
 		Snapshotting:  snapshot,
 		Waiting:       wait,
@@ -120,10 +128,16 @@ func PeriodicFromConfig(g *config.Global, fsf *filters.DatasetMapFilter, in *con
 		return nil, errors.New("interval must be positive")
 	}

+	hookList, err := hooks.ListFromConfig(&in.Hooks)
+	if err != nil {
+		return nil, errors.Wrap(err, "hook config error")
+	}
+
 	args := args{
 		prefix:   in.Prefix,
 		interval: in.Interval,
 		fsf:      fsf,
+		hooks:    hookList,
 		// ctx and log is set in Run()
 	}

@@ -138,6 +152,7 @@ func (s *Snapper) Run(ctx context.Context, snapshotsTaken chan<- struct{}) {
 	s.args.snapshotsTaken = snapshotsTaken
 	s.args.ctx = ctx
 	s.args.log = getLogger(ctx)
+	s.args.dryRun = false // for future expansion

 	u := func(u func(*Snapper)) State {
 		s.mtx.Lock()
@@ -165,7 +180,16 @@ func (s *Snapper) Run(ctx context.Context, snapshotsTaken chan<- struct{}) {
 func onErr(err error, u updater) state {
 	return u(func(s *Snapper) {
 		s.err = err
+		preState := s.state
+		switch s.state {
+		case SyncUp:
+			s.state = SyncUpErrWait
+		case Planning:
+			fallthrough
+		case Snapshotting:
 			s.state = ErrorWait
+		}
+		s.args.log.WithError(err).WithField("pre_state", preState).WithField("post_state", s.state).Error("snapshotting error")
 	}).sf()
 }

@@ -177,7 +201,10 @@ func onMainCtxDone(ctx context.Context, u updater) state {
 }

 func syncUp(a args, u updater) state {
-	fss, err := listFSes(a.fsf)
+	u(func(snapper *Snapper) {
+		snapper.lastInvocation = time.Now()
+	})
+	fss, err := listFSes(a.ctx, a.fsf)
 	if err != nil {
 		return onErr(err, u)
 	}
@@ -185,10 +212,10 @@ func syncUp(a args, u updater) state {
 	if err != nil {
 		return onErr(err, u)
 	}
-	u(func(s *Snapper){
+	u(func(s *Snapper) {
 		s.sleepUntil = syncPoint
 	})
-	t := time.NewTimer(syncPoint.Sub(time.Now()))
+	t := time.NewTimer(time.Until(syncPoint))
 	defer t.Stop()
 	select {
 	case <-t.C:
@@ -204,29 +231,35 @@ func plan(a args, u updater) state {
 	u(func(snapper *Snapper) {
 		snapper.lastInvocation = time.Now()
 	})
-	fss, err := listFSes(a.fsf)
+	fss, err := listFSes(a.ctx, a.fsf)
 	if err != nil {
 		return onErr(err, u)
 	}

-	plan := make(map[*zfs.DatasetPath]snapProgress, len(fss))
+	plan := make(map[*zfs.DatasetPath]*snapProgress, len(fss))
 	for _, fs := range fss {
-		plan[fs] = snapProgress{state: SnapPending}
+		plan[fs] = &snapProgress{state: SnapPending}
 	}
 	return u(func(s *Snapper) {
 		s.state = Snapshotting
 		s.plan = plan
+		s.err = nil
 	}).sf()
 }

 func snapshot(a args, u updater) state {

-	var plan map[*zfs.DatasetPath]snapProgress
+	var plan map[*zfs.DatasetPath]*snapProgress
 	u(func(snapper *Snapper) {
 		plan = snapper.plan
 	})

-	hadErr := false
+	hookMatchCount := make(map[hooks.Hook]int, len(*a.hooks))
+	for _, h := range *a.hooks {
+		hookMatchCount[h] = 0
+	}
+
+	anyFsHadErr := false
 	// TODO channel programs -> allow a little jitter?
 	for fs, progress := range plan {
 		suffix := time.Now().In(time.UTC).Format("20060102_150405_000")
@@ -236,27 +269,71 @@ func snapshot(a args, u updater) state {
 			WithField("fs", fs.ToString()).
 			WithField("snap", snapname)

+		hookEnvExtra := hooks.Env{
+			hooks.EnvFS:       fs.ToString(),
+			hooks.EnvSnapshot: snapname,
+		}
+
+		jobCallback := hooks.NewCallbackHookForFilesystem("snapshot", fs, func(_ context.Context) (err error) {
+			l.Debug("create snapshot")
+			err = zfs.ZFSSnapshot(fs, snapname, false) // TODO propagagte context to ZFSSnapshot
+			if err != nil {
+				l.WithError(err).Error("cannot create snapshot")
+			}
+			return
+		})
+
+		fsHadErr := false
+		var planReport hooks.PlanReport
+		var plan *hooks.Plan
+		{
+			filteredHooks, err := a.hooks.CopyFilteredForFilesystem(fs)
+			if err != nil {
+				l.WithError(err).Error("unexpected filter error")
+				fsHadErr = true
+				goto updateFSState
+			}
+			// account for running hooks
+			for _, h := range filteredHooks {
+				hookMatchCount[h] = hookMatchCount[h] + 1
+			}
+
+			var planErr error
+			plan, planErr = hooks.NewPlan(&filteredHooks, hooks.PhaseSnapshot, jobCallback, hookEnvExtra)
+			if planErr != nil {
+				fsHadErr = true
+				l.WithError(planErr).Error("cannot create job hook plan")
+				goto updateFSState
+			}
+		}
 		u(func(snapper *Snapper) {
 			progress.name = snapname
 			progress.startAt = time.Now()
+			progress.hookPlan = plan
 			progress.state = SnapStarted
 		})
-
-		l.Debug("create snapshot")
-		err := zfs.ZFSSnapshot(fs, snapname, false)
-		if err != nil {
-			hadErr = true
-			l.WithError(err).Error("cannot create snapshot")
+		{
+			l := hooks.GetLogger(a.ctx).WithField("fs", fs.ToString()).WithField("snap", snapname)
+			l.WithField("report", plan.Report().String()).Debug("begin run job plan")
+			plan.Run(hooks.WithLogger(a.ctx, l), a.dryRun)
+			planReport = plan.Report()
+			fsHadErr = planReport.HadError() // not just fatal errors
+			if fsHadErr {
+				l.WithField("report", planReport.String()).Error("end run job plan with error")
+			} else {
+				l.WithField("report", planReport.String()).Info("end run job plan successful")
+			}
 		}
-		doneAt := time.Now()

+	updateFSState:
+		anyFsHadErr = anyFsHadErr || fsHadErr
 		u(func(snapper *Snapper) {
-			progress.doneAt = doneAt
+			progress.doneAt = time.Now()
 			progress.state = SnapDone
-			if err != nil {
+			if fsHadErr {
 				progress.state = SnapError
-				progress.err = err
 			}
+			progress.runResults = planReport
 		})
 	}

@@ -268,12 +345,26 @@ func snapshot(a args, u updater) state {
 		}
 	}

+	for h, mc := range hookMatchCount {
+		if mc == 0 {
+			hookIdx := -1
+			for idx, ah := range *a.hooks {
+				if ah == h {
+					hookIdx = idx
+					break
+				}
+			}
+			a.log.WithField("hook", h.String()).WithField("hook_number", hookIdx+1).Warn("hook did not match any snapshotted filesystems")
+		}
+	}
+
 	return u(func(snapper *Snapper) {
-		if hadErr {
+		if anyFsHadErr {
 			snapper.state = ErrorWait
 			snapper.err = errors.New("one or more snapshots could not be created, check logs for details")
 		} else {
 			snapper.state = Waiting
+			snapper.err = nil
 		}
 	}).sf()
 }
@@ -284,9 +375,15 @@ func wait(a args, u updater) state {
 		lastTick := snapper.lastInvocation
 		snapper.sleepUntil = lastTick.Add(a.interval)
 		sleepUntil = snapper.sleepUntil
+		log := a.log.WithField("sleep_until", sleepUntil).WithField("duration", a.interval)
+		logFunc := log.Debug
+		if snapper.state == ErrorWait || snapper.state == SyncUpErrWait {
+			logFunc = log.Error
+		}
+		logFunc("enter wait-state after error")
 	})

-	t := time.NewTimer(sleepUntil.Sub(time.Now()))
+	t := time.NewTimer(time.Until(sleepUntil))
 	defer t.Stop()

 	select {
@@ -299,8 +396,8 @@ func wait(a args, u updater) state {
 	}
 }

-func listFSes(mf *filters.DatasetMapFilter) (fss []*zfs.DatasetPath, err error) {
-	return zfs.ZFSListMapping(mf)
+func listFSes(ctx context.Context, mf *filters.DatasetMapFilter) (fss []*zfs.DatasetPath, err error) {
+	return zfs.ZFSListMapping(ctx, mf)
 }

 func findSyncPoint(log Logger, fss []*zfs.DatasetPath, prefix string, interval time.Duration) (syncPoint time.Time, err error) {
@@ -366,4 +463,3 @@ func findSyncPoint(log Logger, fss []*zfs.DatasetPath, prefix string, interval t
 	return snaptimes[0].time, nil

 }
-
@@ -3,6 +3,7 @@ package snapper
 import (
 	"context"
 	"fmt"
+
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/filters"
 )
@@ -17,12 +18,20 @@ type PeriodicOrManual struct {
 	s *Snapper
 }

-func (s *PeriodicOrManual) Run(ctx context.Context, wakeUpCommon chan <- struct{}) {
+func (s *PeriodicOrManual) Run(ctx context.Context, wakeUpCommon chan<- struct{}) {
 	if s.s != nil {
 		s.s.Run(ctx, wakeUpCommon)
 	}
 }

+// Returns nil if manual
+func (s *PeriodicOrManual) Report() *Report {
+	if s.s != nil {
+		return s.s.Report()
+	}
+	return nil
+}
+
 func FromConfig(g *config.Global, fsf *filters.DatasetMapFilter, in config.SnapshottingEnum) (*PeriodicOrManual, error) {
 	switch v := in.Ret.(type) {
 	case *config.SnapshottingPeriodic:
@@ -0,0 +1,118 @@
+package snapper
+
+import (
+	"fmt"
+	"sort"
+	"strings"
+	"time"
+
+	"github.com/zrepl/zrepl/daemon/hooks"
+)
+
+type Report struct {
+	State State
+	// valid in state SyncUp and Waiting
+	SleepUntil time.Time
+	// valid in state Err
+	Error string
+	// valid in state Snapshotting
+	Progress []*ReportFilesystem
+}
+
+type ReportFilesystem struct {
+	Path  string
+	State SnapState
+
+	// Valid in SnapStarted and later
+	SnapName      string
+	StartAt       time.Time
+	Hooks         string
+	HooksHadError bool
+
+	// Valid in SnapDone | SnapError
+	DoneAt time.Time
+}
+
+func errOrEmptyString(e error) string {
+	if e != nil {
+		return e.Error()
+	}
+	return ""
+}
+
+func (s *Snapper) Report() *Report {
+	s.mtx.Lock()
+	defer s.mtx.Unlock()
+
+	pReps := make([]*ReportFilesystem, 0, len(s.plan))
+	for fs, p := range s.plan {
+		var hooksStr string
+		var hooksHadError bool
+		if p.hookPlan != nil {
+			hr := p.hookPlan.Report()
+			// FIXME: technically this belongs into client
+			// but we can't serialize hooks.Step ATM
+			rightPad := func(str string, length int, pad string) string {
+				if len(str) > length {
+					return str[:length]
+				}
+				return str + strings.Repeat(pad, length-len(str))
+			}
+			hooksHadError = hr.HadError()
+			rows := make([][]string, len(hr))
+			const numCols = 4
+			lens := make([]int, numCols)
+			for i, e := range hr {
+				rows[i] = make([]string, numCols)
+				rows[i][0] = fmt.Sprintf("%d", i+1)
+				rows[i][1] = e.Status.String()
+				runTime := "..."
+				if e.Status != hooks.StepPending {
+					runTime = e.End.Sub(e.Begin).Round(time.Millisecond).String()
+				}
+				rows[i][2] = runTime
+				rows[i][3] = ""
+				if e.Report != nil {
+					rows[i][3] = e.Report.String()
+				}
+				for j, col := range lens {
+					if len(rows[i][j]) > col {
+						lens[j] = len(rows[i][j])
+					}
+				}
+			}
+			rowsFlat := make([]string, len(hr))
+			for i, r := range rows {
+				colsPadded := make([]string, len(r))
+				for j, c := range r[:len(r)-1] {
+					colsPadded[j] = rightPad(c, lens[j], " ")
+				}
+				colsPadded[len(r)-1] = r[len(r)-1]
+				rowsFlat[i] = strings.Join(colsPadded, " ")
+			}
+			hooksStr = strings.Join(rowsFlat, "\n")
+		}
+		pReps = append(pReps, &ReportFilesystem{
+			Path:          fs.ToString(),
+			State:         p.state,
+			SnapName:      p.name,
+			StartAt:       p.startAt,
+			DoneAt:        p.doneAt,
+			Hooks:         hooksStr,
+			HooksHadError: hooksHadError,
+		})
+	}
+
+	sort.Slice(pReps, func(i, j int) bool {
+		return strings.Compare(pReps[i].Path, pReps[j].Path) == -1
+	})
+
+	r := &Report{
+		State:      s.state,
+		SleepUntil: s.sleepUntil,
+		Error:      errOrEmptyString(s.err),
+		Progress:   pReps,
+	}
+
+	return r
+}
@@ -4,6 +4,16 @@ package snapper

 import "strconv"

