WIP endpoint abstractions + pruning integration / pruner rewrite

docs: update supporters
zfs: ZFSListFilesystemVersions: remove handling of io.ErrUnexpectedEOF
2020-05-29 00:09:43 +02:00 · 2020-05-24 18:39:02 +02:00 · 2020-05-24 18:39:02 +02:00 · 2020-05-24 18:39:02 +02:00 · 2020-05-24 18:39:02 +02:00 · 2020-05-24 18:39:02 +02:00
145 changed files with 4023 additions and 5861 deletions
@@ -1,79 +1,74 @@
-version: 2.1
+version: 2.0
+workflows:
+  version: 2
+  build:
+    jobs:
+      - build-1.11
+      - build-1.12
+      - build-1.13
+      - build-1.14
+      - build-latest
+      - test-build-in-docker
+jobs:

-commands:
-  setup-home-local-bin:
-    steps:
-      - run:
-          shell: /bin/bash -euo pipefail
-          command: |
-            mkdir -p "$HOME/.local/bin"
-            line='export PATH="$HOME/.local/bin:$PATH"'
-            if grep "$line" $BASH_ENV >/dev/null; then
-              echo "$line" >> $BASH_ENV
-            fi
-
-  invoke-lazy-sh:
+  # build-latest serves as the template
+  # we use YAML anchors & aliases to exchange the docker image (and hence Go version used for the build)
+  build-latest: &build-latest
+    description: Builds zrepl
    parameters:
-      subcommand:
+      image:
+        description: "the docker image that the job should use"
        type: string
+    docker:
+      - image: circleci/golang:latest
+    environment:
+      # required by lazy.sh
+      TERM: xterm
+    working_directory: /go/src/github.com/zrepl/zrepl
    steps:
      - run:
-          environment:
-            TERM: xterm
-          command: ./lazy.sh <<parameters.subcommand>>
-
-  apt-update-and-install-common-deps:
-    steps:
-      - run: sudo apt update && sudo apt install gawk make
-
-  restore-cache-gomod:
-    steps:
-      - restore_cache:
-          key: go-mod-v4-{{ checksum "go.sum" }}
-  save-cache-gomod:
-    steps:
-      - save_cache:
-          key: go-mod-v4-{{ checksum "go.sum" }}
-          paths:
-            - "/go/pkg/mod"
-
-  install-godep:
-    steps:
-      - apt-update-and-install-common-deps
-      - invoke-lazy-sh:
-          subcommand: godep
-
-  install-docdep:
-    steps:
-      - apt-update-and-install-common-deps
-      - run: sudo apt install python3 python3-pip libgirepository1.0-dev
-      - invoke-lazy-sh:
-          subcommand: docdep
-
-  download-and-install-minio-client:
-    steps:
-      - setup-home-local-bin
-      - restore_cache:
-          key: minio-client-v2
-      - run:
-          shell: /bin/bash -eo pipefail
+          name: Setup environment variables
          command: |
-            if which mc; then exit 0; fi
-            sudo curl -sSL https://dl.min.io/client/mc/release/linux-amd64/archive/mc.RELEASE.2020-08-20T00-23-01Z \
-              -o "$HOME/.local/bin/mc"
-            sudo chmod +x "$HOME/.local/bin/mc"
-      - save_cache:
-          key: minio-client-v2
-          paths:
-            - "$HOME/.local/bin/mc"
+            # 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 gawk
+      - run: ./lazy.sh devsetup
+
+      - run: make zrepl-bin
+      - run: make vet
+      - run: make lint
+      - run: make release
+      - run: make test-go
+      # cannot run test-platform because circle-ci runs in linux containers
+
+      - store_artifacts:
+          path: ./artifacts/release
+          when: always

-  upload-minio:
-    parameters:
-      src:
-        type: string
-      dst:
-        type: string
-    steps:
      - run:
          shell: /bin/bash -eo pipefail
          when: always
@@ -83,298 +78,78 @@ commands:
              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} ${MINIO_SECRET_KEY}
-
-            # keep in sync with set-github-minio-status
-            jobprefix=zrepl-ci-artifacts/${CIRCLE_SHA1}-pipeline-<<pipeline.number>>/${CIRCLE_JOB}
-
+            
            # Upload artifacts
-            mkdir -p ./artifacts
-            mc cp -r <<parameters.src>> "zrepl-minio/$jobprefix/<<parameters.dst>>"
+            echo "$CIRCLE_BUILD_URL" > ./artifacts/release/cirlceci_build_url
+            mc cp -r artifacts/release "zrepl-minio/zrepl-ci-artifacts/${CIRCLE_SHA1}/${CIRCLE_JOB}/"

-  set-github-minio-status:
-    parameters:
-      context:
-        type: string
-      description:
-        type: string
-      minio-dst:
-        type: string
-    steps:
-      - run:
-          shell: /bin/bash -eo pipefail
-          command: |
-            if [ -n "$CIRCLE_PR_NUMBER" ]; then # CIRCLE_PR_NUMBER is guaranteed to be only present in forked PRs (external)
-              echo "Forked PR detected. Sry, can't trust you with credentials to external artifact store, use CircleCI's instead."
-              exit 0
-            fi
-            set -u # from now on
-
-            # keep in sync with with upload-minio command
-            jobprefix=zrepl-ci-artifacts/${CIRCLE_SHA1}-pipeline-<<pipeline.number>>/${CIRCLE_JOB}
            # Push Artifact Link to GitHub
            REPO="zrepl/zrepl"
            COMMIT="${CIRCLE_SHA1}"
            JOB_NAME="${CIRCLE_JOB}"
-            CONTEXT="<<parameters.context>>"
-            DESCRIPTION="<<parameters.description>>"
-            TARGETURL=https://minio.cschwarz.com/minio/"$jobprefix"/"<<parameters.minio-dst>>"
            curl "https://api.github.com/repos/$REPO/statuses/$COMMIT" \
              -H "Content-Type: application/json" \
              -H "Authorization: token $GITHUB_COMMIT_STATUS_TOKEN" \
              -X POST \
-              -d '{"context":"'"$CONTEXT"'", "state": "success", "description":"'"$DESCRIPTION"'", "target_url":"'"$TARGETURL"'"}'
+              -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"'/"}'

+      - run:
+          shell: /bin/bash -euo pipefail
+          command: |
+            # Trigger Debian Package Build
+            curl -v -X POST https://api.github.com/repos/zrepl/debian-binary-packaging/dispatches \
+              -H 'Accept: application/vnd.github.v3+json' \
+              -H "Authorization: token $ZREPL_DEBIAN_BINARYPACKAGIN_TRIGGER_BUILD_GITHUB_TOKEN" \
+              --data '{"event_type": "push", "client_payload": { "zrepl_main_repo_commit": "'"$CIRCLE_SHA1"'", "go_version": "'"${CIRCLE_JOB##build-}"'" }}'

-  trigger-pipeline:
-    parameters:
-      body_no_shell_subst:
-        type: string
-    steps:
-      - run: |
-          curl -X POST https://circleci.com/api/v2/project/github/zrepl/zrepl/pipeline \
-            -H 'Content-Type: application/json' \
-            -H 'Accept: application/json' \
-            -H "Circle-Token: $ZREPL_BOT_CIRCLE_TOKEN" \
-            --data '<<parameters.body_no_shell_subst>>'
-
-parameters:
-  do_ci:
-    type: boolean
-    default: true
-
-  do_release:
-    type: boolean
-    default: false
-
-  release_docker_baseimage_tag:
-    type: string
-    default: "1.15"
-
-workflows:
-  version: 2
-
-  ci:
-    when: << pipeline.parameters.do_ci >>
-    jobs:
-      - quickcheck-docs
-      - quickcheck-go: &quickcheck-go-smoketest
-          name: quickcheck-go-amd64-linux-1.15
-          goversion: &latest-go-release "1.15"
-          goos: linux
-          goarch: amd64
-      - test-go-on-latest-go-release:
-          goversion: *latest-go-release
-      - quickcheck-go:
-          requires:
-            - quickcheck-go-amd64-linux-1.15 #quickcheck-go-smoketest.name
-          matrix: &quickcheck-go-matrix
-            alias: quickcheck-go-matrix
-            parameters:
-              goversion: [*latest-go-release, "1.11"]
-              goos: ["linux", "freebsd"]
-              goarch: ["amd64", "arm64"]
-            exclude:
-              # don't re-do quickcheck-go-smoketest
-              - goversion: *latest-go-release
-                goos: linux
-                goarch: amd64
-              # not supported by Go 1.11
-              - goversion: "1.11"
-                goos: freebsd
-                goarch: arm64
-
-  release:
-    when: << pipeline.parameters.do_release >>
-    jobs:
-      - release-build
-      - release-deb:
-          requires:
-            - release-build
-      - release-rpm:
-          requires:
-            - release-build
-      - release-upload:
-          requires:
-            - release-build
-            - release-deb
-            - release-rpm
-
-  periodic:
-    triggers:
-      - schedule:
-          cron: "45 15 * * *"
-          filters:
-            branches:
-              only:
-                - master
-                - stable
-                - problame/circleci-build
-    jobs:
-      - periodic-trigger-pipeline
-
-  zrepl.github.io:
-    jobs:
-      - publish-zrepl-github-io:
-          filters:
-            branches:
-              only:
-                - stable
-
-jobs:
-  quickcheck-docs:
+  build-1.11:
+    <<: *build-latest
    docker:
-      - image: cimg/base:2020.08
-    steps:
-      - checkout
-      - install-docdep
-      - run: make docs
+    - image: circleci/golang:1.11