+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[SnapPending-1]
+	_ = x[SnapStarted-2]
+	_ = x[SnapDone-4]
+	_ = x[SnapError-8]
+}
+
 const (
 	_SnapState_name_0 = "SnapPendingSnapStarted"
 	_SnapState_name_1 = "SnapDone"
@@ -4,16 +4,30 @@ package snapper

 import "strconv"

+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[SyncUp-1]
+	_ = x[SyncUpErrWait-2]
+	_ = x[Planning-4]
+	_ = x[Snapshotting-8]
+	_ = x[Waiting-16]
+	_ = x[ErrorWait-32]
+	_ = x[Stopped-64]
+}
+
 const (
-	_State_name_0 = "SyncUpPlanning"
-	_State_name_1 = "Snapshotting"
-	_State_name_2 = "Waiting"
-	_State_name_3 = "ErrorWait"
-	_State_name_4 = "Stopped"
+	_State_name_0 = "SyncUpSyncUpErrWait"
+	_State_name_1 = "Planning"
+	_State_name_2 = "Snapshotting"
+	_State_name_3 = "Waiting"
+	_State_name_4 = "ErrorWait"
+	_State_name_5 = "Stopped"
 )

 var (
-	_State_index_0 = [...]uint8{0, 6, 14}
+	_State_index_0 = [...]uint8{0, 6, 19}
 )

 func (i State) String() string {
@@ -29,6 +43,8 @@ func (i State) String() string {
 		return _State_name_3
 	case i == 32:
 		return _State_name_4
+	case i == 64:
+		return _State_name_5
 	default:
 		return "State(" + strconv.FormatInt(int64(i), 10) + ")"
 	}
@@ -1,25 +0,0 @@
-package streamrpcconfig
-
-import (
-	"github.com/problame/go-streamrpc"
-	"github.com/zrepl/zrepl/config"
-)
-
-func FromDaemonConfig(g *config.Global, in *config.RPCConfig) (*streamrpc.ConnConfig, error) {
-	conf := in
-	if conf == nil {
-		conf = g.RPC
-	}
-	srpcConf := &streamrpc.ConnConfig{
-		RxHeaderMaxLen:       conf.RxHeaderMaxLen,
-		RxStructuredMaxLen:   conf.RxStructuredMaxLen,
-		RxStreamMaxChunkSize: conf.RxStreamChunkMaxLen,
-		TxChunkSize:          conf.TxChunkSize,
-		Timeout: 			  conf.Timeout,
-		SendHeartbeatInterval: conf.SendHeartbeatInterval,
-	}
-	if err := srpcConf.Validate(); err != nil {
-		return nil, err
-	}
-	return srpcConf, nil
-}
@@ -1,30 +0,0 @@
-package connecter
-
-import (
-	"context"
-	"fmt"
-	"github.com/zrepl/zrepl/config"
-	"github.com/zrepl/zrepl/daemon/transport/serve"
-	"net"
-)
-
-type LocalConnecter struct {
-	listenerName string
-	clientIdentity string
-}
-
-func LocalConnecterFromConfig(in *config.LocalConnect) (*LocalConnecter, error) {
-	if in.ClientIdentity == "" {
-		return nil, fmt.Errorf("ClientIdentity must not be empty")
-	}
-	if in.ListenerName == "" {
-		return nil, fmt.Errorf("ListenerName must not be empty")
-	}
-	return &LocalConnecter{listenerName: in.ListenerName, clientIdentity: in.ClientIdentity}, nil
-}
-
-func (c *LocalConnecter) Connect(dialCtx context.Context) (conn net.Conn, err error) {
-	l := serve.GetLocalListener(c.listenerName)
-	return l.Connect(dialCtx, c.clientIdentity)
-}
-
@@ -1,84 +0,0 @@
-package connecter
-
-import (
-	"context"
-	"fmt"
-	"github.com/problame/go-streamrpc"
-	"github.com/zrepl/zrepl/config"
-	"github.com/zrepl/zrepl/daemon/streamrpcconfig"
-	"github.com/zrepl/zrepl/daemon/transport"
-	"net"
-	"time"
-)
-
-
-type HandshakeConnecter struct {
-	connecter streamrpc.Connecter
-}
-
-func (c HandshakeConnecter) Connect(ctx context.Context) (net.Conn, error) {
-	conn, err := c.connecter.Connect(ctx)
-	if err != nil {
-		return nil, err
-	}
-	dl, ok := ctx.Deadline()
-	if !ok {
-		dl = time.Now().Add(10 * time.Second) // FIXME constant
-	}
-	if err := transport.DoHandshakeCurrentVersion(conn, dl); err != nil {
-		conn.Close()
-		return nil, err
-	}
-	return conn, nil
-}
-
-
-
-func FromConfig(g *config.Global, in config.ConnectEnum) (*ClientFactory, error) {
-	var (
-		connecter            streamrpc.Connecter
-		errConnecter, errRPC error
-		connConf             *streamrpc.ConnConfig
-	)
-	switch v := in.Ret.(type) {
-	case *config.SSHStdinserverConnect:
-		connecter, errConnecter = SSHStdinserverConnecterFromConfig(v)
-		connConf, errRPC = streamrpcconfig.FromDaemonConfig(g, v.RPC)
-	case *config.TCPConnect:
-		connecter, errConnecter = TCPConnecterFromConfig(v)
-		connConf, errRPC = streamrpcconfig.FromDaemonConfig(g, v.RPC)
-	case *config.TLSConnect:
-		connecter, errConnecter = TLSConnecterFromConfig(v)
-		connConf, errRPC = streamrpcconfig.FromDaemonConfig(g, v.RPC)
-	case *config.LocalConnect:
-		connecter, errConnecter = LocalConnecterFromConfig(v)
-		connConf, errRPC = streamrpcconfig.FromDaemonConfig(g, v.RPC)
-	default:
-		panic(fmt.Sprintf("implementation error: unknown connecter type %T", v))
-	}
-
-	if errConnecter != nil {
-		return nil, errConnecter
-	}
-	if errRPC != nil {
-		return nil, errRPC
-	}
-
-	config := streamrpc.ClientConfig{ConnConfig: connConf}
-	if err := config.Validate(); err != nil {
-		return nil, err
-	}
-
-	connecter = HandshakeConnecter{connecter}
-
-	return &ClientFactory{connecter: connecter, config: &config}, nil
-}
-
-type ClientFactory struct {
-	connecter streamrpc.Connecter
-	config    *streamrpc.ClientConfig
-}
-
-func (f ClientFactory) NewClient() (*streamrpc.Client, error) {
-	return streamrpc.NewClient(f.connecter, f.config)
-}
@@ -1,136 +0,0 @@
-package transport
-
-import (
-	"bytes"
-	"fmt"
-	"io"
-	"net"
-	"strings"
-	"time"
-	"unicode/utf8"
-)
-
-type HandshakeMessage struct {
-	ProtocolVersion int
-	Extensions []string
-}
-
-func (m *HandshakeMessage) Encode() ([]byte, error) {
-	if m.ProtocolVersion <= 0 || m.ProtocolVersion > 9999 {
-		return nil, fmt.Errorf("protocol version must be in [1, 9999]")
-	}
-	if len(m.Extensions) >= 9999 {
-		return nil, fmt.Errorf("protocol only supports [0, 9999] extensions")
-	}
-	// EXTENSIONS is a count of subsequent \n separated lines that contain protocol extensions
-	var extensions strings.Builder
-	for i, ext := range m.Extensions {
-		if strings.ContainsAny(ext, "\n") {
-			return nil, fmt.Errorf("Extension #%d contains forbidden newline character", i)
-		}
-		if !utf8.ValidString(ext) {
-			return nil, fmt.Errorf("Extension #%d is not valid UTF-8", i)
-		}
-		extensions.WriteString(ext)
-		extensions.WriteString("\n")
-	}
-	withoutLen := fmt.Sprintf("ZREPL_ZFS_REPLICATION PROTOVERSION=%04d EXTENSIONS=%04d\n%s",
-		m.ProtocolVersion, len(m.Extensions), extensions.String())
-	withLen := fmt.Sprintf("%010d %s", len(withoutLen), withoutLen)
-	return []byte(withLen), nil
-}
-
-func (m *HandshakeMessage) DecodeReader(r io.Reader, maxLen int) error {
-	var lenAndSpace [11]byte
-	if _, err := io.ReadFull(r, lenAndSpace[:]); err != nil {
-		return err
-	}
-	if !utf8.Valid(lenAndSpace[:]) {
-		return fmt.Errorf("invalid start of handshake message: not valid UTF-8")
-	}
-	var followLen int
-	n, err := fmt.Sscanf(string(lenAndSpace[:]), "%010d ", &followLen)
-	if n != 1 || err != nil {
-		return fmt.Errorf("could not parse handshake message length")
-	}
-	if followLen > maxLen {
-		return fmt.Errorf("handshake message length exceeds max length (%d vs %d)",
-			followLen, maxLen)
-	}
-
-	var buf bytes.Buffer
-	_, err = io.Copy(&buf, io.LimitReader(r, int64(followLen)))
-	if err != nil {
-		return err
-	}
-
-	var (
-		protoVersion, extensionCount int
-	)
-	n, err = fmt.Fscanf(&buf, "ZREPL_ZFS_REPLICATION PROTOVERSION=%04d EXTENSIONS=%4d\n",
-		&protoVersion, &extensionCount)
-	if n != 2 || err != nil {
-		return fmt.Errorf("could not parse handshake message: %s", err)
-	}
-	if protoVersion < 1 {
-		return fmt.Errorf("invalid protocol version %q", protoVersion)
-	}
-	m.ProtocolVersion = protoVersion
-
-	if extensionCount < 0 {
-		return fmt.Errorf("invalid extension count %q", extensionCount)
-	}
-	if extensionCount == 0 {
-		if buf.Len() != 0 {
-			return fmt.Errorf("unexpected data trailing after header")
-		}
-		m.Extensions = nil
-		return nil
-	}
-	s := buf.String()
-	if strings.Count(s, "\n") != extensionCount {
-		return fmt.Errorf("inconsistent extension count: found %d, header says %d", len(m.Extensions), extensionCount)
-	}
-	exts := strings.Split(s, "\n")
-	if exts[len(exts)-1] != "" {
-		return fmt.Errorf("unexpected data trailing after last extension newline")
-	}
-	m.Extensions = exts[0:len(exts)-1]
-
-	return nil
-}
-
-func DoHandshakeCurrentVersion(conn net.Conn, deadline time.Time) error {
-	// current protocol version is hardcoded here
-	return DoHandshakeVersion(conn, deadline, 1)
-}
-
-func DoHandshakeVersion(conn net.Conn, deadline time.Time, version int) error {
-	ours := HandshakeMessage{
-		ProtocolVersion: version,
-		Extensions: nil,
-	}
-	hsb, err := ours.Encode()
-	if err != nil {
-		return fmt.Errorf("could not encode protocol banner: %s", err)
-	}
-
-	conn.SetDeadline(deadline)
-	_, err = io.Copy(conn, bytes.NewBuffer(hsb))
-	if err != nil {
-		return fmt.Errorf("could not send protocol banner: %s", err)
-	}
-
-	theirs := HandshakeMessage{}
-	if err := theirs.DecodeReader(conn, 16 * 4096); err != nil { // FIXME constant
-		return fmt.Errorf("could not decode protocol banner: %s", err)
-	}
-
-	if theirs.ProtocolVersion != ours.ProtocolVersion {
-		return fmt.Errorf("protocol versions do not match: ours is %d, theirs is %d",
-			ours.ProtocolVersion, theirs.ProtocolVersion)
-	}
-	// ignore extensions, we don't use them
-
-	return nil
-}
@@ -1,147 +0,0 @@
-package serve
-
-import (
-	"github.com/pkg/errors"
-	"github.com/zrepl/zrepl/config"
-	"github.com/zrepl/zrepl/daemon/transport"
-	"net"
-	"github.com/zrepl/zrepl/daemon/streamrpcconfig"
-	"github.com/problame/go-streamrpc"
-	"context"
-	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/zfs"
-	"time"
-)
-
-type contextKey int
-
-const contextKeyLog contextKey = 0
-
-type Logger = logger.Logger
-
-func WithLogger(ctx context.Context, log Logger) context.Context {
-	return context.WithValue(ctx, contextKeyLog, log)
-}
-
-func getLogger(ctx context.Context) Logger {
-	if log, ok := ctx.Value(contextKeyLog).(Logger); ok {
-		return log
-	}
-	return logger.NewNullLogger()
-}
-
-type AuthenticatedConn interface {
-	net.Conn
-	// ClientIdentity must be a string that satisfies ValidateClientIdentity
-	ClientIdentity() string
-}
-
-// A client identity must be a single component in a ZFS filesystem path
-func ValidateClientIdentity(in string) (err error) {
-	path, err := zfs.NewDatasetPath(in)
-	if err != nil {
-		return err
-	}
-	if path.Length() != 1 {
-		return errors.New("client identity must be a single path comonent (not empty, no '/')")
-	}
-	return nil
-}
-
-type authConn struct {
-	net.Conn
-	clientIdentity string
-}
-
-var _ AuthenticatedConn = authConn{}
-
-func (c authConn) ClientIdentity() string {
-	if err := ValidateClientIdentity(c.clientIdentity); err != nil {
-		panic(err)
-	}
-	return c.clientIdentity
-}
-
-// like net.Listener, but with an AuthenticatedConn instead of net.Conn
-type AuthenticatedListener interface {
-	Addr() (net.Addr)
-	Accept(ctx context.Context) (AuthenticatedConn, error)
-	Close() error
-}
-
-type ListenerFactory interface {
-	Listen() (AuthenticatedListener, error)
-}
-
-type HandshakeListenerFactory struct {
-	lf ListenerFactory
-}
-
-func (lf HandshakeListenerFactory) Listen() (AuthenticatedListener, error) {
-	l, err := lf.lf.Listen()
-	if err != nil {
-		return nil, err
-	}
-	return HandshakeListener{l}, nil
-}
-
-type HandshakeListener struct {
-	l AuthenticatedListener
-}
-
-func (l HandshakeListener) Addr() (net.Addr) { return l.l.Addr() }
-
-func (l HandshakeListener) Close() error { return l.l.Close() }
-
-func (l HandshakeListener) Accept(ctx context.Context) (AuthenticatedConn, error) {
-	conn, err := l.l.Accept(ctx)
-	if err != nil {
-		return nil, err
-	}
-	dl, ok := ctx.Deadline()
-	if !ok {
-		dl = time.Now().Add(10*time.Second) // FIXME constant
-	}
-	if err := transport.DoHandshakeCurrentVersion(conn, dl); err != nil {
-		conn.Close()
-		return nil, err
-	}
-	return conn, nil
-}
-
-func FromConfig(g *config.Global, in config.ServeEnum) (lf ListenerFactory, conf *streamrpc.ConnConfig, _ error) {
-
-	var (
-		lfError, rpcErr error
-	)
-	switch v := in.Ret.(type) {
-	case *config.TCPServe:
-		lf, lfError = TCPListenerFactoryFromConfig(g, v)
-		conf, rpcErr = streamrpcconfig.FromDaemonConfig(g, v.RPC)
-	case *config.TLSServe:
-		lf, lfError = TLSListenerFactoryFromConfig(g, v)
-		conf, rpcErr = streamrpcconfig.FromDaemonConfig(g, v.RPC)
-	case *config.StdinserverServer:
-		lf, lfError = MultiStdinserverListenerFactoryFromConfig(g, v)
-		conf, rpcErr = streamrpcconfig.FromDaemonConfig(g, v.RPC)
-	case *config.LocalServe:
-		lf, lfError = LocalListenerFactoryFromConfig(g, v)
-		conf, rpcErr = streamrpcconfig.FromDaemonConfig(g, v.RPC)
-	default:
-		return nil, nil, errors.Errorf("internal error: unknown serve type %T", v)
-	}
-
-	if lfError != nil {
-		return nil, nil, lfError
-	}
-	if rpcErr != nil {
-		return nil, nil, rpcErr
-	}
-
-	lf = HandshakeListenerFactory{lf}
-
-	return lf, conf, nil
-
-}
-
-
@@ -1,83 +0,0 @@
-package serve
-
-import (
-	"crypto/tls"
-	"crypto/x509"
-	"fmt"
-	"github.com/pkg/errors"
-	"github.com/zrepl/zrepl/config"
-	"github.com/zrepl/zrepl/tlsconf"
-	"net"
-	"time"
-	"context"
-)
-
-type TLSListenerFactory struct {
-	address          string
-	clientCA         *x509.CertPool
-	serverCert       tls.Certificate
-	handshakeTimeout time.Duration
-	clientCNs map[string]struct{}
-}
-
-func TLSListenerFactoryFromConfig(c *config.Global, in *config.TLSServe) (lf *TLSListenerFactory, err error) {
-	lf = &TLSListenerFactory{
-		address: in.Listen,
-		handshakeTimeout: in.HandshakeTimeout,
-	}
-
-	if in.Ca == "" || in.Cert == "" || in.Key == "" {
-		return nil, errors.New("fields 'ca', 'cert' and 'key'must be specified")
-	}
-
-	lf.clientCA, err = tlsconf.ParseCAFile(in.Ca)
-	if err != nil {
-		return nil, errors.Wrap(err, "cannot parse ca file")
-	}
-
-	lf.serverCert, err = tls.LoadX509KeyPair(in.Cert, in.Key)
-	if err != nil {
-		return nil, errors.Wrap(err, "cannot parse cer/key pair")
-	}
-
-	lf.clientCNs = make(map[string]struct{}, len(in.ClientCNs))
-	for i, cn := range in.ClientCNs {
-		if err := ValidateClientIdentity(cn); err != nil {
-			return nil, errors.Wrapf(err, "unsuitable client_cn #%d %q", i, cn)
-		}
-		// dupes are ok fr now
-		lf.clientCNs[cn] = struct{}{}
-	}
-
-	return lf, nil
-}
-
-func (f *TLSListenerFactory) Listen() (AuthenticatedListener, error) {
-	l, err := net.Listen("tcp", f.address)
-	if err != nil {
-		return nil, err
-	}
-	tl := tlsconf.NewClientAuthListener(l, f.clientCA, f.serverCert, f.handshakeTimeout)
-	return tlsAuthListener{tl, f.clientCNs}, nil
-}
-
-type tlsAuthListener struct {
-	*tlsconf.ClientAuthListener
-	clientCNs map[string]struct{}
-}
-
-func (l tlsAuthListener) Accept(ctx context.Context) (AuthenticatedConn, error) {
-	c, cn, err := l.ClientAuthListener.Accept()
-	if err != nil {
-		return nil, err
-	}
-	if _, ok := l.clientCNs[cn]; !ok {
-		if err := c.Close(); err != nil {
-			getLogger(ctx).WithError(err).Error("error closing connection with unauthorized common name")
-		}
-		return nil, fmt.Errorf("unauthorized client common name %q from %s", cn, c.RemoteAddr())
-	}
-	return authConn{c, cn}, nil
-}
-
-
@@ -0,0 +1,595 @@
+{
+  "annotations": {
+    "list": [
+      {
+        "$$hashKey": "object:3351",
+        "builtIn": 1,
+        "datasource": "-- Grafana --",
+        "enable": true,
+        "hide": true,
+        "iconColor": "rgba(0, 211, 255, 1)",
+        "name": "Annotations & Alerts",
+        "type": "dashboard"
+      }
+    ]
+  },
+  "editable": true,
+  "gnetId": null,
+  "graphTooltip": 1,
+  "id": 7,
+  "iteration": 1552746418826,
+  "links": [],
+  "panels": [
+    {
+      "content": "# zrepl Prometheus Metrics\n\nzrepl exposes Prometheus metrics and ships with this Grafana dashboard.\nThe exported metrics are suitable for health checks:\n\n* The log should generally be warning & error-free\n  * The `Log Messages that require attention` graph visualizes log message counts indicating problems.\n* The number of goroutines should not grow unboundedly over time.\n  * During replication, the number of goroutines can be way higher than during idle time.\n  * If the goroutine count grows with each replication, there is clearly a goroutine leak. Please open a bug report.\n* The sys memory consumption should not grow unboundedly over time.\n  * Note that the Go runtime pre-allocates some of its heap from the OS.\n  * zrepl actually uses much less memory than allocated from the OS.\n  * Since Go 1.11, Go pre-allocates more aggressively.\n* Monitor that some data is replicated, although that metric does not guarantee that replication was successful.\n\n**In general, note that the exported metrics are not stable unless declared otherwise.**",
+      "gridPos": {
+        "h": 9,
+        "w": 24,
+        "x": 0,
+        "y": 0
+      },
+      "id": 35,
+      "mode": "markdown",
+      "title": "Panel Title",
+      "type": "text"
+    },
+    {
+      "aliasColors": {},
+      "bars": false,
+      "dashLength": 10,
+      "dashes": false,
+      "datasource": null,
+      "fill": 1,
+      "gridPos": {
+        "h": 9,
+        "w": 12,
+        "x": 0,
+        "y": 9
+      },
+      "id": 15,
+      "legend": {
+        "avg": false,
+        "current": false,
+        "max": false,
+        "min": false,
+        "show": true,
+        "total": false,
+        "values": false
+      },
+      "lines": true,
+      "linewidth": 1,
+      "links": [],
+      "nullPointMode": "null",
+      "percentage": false,
+      "pointradius": 5,
+      "points": false,
+      "renderer": "flot",
+      "seriesOverrides": [],
+      "spaceLength": 10,
+      "stack": true,
+      "steppedLine": false,
+      "targets": [
+        {
+          "$$hashKey": "object:3436",
+          "expr": "up{job='$prom_job_name'}",
+          "format": "time_series",
+          "intervalFactor": 1,
+          "refId": "A"
+        }
+      ],
+      "thresholds": [],
+      "timeFrom": null,
+      "timeShift": null,
+      "title": "zrepl Instances Up",
+      "tooltip": {
+        "shared": true,
+        "sort": 0,
+        "value_type": "individual"
+      },
+      "type": "graph",
+      "xaxis": {
+        "buckets": null,
+        "mode": "time",
+        "name": null,
+        "show": true,
+        "values": []
+      },
+      "yaxes": [
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": "5",
+          "min": "0",
+          "show": true
+        },
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": null,
+          "show": true
+        }
+      ]
+    },
+    {
+      "aliasColors": {},
+      "bars": false,
+      "dashLength": 10,
+      "dashes": false,
+      "datasource": null,
+      "fill": 1,
+      "gridPos": {
+        "h": 9,
+        "w": 12,
+        "x": 12,
+        "y": 9
+      },
+      "id": 22,
+      "legend": {
+        "avg": false,
+        "current": false,
+        "max": false,
+        "min": false,
+        "show": true,
+        "total": false,
+        "values": false
+      },
+      "lines": true,
+      "linewidth": 1,
+      "links": [],
+      "nullPointMode": "null",
+      "percentage": false,
+      "pointradius": 5,
+      "points": false,
+      "renderer": "flot",
+      "seriesOverrides": [],
+      "spaceLength": 10,
+      "stack": false,
+      "steppedLine": false,
+      "targets": [
+        {
+          "expr": "increase(zrepl_daemon_log_entries{job='$prom_job_name',level=~'warn|error'}[$__interval])",
+          "format": "time_series",
+          "intervalFactor": 1,
+          "refId": "A"
+        }
+      ],
+      "thresholds": [],
+      "timeFrom": null,
+      "timeShift": null,
+      "title": "Log Messages that require attention",
+      "tooltip": {
+        "shared": true,
+        "sort": 0,
+        "value_type": "individual"
+      },
+      "type": "graph",
+      "xaxis": {
+        "buckets": null,
+        "mode": "time",
+        "name": null,
+        "show": true,
+        "values": []
+      },
+      "yaxes": [
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": "0",
+          "show": true
+        },
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": null,
+          "show": true
+        }
+      ]
+    },
+    {
+      "aliasColors": {},
+      "bars": false,
+      "dashLength": 10,
+      "dashes": false,
+      "datasource": null,
+      "fill": 1,
+      "gridPos": {
+        "h": 9,
+        "w": 12,
+        "x": 0,
+        "y": 18
+      },
+      "id": 33,
+      "legend": {
+        "avg": false,
+        "current": false,
+        "max": false,
+        "min": false,
+        "show": true,
+        "total": false,
+        "values": false
+      },
+      "lines": true,
+      "linewidth": 1,
+      "links": [],
+      "nullPointMode": "null",
+      "percentage": false,
+      "pointradius": 5,
+      "points": false,
+      "renderer": "flot",
+      "seriesOverrides": [
+        {
+          "alias": "/replicated bytes in last.*/",
+          "yaxis": 2
+        }
+      ],
+      "spaceLength": 10,
+      "stack": false,
+      "steppedLine": false,
+      "targets": [
+        {
+          "expr": "sum(rate(zrepl_replication_bytes_replicated{job='$prom_job_name'}[$__interval])) by (zrepl_job)",
+          "format": "time_series",
+          "hide": false,
+          "interval": "",
+          "intervalFactor": 1,
+          "legendFormat": "replication data rate zrepl_job={{zrepl_job}}",
+          "refId": "A"
+        },
+        {
+          "expr": "sum(increase(zrepl_replication_bytes_replicated{job='$prom_job_name'}[10m])) by (zrepl_job)",
+          "format": "time_series",
+          "hide": false,
+          "interval": "",
+          "intervalFactor": 1,
+          "legendFormat": "replicated bytes in last 10min zrepl_job={{zrepl_job}}",
+          "refId": "B"
+        }
+      ],
+      "thresholds": [],
+      "timeFrom": null,
+      "timeShift": null,
+      "title": "Replication Data Rate and Volume(integrates last 10min)",
+      "tooltip": {
+        "shared": true,
+        "sort": 0,
+        "value_type": "individual"
+      },
+      "type": "graph",
+      "xaxis": {
+        "buckets": null,
+        "mode": "time",
+        "name": null,
+        "show": true,
+        "values": []
+      },
+      "yaxes": [
+        {
+          "format": "Bps",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": null,
+          "show": true
+        },
+        {
+          "format": "bytes",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": null,
+          "show": true
+        }
+      ]
+    },
+    {
+      "aliasColors": {},
+      "bars": false,
+      "dashLength": 10,
+      "dashes": false,
+      "datasource": null,
+      "fill": 1,
+      "gridPos": {
+        "h": 9,
+        "w": 12,
+        "x": 12,
+        "y": 18
+      },
+      "id": 23,
+      "legend": {
+        "avg": false,
+        "current": false,
+        "max": false,
+        "min": false,
+        "show": true,
+        "total": false,
+        "values": false
+      },
+      "lines": true,
+      "linewidth": 1,
+      "links": [],
+      "nullPointMode": "null",
+      "percentage": false,
+      "pointradius": 5,
+      "points": false,
+      "renderer": "flot",
+      "seriesOverrides": [],
+      "spaceLength": 10,
+      "stack": false,
+      "steppedLine": false,
+      "targets": [
+        {
+          "expr": "sum(increase(zrepl_daemon_log_entries{job='$prom_job_name',zrepl_job=~\"^[^_].*\"}[$__interval])) by (instance,zrepl_job)",
+          "format": "time_series",
+          "intervalFactor": 1,
+          "refId": "A"
+        }
+      ],
+      "thresholds": [],
+      "timeFrom": null,
+      "timeShift": null,
+      "title": "Log Activity (without internal jobs)",
+      "tooltip": {
+        "shared": true,
+        "sort": 0,
+        "value_type": "individual"
+      },
+      "type": "graph",
+      "xaxis": {
+        "buckets": null,
+        "mode": "time",
+        "name": null,
+        "show": true,
+        "values": []
+      },
+      "yaxes": [
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": "0",
+          "show": true
+        },
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": null,
+          "show": true
+        }
+      ]
+    },
+    {
+      "aliasColors": {},
+      "bars": false,
+      "dashLength": 10,
+      "dashes": false,
+      "datasource": null,
+      "fill": 1,
+      "gridPos": {
+        "h": 9,
+        "w": 12,
+        "x": 0,
+        "y": 27
+      },
+      "id": 17,
+      "legend": {
+        "avg": false,
+        "current": false,
+        "max": false,
+        "min": false,
+        "show": true,
+        "total": false,
+        "values": false
+      },
+      "lines": true,
+      "linewidth": 1,
+      "links": [],
+      "nullPointMode": "null",
+      "percentage": false,
+      "pointradius": 5,
+      "points": false,
+      "renderer": "flot",
+      "seriesOverrides": [],
+      "spaceLength": 10,
+      "stack": false,
+      "steppedLine": false,
+      "targets": [
+        {
+          "$$hashKey": "object:3535",
+          "expr": "go_memstats_sys_bytes{job='$prom_job_name'}",
+          "format": "time_series",
+          "hide": false,
+          "intervalFactor": 1,
+          "refId": "A"
+        }
+      ],
+      "thresholds": [],
+      "timeFrom": null,
+      "timeShift": null,
+      "title": "Memory Allocated by the Go runtime from the OS (should not grow unboundedly)",
+      "tooltip": {
+        "shared": true,
+        "sort": 0,
+        "value_type": "individual"
+      },
+      "type": "graph",
+      "xaxis": {
+        "buckets": null,
+        "mode": "time",
+        "name": null,
+        "show": true,
+        "values": []
+      },
+      "yaxes": [
+        {
+          "format": "bytes",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": "0",
+          "show": true
+        },
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": null,
+          "show": true
+        }
+      ]
+    },
+    {
+      "aliasColors": {},
+      "bars": false,
+      "dashLength": 10,
+      "dashes": false,
+      "datasource": null,
+      "fill": 1,
+      "gridPos": {
+        "h": 9,
+        "w": 12,
+        "x": 12,
+        "y": 27
+      },
+      "id": 19,
+      "legend": {
+        "avg": false,
+        "current": false,
+        "max": false,
+        "min": false,
+        "show": true,
+        "total": false,
+        "values": false
+      },
+      "lines": true,
+      "linewidth": 1,
+      "links": [],
+      "nullPointMode": "null",
+      "percentage": false,
+      "pointradius": 5,
+      "points": false,
+      "renderer": "flot",
+      "seriesOverrides": [],
+      "spaceLength": 10,
+      "stack": false,
+      "steppedLine": false,
+      "targets": [
+        {
+          "expr": "go_goroutines{job='$prom_job_name'}",
+          "format": "time_series",
+          "intervalFactor": 1,
+          "refId": "A"
+        }
+      ],
+      "thresholds": [],
+      "timeFrom": null,
+      "timeShift": null,
+      "title": "number of goroutines (should not grow unboundedly)",
+      "tooltip": {
+        "shared": true,
+        "sort": 0,
+        "value_type": "individual"
+      },
+      "type": "graph",
+      "xaxis": {
+        "buckets": null,
+        "mode": "time",
+        "name": null,
+        "show": true,
+        "values": []
+      },
+      "yaxes": [
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": "0",
+          "show": true
+        },
+        {
+          "format": "short",
+          "label": null,
+          "logBase": 1,
+          "max": null,
+          "min": null,
+          "show": true
+        }
+      ]
+    }
+  ],
+  "refresh": "1m",
+  "schemaVersion": 16,
+  "style": "dark",
+  "tags": [],
+  "templating": {
+    "list": [
+      {
+        "allValue": null,
+        "current": {
+          "text": "zrepl",
+          "value": "zrepl"
+        },
+        "datasource": "prometheus",
+        "hide": 0,
+        "includeAll": false,
+        "label": "Prometheus Job Name",
+        "multi": false,
+        "name": "prom_job_name",
+        "options": [],
+        "query": "label_values(up, job)",
+        "refresh": 1,
+        "regex": "",
+        "sort": 1,
+        "tagValuesQuery": "",
+        "tags": [],
+        "tagsQuery": "",
+        "type": "query",
+        "useTags": false
+      }
+    ]
+  },
+  "time": {
+    "from": "now-2d",
+    "to": "now"
+  },
+  "timepicker": {
+    "refresh_intervals": [
+      "5s",
+      "10s",
+      "30s",
+      "1m",
+      "5m",
+      "15m",
+      "30m",
+      "1h",
+      "2h",
+      "1d"
+    ],
+    "time_options": [
+      "5m",
+      "15m",
+      "1h",
+      "6h",
+      "12h",
+      "24h",
+      "2d",
+      "7d",
+      "30d"
+    ]
+  },
+  "timezone": "",
+  "title": "zrepl 0.1",
+  "uid": "xTljn4qmk",
+  "version": 6
+}
@@ -0,0 +1,43 @@
+[Unit]
+Description=zrepl daemon
+Documentation=https://zrepl.github.io
+
+[Service]
+Type=simple
+ExecStart=/usr/local/bin/zrepl --config /etc/zrepl/zrepl.yml daemon
+RuntimeDirectory=zrepl
+RuntimeDirectoryMode=0700
+
+ProtectSystem=strict
+#PrivateDevices=yes # TODO ZFS needs access to /dev/zfs, could we limit this?
+ProtectKernelTunables=yes
+ProtectControlGroups=yes
+PrivateTmp=yes
+#PrivateUsers=yes # TODO Does not work, why?
+ProtectKernelModules=true
+RestrictAddressFamilies=AF_UNIX AF_INET AF_INET6
+RestrictNamespaces=true
+RestrictRealtime=yes
+SystemCallArchitectures=native
+
+# BEGIN ProtectHome
+# DEBIAN STRETCH
+ProtectHome=read-only
+# FEDORA 28 / 29
+# ProtectHome=tmpfs
+# END ProtectHome
+
+# BEGIN SystemCallFilter
+## BEGIN DEBIAN STRETCH
+SystemCallFilter=~ @mount @cpu-emulation @keyring @module @obsolete @privileged @raw-io @debug @clock @resources
+## END DEBIAN STRETCH
+## BEGIN FEDORA 28/29 
+## Syscall blacklist (should be fairly stable)
+#SystemCallFilter=~ @mount @aio @cpu-emulation @keyring @memlock @module @obsolete @privileged @raw-io @reboot @setuid @swap @sync @timer @debug @clock @chown @resources
+## Syscall whitelist (not sure how stable)
+#SystemCallFilter=@default @file-system @process @basic-io @ipc @network-io @signal @io-event brk mprotect sched_getaffinity ioctl getrandom
+## END END FEDORA 28/29
+# END SystemCallFilter
+
+[Install]
+WantedBy=multi-user.target
@@ -1,8 +1,11 @@
+.. include:: global.rst.inc
+
 .. |break_config| replace:: **[CONFIG]**
 .. |break| replace:: **[BREAK]**
 .. |bugfix| replace:: [BUG]
 .. |docs| replace:: [DOCS]
 .. |feature| replace:: [FEATURE]
+.. |mig| replace:: **[MIGRATION]**