-      - store_artifacts:
-          path: artifacts
-      - download-and-install-minio-client
-      - upload-minio:
-          src: artifacts
-          dst: ""
-      - set-github-minio-status:
-          context: artifacts/${CIRCLE_JOB}
-          description: artifacts of CI job ${CIRCLE_JOB}
-          minio-dst: ""
-
-  quickcheck-go:
-    parameters:
-      goversion:
-        type: string
-      goos:
-        type: string
-      goarch:
-        type: string
+  build-1.12:
+    <<: *build-latest
    docker:
-      - image: circleci/golang:<<parameters.goversion>>
+    - image: circleci/golang:1.12
+
+  build-1.13:
+    <<: *build-latest
+    docker:
+    - image: circleci/golang:1.13
+
+  build-1.14:
+    <<: *build-latest
+    docker:
+    - image: circleci/golang:1.14
+
+  # this job tries to mimic the build-in-docker instructions
+  # given in docs/installation.rst
+  #
+  # However, CircleCI doesn't support volume mounts, so we have to copy
+  # the source into the build-container by modifying the Dockerfile
+  test-build-in-docker:
+    description: Check that build-in-docker works
+    docker:
+      - image: circleci/golang:latest
    environment:
-      GOOS: <<parameters.goos>>
-      GOARCH: <<parameters.goarch>>
-
+    working_directory: /go/src/github.com/zrepl/zrepl
    steps:
-      - checkout
-
-      - restore-cache-gomod
-      - run: go mod download
-      - run: cd build && go mod download
-      - save-cache-gomod
-
-      - install-godep
-      - run: make formatcheck
-      - run: make generate-platform-test-list
-      - run: make zrepl-bin test-platform-bin
-      - run: make vet
-      - run: make lint
-
-      - download-and-install-minio-client
-      - run: rm -f artifacts/generate-platform-test-list
-      - store_artifacts:
-          path: artifacts
-      - upload-minio:
-          src: artifacts
-          dst: ""
-      - set-github-minio-status:
-          context: artifacts/${CIRCLE_JOB}
-          description: artifacts of CI job ${CIRCLE_JOB}
-          minio-dst: ""
-
-  test-go-on-latest-go-release:
-    parameters:
-      goversion:
-        type: string
-    docker:
-      - image: circleci/golang:<<parameters.goversion>>
-    steps:
-      - checkout
-      - restore-cache-gomod
-      - run: make test-go
-      # don't save-cache-gomod here, test-go doesn't pull all the dependencies
-
-  release-build:
-    machine: true
-    steps:
-      - checkout
-      - run: make release-docker RELEASE_DOCKER_BASEIMAGE_TAG=<<pipeline.parameters.release_docker_baseimage_tag>>
-      - persist_to_workspace:
-          root: .
-          paths: [.]
-  release-deb:
-    machine: true
-    steps:
-      - attach_workspace:
-          at: .
-      - run: make debs-docker
-      - persist_to_workspace:
-          root: .
-          paths:
-            - "artifacts/*.deb"
-
-  release-rpm:
-    machine: true
-    steps:
-      - attach_workspace:
-          at: .
-      - run: make rpms-docker
-      - persist_to_workspace:
-          root: .
-          paths:
-            - "artifacts/*.rpm"
-
-  release-upload:
-    docker:
-      - image: cimg/base:2020.08
-    steps:
-      - attach_workspace:
-          at: .
-      - store_artifacts:
-          path: artifacts
-      - download-and-install-minio-client
-      - upload-minio:
-          src: artifacts
-          dst: ""
-      - set-github-minio-status:
-          context: artifacts/release
-          description: CI-generated release artifacts
-          minio-dst: ""
-
-  periodic-trigger-pipeline:
-    docker:
-      - image: cimg/base:2020.08
-    steps:
-      - trigger-pipeline:
-          body_no_shell_subst: '{"branch":"master", "parameters": { "do_ci": true, "do_release": true }}'
-      - trigger-pipeline:
-          body_no_shell_subst: '{"branch":"stable", "parameters": { "do_ci": true, "do_release": true }}'
-
-  publish-zrepl-github-io:
-    docker:
-      - image: cimg/python:3.7
-    steps:
-      - checkout
-      - invoke-lazy-sh:
-          subcommand: docdep
-      - run:
-          command: |
-            git config --global user.email "me@cschwarz.com"
-            git config --global user.name "zrepl-github-io-ci"
-
-      # https://circleci.com/docs/2.0/add-ssh-key/#adding-multiple-keys-with-blank-hostnames
-      - run: ssh-add -D
-      # the default circleci ssh config only additional ssh keys for Host !github.com
-      - run:
-          command: |
-            cat > ~/.ssh/config \<<EOF
-            Host *
-              IdentityFile /home/circleci/.ssh/id_rsa_458e62c517f6c480e40452126ce47421
-            EOF
-      - add_ssh_keys:
-          fingerprints:
-            # deploy key for zrepl.github.io
-            - "45:8e:62:c5:17:f6:c4:80:e4:04:52:12:6c:e4:74:21"
-
-      - run: bash -x docs/publish.sh -c -a
+    - checkout
+    - setup_remote_docker
+    - run:
+        name: (hacky) circleci doesn't allow volume mounts, so copy src to container
+        command: echo "ADD . /src" >> build.Dockerfile
+    - run:
+        name: (hacky) commit modified Dockerfile to avoid failing git clean check in Makefile
+        command: git -c user.name='circleci' -c user.email='circleci@localhost' commit -m 'CIRCLECI modified Dockerfile with zrepl src' --author 'autoauthor <circleci@localhost>' -- build.Dockerfile
+    - run:
+        name: build the build image (build deps)
+        command: docker build -t zrepl_build -f build.Dockerfile .
+    - run:
+        name: try compiling
+        command: docker run -it zrepl_build make release
@@ -1,10 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-COMMIT="$1"
-GO_VERSION="$2"
-
-curl -v -X POST https://api.github.com/repos/zrepl/debian-binary-packaging/dispatches \
-    -H 'Accept: application/vnd.github.v3+json' \
-    -H "Authorization: token $GITHUB_ACCESS_TOKEN" \
-    --data '{"event_type": "push", "client_payload": { "zrepl_main_repo_commit": "'"$COMMIT"'", "go_version": "'"$GO_VERSION"'" }}'
@@ -1,4 +1,3 @@
-github: problame
 patreon: zrepl
 liberapay: zrepl
 custom: https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96
@@ -0,0 +1,70 @@
+dist: xenial
+services:
+- docker
+
+env: # for allow_failures: https://docs.travis-ci.com/user/customizing-the-build/
+matrix:
+  include:
+
+  - language: go
+    name: "Build in Docker (docs/installation.rst)"
+    script:
+    - sudo docker build -t zrepl_build -f build.Dockerfile .
+    - |
+      sudo docker run -it --rm \
+        -v "${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"
@@ -28,8 +28,6 @@ GO_BUILDFLAGS := $(GO_MOD_READONLY) $(GO_EXTRA_BUILDFLAGS)
 GO_BUILD := $(GO_ENV_VARS) $(GO) build $(GO_BUILDFLAGS) -ldflags $(GO_LDFLAGS)
 GOLANGCI_LINT := golangci-lint
 GOCOVMERGE := gocovmerge
-RELEASE_DOCKER_BASEIMAGE_TAG ?= 1.15
-RELEASE_DOCKER_BASEIMAGE ?= golang:$(RELEASE_DOCKER_BASEIMAGE_TAG)

 ifneq ($(GOARM),)
 	ZREPL_TARGET_TUPLE := $(GOOS)-$(GOARCH)v$(GOARM)
@@ -59,85 +57,6 @@ ifeq (SIGN, 1)
 endif
 	@echo "ZREPL RELEASE ARTIFACTS AVAILABLE IN artifacts/release"

-release-docker: $(ARTIFACTDIR)
-	sed 's/FROM.*!SUBSTITUTED_BY_MAKEFILE/FROM $(RELEASE_DOCKER_BASEIMAGE)/' build.Dockerfile > artifacts/release-docker.Dockerfile
-	docker build -t zrepl_release --pull -f artifacts/release-docker.Dockerfile .
-	docker run --rm -i -v $(CURDIR):/src -u $$(id -u):$$(id -g) \
-		zrepl_release \
-		make release GOOS=$(GOOS) GOARCH=$(GOARCH) GOARM=$(GOARM)
-
-debs-docker:
-	$(MAKE) _debs_or_rpms_docker _DEB_OR_RPM=deb
-rpms-docker:
-	$(MAKE) _debs_or_rpms_docker _DEB_OR_RPM=rpm
-_debs_or_rpms_docker: # artifacts/_zrepl.zsh_completion artifacts/bash_completion docs zrepl-bin
-	$(MAKE) $(_DEB_OR_RPM)-docker GOOS=linux GOARCH=amd64
-	$(MAKE) $(_DEB_OR_RPM)-docker GOOS=linux GOARCH=arm64
-	$(MAKE) $(_DEB_OR_RPM)-docker GOOS=linux GOARCH=arm GOARM=7
-	$(MAKE) $(_DEB_OR_RPM)-docker GOOS=linux GOARCH=386
-
-rpm: $(ARTIFACTDIR) # artifacts/_zrepl.zsh_completion artifacts/bash_completion docs zrepl-bin
-	$(eval _ZREPL_RPM_VERSION := $(subst -,.,$(_ZREPL_VERSION)))
-	$(eval _ZREPL_RPM_TOPDIR_ABS := $(CURDIR)/$(ARTIFACTDIR)/rpmbuild)
-	rm -rf "$(_ZREPL_RPM_TOPDIR_ABS)"
-	mkdir "$(_ZREPL_RPM_TOPDIR_ABS)"
-	mkdir -p "$(_ZREPL_RPM_TOPDIR_ABS)"/{SPECS,RPMS,BUILD,BUILDROOT}
-	sed "s/^Version:.*/Version:          $(_ZREPL_RPM_VERSION)/g" \
-	    packaging/rpm/zrepl.spec >  $(_ZREPL_RPM_TOPDIR_ABS)/SPECS/zrepl.spec
-
-	# see /usr/lib/rpm/platform
-ifeq ($(GOARCH),amd64)
-	$(eval _ZREPL_RPMBUILD_TARGET := x86_64)
-else ifeq ($(GOARCH), 386)
-	$(eval _ZREPL_RPMBUILD_TARGET := i386)
-else ifeq ($(GOARCH), arm64)
-	$(eval _ZREPL_RPMBUILD_TARGET := aarch64)
-else ifeq ($(GOARCH), arm)
-	$(eval _ZREPL_RPMBUILD_TARGET := armv7hl)
-else
-	$(eval _ZREPL_RPMBUILD_TARGET := $(GOARCH))
-endif
-	rpmbuild \
-		--build-in-place \
-		--define "_sourcedir $(CURDIR)" \
-		--define "_topdir $(_ZREPL_RPM_TOPDIR_ABS)" \
-		--define "_zrepl_binary_filename zrepl-$(ZREPL_TARGET_TUPLE)" \
-		--target $(_ZREPL_RPMBUILD_TARGET) \
-		-bb "$(_ZREPL_RPM_TOPDIR_ABS)"/SPECS/zrepl.spec
-	cp "$(_ZREPL_RPM_TOPDIR_ABS)"/RPMS/$(_ZREPL_RPMBUILD_TARGET)/zrepl-$(_ZREPL_RPM_VERSION)*.rpm $(ARTIFACTDIR)/
-
-rpm-docker:
-	docker build -t zrepl_rpm_pkg --pull -f packaging/rpm/Dockerfile .
-	docker run --rm -i -v $(CURDIR):/build/src -u $$(id -u):$$(id -g) \
-		zrepl_rpm_pkg \
-		make rpm GOOS=$(GOOS) GOARCH=$(GOARCH) GOARM=$(GOARM)
-
-deb: $(ARTIFACTDIR) # artifacts/_zrepl.zsh_completion artifacts/bash_completion docs zrepl-bin
-
-	cp packaging/deb/debian/changelog.template packaging/deb/debian/changelog
-	sed -i 's/DATE_DASH_R_OUTPUT/$(shell date -R)/' packaging/deb/debian/changelog
-	VERSION="$(subst -,.,$(_ZREPL_VERSION))"; \
-		export VERSION="$${VERSION#v}"; \
-		sed -i 's/VERSION/'"$$VERSION"'/' packaging/deb/debian/changelog
-
-ifeq ($(GOARCH), arm)
-	$(eval  DEB_HOST_ARCH := armhf)
-else ifeq ($(GOARCH), 386)
-	$(eval DEB_HOST_ARCH := i386)
-else
-	$(eval DEB_HOST_ARCH := $(GOARCH))
-endif
-
-	export ZREPL_DPKG_ZREPL_BINARY_FILENAME=zrepl-$(ZREPL_TARGET_TUPLE); \
-		dpkg-buildpackage -b --no-sign --host-arch $(DEB_HOST_ARCH)
-	cp ../*.deb artifacts/
-
-deb-docker:
-	docker build -t zrepl_debian_pkg --pull -f packaging/deb/Dockerfile .
-	docker run --rm -i -v $(CURDIR):/build/src -u $$(id -u):$$(id -g) \
-		zrepl_debian_pkg \
-		make deb GOOS=$(GOOS) GOARCH=$(GOARCH) GOARM=$(GOARM)
-
 # expects `release` target to have run before
 NOARCH_TARBALL := $(ARTIFACTDIR)/zrepl-noarch.tar
 wrapup-and-checksum:
@@ -228,41 +147,26 @@ zrepl-bin:
 generate-platform-test-list:
 	$(GO_BUILD) -o $(ARTIFACTDIR)/generate-platform-test-list ./platformtest/tests/gen

-COVER_PLATFORM_BIN_PATH := $(ARTIFACTDIR)/platformtest-cover-$(ZREPL_TARGET_TUPLE)
-cover-platform-bin:
+test-platform-bin:
 	$(GO_ENV_VARS) $(GO) test $(GO_BUILDFLAGS) \
-		-c -o "$(COVER_PLATFORM_BIN_PATH)" \
+		-c -o "$(ARTIFACTDIR)/platformtest-$(ZREPL_TARGET_TUPLE)" \
 		-covermode=atomic -cover -coverpkg github.com/zrepl/zrepl/... \
 		./platformtest/harness
-cover-platform:
-	# do not track dependency on cover-platform-bin to allow build of binary outside of test VM
-	export _TEST_PLATFORM_CMD="$(COVER_PLATFORM_BIN_PATH) \
-		-test.coverprofile \"$(ARTIFACTDIR)/platformtest.cover\" \
-		-test.v \
-		__DEVEL--i-heard-you-like-tests"; \
-		$(MAKE) _test-or-cover-platform-impl
-
-TEST_PLATFORM_BIN_PATH := $(ARTIFACTDIR)/platformtest-$(ZREPL_TARGET_TUPLE)
-test-platform-bin:
-	$(GO_BUILD) -o "$(TEST_PLATFORM_BIN_PATH)" ./platformtest/harness
-test-platform:
-	export _TEST_PLATFORM_CMD="\"$(TEST_PLATFORM_BIN_PATH)\""; \
-		$(MAKE) _test-or-cover-platform-impl

 ZREPL_PLATFORMTEST_POOLNAME := zreplplatformtest
 ZREPL_PLATFORMTEST_IMAGEPATH := /tmp/zreplplatformtest.pool.img
 ZREPL_PLATFORMTEST_MOUNTPOINT := /tmp/zreplplatformtest.pool
 ZREPL_PLATFORMTEST_ZFS_LOG := /tmp/zreplplatformtest.zfs.log
 # ZREPL_PLATFORMTEST_STOP_AND_KEEP := -failure.stop-and-keep-pool
-ZREPL_PLATFORMTEST_ARGS :=
-_test-or-cover-platform-impl: $(ARTIFACTDIR)
-ifndef _TEST_PLATFORM_CMD
-	$(error _TEST_PLATFORM_CMD is undefined, caller 'cover-platform' or 'test-platform' should have defined it)
-endif
+ZREPL_PLATFORMTEST_ARGS := 
+test-platform: $(ARTIFACTDIR) # do not track dependency on test-platform-bin to allow build of platformtest outside of test VM
 	rm -f "$(ZREPL_PLATFORMTEST_ZFS_LOG)"
 	rm -f "$(ARTIFACTDIR)/platformtest.cover"
 	platformtest/logmockzfs/logzfsenv "$(ZREPL_PLATFORMTEST_ZFS_LOG)" `which zfs` \
-		$(_TEST_PLATFORM_CMD) \
+	"$(ARTIFACTDIR)/platformtest-$(ZREPL_TARGET_TUPLE)" \
+		-test.coverprofile "$(ARTIFACTDIR)/platformtest.cover" \
+		-test.v \
+		__DEVEL--i-heard-you-like-tests \
 		-poolname "$(ZREPL_PLATFORMTEST_POOLNAME)" \
 		-imagepath "$(ZREPL_PLATFORMTEST_IMAGEPATH)" \
 		-mountpoint "$(ZREPL_PLATFORMTEST_MOUNTPOINT)" \
@@ -274,31 +178,22 @@ cover-merge: $(ARTIFACTDIR)
 cover-html: cover-merge
 	$(GO) tool cover -html "$(ARTIFACTDIR)/merged.cover" -o "$(ARTIFACTDIR)/merged.cover.html"

-cover-full:
+test-full:
 	test "$$(id -u)" = "0" || echo "MUST RUN AS ROOT" 1>&2
 	$(MAKE) test-go COVER=1
-	$(MAKE) cover-platform-bin
-	$(MAKE) cover-platform
+	$(MAKE) test-platform
 	$(MAKE) cover-html

 ##################### DEV TARGETS #####################
 # not part of the build, must do that manually
-.PHONY: generate formatcheck format
+.PHONY: generate format

 generate: generate-platform-test-list
 	protoc -I=replication/logic/pdu --go_out=plugins=grpc:replication/logic/pdu replication/logic/pdu/pdu.proto
 	$(GO_ENV_VARS) $(GO) generate $(GO_BUILDFLAGS) -x ./...

-GOIMPORTS := goimports -srcdir . -local 'github.com/zrepl/zrepl'
-FINDSRCFILES := find . -type f -name '*.go' -not -path "./vendor/*" -not -name '*.pb.go' -not -name '*_enumer.go'
-
-formatcheck:
-	@# goimports doesn't have a knob to exit with non-zero status code if formatting is needed
-	@# see https://go-review.googlesource.com/c/tools/+/237378
-	@ affectedfiles=$$($(GOIMPORTS) -l $(shell $(FINDSRCFILES)) | tee /dev/stderr | wc -l); test "$$affectedfiles" = 0
-
 format:
-	@ $(GOIMPORTS) -w -d $(shell  $(FINDSRCFILES))
+	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')

 ##################### NOARCH #####################
 .PHONY: noarch $(ARTIFACTDIR)/bash_completion $(ARTIFACTDIR)/_zrepl.zsh_completion $(ARTIFACTDIR)/go_env.txt docs docs-clean
@@ -322,14 +217,11 @@ $(ARTIFACTDIR)/_zrepl.zsh_completion:

 $(ARTIFACTDIR)/go_env.txt:
 	$(GO_ENV_VARS) $(GO) env > $@
-	$(GO) version >> $@

 docs: $(ARTIFACTDIR)/docs
-	# https://www.sphinx-doc.org/en/master/man/sphinx-build.html
 	make -C docs \
 		html \
 		BUILDDIR=../artifacts/docs \
-		SPHINXOPTS="-W --keep-going -n"

 docs-clean:
 	make -C docs \
@@ -1,9 +1,8 @@
 [![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 Patreon](https://img.shields.io/endpoint.svg?url=https%3A%2F%2Fshieldsio-patreon.vercel.app%2Fapi%3Fusername%3Dzrepl%26type%3Dpatrons&style=flat&color=yellow)](https://patreon.com/zrepl)
-[![Donate via GitHub Sponsors](https://img.shields.io/static/v1?label=Sponsor&message=%E2%9D%A4&logo=GitHub&style=flat&color=yellow)](https://github.com/sponsors/problame)
-[![Donate via Liberapay](https://img.shields.io/liberapay/patrons/zrepl.svg?logo=liberapay)](https://liberapay.com/zrepl/donate)
+[![Donate via Patreon](https://img.shields.io/endpoint.svg?url=https%3A%2F%2Fshieldsio-patreon.herokuapp.com%2Fzrepl%2Fpledges&style=flat&color=yellow)](https://www.patreon.com/zrepl)
+[![Donate via Liberapay](https://img.shields.io/liberapay/receives/zrepl.svg?logo=liberapay)](https://liberapay.com/zrepl/donate)
 [![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)
 [![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)

@@ -21,7 +20,7 @@ zrepl is a one-stop ZFS backup & replication solution.

 ## Feature Requests

-1. Does your feature request require default values / some kind of configuration?
+1. Does you feature request require default values / some kind of configuration?
   If so, think of an expressive configuration example.
 2. Think of at least one use case that generalizes from your concrete application.
 3. Open an issue on GitHub with example conf & use case attached.
@@ -30,12 +29,20 @@ zrepl is a one-stop ZFS backup & replication solution.
 The above does not apply if you already implemented everything.
 Check out the *Coding Workflow* section below for details.

-## Building, Releasing, Downstream-Packaging
+## Package Maintainer Information

-This section provides an overview of the zrepl build & release process.
-Check out `docs/installation/compile-from-source.rst` for build-from-source instructions.
+* Follow the steps in `docs/installation.rst -> Compiling from Source` and read the Makefile / shell scripts used in this process.
+* 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 `sudo make test-platform` **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.

-### Overview
+## Developer Documentation

 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.
@@ -52,61 +59,60 @@ An HTML report can be generated using `make cover-html`.