 .. _changelog:

@@ -19,10 +22,52 @@ We use the following annotations for classifying changes:
 * |break| Change that breaks interoperability or persistent state representation with previous releases.
  As a package maintainer, make sure to warn your users about config breakage somehow.
  Note that even updating the package on both sides might not be sufficient, e.g. if persistent state needs to be migrated to a new format.
+* |mig| Migration that must be run by the user.
 * |feature| Change that introduces new functionality.
 * |bugfix| Change that fixes a bug, no regressions or incompatibilities expected.
 * |docs| Change to the documentation.

+0.2
+---
+
+* |feature| :ref:`Pre- and Post-Snapshot Hooks <job-snapshotting-hooks>`
+  with built-in support for MySQL and Postgres checkpointing
+  as well as custom scripts (thanks, `@overhacked <https://github.com/overhacked>`_!)
+* |feature| Use ``zfs destroy pool/fs@snap1,snap2,...`` CLI feature if available
+* |feature| Linux ARM64 Docker build support & binary builds
+* |feature| ``zrepl status`` now displays snapshotting reports
+* |feature| ``zrepl status --job <JOBNAME>`` filter flag
+* |bugfix| i386 build
+* |bugfix| early validation of host:port tuples in config
+* |bugfix| ``zrepl status`` now supports ``TERM=screen`` (tmux on FreeBSD / FreeNAS)
+* |bugfix| ignore *connection reset by peer* errors when shutting down connections
+* |bugfix| correct error messages when receive-side pool or ``root_fs`` dataset is not imported
+* |bugfix| fail fast for misconfigured local transport
+* |bugfix| race condition in replication report generation would crash the daemon when running ``zrepl status``
+* |bugfix| rpc goroutine leak in ``push`` mode if zfs recv fails on the ``sink`` side
+* [MAINTAINER NOTICE] Go modules for dependency management both inside and outside of GOPATH
+  (``lazy.sh`` and ``Makefile`` force ``GO111MODULE=on``)
+* [MAINTAINER NOTICE] ``make platformtest`` target to check zrepl's ZFS abstractions (screen scraping, etc.).
+  These tests only work on a system with ZFS installed, and must be run as root because they create a file-backed pool for each test case.
+  The pool name ``zreplplatformtest`` is reserved for this use case.
+  Only run ``make platformtest`` on test systems, e.g. a FreeBSD VM image.
+
+.. NOTE::
+   |  zrepl is a spare-time project primarily developed by `Christian Schwarz <https://cschwarz.com>`_.
+   |  You can support maintenance and feature development through one of the following services:
+   |  |Donate via Patreon| |Donate via Liberapay| |Donate via PayPal|
+   |  Note that PayPal processing fees are relatively high for small donations.
+   |  For SEPA wire transfer and **commerical support**, please `contact Christian directly <https://cschwarz.com>`_.
+
+
+0.1.1
+-----
+
+* |bugfix| :issue:`162` :commit:`d6304f4` : fix I/O timeout errors on variable receive rate
+
+  * A significant reduction or sudden stall of the receive rate (e.g. recv pool has other I/O to do)
+    would cause a ``writev I/O timeout`` error after approximately ten seconds.
+
 0.1
 ---

@@ -40,38 +85,59 @@ It breaks both configuration and transport format, and thus requires manual inte
 Notes to Package Maintainers
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

+* Notify users about config changes and migrations (see changes attributed with |break| and |mig| below)
 * If the daemon crashes, the stack trace produced by the Go runtime and possibly diagnostic output of zrepl will be written to stderr.
  This behavior is independent from the ``stdout`` outlet type.
-  Please make sure the stderr output of the daemon is captured to a file.
-  Rotation should not be necessary because stderr is not written to under normal circumstances.
+  Please make sure the stderr output of the daemon is captured somewhere.
  To conserve precious stack traces, make sure that multiple service restarts do not directly discard previous stderr output.
 * Make it obvious for users how to set the ``GOTRACEBACK`` environment variable to ``GOTRACEBACK=crash``.
  This functionality will cause SIGABRT on panics and can be used to capture a coredump of the panicking process.
  To that extend, make sure that your package build system, your OS's coredump collection and the Go delve debugger work together.
  Use your build system to package the Go program in `this tutorial on Go coredumps and the delve debugger <https://rakyll.org/coredumps/>`_ , and make sure the symbol resolution etc. work on coredumps captured from the binary produced by your build system. (Special focus on symbol stripping, etc.)
+* Consider using the ``zrepl configcheck`` subcommand in startup scripts to abort a restart that would fail due to an invalid config.

 Changes
 ~~~~~~~

+* |break| |mig| Placeholder property representation changed
+
+  * The :ref:`placeholder property <replication-placeholder-property>` now uses ``on|off`` as values
+    instead of hashes of the dataset path. This permits renames of the sink filesystem without
+    updating all placeholder properties.
+  * Relevant for 0.0.X-0.1-rc* to 0.1 migrations
+  * Make sure your config is valid with ``zrepl configcheck``
+  * Run ``zrepl migrate 0.0.X:0.1:placeholder``
+
 * |feature| :issue:`55` : Push replication (see :ref:`push job <job-push>` and :ref:`sink job <job-sink>`)
 * |feature| :ref:`TCP Transport <transport-tcp>`
 * |feature| :ref:`TCP + TLS client authentication transport <transport-tcp+tlsclientauth>`
-* |feature| :issue:`78` TODO MERGE COMMIT Replication protocol rewrite
+* |feature| :issue:`111`: RPC protocol rewrite

-  * Uses ``github.com/problame/go-streamrpc`` for RPC layer
-  * |break| zrepl 0.1 and restart on both sides of a replication setup is required
-  * |feature| :issue:`83`:  Improved error handling of network-level errors (zrepl retries instead of failing the entire job)
-  * |bugfix| :issue:`75` :issue:`81`: use connection timeouts and protocol-level heartbeats
-  * |break| |break_config|: mappings are no longer supported
+  * |break| Protocol breakage; Update and restart of all zrepl daemons is required.
+  * Use `gRPC <https://grpc.io/>`_ for control RPCs and a custom protocol for bulk data transfer.
+  * Automatic retries for network-temporary errors
+
+    * Limited to errors during replication for this release.
+      Addresses the common problem of ISP-forced reconnection at night, but will become
+      way more useful with resumable send & recv support.
+      Pruning errors are handled per FS, i.e., a prune RPC is attempted at least once per FS.
+
+* |feature| Proper timeout handling for the :ref:`SSH transport <transport-ssh+stdinserver>`
+
+  * |break| Requires Go 1.11 or later.
+  
+* |break| |break_config|: mappings are no longer supported

  * Receiving sides (``pull`` and ``sink`` job) specify a single ``root_fs``.
    Received filesystems are then stored *per client* in ``${root_fs}/${client_identity}``.
+    See :ref:`job-overview` for details.

 * |feature| |break| |break_config| Manual snapshotting + triggering of replication

  * |feature| :issue:`69`: include manually created snapshots in replication
  * |break_config| ``manual`` and ``periodic`` :ref:`snapshotting types <job-snapshotting-spec>`
-  * |feature| ``zrepl wakeup JOB`` subcommand to trigger *just* replication
+  * |feature| ``zrepl signal wakeup JOB`` subcommand to trigger replication + pruning
+  * |feature| ``zrepl signal reset JOB`` subcommand to abort current replication + pruning

 * |feature| |break| |break_config| New pruning system

@@ -85,16 +151,22 @@ Changes
  * |feature| |break| Bookmark pruning is no longer necessary

    * Per filesystem, zrepl creates a single bookmark (``#zrepl_replication_cursor``) and moves it forward with the most recent successfully replicated snapshot on the receiving side.
-    * Old bookmarks created prior to zrepl 0.1 (named like their corresponding snapshot) must be deleted manually.
+    * Old bookmarks created by prior versions of zrepl (named like their corresponding snapshot) must be deleted manually.
    * |break_config| ``keep_bookmarks`` parameter of the ``grid`` keep rule has been removed

 * |feature| ``zrepl status`` for live-updating replication progress (it's really cool!)
+* |feature| :ref:`Snapshot- & pruning-only job type <job-snap>` (for local snapshot management)
 * |feature| :issue:`67`: Expose `Prometheus <https://prometheus.io>`_ metrics via HTTP (:ref:`config docs <monitoring-prometheus>`)

+  * Compatible Grafana dashboard shipping in ``dist/grafana``
+
 * |break_config| Logging outlet types must be specified using the ``type`` instead of ``outlet`` key
 * |break| :issue:`53`: CLI: ``zrepl control *`` subcommands have been made direct subcommands of ``zrepl *``
 * |bugfix| Goroutine leak on ssh transport connection timeouts
 * |bugfix| :issue:`81` :issue:`77` : handle failed accepts correctly (``source`` job)
+* |bugfix| :issue:`100`: fix incompatibility with ZoL 0.8
+* |feature| :issue:`115`: logging: configurable syslog facility
+* |feature| Systemd unit file in ``dist/systemd``

 .. |lastrelease| replace:: 0.0.3

@@ -104,8 +176,3 @@ Previous Releases
 .. NOTE::
    Due to limitations in our documentation system, we only show the changelog since the last release and the time this documentation is built.
    For the changelog of previous releases, use the version selection in the hosted version of these docs at `zrepl.github.io <https://zrepl.github.io>`_.
-    
-    
-
-
-W
@@ -7,10 +7,11 @@ Configuration

 .. toctree::

-    configuration/preface
+    configuration/overview
    configuration/jobs
    configuration/transports
    configuration/filter_syntax
+    configuration/snapshotting
    configuration/prune
    configuration/logging
    configuration/monitoring
@@ -1,159 +1,9 @@
 .. include:: ../global.rst.inc

-.. |serve-transport| replace:: :ref:`serve specification<transport>`
-.. |connect-transport| replace:: :ref:`connect specification<transport>`
-.. |snapshotting-spec| replace:: :ref:`snapshotting specification <job-snapshotting-spec>`
-.. |pruning-spec| replace:: :ref:`pruning specification <prune>`
-.. |filter-spec| replace:: :ref:`filter specification<pattern-filter>`
-
 .. _job:

-Job Types & Replication
-=======================
-
-Overview & Terminology
----------------------
-
-A *job* is the unit of activity tracked by the zrepl daemon and configured in the |mainconfig|.
-Every job has a unique ``name``, a ``type`` and type-dependent fields which are documented on this page.
-
-Replication always happens between a pair of jobs: one is the **active side**, and one the **passive side**.
-The active side executes the replication logic whereas the passive side responds to requests after checking the active side's permissions.
-For communication, the active side connects to the passive side using a :ref:`transport <transport>` and starts issuing remote procedure calls (RPCs).
-
-The following table shows how different job types can be combined to achieve both push and pull mode setups:
-
-+-----------------------+--------------+----------------------------------+-----------------------------------------------+
-| Setup name            | active side  | passive side                     | use case                                      |
-+=======================+==============+==================================+===============================================+
-| Push mode             | ``push``     | ``sink``                         | * Laptop backup                               |
-|                       |              |                                  | * NAS behind NAT to offsite                   |
-+-----------------------+--------------+----------------------------------+-----------------------------------------------+
-| Pull mode             | ``pull``     | ``source``                       | * Central backup-server for many nodes        |
-|                       |              |                                  | * Remote server to NAS behind NAT             |
-+-----------------------+--------------+----------------------------------+-----------------------------------------------+
-| Local replication     | | ``push`` + ``sink`` in one config             | * Backup FreeBSD boot pool                    |
-|                       | | with :ref:`local transport <transport-local>` |                                               |
-+-----------------------+--------------+----------------------------------+-----------------------------------------------+
-
-How the Active Side Works
-~~~~~~~~~~~~~~~~~~~~~~~~~
-
-The active side (:ref:`push <job-push>` and :ref:`pull <job-pull>` job) executes the replication and pruning logic:
-
-* Wakeup because of finished snapshotting (``push`` job) or pull interval ticker (``pull`` job).
-* Connect to the corresponding passive side using a :ref:`transport <transport>` and instantiate an RPC client.
-* Replicate data from the sending to the receiving side.
-* Prune on sender & receiver.
-
-.. TIP::
-  The progress of the active side can be watched live using the ``zrepl status`` subcommand.
-
-How the Passive Side Works
-~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-The passive side (:ref:`sink <job-sink>` and :ref:`source <job-source>`) waits for connections from the corresponding active side,
-using the transport listener type specified in the ``serve`` field of the job configuration.
-Each transport listener provides a client's identity to the passive side job.
-It uses the client identity for access control:
-
-* The ``sink`` job only allows pushes to those ZFS filesystems to the active side that are located below ``root_fs/${client_identity}``.
-* The ``source`` job has a whitelist of client identities that are allowed pull access.
-
-.. TIP::
-   The implementation of the ``sink`` job requires that the connecting client identities be a valid ZFS filesystem name components.
-
-How Replication Works
-~~~~~~~~~~~~~~~~~~~~~
-
-One of the major design goals of the replication module is to avoid any duplication of the nontrivial logic.
-As such, the code works on abstract senders and receiver **endpoints**, where typically one will be implemented by a local program object and the other is an RPC client instance.
-Regardless of push- or pull-style setup, the logic executes on the active side, i.e. in the ``push`` or ``pull`` job.
-
-The following steps take place during replication and can be monitored using the ``zrepl status`` subcommand:
-
-* Plan the replication:
-
-  * Compare sender and receiver filesystem snapshots
-  * Build the **replication plan**
-
-    * Per filesystem, compute a diff between sender and receiver snapshots
-    * Build a list of replication steps
-
-      * If possible, use incremental sends (``zfs send -i``)
-      * Otherwise, use full send of most recent snapshot on sender
-      * Give up on filesystems that cannot be replicated without data loss
-
-  * Retry on errors that are likely temporary (i.e. network failures).
-  * Give up on filesystems where a permanent error was received over RPC.
-
-* Execute the plan
-
-  * Perform replication steps in the following order:
-    Among all filesystems with pending replication steps, pick the filesystem whose next replication step's snapshot is the oldest.
-  * After a successful replication step, update the replication cursor bookmark (see below)
-   
-The idea behind the execution order of replication steps is that if the sender snapshots all filesystems simultaneously at fixed intervals, the receiver will have all filesystems snapshotted at time ``T1`` before the first snapshot at ``T2 = T1 + $interval`` is replicated.
-
-.. _replication-cursor-bookmark:
-
-The **replication cursor bookmark** ``#zrepl_replication_cursor`` is kept per filesystem on the sending side of a replication setup:
-It is a bookmark of the most recent successfully replicated snapshot to the receiving side.
-It is is used by the :ref:`not_replicated <prune-keep-not-replicated>` keep rule to identify all snapshots that have not yet been replicated to the receiving side.
-Regardless of whether that keep rule is used, the bookmark ensures that replication can always continue incrementally.
-
-.. ATTENTION::
-
-    Currently, zrepl does not replicate filesystem properties.
-    Whe receiving a filesystem, it is never mounted (`-u` flag)  and `mountpoint=none` is set.
-    This is temporary and being worked on :issue:`24`.
-
-
-.. _job-snapshotting-spec:
-
-Taking Snaphots
---------------
-
-The ``push`` and ``source`` jobs can automatically take periodic snapshots of the filesystems matched by the ``filesystems`` filter field.
-The snapshot names are composed of a user-defined prefix followed by a UTC date formatted like ``20060102_150405_000``.
-We use UTC because it will avoid name conflicts when switching time zones or between summer and winter time.
-
-For ``push`` jobs, replication is automatically triggered after all filesystems have been snapshotted.
-
-::
-
-    jobs:
-    - type: push
-      filesystems: {
-        "<": true,
-        "tmp": false
-      }
-      snapshotting:
-        type: periodic
-        prefix: zrepl_
-        interval: 10m
-      ...
-
-There is also a ``manual`` snapshotting type, which covers the following use cases:
-
-* Existing infrastructure for automatic snapshots: you only want to use zrepl for replication.
-* Run scripts before and after taking snapshots (like locking database tables).
-  We are working on better integration for this use case: see :issue:`74`.
-
-Note that you will have to trigger replication manually using the ``zrepl wakeup JOB`` subcommand in that case.
-
-::
-
-   jobs:
-   - type: push
-     filesystems: {
-       "<": true,
-       "tmp": false
-     }
-     snapshotting:
-       type: manual
-     ...
-
+Job Types in Detail
+===================

 .. _job-push:

@@ -199,8 +49,8 @@ Job Type ``sink``
    * - ``serve``
      - |serve-transport|
    * - ``root_fs``
-      - ZFS dataset path are received to
-        ``$root_fs/$client_identity``
+      - ZFS filesystems are received to
+        ``$root_fs/$client_identity/$source_path``

 Example config: :sampleconf:`/sink.yml`

@@ -222,10 +72,11 @@ Job Type ``pull``
    * - ``connect``
      - |connect-transport|
    * - ``root_fs``
-      - ZFS dataset path are received to
-        ``$root_fs/$client_identity``
+      - ZFS filesystems are received to
+        ``$root_fs/$source_path``
    * - ``interval``
-      - Interval at which to pull from the source job
+      - | Interval at which to pull from the source job (e.g. ``10m``).
+        | ``manual`` disables periodic pulling, replication then only happens on :ref:`wakeup <cli-signal-wakeup>`.
    * - ``pruning``
      - |pruning-spec|

@@ -255,6 +106,7 @@ Job Type ``source``

 Example config: :sampleconf:`/source.yml`

+
 .. _replication-local:

 Local replication
@@ -265,3 +117,28 @@ If you have the need for local replication (most likely between two local storag
 Example config: :sampleconf:`/local.yml`.


+.. _job-snap:
+
+Job Type ``snap`` (snapshot & prune only)
+-----------------------------------------
+
+Job type that only takes snapshots and performs pruning on the local machine.
+
+.. list-table::
+    :widths: 20 80
+    :header-rows: 1
+
+    * - Parameter
+      - Comment
+    * - ``type``
+      - = ``snap``
+    * - ``name``
+      - unique name of the job
+    * - ``filesystems``
+      - |filter-spec| for filesystems to be snapshotted
+    * - ``snapshotting``
+      - |snapshotting-spec|
+    * - ``pruning``
+      - |pruning-spec|
+
+Example config: :sampleconf:`/snap.yml`
@@ -7,7 +7,7 @@ Logging

 zrepl uses structured logging to provide users with easily processable log messages.

-Logging outlets are configured in the ``global`` section of the |mainconfig|.
+Logging outlets are configured in the ``global`` section of the config file.

 ::

@@ -147,6 +147,8 @@ Can only be specified once.
      -  minimum  :ref:`log level <logging-levels>`
    * - ``format``
      - output :ref:`format <logging-formats>`
+    * - ``facility``
+      - Which syslog facility to use (default = ``local0``)
    * - ``retry_interval``
      - Interval between reconnection attempts to syslog (default = 0)

@@ -6,16 +6,16 @@ Miscellaneous
 Runtime Directories & UNIX Sockets
 ----------------------------------

-zrepl daemon creates various UNIX sockets to allow communicating with it:
+The zrepl daemon needs to open various UNIX sockets in a runtime directory:

-* the :ref:`transport-ssh+stdinserver` transport connects to a socket named after ``client_identity`` parameter
-* the ``control`` CLI subcommand connects to a defined control socket
+* a ``control`` socket that the CLI commands use to interact with the daemon
+* the :ref:`transport-ssh+stdinserver` listener opens one socket per configured client, named after ``client_identity`` parameter

-There is no further authentication on these sockets.
-Therefore we have to make sure they can only be created and accessed by ``zrepl daemon``.
-In fact, ``zrepl daemon`` will not bind a socket to a path in a directory that is world-accessible.
+There is no authentication on these sockets except the UNIX permissions.
+The zrepl daemon will refuse to bind any of the above sockets in a directory that is world-accessible.

-The directories can be configured in the main configuration file, the defaults are provided below:
+The following sections of the ``global`` config shows the default paths.
+The shell script below shows how the default runtime directory can be created.

 ::

@@ -27,6 +27,12 @@ The directories can be configured in the main configuration file, the defaults a
          sockdir: /var/run/zrepl/stdinserver


+::
+
+    mkdir -p /var/run/zrepl/stdinserver
+    chmod -R 0700 /var/run/zrepl
+
+
 Durations & Intervals
 ---------------------

@@ -5,18 +5,24 @@
 Monitoring
 ==========

-Monitoring endpoints are configured in the ``global.monitoring`` section of the |mainconfig|.
+Monitoring endpoints are configured in the ``global.monitoring`` section of the config file.

 .. _monitoring-prometheus:

-Prometheus
----------
+Prometheus & Grafana
+--------------------

 zrepl can expose `Prometheus metrics <https://prometheus.io/docs/instrumenting/exposition_formats/>`_ via HTTP.
 The ``listen`` attribute is a `net.Listen <https://golang.org/pkg/net/#Listen>`_  string for tcp, e.g. ``:9091`` or ``127.0.0.1:9091``.
-
 The Prometheues monitoring job appears in the ``zrepl control`` job list and may be specified **at most once**.
-At the time of writing, there is no stability guarantee on the exported metrics.
+
+zrepl also ships with an importable `Grafana <https://grafana.com>`_ dashboard that consumes the Prometheus metrics:
+see :repomasterlink:`dist/grafana`.
+The dashboard also contains some advice on which metrics are important to monitor.
+
+.. NOTE::
+
+  At the time of writing, there is no stability guarantee on the exported metrics.

 ::

@@ -0,0 +1,191 @@
+
+Overview & Terminology
+======================
+
+All work zrepl does is performed by the zrepl daemon which is configured in a single YAML configuration file loaded on startup.
+The following paths are considered:
+
+* If set, the location specified via the global ``--config`` flag
+* ``/etc/zrepl/zrepl.yml``
+* ``/usr/local/etc/zrepl/zrepl.yml``
+
+The ``zrepl configcheck`` subcommand can be used to validate the configuration.
+The command will output nothing and exit with zero status code if the configuration is valid.
+The error messages vary in quality and usefulness: please report confusing config errors to the tracking :issue:`155`.
+Full example configs such as in the :ref:`tutorial` or the :sampleconf:`/` directory might also be helpful.
+However, copy-pasting examples is no substitute for reading documentation!
+
+Config File Structure
+---------------------
+
+.. code-block:: yaml
+
+   global: ...
+   jobs:
+   - name: backup
+     type: push
+   - ...
+
+zrepl is confgured using a single YAML configuration file with two main sections: ``global`` and ``jobs``.
+The ``global`` section is filled with sensible defaults and is covered later in this chapter.
+The ``jobs`` section is a list of jobs which we are goind to explain now.
+
+.. _job-overview:
+
+Jobs \& How They Work Together
+------------------------------
+
+A *job* is the unit of activity tracked by the zrepl daemon.
+The ``type`` of a job determines its role in a replication setup and in snapshot management.
+Jobs are identified by their ``name``, both in log files and the ``zrepl status`` command.
+
+Replication always happens between a pair of jobs: one is the **active side**, and one the **passive side**.
+The active side connects to the passive side using a :ref:`transport <transport>` and starts executing the replication logic.
+The passive side responds to requests from the active side after checking its persmissions.
+
+The following table shows how different job types can be combined to achieve **both push and pull mode setups**.
+Note that snapshot-creation denoted by "(snap)" is orthogonal to whether a job is active or passive.
+
+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
+| Setup name            | active side  | passive side                     | use case                                                                           |
+=======================+==============+==================================+====================================================================================+
+| Push mode             | ``push``     | ``sink``                         | * Laptop backup                                                                    |
+|                       | (snap)       |                                  | * NAS behind NAT to offsite                                                        |
+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
+| Pull mode             | ``pull``     | ``source``                       | * Central backup-server for many nodes                                             |
+|                       |              | (snap)                           | * Remote server to NAS behind NAT                                                  |
+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
+| Local replication     | | ``push`` + ``sink`` in one config             | * Backup FreeBSD boot pool                                                         |
+|                       | | with :ref:`local transport <transport-local>` |                                                                                    |
+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
+| Snap & prune-only     | ``snap``     | N/A                              | * | Snapshots & pruning but no replication                                         |
+|                       | (snap)       |                                  |   | required                                                                       |
+|                       |              |                                  | * Workaround for :ref:`source-side pruning <prune-workaround-source-side-pruning>` |
+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
+
+How the Active Side Works
+-------------------------
+
+The active side (:ref:`push <job-push>` and :ref:`pull <job-pull>` job) executes the replication and pruning logic:
+
+* Wakeup because of finished snapshotting (``push`` job) or pull interval ticker (``pull`` job).
+* Connect to the corresponding passive side using a :ref:`transport <transport>` and instantiate an RPC client.
+* Replicate data from the sending to the receiving side (see below).
+* Prune on sender & receiver.
+
+.. TIP::
+  The progress of the active side can be watched live using the ``zrepl status`` subcommand.
+
+How the Passive Side Works