 **Code generation** is triggered by `make generate`. Generated code is committed to the source tree.

-### Build & Release Process
+### Project Structure

-**The `Makefile` is catering to the needs of developers & CI, not distro packagers**.
-It provides phony targets for
-* local development (building, running tests, etc)
-* building a release in Docker (used by the CI & release management)
-* building .deb and .rpm packages out of the release artifacts.
+```
+├── 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
+│   ├── hooks               # snapshot hooks
+│   ├── job                 # job implementations
+│   ├── logging             # logging outlets + formatters
+│   ├── nethelpers
+│   ├── prometheus
+│   ├── pruner              # pruner implementation
+│   ├── snapper             # snapshotter implementation
+├── dist                    # supplemental material for users & package maintainers
+├── docs                    # sphinx-based documentation
+│   ├── **/*.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
+├── endpoint                # implementation of replication endpoints (=> package replication)
+├── logger                  # our own logger package
+├── platformtest            # test suite for our zfs abstractions (error classification, etc)
+├── 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               # transport implementations
+│   ├── fromconfig
+│   ├── local
+│   ├── ssh
+│   ├── tcp
+│   └── tls
+├── util
+├── version                 # abstraction for versions (filled during build by Makefile)
+└── zfs                     # zfs(8) wrappers
+```

-**Build tooling & dependencies** are documented as code in `lazy.sh`.
-Go dependencies are then fetched by the go command and pip dependencies are pinned through a `requirements.txt`.
-
-**We use CircleCI for continuous integration**.
-There are two workflows:
-
-* `ci` runs for every commit / branch / tag pushed to GitHub.
-  It is supposed to run very fast (<5min and provides quick feedback to developers).
-  It runs formatting checks, lints and tests on the most important OSes / architectures.
-  Artifacts are published to minio.cschwarz.com (see GitHub Commit Status).
-
-* `release` runs
-  * on manual triggers through the CircleCI API (in order to produce a release)
-  * periodically on `master`
-  Artifacts are published to minio.cschwarz.com (see GitHub Commit Status).
-
-**Releases** are issued via Git tags + GitHub Releases feature.
-The procedure to issue a release is as follows:
-* Issue the source release:
-  * Git tag the release on the `master` branch.
-  * Push the tag.
-  * Run `./docs/publish.sh` to re-build & push zrepl.github.io.
-* Issue the official binary release:
-  * Run the `release` pipeline (triggered via CircleCI API)
-  * Download the artifacts to the release manager's machine.
-  * Create a GitHub release, edit the changelog, upload all the release artifacts, including .rpm and .deb files.
-  * Issue the GitHub release.
-  * Add the .rpm and .deb files to the official zrepl repos, publish those.
-
-**Official binary releases are not re-built when Go receives an update. If the Go update is critical to zrepl (e.g. a Go security update that affects zrepl), we'd issue a new source release**.
-The rationale for this is that whereas distros provide a mechanism for this (`$zrepl_source_release-$distro_package_revision`), GitHub Releases doesn't which means we'd need to update the existing GitHub release's assets, which nobody would notice (no RSS feed updates, etc.).
-Downstream packagers can read the changelog to determine whether they want to push that minor release into their distro or simply skip it.
-
-### Additional Notes to Distro Package Maintainers
-
-* Use `sudo make test-platform-bin && sudo make test-platform` **on a test system** to validate that zrepl's abstractions on top of ZFS work with the system ZFS.
-* 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/`.
-* Have a look at the `Makefile`'s `ZREPL_VERSION` variable and how it passed to Go's `ldFlags`.
-  This is how `zrepl version` knows what version number to show.
-  Your build system should set the `ldFlags` flags appropriately and add a prefix or suffix that indicates that the given zrepl binary is a distro build, not an official one.
-* Make sure you are informed about new zrepl versions, e.g. by subscribing to GitHub's release RSS feed.
-
-
-## Contributing Code
+### Coding Workflow

 * Open an issue when starting to hack on a new feature
 * Commits should reference the issue they are related to
@@ -116,6 +122,9 @@ Downstream packagers can read the changelog to determine whether they want to pu

 Backward-incompatible changes must be documented in the git commit message and are listed in `docs/changelog.rst`.

+* Config-breaking changes must contain a line `BREAK CONFIG` in the commit message
+* Other breaking changes must contain a line `BREAK` in the commit message
+
 ### Glossary & Naming Inconsistencies

 In ZFS, *dataset* refers to the objects *filesystem*, *ZVOL* and *snapshot*. <br />
@@ -132,3 +141,16 @@ variables and types are often named *dataset* when they in fact refer to a *file
 There will not be a big refactoring (an attempt was made, but it's destroying too much history without much gain).

 However, new contributions & patches should fix naming without further notice in the commit message.
+
+### RPC debugging
+
+Optionally, there are various RPC-related environment variables, that if set to something != `""` will produce additional debug output on stderr:
+
+https://github.com/zrepl/zrepl/blob/master/rpc/rpc_debug.go#L11
+
+https://github.com/zrepl/zrepl/blob/master/rpc/dataconn/dataconn_debug.go#L11
+
+https://github.com/zrepl/zrepl/blob/master/rpc/dataconn/stream/stream_debug.go#L11
+
+https://github.com/zrepl/zrepl/blob/master/rpc/dataconn/heartbeatconn/heartbeatconn_debug.go#L11
+
@@ -1,4 +1,4 @@
-FROM !SUBSTITUTED_BY_MAKEFILE
+FROM golang:latest

 RUN apt-get update && apt-get install -y \
    python3-pip \
@@ -7,7 +7,6 @@ import (

 	"github.com/spf13/cobra"
 	"github.com/spf13/pflag"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"

 	"github.com/zrepl/zrepl/config"
@@ -480,11 +480,6 @@ func (t *tui) renderReplicationReport(rep *report.Report, history *bytesProgress
 			t.newline()
 		}

-		if len(latest.Filesystems) == 0 {
-			t.write("NOTE: no filesystems were considered for replication!")
-			t.newline()
-		}
-
 		var maxFSLen int
 		for _, fs := range latest.Filesystems {
 			if len(fs.Info.Name) > maxFSLen {
@@ -63,7 +63,7 @@ func doZabsList(ctx context.Context, sc *cli.Subcommand, args []string) error {
 				defer line.Lock().Unlock()
 				if zabsListFlags.Json {
 					enc.SetIndent("", "  ")
-					if err := enc.Encode(a); err != nil {
+					if err := enc.Encode(abstractions); err != nil {
 						panic(err)
 					}
 					fmt.Println()
@@ -52,12 +52,11 @@ func (j JobEnum) Name() string {
 }

 type ActiveJob struct {
-	Type        string                `yaml:"type"`
-	Name        string                `yaml:"name"`
-	Connect     ConnectEnum           `yaml:"connect"`
-	Pruning     PruningSenderReceiver `yaml:"pruning"`
-	Debug       JobDebugSettings      `yaml:"debug,optional"`
-	Replication *Replication          `yaml:"replication,optional,fromdefaults"`
+	Type    string                `yaml:"type"`
+	Name    string                `yaml:"name"`
+	Connect ConnectEnum           `yaml:"connect"`
+	Pruning PruningSenderReceiver `yaml:"pruning"`
+	Debug   JobDebugSettings      `yaml:"debug,optional"`
 }

 type PassiveJob struct {
@@ -77,7 +76,13 @@ type SnapJob struct {
 }

 type SendOptions struct {
-	Encrypted bool `yaml:"encrypted,optional,default=false"`
+	Encrypted bool `yaml:"encrypted"`
+}
+
+var _ yaml.Defaulter = (*SendOptions)(nil)
+
+func (l *SendOptions) SetDefault() {
+	*l = SendOptions{Encrypted: false}
 }

 type RecvOptions struct {
@@ -88,13 +93,10 @@ type RecvOptions struct {
 	// Reencrypt bool `yaml:"reencrypt"`
 }

-type Replication struct {
-	Protection *ReplicationOptionsProtection `yaml:"protection,optional,fromdefaults"`
-}
+var _ yaml.Defaulter = (*RecvOptions)(nil)

-type ReplicationOptionsProtection struct {
-	Initial     string `yaml:"initial,optional,default=guarantee_resumability"`
-	Incremental string `yaml:"incremental,optional,default=guarantee_resumability"`
+func (l *RecvOptions) SetDefault() {
+	*l = RecvOptions{}
 }

 type PushJob struct {
@@ -104,9 +106,6 @@ type PushJob struct {
 	Send         *SendOptions      `yaml:"send,fromdefaults,optional"`
 }

-func (j *PushJob) GetFilesystems() FilesystemsFilter { return j.Filesystems }
-func (j *PushJob) GetSendOptions() *SendOptions      { return j.Send }
-
 type PullJob struct {
 	ActiveJob `yaml:",inline"`
 	RootFS    string                   `yaml:"root_fs"`
@@ -114,10 +113,6 @@ type PullJob struct {
 	Recv      *RecvOptions             `yaml:"recv,fromdefaults,optional"`
 }

-func (j *PullJob) GetRootFS() string             { return j.RootFS }
-func (j *PullJob) GetAppendClientIdentity() bool { return false }
-func (j *PullJob) GetRecvOptions() *RecvOptions  { return j.Recv }
-
 type PositiveDurationOrManual struct {
 	Interval time.Duration
 	Manual   bool
@@ -155,10 +150,6 @@ type SinkJob struct {
 	Recv       *RecvOptions `yaml:"recv,optional,fromdefaults"`
 }

-func (j *SinkJob) GetRootFS() string             { return j.RootFS }
-func (j *SinkJob) GetAppendClientIdentity() bool { return true }
-func (j *SinkJob) GetRecvOptions() *RecvOptions  { return j.Recv }
-
 type SourceJob struct {
 	PassiveJob   `yaml:",inline"`
 	Snapshotting SnapshottingEnum  `yaml:"snapshotting"`
@@ -166,9 +157,6 @@ type SourceJob struct {
 	Send         *SendOptions      `yaml:"send,optional,fromdefaults"`
 }

-func (j *SourceJob) GetFilesystems() FilesystemsFilter { return j.Filesystems }
-func (j *SourceJob) GetSendOptions() *SendOptions      { return j.Send }
-
 type FilesystemsFilter map[string]bool

 type SnapshottingEnum struct {
@@ -321,10 +309,14 @@ type PruneKeepNotReplicated struct {
 	KeepSnapshotAtCursor bool   `yaml:"keep_snapshot_at_cursor,optional,default=true"`
 }

+type PruneKeepStepHolds struct {
+	Type             string   `yaml:"type"`
+	AdditionalJobIds []string `yaml:"additional_job_ids,optional"`
+}
+
 type PruneKeepLastN struct {
 	Type  string `yaml:"type"`
 	Count int    `yaml:"count"`
-	Regex string `yaml:"regex,optional"`
 }

 type PruneKeepRegex struct { // FIXME rename to KeepRegex
@@ -493,6 +485,7 @@ func (t *ServeEnum) UnmarshalYAML(u func(interface{}, bool) error) (err error) {

 func (t *PruningEnum) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
 	t.Ret, err = enumUnmarshal(u, map[string]interface{}{
+		"step_holds":     &PruneKeepStepHolds{},
 		"not_replicated": &PruneKeepNotReplicated{},
 		"last_n":         &PruneKeepLastN{},
 		"grid":           &PruneGrid{},
@@ -1,90 +0,0 @@
-package config
-
-import (
-	"fmt"
-	"testing"
-
-	"github.com/stretchr/testify/require"
-	"github.com/zrepl/yaml-config"
-)
-
-type A struct {
-	B  *B     `yaml:"b,optional,fromdefaults"`
-	A1 string `yaml:"a1,optional"`
-}
-
-type B struct {
-	C *C     `yaml:"c,optional,fromdefaults"`
-	D string `yaml:"d,default=ddd"`
-	E string `yaml:"e,optional"`
-}
-
-type C struct {
-	Q string `yaml:"q,optional"`
-	R string `yaml:"r,default=r"`
-}
-
-func TestDepFromDefaults(t *testing.T) {
-
-	type testcase struct {
-		name   string
-		yaml   string
-		expect *A
-	}
-
-	tcs := []testcase{
-		{
-			name: "empty",
-			yaml: `{}`,
-			expect: &A{
-				B: &B{
-					C: &C{
-						R: "r",
-					},
-					D: "ddd",
-				},
-			},
-		},
-		{
-			name: "a1 set",
-			yaml: `{"a1":"blah"}`,
-			expect: &A{
-				A1: "blah",
-				B: &B{
-					C: &C{
-						R: "r",
-					},
-					D: "ddd",
-				},
-			},
-		},
-		{
-			name: "D set",
-			yaml: `
-b:
-  d: 4d
-`,
-			expect: &A{
-				B: &B{
-					D: "4d",
-					C: &C{
-						R: "r",
-					},
-				},
-			},
-		},
-	}
-
-	for tci := range tcs {
-		t.Run(fmt.Sprintf("%d-%s", tci, tcs[tci].name), func(t *testing.T) {
-			tc := tcs[tci]
-
-			var a A
-			err := yaml.UnmarshalStrict([]byte(tc.yaml), &a)
-			require.NoError(t, err)
-
-			require.Equal(t, tc.expect, &a)
-		})
-	}
-
-}
@@ -60,9 +60,9 @@ jobs:
 	})

 	t.Run("encrypted_unspecified", func(t *testing.T) {
-		c = testValidConfig(t, fill(encrypted_unspecified))
-		encrypted := c.Jobs[0].Ret.(*PushJob).Send.Encrypted
-		assert.Equal(t, false, encrypted)
+		c, err := testConfig(t, fill(encrypted_unspecified))
+		assert.Error(t, err)
+		assert.Nil(t, c)
 	})

 	t.Run("send_not_specified", func(t *testing.T) {
@@ -37,7 +37,7 @@ func (t *RetentionIntervalList) UnmarshalYAML(u func(interface{}, bool) error) (
 		return err
 	}

-	intervals, err := ParseRetentionIntervalSpec(in)
+	intervals, err := parseRetentionGridIntervalsString(in)
 	if err != nil {
 		return err
 	}
@@ -102,7 +102,7 @@ func parseRetentionGridIntervalString(e string) (intervals []RetentionInterval,

 }

-func ParseRetentionIntervalSpec(s string) (intervals []RetentionInterval, err error) {
+func parseRetentionGridIntervalsString(s string) (intervals []RetentionInterval, err error) {

 	ges := strings.Split(s, "|")
 	intervals = make([]RetentionInterval, 0, 7*len(ges))
@@ -1,88 +0,0 @@
-# This config serves as an example for a local zrepl installation that
-# backups the entire zpool `system` to `backuppool/zrepl/sink`
-#
-# The requirements covered by this setup are described in the zrepl documentation's
-# quick start section which inlines this example.
-#
-# CUSTOMIZATIONS YOU WILL LIKELY WANT TO APPLY:
-# - adjust the name of the production pool `system` in the `filesystems` filter of jobs `snapjob` and `push_to_derive`
-# - adjust the name of the backup pool `backuppool` in the `backuppool_sink` job
-# - adjust the occurences of `myhostname` to the name of the system you are backing up (cannot be easily changed once you start replicating)
-# - make sure the `zrepl_` prefix is not being used by any other zfs tools you might have installed (it likely isn't)
-
-jobs:
-
-# this job takes care of snapshot creation + pruning
- name: snapjob
-  type: snap
-  filesystems: {
-      "system<": true,
-  }
-  # create snapshots with prefix `zrepl_` every 15 minutes
-  snapshotting:
-    type: periodic
-    interval: 15m
-    prefix: zrepl_
-  pruning:
-    keep:
-    # fade-out scheme for snapshots starting with `zrepl_`
-    # - keep all created in the last hour
-    # - then destroy snapshots such that we keep 24 each 1 hour apart
-    # - then destroy snapshots such that we keep 14 each 1 day apart
-    # - then destroy all older snapshots
-    - type: grid
-      grid: 1x1h(keep=all) | 24x1h | 14x1d
-      regex: "^zrepl_.*"
-    # keep all snapshots that don't have the `zrepl_` prefix
-    - type: regex
-      negate: true
-      regex: "^zrepl_.*"
-
-# This job pushes to the local sink defined in job `backuppool_sink`.
-# We trigger replication manually from the command line / udev rules using
-#  `zrepl signal wakeup push_to_drive`
- type: push
-  name: push_to_drive
-  connect:
-    type: local
-    listener_name: backuppool_sink
-    client_identity: myhostname
-  filesystems: {
-      "system<": true
-  }
-  send:
-    encrypted: true
-  replication:
-    protection:
-      initial: guarantee_resumability
-      # Downgrade protection to guarantee_incremental which uses zfs bookmarks instead of zfs holds.
-      # Thus, when we yank out the backup drive during replication
-      # - we might not be able to resume the interrupted replication step because the partially received `to` snapshot of a `from`->`to` step may be pruned any time
-      # - but in exchange we get back the disk space allocated by `to` when we prune it
-      # - and because we still have the bookmarks created by `guarantee_incremental`, we can still do incremental replication of `from`->`to2` in the future
-      incremental: guarantee_incremental
-  snapshotting:
-    type: manual
-  pruning:
-    # no-op prune rule on sender (keep all snapshots), job `snapshot` takes care of this
-    keep_sender:
-    - type: regex
-      regex: ".*"
-    # retain 
-    keep_receiver:
-    # longer retention on the backup drive, we have more space there
-    - type: grid
-      grid: 1x1h(keep=all) | 24x1h | 360x1d
-      regex: "^zrepl_.*"
-    # retain all non-zrepl snapshots on the backup drive
-    - type: regex
-      negate: true
-      regex: "^zrepl_.*"
-
-# This job receives from job `push_to_drive` into `backuppool/zrepl/sink/myhostname`
- type: sink
-  name: backuppool_sink
-  root_fs: "backuppool/zrepl/sink"
-  serve:
-    type: local
-    listener_name: backuppool_sink
@@ -1,12 +0,0 @@
-jobs:
- name: sink
-  type: sink
-  serve:
-      type: tls
-      listen: ":8888"
-      ca: "/etc/zrepl/prod.crt"
-      cert: "/etc/zrepl/backups.crt"
-      key: "/etc/zrepl/backups.key"
-      client_cns:
-        - "prod"
-  root_fs: "storage/zrepl/sink"
@@ -1,28 +0,0 @@
-jobs:
- name: prod_to_backups
-  type: push
-  connect:
-    type: tls
-    address: "backups.example.com:8888"
-    ca: /etc/zrepl/backups.crt
-    cert: /etc/zrepl/prod.crt
-    key:  /etc/zrepl/prod.key
-    server_cn: "backups"
-  filesystems: {
-    "zroot/var/db": true,
-    "zroot/usr/home<": true,
-    "zroot/usr/home/paranoid": false
-  }
-  snapshotting:
-    type: periodic
-    prefix: zrepl_
-    interval: 10m
-  pruning:
-    keep_sender:
-    - type: not_replicated
-    - type: last_n
-      count: 10
-    keep_receiver:
-    - type: grid
-      grid: 1x1h(keep=all) | 24x1h | 30x1d | 6x30d
-      regex: "^zrepl_"
@@ -120,13 +120,7 @@ func (j *controlJob) Run(ctx context.Context) {
 		jsonResponder{log, func() (interface{}, error) {
 			jobs := j.jobs.status()
 			globalZFS := zfscmd.GetReport()
-			envconstReport := envconst.GetReport()
-			s := Status{
-				Jobs: jobs,
-				Global: GlobalStatus{
-					ZFSCmds:  globalZFS,
-					Envconst: envconstReport,
-				}}
+			s := Status{Jobs: jobs, Global: GlobalStatus{ZFSCmds: globalZFS}}
 			return s, nil
 		}})

@@ -12,10 +12,8 @@ import (

 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"
 	"github.com/zrepl/zrepl/endpoint"
-	"github.com/zrepl/zrepl/util/envconst"

 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/job"
@@ -96,7 +94,6 @@ func Run(ctx context.Context, conf *config.Config) error {
 	}

 	// register global (=non job-local) metrics
-	version.PrometheusRegister(prometheus.DefaultRegisterer)
 	zfscmd.RegisterMetrics(prometheus.DefaultRegisterer)
 	trace.RegisterMetrics(prometheus.DefaultRegisterer)
 	endpoint.RegisterMetrics(prometheus.DefaultRegisterer)
@@ -153,8 +150,7 @@ type Status struct {
 }

 type GlobalStatus struct {
-	ZFSCmds  *zfscmd.Report
-	Envconst *envconst.Report
+	ZFSCmds *zfscmd.Report
 }

 func (s *jobs) status() map[string]*job.Status {
@@ -10,7 +10,6 @@ import (
 	"text/template"

 	"github.com/stretchr/testify/require"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"

 	"github.com/zrepl/zrepl/config"
@@ -9,10 +9,10 @@ import (
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/common/log"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"

 	"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/pruner"
@@ -35,10 +35,9 @@ type ActiveSide struct {

 	prunerFactory *pruner.PrunerFactory

-	promRepStateSecs      *prometheus.HistogramVec // labels: state
-	promPruneSecs         *prometheus.HistogramVec // labels: prune_side
-	promBytesReplicated   *prometheus.CounterVec   // labels: filesystem
-	promReplicationErrors prometheus.Gauge
+	promRepStateSecs    *prometheus.HistogramVec // labels: state
+	promPruneSecs       *prometheus.HistogramVec // labels: prune_side
+	promBytesReplicated *prometheus.CounterVec   // labels: filesystem

 	tasksMtx sync.Mutex
 	tasks    activeSideTasks
@@ -146,24 +145,22 @@ func (m *modePush) ResetConnectBackoff() {

 func modePushFromConfig(g *config.Global, in *config.PushJob, jobID endpoint.JobID) (*modePush, error) {
 	m := &modePush{}
-	var err error

-	m.senderConfig, err = buildSenderConfig(in, jobID)
+	fsf, err := filters.DatasetMapFilterFromConfig(in.Filesystems)
 	if err != nil {
-		return nil, errors.Wrap(err, "sender config")
+		return nil, errors.Wrap(err, "cannot build filesystem filter")
 	}