+--------------------------
+
+The passive side (:ref:`sink <job-sink>` and :ref:`source <job-source>`) waits for connections from the corresponding active side,
+using the transport listener type specified in the ``serve`` field of the job configuration.
+Each transport listener provides a client's identity to the passive side job.
+It uses the client identity for access control:
+
+* The ``sink`` job maps requests from different client identities to their respective sub-filesystem tree ``root_fs/${client_identity}``.
+* The ``source`` job has a whitelist of client identities that are allowed pull access.
+
+.. TIP::
+   The implementation of the ``sink`` job requires that the connecting client identities be a valid ZFS filesystem name components.
+
+How Replication Works
+---------------------
+
+One of the major design goals of the replication module is to avoid any duplication of the nontrivial logic.
+As such, the code works on abstract senders and receiver **endpoints**, where typically one will be implemented by a local program object and the other is an RPC client instance.
+Regardless of push- or pull-style setup, the logic executes on the active side, i.e. in the ``push`` or ``pull`` job.
+
+The following steps take place during replication and can be monitored using the ``zrepl status`` subcommand:
+
+* Plan the replication:
+
+  * Compare sender and receiver filesystem snapshots
+  * Build the **replication plan**
+
+    * Per filesystem, compute a diff between sender and receiver snapshots
+    * Build a list of replication steps
+
+      * If possible, use incremental sends (``zfs send -i``)
+      * Otherwise, use full send of most recent snapshot on sender
+      * Give up on filesystems that cannot be replicated without data loss
+
+  * Retry on errors that are likely temporary (i.e. network failures).
+  * Give up on filesystems where a permanent error was received over RPC.
+
+* Execute the plan
+
+  * Perform replication steps in the following order:
+    Among all filesystems with pending replication steps, pick the filesystem whose next replication step's snapshot is the oldest.
+  * Create placeholder filesystems on the receiving side to mirror the dataset paths on the sender to ``root_fs/${client_identity}``.
+  * After a successful replication step, update the replication cursor bookmark (see below).
+   
+The idea behind the execution order of replication steps is that if the sender snapshots all filesystems simultaneously at fixed intervals, the receiver will have all filesystems snapshotted at time ``T1`` before the first snapshot at ``T2 = T1 + $interval`` is replicated.
+
+.. _replication-cursor-bookmark:
+
+The **replication cursor bookmark** ``#zrepl_replication_cursor`` is kept per filesystem on the sending side of a replication setup:
+It is a bookmark of the most recent successfully replicated snapshot to the receiving side.
+It is is used by the :ref:`not_replicated <prune-keep-not-replicated>` keep rule to identify all snapshots that have not yet been replicated to the receiving side.
+Regardless of whether that keep rule is used, the bookmark ensures that replication can always continue incrementally.
+Note that there is only one cursor bookmark per filesystem, which prohibits multiple jobs to replicate the same filesystem (:ref:`see below<jobs-multiple-jobs>`).
+
+.. _replication-placeholder-property:
+
+**Placeholder filesystems** on the receiving side are regular ZFS filesystems with the placeholder property ``zrepl:placeholder=on``.
+Placeholders allow the receiving side to mirror the sender's ZFS dataset hierachy without replicating every filesystem at every intermediary dataset path component.
+Consider the following example: ``S/H/J`` shall be replicated to ``R/sink/job/S/H/J``, but neither ``S/H`` nor ``S`` shall be replicated.
+ZFS requires the existence of ``R/sink/job/S`` and ``R/sink/job/S/H`` in order to receive into ``R/sink/job/S/H/J``.
+Thus, zrepl creates the parent filesystems as placeholders on the receiving side.
+If at some point ``S/H`` and ``S`` shall be replicated, the receiving side invalidates the placeholder flag automatically.
+The ``zrepl test placeholder`` command can be used to check whether a filesystem is a placeholder.
+
+.. ATTENTION::
+
+    Currently, zrepl does not replicate filesystem properties.
+    Whe receiving a filesystem, it is never mounted (`-u` flag)  and `mountpoint=none` is set.
+    This is temporary and being worked on :issue:`24`.
+
+
+.. _jobs-multiple-jobs:
+
+Multiple Jobs & More than 2 Machines
+------------------------------------
+
+.. ATTENTION::
+
+  When using multiple jobs across single or multiple machines, the following rules are critical to avoid race conditions & data loss:
+
+  1. The sets of ZFS filesystems matched by the ``filesystems`` filter fields must be disjoint across all jobs configured on a machine.
+  2. The ZFS filesystem subtrees of jobs with ``root_fs`` must be disjoint.
+  3. Across all zrepl instances on all machines in the replication domain, there must be a 1:1 correspondence between active and passive jobs.
+
+  Explanations & exceptions to above rules are detailed below.
+
+If you would like to see improvements to multi-job setups, please `open an issue on GitHub <https://github.com/zrepl/zrepl/issues/new>`_.
+
+No Overlapping
+~~~~~~~~~~~~~~
+
+Jobs run independently of each other.
+If two jobs match the same filesystem with their ``filesystems`` filter, they will operate on that filesystem independently and potentially in parallel.
+For example, if job A prunes snapshots that job B is planning to replicate, the replication will fail because B asssumed the snapshot to still be present.
+More subtle race conditions can occur with the :ref:`replication cursor bookmark <replication-cursor-bookmark>`, which currently only exists once per filesystem.
+
+N push jobs to 1 sink
+~~~~~~~~~~~~~~~~~~~~~
+
+The :ref:`sink job <job-sink>` namespaces by client identity.
+It is thus safe to push to one sink job with different client identities.
+If the push jobs have the same client identity, the filesystems matched by the push jobs must be disjoint to avoid races.
+
+N pull jobs from 1 source
+~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Multiple pull jobs pulling from the same source have potential for race conditions during pruning:
+each pull job prunes the source side independently, causing replication-prune and prune-prune races.
+
+There is currently no way for a pull job to filter which snapshots it should attempt to replicate.
+Thus, it is not possibe to just manually assert that the prune rules of all pull jobs are disjoint to avoid replication-prune and prune-prune races.
+
@@ -1,39 +0,0 @@
-
-.. _configuration_preface:
-
-=======
-Preface
-=======
-
-----------------------
-Configuration File Path
-----------------------
-
-zrepl searches for its main configuration file in the following locations (in that order):
-
-* If set, the location specified via the global ``--config`` flag
-* ``/etc/zrepl/zrepl.yml``
-* ``/usr/local/etc/zrepl/zrepl.yml``
-
-The examples in the :ref:`tutorial` or the :sampleconf:`/` directory should provide a good starting point.
-
-------------------
-Runtime Directories
-------------------
-
-zrepl requires runtime directories for various UNIX sockets --- they are documented in the :ref:`config file<conf-runtime-directories>`.
-Your package maintainer / init script should take care of creating them.
-Alternatively, for default settings, the following should to the trick.
-
-::
-
-    mkdir -p /var/run/zrepl/stdinserver
-    chmod -R 0700 /var/run/zrepl
-
-
----------
-Validating
----------
-
-The config can be validated using the ``zrepl configcheck`` subcommand.
-
@@ -10,10 +10,12 @@ Typically, the requirements to temporal resolution and maximum retention time di
 For example, when using zrepl to back up a busy database server, you will want high temporal resolution (snapshots every 10 min) for the last 24h in case of administrative disasters, but cannot afford to store them for much longer because you might have high turnover volume in the database.
 On the receiving side, you may have more disk space available, or need to comply with other backup retention policies.

-zrepl uses a set of  **keep rules** to determine which snapshots shall be kept per filesystem.
+zrepl uses a set of  **keep rules** per sending and receiving side to determine which snapshots shall be kept per filesystem.
 **A snapshot that is not kept by any rule is destroyed.**
 The keep rules are **evaluated on the active side** (:ref:`push <job-push>` or :ref:`pull job <job-pull>`) of the replication setup, for both active and passive side, after replication completed or was determined to have failed permanently.

+
+
 Example Configuration:

 ::
@@ -49,18 +51,10 @@ Example Configuration:
    You might have **existing snapshots** of filesystems affected by pruning which you want to keep, i.e. not be destroyed by zrepl.
    Make sure to actually add the necessary ``regex`` keep rules on both sides, like with ``manual`` in the example above.

-.. ATTENTION::
-
-    It is currently not possible to define pruning on a source job.
-    The source job creates snapshots, which means that extended replication downtime will fill up the source's zpool with snapshots, since pruning is directed by the corresponding active side (pull job).
-    If this is a potential risk for you, consider using :ref:`push mode <job-push>`.
-
-
 .. _prune-keep-not-replicated:

 Policy ``not_replicated``
 -------------------------
-
 ::

   jobs:
@@ -100,7 +94,7 @@ the left edge of the leftmost (first) interval is the ``creation`` date of the y
 All intervals to its right describe time intervals further in the past.

 Each interval carries a maximum number of snapshots to keep.
-It is secified via ``(keep=N)``, where ``N`` is either ``all`` (all snapshots are kept) or a positive integer.
+It is specified via ``(keep=N)``, where ``N`` is either ``all`` (all snapshots are kept) or a positive integer.
 The default value is **keep=1**.

 The following procedure happens during pruning:
@@ -146,12 +140,75 @@ Policy ``regex``
     - type: push
       pruning:
         keep_receiver:
+         # keep all snapshots with prefix zrepl_ or manual_
         - type: regex
           regex: "^(zrepl|manual)_.*"
-    ...
+
+     - type: push
+       snapshotting:
+         prefix: zrepl_
+       pruning:
+         keep_sender:
+         # keep all snapshots that were not created by zrepl
+         - type: regex
+           negate: true
+           regex: "^zrepl_.*"

 ``regex`` keeps all snapshots whose names are matched by the regular expressionin ``regex``.
 Like all other regular expression fields in prune policies, zrepl uses Go's `regexp.Regexp <https://golang.org/pkg/regexp/#Compile>`_ Perl-compatible regular expressions (`Syntax <https://golang.org/pkg/regexp/syntax>`_).
+The optional `negate` boolean field inverts the semantics: Use it if you want to keep all snapshots that *do not* match the given regex.

+.. _prune-workaround-source-side-pruning:

+Source-side snapshot pruning
+----------------------------

+A :ref:`source jobs<job-source>` takes snapshots on the system it runs on.
+The corresponding :ref:`pull job <job-pull>` on the replication target connects to the source job and replicates the snapshots.
+Afterwards, the pull job coordinates pruning on both sender (the source job side) and receiver (the pull job side).
+
+There is no built-in way to define and execute pruning on the source side independently of the pull side.
+The source job will continue taking snapshots which will not be pruned until the pull side connects.
+This means that **extended replication downtime will fill up the source's zpool with snapshots**.
+
+If the above is a conceivable situation for you, consider using :ref:`push mode <job-push>`, where pruning happens on the same side where snapshots are taken.
+
+Workaround using ``snap`` job
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+As a workaround (see GitHub :issue:`102` for development progress), a pruning-only :ref:`snap job <job-snap>` can be defined on the source side:
+The snap job is in charge of snapshot creation & destruction, whereas the source job's role is reduced to just serving snapshots.
+However, since, jobs are run independently, it is possible that the snap job will prune snapshots that are queued for replication / destruction by the remote pull job that connects to the source job.
+Symptoms of such race conditions are spurious replication and destroy errors.
+
+Example configuration:
+
+::
+
+  # source side
+  jobs:
+  - type: snap
+    snapshotting:
+      type: periodic
+    pruning:
+      keep:
+        # source side pruning rules go here
+    ...
+
+  - type: source
+    snapshotting:
+      type: manual
+    root_fs: ...
+
+  # pull side
+  jobs:
+  - type: pull
+    pruning:
+      keep_sender:
+        # let the source-side snap job do the pruning
+        - type: regex
+          regex: ".*"
+        ...
+      keep_receiver:
+        # feel free to prune on the pull side as desired
+        ...
@@ -0,0 +1,192 @@
+.. include:: ../global.rst.inc
+
+.. _job-snapshotting-spec:
+
+Taking Snaphots
+===============
+
+The ``push``, ``source`` and ``snap`` jobs can automatically take periodic snapshots of the filesystems matched by the ``filesystems`` filter field.
+The snapshot names are composed of a user-defined prefix followed by a UTC date formatted like ``20060102_150405_000``.
+We use UTC because it will avoid name conflicts when switching time zones or between summer and winter time.
+
+For ``push`` jobs, replication is automatically triggered after all filesystems have been snapshotted.
+
+::
+
+    jobs:
+    - type: push
+      filesystems: {
+        "<": true,
+        "tmp": false
+      }
+      snapshotting:
+        type: periodic
+        prefix: zrepl_
+        interval: 10m
+        hooks: ...
+      ...
+
+There is also a ``manual`` snapshotting type, which covers the following use cases:
+
+* Existing infrastructure for automatic snapshots: you only want to use this zrepl job for replication.
+* Handling snapshotting through a separate ``snap`` job.
+
+Note that you will have to trigger replication manually using the ``zrepl signal wakeup JOB`` subcommand in that case.
+
+::
+
+   jobs:
+   - type: push
+     filesystems: {
+       "<": true,
+       "tmp": false
+     }
+     snapshotting:
+       type: manual
+     ...
+
+.. _job-snapshotting-hooks:
+
+Pre- and Post-Snapshot Hooks
+----------------------------
+
+Jobs with `periodic snapshots <job-snapshotting-spec_>`_ can run hooks before and/or after taking the snapshot specified in ``snapshotting.hooks``:
+Hooks are called per filesystem before and after the snapshot is taken (pre- and post-edge).
+Pre-edge invocations are in configuration order, post-edge invocations in reverse order, i.e. like a stack.
+If a pre-snapshot invocation fails, ``err_is_fatal=true`` cuts off subsequent hooks, does not take a snapshot, and only invokes post-edges corresponding to previous successful pre-edges.
+``err_is_fatal=false`` logs the failed pre-edge invocation but does not affect subsequent hooks nor snapshotting itself.
+Post-edges are only invoked for hooks whose pre-edges ran without error.
+Note that hook failures for one filesystem never affect other filesystems.
+
+The optional ``timeout`` parameter specifies a period after which zrepl will kill the hook process and report an error.
+The default is 30 seconds and may be specified in any units understood by `time.ParseDuration <https://golang.org/pkg/time/#ParseDuration>`_.
+
+The optional ``filesystems`` filter which limits the filesystems the hook runs for. This uses the same |filter-spec| as jobs.
+
+Most hook types take additional parameters, please refer to the respective subsections below.
+
+.. list-table::
+    :widths: 20 10 70
+    :header-rows: 1
+
+    * - Hook ``type``
+      - Details
+      - Description
+    * - ``command``
+      - :ref:`Details <job-hook-type-command>`
+      - Arbitrary pre- and post snapshot scripts.
+    * - ``postgres-checkpoint``
+      - :ref:`Details <job-hook-type-postgres-checkpoint>`
+      - Execute Postgres ``CHECKPOINT`` SQL command before snapshot.
+    * - ``mysql-lock-tables``
+      - :ref:`Details <job-hook-type-mysql-lock-tables>`
+      - Flush and read-Lock MySQL tables while taking the snapshot.
+      
+.. _job-hook-type-command:
+
+``command`` Hooks
+~~~~~~~~~~~~~~~~~
+
+::
+
+
+    jobs:
+    - type: push
+      filesystems: {
+        "<": true,
+        "tmp": false
+      }
+      snapshotting:
+        type: periodic
+        prefix: zrepl_
+        interval: 10m
+        hooks:
+        - type: command
+          path: /etc/zrepl/hooks/zrepl-notify.sh
+          timeout: 30s
+          err_is_fatal: false
+        - type: command
+          path: /etc/zrepl/hooks/special-snapshot.sh
+          filesystems: {
+            "tank/special": true
+          }
+      ...
+
+
+The hook type ``command`` is the only currently supported hook type.
+Future versions of zrepl may support other hook types.
+The ``path`` to the hook executables must be absolute (e.g. ``/etc/zrepl/hooks/zrepl-notify.sh``).
+No arguments may be specified; create a wrapper script if zrepl must call an executable that requires arguments.
+zrepl will call the hook both before and after the snapshot, but with different values of the ``ZREPL_HOOKTYPE`` environment variable; see below.
+The process standard output is logged at level INFO. Standard error is logged at level WARN.
+
+zrepl sets a number of environment variables for the hook processes:
+
+* ``ZREPL_HOOKTYPE``: either "pre_snapshot" or "post_snapshot"
+* ``ZREPL_FS``: the ZFS filesystem name being snapshotted
+* ``ZREPL_SNAPNAME``: the zrepl-generated snapshot name (e.g. ``zrepl_20380119_031407_000``)
+* ``ZREPL_DRYRUN``: set to ``"true"`` if a dry run is in progress so scripts can print, but not run, their commands
+
+An empty template hook can be found in :sampleconf:`hooks/template.sh`.
+
+.. _job-hook-type-postgres-checkpoint:
+
+``postgres-checkpoint`` Hook
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Connects to a Postgres server and executes the ``CHECKPOINT`` statement pre-snapshot.
+Checkpointing applies the WAL contents to all data files and syncs the data files to disk.
+This is not required for a consistent database backup: it merely forward-pays the "cost" of WAL replay to the time of snapshotting instead of at restore.
+However, the Postgres manual recommends against checkpointing during normal operation.
+Further, the operation requires Postgres superuser privileges.
+zrepl users must decide on their own whether this hook is useful for them (it likely isn't).
+
+.. ATTENTION::
+    Note that WALs and Postgres data directory (with all database data files) must be on the same filesystem to guarantee a correct point-in-time backup with the ZFS snapshot.
+
+DSN syntax documented here: `<https://godoc.org/github.com/lib/pq>`_
+
+.. code-block:: sql
+
+   CREATE USER zrepl_checkpoint PASSWORD yourpasswordhere;
+   ALTER ROLE zrepl_checkpoint SUPERUSER;
+
+.. code-block:: yaml
+
+  - type: postgres-checkpoint
+    dsn: "host=localhost port=5432 user=postgres password=yourpasswordhere sslmode=disable"
+    filesystems: {
+        "p1/postgres/data11": true
+    }
+
+.. _job-hook-type-mysql-lock-tables:
+
+``mysql-lock-tables`` Hook
+~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Connects to MySQL and executes
+
+* pre-snapshot ``FLUSH TABLES WITH READ LOCK`` to lock all tables in all databases in the MySQL server we connect to (`docs <https://dev.mysql.com/doc/refman/8.0/en/flush.html#flush-tables-with-read-lock>`_)
+* post-snapshot ``UNLOCK TABLES``  reverse above operation.
+
+Above procedure is documented in the `MySQL manual <https://dev.mysql.com/doc/mysql-backup-excerpt/5.7/en/backup-methods.html>`_
+as a means to produce a consistent backup of a MySQL DBMS installation (i.e., all databases).
+
+`DSN syntax <https://github.com/go-sql-driver/mysql#dsn-data-source-name>`_: ``[username[:password]@][protocol[(address)]]/dbname[?param1=value1&...&paramN=valueN]``
+
+.. ATTENTION::
+    All MySQL databases must be on the same ZFS filesystem to guarantee a consistent point-in-time backup with the ZFS snapshot.
+
+.. code-block:: sql
+
+   CREATE USER zrepl_lock_tables IDENTIFIED BY 'yourpasswordhere';
+   GRANT RELOAD ON *.* TO zrepl_lock_tables;
+   FLUSH PRIVILEGES;
+
+.. code-block:: yaml
+
+  - type: mysql-lock-tables
+    dsn: "zrepl_lock_tables:yourpasswordhere@tcp(localhost)/"
+    filesystems: {
+      "tank/mysql": true
+    }
@@ -10,6 +10,8 @@ On the passive (serving) side, the transport also provides the **client identity
 this string is used for access control and separation of filesystem sub-trees in :ref:`sink jobs <job-sink>`.
 Transports are specified in the ``connect`` or ``serve`` section of a job definition.

+.. contents::
+
 .. ATTENTION::

    The **client identities must be valid ZFS dataset path components**
@@ -27,6 +29,8 @@ This transport may also be used in conjunction with network-layer encryption and
 To make the IP-based client authentication effective, such solutions should provide authenticated IP addresses.
 Some options to consider:

+.. _transport-tcp-tunneling:
+
 * `WireGuard <https://www.wireguard.com/>`_: Linux-focussed, in-kernel TLS
 * `OpenVPN <https://openvpn.net/>`_: Cross-platform VPN, uses tun on \*nix
 * `IPSec <https://en.wikipedia.org/wiki/IPsec>`_: Properly standardized, in-kernel network-layer VPN
@@ -72,7 +76,9 @@ Connect

 The ``tls`` transport uses TCP + TLS with client authentication using client certificates.
 The client identity is the common name (CN) presented in the client certificate.
+
 It is recommended to set up a dedicated CA infrastructure for this transport, e.g. using OpenVPN's `EasyRSA <https://github.com/OpenVPN/easy-rsa>`_.
+For a simple 2-machine setup, see the :ref:`instructions below<transport-tcp+tlsclientauth-2machineopenssl>`.

 The implementation uses `Go's TLS library <https://golang.org/pkg/crypto/tls/>`_.
 Since Go binaries are statically linked, you or your distribution need to recompile zrepl when vulnerabilities in that library are disclosed.
@@ -80,6 +86,10 @@ Since Go binaries are statically linked, you or your distribution need to recomp
 All file paths are resolved relative to the zrepl daemon's working directory.
 Specify absolute paths if you are unsure what directory that is (or find out from your init system).

+If intermediate CAs are used, the **full chain** must be present in either in the ``ca`` file or the individual ``cert`` files.
+Regardless, the client's certificate must be first in the ``cert`` file, with each following certificate directly certifying the one preceding it (see `TLS's specification <https://tools.ietf.org/html/rfc5246#section-7.4.2>`_).
+This is the common default when using a CA management tool.
+
 Serve
 ~~~~~