-	replicationConfig, err := logic.ReplicationConfigFromConfig(in.Replication)
-	if err != nil {
-		return nil, errors.Wrap(err, "field `replication`")
+	m.senderConfig = &endpoint.SenderConfig{
+		FSF:     fsf,
+		Encrypt: &zfs.NilBool{B: in.Send.Encrypted},
+		JobID:   jobID,
 	}
-
 	m.plannerPolicy = &logic.PlannerPolicy{
-		EncryptedSend:     logic.TriFromBool(in.Send.Encrypted),
-		ReplicationConfig: *replicationConfig,
+		EncryptedSend: logic.TriFromBool(in.Send.Encrypted),
 	}

-	if m.snapper, err = snapper.FromConfig(g, m.senderConfig.FSF, in.Snapshotting); err != nil {
+	if m.snapper, err = snapper.FromConfig(g, fsf, in.Snapshotting); err != nil {
 		return nil, errors.Wrap(err, "cannot build snapper")
 	}

@@ -175,6 +172,7 @@ type modePull struct {
 	receiver       *endpoint.Receiver
 	receiverConfig endpoint.ReceiverConfig
 	sender         *rpc.Client
+	rootFS         *zfs.DatasetPath
 	plannerPolicy  *logic.PlannerPolicy
 	interval       config.PositiveDurationOrManual
 }
@@ -248,19 +246,26 @@ func modePullFromConfig(g *config.Global, in *config.PullJob, jobID endpoint.Job
 	m = &modePull{}
 	m.interval = in.Interval

-	replicationConfig, err := logic.ReplicationConfigFromConfig(in.Replication)
+	m.rootFS, err = zfs.NewDatasetPath(in.RootFS)
 	if err != nil {
-		return nil, errors.Wrap(err, "field `replication`")
+		return nil, errors.New("RootFS is not a valid zfs filesystem path")
+	}
+	if m.rootFS.Length() <= 0 {
+		return nil, errors.New("RootFS must not be empty") // duplicates error check of receiver
 	}

 	m.plannerPolicy = &logic.PlannerPolicy{
-		EncryptedSend:     logic.DontCare,
-		ReplicationConfig: *replicationConfig,
+		EncryptedSend: logic.DontCare,
 	}

-	m.receiverConfig, err = buildReceiverConfig(in, jobID)
-	if err != nil {
-		return nil, err
+	m.receiverConfig = endpoint.ReceiverConfig{
+		JobID:                      jobID,
+		RootWithoutClientComponent: m.rootFS,
+		AppendClientIdentity:       false, // !
+		UpdateLastReceivedHold:     true,
+	}
+	if err := m.receiverConfig.Validate(); err != nil {
+		return nil, errors.Wrap(err, "cannot build receiver config")
 	}

 	return m, nil
@@ -301,14 +306,6 @@ func activeSide(g *config.Global, in *config.ActiveJob, configJob interface{}) (
 		ConstLabels: prometheus.Labels{"zrepl_job": j.name.String()},
 	}, []string{"filesystem"})

-	j.promReplicationErrors = prometheus.NewGauge(prometheus.GaugeOpts{
-		Namespace:   "zrepl",
-		Subsystem:   "replication",
-		Name:        "filesystem_errors",
-		Help:        "number of filesystems that failed replication in the latest replication attempt, or -1 if the job failed before enumerating the filesystems",
-		ConstLabels: prometheus.Labels{"zrepl_job": j.name.String()},
-	})
-
 	j.connecter, err = fromconfig.ConnecterFromConfig(g, in.Connect)
 	if err != nil {
 		return nil, errors.Wrap(err, "cannot build client")
@@ -333,7 +330,6 @@ func (j *ActiveSide) RegisterMetrics(registerer prometheus.Registerer) {
 	registerer.MustRegister(j.promRepStateSecs)
 	registerer.MustRegister(j.promPruneSecs)
 	registerer.MustRegister(j.promBytesReplicated)
-	registerer.MustRegister(j.promReplicationErrors)
 }

 func (j *ActiveSide) Name() string { return j.name.String() }
@@ -368,7 +364,7 @@ func (j *ActiveSide) OwnedDatasetSubtreeRoot() (rfs *zfs.DatasetPath, ok bool) {
 		_ = j.mode.(*modePush) // make sure we didn't introduce a new job type
 		return nil, false
 	}
-	return pull.receiverConfig.RootWithoutClientComponent.Copy(), true
+	return pull.rootFS.Copy(), true
 }

 func (j *ActiveSide) SenderConfig() *endpoint.SenderConfig {
@@ -466,10 +462,6 @@ func (j *ActiveSide) do(ctx context.Context) {
 		GetLogger(ctx).Info("start replication")
 		repWait(true) // wait blocking
 		repCancel()   // always cancel to free up context resources
-
-		replicationReport := j.tasks.replicationReport()
-		j.promReplicationErrors.Set(float64(replicationReport.GetFailedFilesystemsCountInLatestAttempt()))
-
 		endSpan()
 	}

@@ -5,7 +5,6 @@ import (

 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
-
 	"github.com/zrepl/zrepl/endpoint"
 	"github.com/zrepl/zrepl/transport"
 )
@@ -1,56 +0,0 @@
-package job
-
-import (
-	"github.com/pkg/errors"
-
-	"github.com/zrepl/zrepl/config"
-	"github.com/zrepl/zrepl/daemon/filters"
-	"github.com/zrepl/zrepl/endpoint"
-	"github.com/zrepl/zrepl/zfs"
-)
-
-type SendingJobConfig interface {
-	GetFilesystems() config.FilesystemsFilter
-	GetSendOptions() *config.SendOptions // must not be nil
-}
-
-func buildSenderConfig(in SendingJobConfig, jobID endpoint.JobID) (*endpoint.SenderConfig, error) {
-
-	fsf, err := filters.DatasetMapFilterFromConfig(in.GetFilesystems())
-	if err != nil {
-		return nil, errors.Wrap(err, "cannot build filesystem filter")
-	}
-
-	return &endpoint.SenderConfig{
-		FSF:     fsf,
-		Encrypt: &zfs.NilBool{B: in.GetSendOptions().Encrypted},
-		JobID:   jobID,
-	}, nil
-}
-
-type ReceivingJobConfig interface {
-	GetRootFS() string
-	GetAppendClientIdentity() bool
-	GetRecvOptions() *config.RecvOptions
-}
-
-func buildReceiverConfig(in ReceivingJobConfig, jobID endpoint.JobID) (rc endpoint.ReceiverConfig, err error) {
-	rootFs, err := zfs.NewDatasetPath(in.GetRootFS())
-	if err != nil {
-		return rc, errors.New("root_fs is not a valid zfs filesystem path")
-	}
-	if rootFs.Length() <= 0 {
-		return rc, errors.New("root_fs must not be empty") // duplicates error check of receiver
-	}
-
-	rc = endpoint.ReceiverConfig{
-		JobID:                      jobID,
-		RootWithoutClientComponent: rootFs,
-		AppendClientIdentity:       in.GetAppendClientIdentity(),
-	}
-	if err := rc.Validate(); err != nil {
-		return rc, errors.Wrap(err, "cannot build receiver config")
-	}
-
-	return rc, nil
-}
@@ -2,16 +2,12 @@ package job

 import (
 	"fmt"
-	"path"
-	"path/filepath"
 	"testing"

-	"github.com/kr/pretty"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"

 	"github.com/zrepl/zrepl/config"
-	"github.com/zrepl/zrepl/transport/tls"
 )

 func TestValidateReceivingSidesDoNotOverlap(t *testing.T) {
@@ -112,35 +108,3 @@ jobs:
 	}

 }
-
-func TestSampleConfigsAreBuiltWithoutErrors(t *testing.T) {
-	paths, err := filepath.Glob("../../config/samples/*")
-	if err != nil {
-		t.Errorf("glob failed: %+v", err)
-	}
-
-	for _, p := range paths {
-
-		if path.Ext(p) != ".yml" {
-			t.Logf("skipping file %s", p)
-			continue
-		}
-
-		t.Run(p, func(t *testing.T) {
-			c, err := config.ParseConfig(p)
-			if err != nil {
-				t.Errorf("error parsing %s:\n%+v", p, err)
-			}
-
-			t.Logf("file: %s", p)
-			t.Log(pretty.Sprint(c))
-
-			tls.FakeCertificateLoading(t)
-			jobs, err := JobsFromConfig(c)
-			t.Logf("jobs: %#v", jobs)
-			assert.NoError(t, err)
-		})
-
-	}
-
-}
@@ -6,10 +6,10 @@ import (

 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"

 	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
 	"github.com/zrepl/zrepl/daemon/logging"
 	"github.com/zrepl/zrepl/daemon/snapper"
 	"github.com/zrepl/zrepl/endpoint"
@@ -48,9 +48,19 @@ func (m *modeSink) SnapperReport() *snapper.Report { return nil }
 func modeSinkFromConfig(g *config.Global, in *config.SinkJob, jobID endpoint.JobID) (m *modeSink, err error) {
 	m = &modeSink{}

-	m.receiverConfig, err = buildReceiverConfig(in, jobID)
+	rootDataset, err := zfs.NewDatasetPath(in.RootFS)
 	if err != nil {
-		return nil, err
+		return nil, errors.New("root dataset is not a valid zfs filesystem path")
+	}
+
+	m.receiverConfig = endpoint.ReceiverConfig{
+		JobID:                      jobID,
+		RootWithoutClientComponent: rootDataset,
+		AppendClientIdentity:       true, // !
+		UpdateLastReceivedHold:     true,
+	}
+	if err := m.receiverConfig.Validate(); err != nil {
+		return nil, errors.Wrap(err, "cannot build receiver config")
 	}

 	return m, nil
@@ -64,13 +74,17 @@ type modeSource struct {
 func modeSourceFromConfig(g *config.Global, in *config.SourceJob, jobID endpoint.JobID) (m *modeSource, err error) {
 	// FIXME exact dedup of modePush
 	m = &modeSource{}
-
-	m.senderConfig, err = buildSenderConfig(in, jobID)
+	fsf, err := filters.DatasetMapFilterFromConfig(in.Filesystems)
 	if err != nil {
-		return nil, errors.Wrap(err, "send options")
+		return nil, errors.Wrap(err, "cannot build filesystem filter")
+	}
+	m.senderConfig = &endpoint.SenderConfig{
+		FSF:     fsf,
+		Encrypt: &zfs.NilBool{B: in.Send.Encrypted},
+		JobID:   jobID,
 	}

-	if m.snapper, err = snapper.FromConfig(g, m.senderConfig.FSF, in.Snapshotting); err != nil {
+	if m.snapper, err = snapper.FromConfig(g, fsf, in.Snapshotting); err != nil {
 		return nil, errors.Wrap(err, "cannot build snapper")
 	}

@@ -7,7 +7,6 @@ import (

 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"

 	"github.com/zrepl/zrepl/config"
@@ -9,7 +9,6 @@ import (

 	"github.com/mattn/go-isatty"
 	"github.com/pkg/errors"
-
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/logging/trace"
 	"github.com/zrepl/zrepl/logger"
@@ -93,15 +93,11 @@ package trace
 import (
 	"context"
 	"fmt"
-	"os"
-	runtimedebug "runtime/debug"
 	"strings"
 	"time"

-	"github.com/kr/pretty"
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
-
 	"github.com/zrepl/zrepl/util/chainlock"
 )

@@ -124,19 +120,15 @@ func RegisterMetrics(r prometheus.Registerer) {
 }

 type traceNode struct {
-	// NOTE: members with prefix debug are only valid if the package variable debugEnabled is true.
-
-	id                 string
-	annotation         string
-	parentTask         *traceNode
-	debugCreationStack string // debug: stack trace of when the traceNode was created
+	id         string
+	annotation string
+	parentTask *traceNode

 	mtx chainlock.L

-	activeChildTasks      int32               // only for task nodes, insignificant for span nodes
-	debugActiveChildTasks map[*traceNode]bool // debug: set of active child tasks
-	parentSpan            *traceNode
-	activeChildSpan       *traceNode // nil if task or span doesn't have an active child span
+	activeChildTasks int32 // only for task nodes, insignificant for span nodes
+	parentSpan       *traceNode
+	activeChildSpan  *traceNode // nil if task or span doesn't have an active child span

 	startedAt time.Time
 	endedAt   time.Time
@@ -145,14 +137,6 @@ type traceNode struct {
 func (s *traceNode) StartedAt() time.Time { return s.startedAt }
 func (s *traceNode) EndedAt() time.Time   { return s.endedAt }

-// caller must hold mtx
-func (s *traceNode) debugString() string {
-	if !debugEnabled {
-		return "<debugEnabled=false>"
-	}
-	return pretty.Sprint(s)
-}
-
 // Returned from WithTask or WithSpan.
 // Must be called once the task or span ends.
 // See package-level docs for nesting rules.
@@ -212,16 +196,8 @@ func WithTask(ctx context.Context, taskName string) (context.Context, DoneFunc)
 		endedAt:   time.Time{},
 	}

-	if debugEnabled {
-		this.debugCreationStack = string(runtimedebug.Stack())
-		this.debugActiveChildTasks = map[*traceNode]bool{}
-	}
-
 	if parentTask != nil {
 		this.parentTask.activeChildTasks++
-		if debugEnabled {
-			this.parentTask.debugActiveChildTasks[this] = true
-		}
 		parentTask.mtx.Unlock()
 	}

@@ -242,11 +218,7 @@ func WithTask(ctx context.Context, taskName string) (context.Context, DoneFunc)
 			defer this.mtx.Lock().Unlock()

 			if this.activeChildTasks != 0 {
-				if debugEnabled {
-					// the debugString can be quite long and panic won't print it completely
-					fmt.Fprintf(os.Stderr, "going to panic due to activeChildTasks:\n%s\n", this.debugString())
-				}
-				panic(errors.WithMessagef(ErrTaskStillHasActiveChildTasks, "end task: %v active child tasks\n", this.activeChildTasks))
+				panic(errors.Wrapf(ErrTaskStillHasActiveChildTasks, "end task: %v active child tasks", this.activeChildTasks))
 			}

 			// support idempotent task ends
@@ -257,15 +229,8 @@ func WithTask(ctx context.Context, taskName string) (context.Context, DoneFunc)

 			if this.parentTask != nil {
 				this.parentTask.activeChildTasks--
-				if debugEnabled {
-					delete(this.parentTask.debugActiveChildTasks, this)
-				}
 				if this.parentTask.activeChildTasks < 0 {
-					if debugEnabled {
-						// the debugString can be quite long and panic won't print it completely
-						fmt.Fprintf(os.Stderr, "going to panic due to activeChildTasks < 0:\n%s\n", this.parentTask.debugString())
-					}
-					panic(fmt.Sprintf("impl error: parent task with negative activeChildTasks count: %v", this.parentTask.activeChildTasks))
+					panic("impl error: parent task with negative activeChildTasks count")
 				}
 			}
 			return false
@@ -314,10 +279,6 @@ func WithSpan(ctx context.Context, annotation string) (context.Context, DoneFunc
 		endedAt:   time.Time{},
 	}

-	if debugEnabled {
-		this.debugCreationStack = string(runtimedebug.Stack())
-	}
-
 	parentSpan.mtx.HoldWhile(func() {
 		if parentSpan.activeChildSpan != nil {
 			panic(ErrAlreadyActiveChildSpan)
@@ -3,11 +3,9 @@ package trace
 import (
 	"fmt"
 	"os"
-
-	"github.com/zrepl/zrepl/util/envconst"
 )

-var debugEnabled = envconst.Bool("ZREPL_TRACE_DEBUG_ENABLED", false)
+const debugEnabled = false