@@ -92,7 +102,7 @@ Serve
          type: tls
          listen: ":8888"
          ca:   /etc/zrepl/ca.crt
-          cert: /etc/zrepl/prod.crt
+          cert: /etc/zrepl/prod.fullchain
          key:  /etc/zrepl/prod.key
          client_cns:
            - "laptop1"
@@ -112,7 +122,7 @@ Connect
        type: tls
        address: "server1.foo.bar:8888"
        ca:   /etc/zrepl/ca.crt
-        cert: /etc/zrepl/backupserver.crt
+        cert: /etc/zrepl/backupserver.fullchain
        key:  /etc/zrepl/backupserver.key
        server_cn: "server1"
        dial_timeout: # optional, default 10s
@@ -122,13 +132,91 @@ The ``server_cn`` specifies the expected common name (CN) of the server's certif
 It overrides the hostname specified in ``address``.
 The connection fails if either do not match.

+.. _transport-tcp+tlsclientauth-2machineopenssl:
+
+Self-Signed Certificates
+~~~~~~~~~~~~~~~~~~~~~~~~
+
+Tools like `EasyRSA <https://github.com/OpenVPN/easy-rsa>`_ make it easy to manage CA infrastructure for multiple clients, e.g. a central zrepl backup server (in sink mode).
+However, for a two-machine setup, self-signed certificates distributed using an out-of-band mechanism will also work just fine:
+
+Suppose you have a push-mode setup, with `backups.example.com` running the :ref:`sink job <job-sink>`, and `prod.example.com` running the :ref:`push job <job-push>`.
+Run the following OpenSSL commands on each host, substituting HOSTNAME in both filenames and the interactive input prompt by OpenSSL:
+
+.. code-block:: bash
+   :emphasize-lines: 1-5,24
+
+   openssl req -x509 -sha256 -nodes \
+      -newkey rsa:4096 \
+      -days 365 \
+      -keyout HOSTNAME.key \
+      -out HOSTNAME.crt
+
+   #Generating a 4096 bit RSA private key
+   #................++++
+   #.++++
+   #writing new private key to 'backups.key'
+   #-----
+   #You are about to be asked to enter information that will be incorporated
+   #into your certificate request.
+   #What you are about to enter is what is called a Distinguished Name or a DN.
+   #There are quite a few fields but you can leave some blank
+   #For some fields there will be a default value,
+   #If you enter '.', the field will be left blank.
+   #-----
+   #Country Name (2 letter code) [XX]:
+   #State or Province Name (full name) []:
+   #Locality Name (eg, city) [Default City]:
+   #Organization Name (eg, company) [Default Company Ltd]:
+   #Organizational Unit Name (eg, section) []:
+   #Common Name (eg, your name or your server's hostname) []:HOSTNAME
+   #Email Address []:
+
+Now copy each machine's ``HOSTNAME.crt`` to the other machine's ``/etc/zrepl/HOSTNAME.crt``, for example using `scp`.
+The serve & connect configuration will thus look like the following:
+
+::
+
+   # on backups.example.com
+   - type: sink
+     serve:
+       type: tls
+       listen: ":8888"
+       ca: "/etc/zrepl/prod.example.com.crt"
+       cert: "/etc/zrepl/backups.example.com.crt"
+       key: "/etc/zrepl/backups.example.com.key"
+       client_cns:
+         - "prod.example.com"
+     ...
+
+   # on prod.example.com
+   - type: push
+     connect:
+       type: tls
+       address:"backups.example.com:8888"
+       ca: /etc/zrepl/backups.example.com.crt
+       cert: /etc/zrepl/prod.example.com.crt
+       key:  /etc/zrepl/prod.example.com.key
+       server_cn: "backups.example.com"
+     ...
+
+
 .. _transport-ssh+stdinserver:

 ``ssh+stdinserver`` Transport
 -----------------------------

-``ssh+stdinserver`` is inspired by `git shell <https://git-scm.com/docs/git-shell>`_ and `Borg Backup <https://borgbackup.readthedocs.io/en/stable/deployment.html>`_.
-It is provided by the Go package ``github.com/problame/go-netssh``.
+``ssh+stdinserver`` uses the ``ssh`` command and some features of the server-side SSH ``authorized_keys`` file.
+It is less efficient than other transports because the data passes through two more pipes.
+However, it is fairly convenient to set up and allows the zrepl daemon to not be directly exposed to the internet, because all traffic passes through the system's SSH server.
+
+The concept is inspired by `git shell <https://git-scm.com/docs/git-shell>`_ and `Borg Backup <https://borgbackup.readthedocs.io/en/stable/deployment.html>`_.
+The implementation is provided by the Go package ``github.com/problame/go-netssh``.
+
+.. NOTE::
+
+   ``ssh+stdinserver`` generally provides inferior error detection and handling compared to the ``tcp`` and ``tls`` transports.
+   When encountering such problems, consider using  ``tcp`` or ``tls`` transports, or help improve package go-netssh.

 .. _transport-ssh+stdinserver-serve:

@@ -257,5 +345,6 @@ The ``client_identity`` is used by the sink as documented above.
        type: local
        listener_name: localsink
        client_identity: local_backup
+        dial_timeout: 2s # optional, 0 for no timeout
      ...

@@ -1,3 +1,26 @@
 .. global list of substitutions

-.. |mainconfig| replace:: :ref:`main configuration file <mainconfigfile>`
+.. 
+  .. |macroname| replace:: macro expansion
+
+.. |GitHub license| image:: https://img.shields.io/github/license/zrepl/zrepl.svg
+   :target: https://github.com/zrepl/zrepl/blob/master/LICENSE
+.. |Language: Go| image:: https://img.shields.io/badge/language-Go-6ad7e5.svg
+   :target: https://golang.org/
+.. |User Docs| image:: https://img.shields.io/badge/docs-web-blue.svg
+   :target: https://zrepl.github.io
+.. |Donate via PayPal| image:: https://img.shields.io/badge/donate-paypal-yellow.svg
+   :target: https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96
+.. |Donate via Liberapay| image:: https://img.shields.io/liberapay/receives/zrepl.svg?logo=liberapay
+   :target: https://liberapay.com/zrepl/donate
+.. |Donate via Patreon| image:: https://img.shields.io/endpoint.svg?url=https%3A%2F%2Fshieldsio-patreon.herokuapp.com%2Fzrepl%2Fpledges&style=flat&color=yellow
+   :target: https://www.patreon.com/zrepl
+.. |Twitter| image:: https://img.shields.io/twitter/url/https/github.com/zrepl/zrepl.svg?style=social
+   :target: https://twitter.com/intent/tweet?text=Wow:&url=https%3A%2F%2Fgithub.com%2Fzrepl%2Fzrepl
+
+
+.. |serve-transport| replace:: :ref:`serve specification<transport>`
+.. |connect-transport| replace:: :ref:`connect specification<transport>`
+.. |snapshotting-spec| replace:: :ref:`snapshotting specification <job-snapshotting-spec>`
+.. |pruning-spec| replace:: :ref:`pruning specification <prune>`
+.. |filter-spec| replace:: :ref:`filter specification<pattern-filter>`
@@ -5,6 +5,9 @@

 .. include:: global.rst.inc

+|GitHub license| |Language: Go| |Twitter| |Donate via Patreon| |Donate via Liberapay| |Donate via PayPal| 
+
+
 zrepl - ZFS replication
 -----------------------

@@ -12,12 +15,32 @@ zrepl - ZFS replication

 .. raw:: html

-   <div style="margin-bottom: 1em; background: #2e3436; min-height: 6em; max-width: 100%">
-     <a href="https://raw.githubusercontent.com/wiki/zrepl/zrepl/zrepl_0.1_status.mp4" target="_new" >
-       <video title="zrepl status subcommand" loop autoplay style="width: 100%; display: block;" src="https://raw.githubusercontent.com/wiki/zrepl/zrepl/zrepl_0.1_status.mp4"></video>
-     </a>
+   <div style="margin-bottom: 1em; background: #2e3436; color: white; font-size: 0.8em; min-height: 6em; max-width: 100%; overflow: auto;">
+     <pre>
+     Job: prod_to_backups
+     Type: push
+     Replication:
+         Attempt #1
+         Status: fan-out-filesystems
+         Progress: [=========================\----] 246.7 MiB / 264.7 MiB @ 11.5 MiB/s
+           zroot              STEPPING (step 1/2, 624 B/1.2 KiB) next: @a => @b
+           zroot/ROOT         DONE (step 2/2, 1.2 KiB/1.2 KiB)
+           zroot/ROOT/default STEPPING (step 1/2, 123.4 MiB/129.3 MiB) next: @a => @b
+           zroot/tmp          STEPPING (step 1/2, 29.9 KiB/44.2 KiB) next: @a => @b
+           zroot/usr          STEPPING (step 1/2, 624 B/1.2 KiB) next: @a => @b
+           zroot/usr/home     STEPPING (step 1/2, 123.3 MiB/135.3 MiB) next: @a => @b
+           zroot/var          STEPPING (step 1/2, 624 B/1.2 KiB) next: @a => @b
+           zroot/var/audit    DONE (step 2/2, 1.2 KiB/1.2 KiB)
+           zroot/var/crash    DONE (step 2/2, 1.2 KiB/1.2 KiB)
+           zroot/var/log      STEPPING (step 1/2, 22.0 KiB/29.2 KiB) next: @a => @b
+           zroot/var/mail     STEPPING (step 1/2, 624 B/1.2 KiB) next: @a => @b
+     Pruning Sender:
+         ...
+     Pruning Receiver:
+     </pre>
   </div>

+
 Getting started
 ~~~~~~~~~~~~~~~

@@ -29,17 +52,19 @@ Main Features
 * **Filesystem replication**

  * [x] Pull & Push mode
-  * [x] Multiple transport :ref:`transports <transport>`: TCP, TCP + TLS client auth, SSH
+  * [x] Multiple :ref:`transport modes <transport>`: TCP, TCP + TLS client auth, SSH

  * Advanced replication features

+    * [x] Automatic retries for temporary network errors
    * [ ] Resumable send & receive
    * [ ] Compressed send & receive
    * [ ] Raw encrypted send & receive

 * **Automatic snapshot management**

-  * [x] Periodic filesystem snapshots
+  * [x] Periodic :ref:`filesystem snapshots <job-snapshotting-spec>`
+  * [x] Support for :ref:`pre- and post-snapshot hooks <job-snapshotting-hooks>` with builtins for MySQL & Postgres
  * [x] Flexible :ref:`pruning rule system <prune>`

    * [x] Age-based fading (grandfathering scheme)
@@ -72,12 +97,14 @@ Contributing

 We are happy about any help we can get!

+* :ref:`Financial Support <supporters>`
 * Explore the codebase

  * These docs live in the ``docs/`` subdirectory

 * Document any non-obvious / confusing / plain broken behavior you encounter when setting up zrepl for the first time
-* Check the *Issues* and *Projects* sections for things to do
+* Check the *Issues* and *Projects* sections for things to do.
+  The `good first issues <https://github.com/zrepl/zrepl/labels/good%20first%20issue>`_ and `docs <https://github.com/zrepl/zrepl/labels/docs>`_ are suitable starting points.

 .. admonition:: Development Workflow
    :class: note
@@ -86,6 +113,8 @@ We are happy about any help we can get!
    Make sure to read the `Developer Documentation section <https://github.com/zrepl/zrepl>`_ and open new issues or pull requests there.


+
+
 Table of Contents
 ~~~~~~~~~~~~~~~~~

@@ -101,4 +130,5 @@ Table of Contents
   pr
   changelog
   GitHub Repository & Issue Tracker <https://github.com/zrepl/zrepl>
+   supporters

--- a/Show More
+++ b/Show More