 func debug(format string, args ...interface{}) {
 	if !debugEnabled {
@@ -0,0 +1,72 @@
+// Code generated by "enumer -type=FSState -json"; DO NOT EDIT.
+
+//
+package pruner
+
+import (
+	"encoding/json"
+	"fmt"
+)
+
+const _FSStateName = "FSStateInitializedFSStatePlanningFSStatePlanErrFSStateExecutingFSStateExecuteErrFSStateExecuteSuccess"
+
+var _FSStateIndex = [...]uint8{0, 18, 33, 47, 63, 80, 101}
+
+func (i FSState) String() string {
+	if i < 0 || i >= FSState(len(_FSStateIndex)-1) {
+		return fmt.Sprintf("FSState(%d)", i)
+	}
+	return _FSStateName[_FSStateIndex[i]:_FSStateIndex[i+1]]
+}
+
+var _FSStateValues = []FSState{0, 1, 2, 3, 4, 5}
+
+var _FSStateNameToValueMap = map[string]FSState{
+	_FSStateName[0:18]:   0,
+	_FSStateName[18:33]:  1,
+	_FSStateName[33:47]:  2,
+	_FSStateName[47:63]:  3,
+	_FSStateName[63:80]:  4,
+	_FSStateName[80:101]: 5,
+}
+
+// FSStateString retrieves an enum value from the enum constants string name.
+// Throws an error if the param is not part of the enum.
+func FSStateString(s string) (FSState, error) {
+	if val, ok := _FSStateNameToValueMap[s]; ok {
+		return val, nil
+	}
+	return 0, fmt.Errorf("%s does not belong to FSState values", s)
+}
+
+// FSStateValues returns all values of the enum
+func FSStateValues() []FSState {
+	return _FSStateValues
+}
+
+// IsAFSState returns "true" if the value is listed in the enum definition. "false" otherwise
+func (i FSState) IsAFSState() bool {
+	for _, v := range _FSStateValues {
+		if i == v {
+			return true
+		}
+	}
+	return false
+}
+
+// MarshalJSON implements the json.Marshaler interface for FSState
+func (i FSState) MarshalJSON() ([]byte, error) {
+	return json.Marshal(i.String())
+}
+
+// UnmarshalJSON implements the json.Unmarshaler interface for FSState
+func (i *FSState) UnmarshalJSON(data []byte) error {
+	var s string
+	if err := json.Unmarshal(data, &s); err != nil {
+		return fmt.Errorf("FSState should be a string, got %s", data)
+	}
+
+	var err error
+	*i, err = FSStateString(s)
+	return err
+}
@@ -0,0 +1,454 @@
+package pruner
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"os"
+
+	"github.com/pkg/errors"
+	"github.com/zrepl/zrepl/daemon/logging"
+	"github.com/zrepl/zrepl/daemon/logging/trace"
+	"github.com/zrepl/zrepl/endpoint"
+	"github.com/zrepl/zrepl/pruning"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type Pruner struct {
+	fsfilter  endpoint.FSFilter
+	jid       endpoint.JobID
+	side      Side
+	keepRules []pruning.KeepRule
+
+	// all channels consumed by the run loop
+	reportReqs chan reportRequest
+	stopReqs   chan stopRequest
+	done       chan struct{}
+	fsListRes  chan fsListRes
+
+	state State
+
+	listFilesystemsError error       // only in state StateListFilesystemsError
+	fsPruners            []*FSPruner // only in state StateFanOutFilesystems
+}
+
+//go:generate enumer -type=State -json
+type State int
+
+const (
+	StateInitialized State = iota
+	StateListFilesystems
+	StateListFilesystemsError
+	StateFanOutFilesystems
+	StateDone
+)
+
+type Report struct {
+	State                State
+	ListFilesystemsError error       // only valid in StateListFilesystemsError
+	Filesystems          []*FSReport // valid from StateFanOutFilesystems
+}
+
+type reportRequest struct {
+	ctx   context.Context
+	reply chan *Report
+}
+
+type runRequest struct {
+	complete chan struct{}
+}
+
+type stopRequest struct {
+	complete chan struct{}
+}
+
+type fsListRes struct {
+	filesystems []*zfs.DatasetPath
+	err         error
+}
+
+type Side interface {
+	// may return both nil, indicating there is no replication position
+	GetReplicationPosition(ctx context.Context, fs string) (*zfs.FilesystemVersion, error)
+	isSide() Side
+}
+
+func NewPruner(fsfilter endpoint.FSFilter, jid endpoint.JobID, side Side, keepRules []pruning.KeepRule) *Pruner {
+	return &Pruner{
+		fsfilter,
+		jid,
+		side,
+		keepRules,
+		make(chan reportRequest),
+		make(chan stopRequest),
+		make(chan struct{}),
+		make(chan fsListRes),
+		StateInitialized,
+		nil,
+		nil,
+	}
+}
+
+func (p *Pruner) Run(ctx context.Context) *Report {
+
+	ctx, cancel := context.WithCancel(ctx)
+	defer cancel()
+
+	if p.state != StateInitialized {
+		panic("Run can onl[y be called once")
+	}
+
+	go func() {
+		fss, err := zfs.ZFSListMapping(ctx, p.fsfilter)
+		p.fsListRes <- fsListRes{fss, err}
+	}()
+
+	for {
+		select {
+		case res := <-p.fsListRes:
+			if res.err != nil {
+				p.state = StateListFilesystemsError
+				p.listFilesystemsError = res.err
+				close(p.done)
+				continue
+			}
+
+			p.state = StateFanOutFilesystems
+
+			p.fsPruners = make([]*FSPruner, len(res.filesystems))
+			_, add, end := trace.WithTaskGroup(ctx, "pruner-fan-out-fs")
+			for i, fs := range res.filesystems {
+				p.fsPruners[i] = NewFSPruner(p.jid, p.side, p.keepRules, fs)
+				add(func(ctx context.Context) {
+					p.fsPruners[i].Run(ctx)
+				})
+			}
+			go func() {
+				end()
+				close(p.done)
+			}()
+
+		case req := <-p.stopReqs:
+			cancel()
+			go func() {
+				<-p.done
+				close(req.complete)
+			}()
+		case req := <-p.reportReqs:
+			req.reply <- p.report(req.ctx)
+		case <-p.done:
+			p.state = StateDone
+			return p.report(ctx)
+		}
+	}
+}
+
+func (p *Pruner) Report(ctx context.Context) *Report {
+	req := reportRequest{
+		ctx:   ctx,
+		reply: make(chan *Report, 1),
+	}
+	select {
+	case p.reportReqs <- req:
+		return <-req.reply
+	case <-ctx.Done():
+		return nil
+	case <-p.done:
+		return nil
+	}
+}
+
+func (p *Pruner) report(ctx context.Context) *Report {
+	fsreports := make([]*FSReport, len(p.fsPruners))
+	for i := range fsreports {
+		fsreports[i] = p.fsPruners[i].report()
+	}
+	return &Report{
+		State:                p.state,
+		ListFilesystemsError: p.listFilesystemsError,
+		Filesystems:          fsreports,
+	}
+}
+
+// implements pruning.Snapshot
+type snapshot struct {
+	replicated bool
+	stepHolds  []pruning.StepHold
+	zfs.FilesystemVersion
+
+	state     SnapState
+	destroyOp *zfs.DestroySnapOp
+}
+
+//go:generate enumer -type=SnapState -json
+type SnapState int
+
+const (
+	SnapStateInitialized SnapState = iota
+	SnapStateKeeping
+	SnapStateDeletePending
+	SnapStateDeleteAttempted
+)
+
+// implements pruning.StepHold
+type stepHold struct {
+	endpoint.Abstraction
+}
+
+func (s snapshot) Replicated() bool              { return s.replicated }
+func (s snapshot) StepHolds() []pruning.StepHold { return s.stepHolds }
+
+func (s stepHold) GetJobID() endpoint.JobID { return *s.Abstraction.GetJobID() }
+
+type FSPruner struct {
+	jid       endpoint.JobID
+	side      Side
+	keepRules []pruning.KeepRule
+	fsp       *zfs.DatasetPath
+
+	state FSState
+
+	// all channels consumed by the run loop
+	planned    chan fsPlanRes
+	executed   chan fsExecuteRes
+	done       chan struct{}
+	reportReqs chan fsReportReq
+
+	keepList    []*snapshot // valid in FSStateExecuting and forward
+	destroyList []*snapshot // valid in FSStateExecuting and forward, field .destroyOp is invalid until FSStateExecuting is left
+
+}
+
+type fsPlanRes struct {
+	keepList    []*snapshot
+	destroyList []*snapshot
+	err         error
+}
+
+type fsExecuteRes struct {
+	completedDestroyOps []*zfs.DestroySnapOp // same len() as FSPruner.destroyList
+}
+
+type fsReportReq struct {
+	res chan *FSReport
+}
+
+type FSReport struct {
+	State    FSState
+	KeepList []*SnapReport
+	Destroy  []*SnapReport
+}
+
+type SnapReport struct {
+	State         SnapState
+	Name          string
+	Replicated    bool
+	StepHoldCount int
+	DestroyError  error
+}
+
+//go:generate enumer -type=FSState -json
+type FSState int
+
+const (
+	FSStateInitialized FSState = iota
+	FSStatePlanning
+	FSStatePlanErr
+	FSStateExecuting
+	FSStateExecuteErr
+	FSStateExecuteSuccess
+)
+
+func (s FSState) IsTerminal() bool {
+	return s == FSStatePlanErr || s == FSStateExecuteErr || s == FSStateExecuteSuccess
+}
+
+func NewFSPruner(jid endpoint.JobID, side Side, keepRules []pruning.KeepRule, fsp *zfs.DatasetPath) *FSPruner {
+	return &FSPruner{
+		jid, side, keepRules, fsp,
+		FSStateInitialized,
+		make(chan fsPlanRes),
+		make(chan fsExecuteRes),
+		make(chan struct{}),
+		make(chan fsReportReq),
+		nil, nil,
+	}
+}
+
+func (p *FSPruner) Run(ctx context.Context) *FSReport {
+
+	defer func() {
+	}()
+
+	p.state = FSStatePlanning
+
+	go func() { p.planned <- p.plan(ctx) }()
+
+out:
+	for !p.state.IsTerminal() {
+		select {
+		case res := <-p.planned:
+
+			if res.err != nil {
+				p.state = FSStatePlanErr
+				continue
+			}
+			p.state = FSStateExecuting
+			p.keepList = res.keepList
+			p.destroyList = res.destroyList
+
+			go func() { p.executed <- p.execute(ctx, p.destroyList) }()
+
+		case res := <-p.executed:
+
+			if len(res.completedDestroyOps) != len(p.destroyList) {
+				panic("impl error: completedDestroyOps is a vector corresponding to entries in p.destroyList")
+			}
+
+			var erronous []*zfs.DestroySnapOp
+			for i, op := range res.completedDestroyOps {
+				if *op.ErrOut != nil {
+					erronous = append(erronous, op)
+				}
+				p.destroyList[i].destroyOp = op
+				p.destroyList[i].state = SnapStateDeleteAttempted
+			}
+			if len(erronous) > 0 {
+				p.state = FSStateExecuteErr
+			} else {
+				p.state = FSStateExecuteSuccess
+			}
+
+			close(p.done)
+
+		case <-p.reportReqs:
+			panic("unimp")
+		case <-p.done:
+			break out
+		}
+	}
+
+	// TODO render last FS report
+	return nil
+}
+
+func (p *FSPruner) plan(ctx context.Context) fsPlanRes {
+	fs := p.fsp.ToString()
+	vs, err := zfs.ZFSListFilesystemVersions(ctx, p.fsp, zfs.ListFilesystemVersionsOptions{})
+	if err != nil {
+		return fsPlanRes{err: errors.Wrap(err, "list filesystem versions")}
+	}
+
+	allJobsStepHolds, absErrs, err := endpoint.ListAbstractions(ctx, endpoint.ListZFSHoldsAndBookmarksQuery{
+		FS: endpoint.ListZFSHoldsAndBookmarksQueryFilesystemFilter{
+			FS: &fs,
+		},
+		What: endpoint.AbstractionTypeSet{
+			endpoint.AbstractionStepHold: true,
+		},
+		Concurrency: 1,
+	})
+	if err != nil {
+		return fsPlanRes{err: errors.Wrap(err, "list abstractions")}
+	}
+	if len(absErrs) > 0 {
+		logging.GetLogger(ctx, logging.SubsysPruning).WithError(endpoint.ListAbstractionsErrors(absErrs)).
+			Error("error listing some step holds, prune attempt might fail with 'dataset is busy' errors")
+	}
+
+	repPos, err := p.side.GetReplicationPosition(ctx, p.fsp.ToString())
+	if err != nil {
+		return fsPlanRes{err: errors.Wrap(err, "get replication position")}
+	}
+
+	vsAsSnaps := make([]pruning.Snapshot, len(vs))
+	for i := range vs {
+		var repPosCreateTxgOrZero uint64
+		if repPos != nil {
+			repPosCreateTxgOrZero = repPos.GetCreateTXG()
+		}
+		s := &snapshot{
+			state:             SnapStateInitialized,
+			FilesystemVersion: vs[i],
+			replicated:        vs[i].GetCreateTXG() <= repPosCreateTxgOrZero,
+		}
+		for _, h := range allJobsStepHolds {
+			if zfs.FilesystemVersionEqualIdentity(vs[i], h.GetFilesystemVersion()) {
+				s.stepHolds = append(s.stepHolds, stepHold{h})
+			}
+		}
+		vsAsSnaps[i] = s
+	}
+
+	downcastToSnapshots := func(l []pruning.Snapshot) (r []*snapshot) {
+		r = make([]*snapshot, len(l))
+		for i, e := range l {
+			r[i] = e.(*snapshot)
+		}
+		return r
+	}
+	pruningResult := pruning.PruneSnapshots(vsAsSnaps, p.keepRules)
+	remove, keep := downcastToSnapshots(pruningResult.Remove), downcastToSnapshots(pruningResult.Keep)
+	if len(remove)+len(keep) != len(vsAsSnaps) {
+		for _, s := range vsAsSnaps {
+			r, _ := json.MarshalIndent(s.(*snapshot).report(), "", "  ")
+			fmt.Fprintf(os.Stderr, "%s\n", string(r))
+		}
+		panic("indecisive")
+	}
+
+	for _, s := range remove {
+		s.state = SnapStateDeletePending
+	}
+	for _, s := range keep {
+		s.state = SnapStateKeeping
+	}
+
+	return fsPlanRes{keepList: keep, destroyList: remove, err: nil}
+}
+
+func (p *FSPruner) execute(ctx context.Context, destroyList []*snapshot) fsExecuteRes {
+	ops := make([]*zfs.DestroySnapOp, len(destroyList))
+	for i, fsv := range p.destroyList {
+		ops[i] = &zfs.DestroySnapOp{
+			Filesystem: p.fsp.ToString(),
+			Name:       fsv.GetName(),
+			ErrOut:     new(error),
+		}
+	}
+	zfs.ZFSDestroyFilesystemVersions(ctx, ops)
+
+	return fsExecuteRes{completedDestroyOps: ops}
+}
+
+func (p *FSPruner) report() *FSReport {
+	return &FSReport{
+		State:    p.state,
+		KeepList: p.reportRenderSnapReports(p.keepList),
+		Destroy:  p.reportRenderSnapReports(p.destroyList),
+	}
+}
+
+func (p *FSPruner) reportRenderSnapReports(l []*snapshot) (r []*SnapReport) {
+	r = make([]*SnapReport, len(l))
+	for i := range l {
+		r[i] = l[i].report()
+	}
+	return r
+}
+
+func (s *snapshot) report() *SnapReport {
+	var snapErr error
+	if s.state == SnapStateDeleteAttempted {
+		if *s.destroyOp.ErrOut != nil {
+			snapErr = (*s.destroyOp.ErrOut)
+		}
+	}
+	return &SnapReport{
+		State:         s.state,
+		Name:          s.Name,
+		Replicated:    s.Replicated(),
+		StepHoldCount: len(s.stepHolds),
+		DestroyError:  snapErr,
+	}
+}
@@ -0,0 +1,27 @@
+package pruner
+
+import (
+	"context"
+
+	"github.com/zrepl/zrepl/endpoint"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type SideSender struct {
+	jobID endpoint.JobID
+}
+
+func NewSideSender(jid endpoint.JobID) *SideSender {
+	return &SideSender{jid}
+}
+
+func (s *SideSender) isSide() Side { return nil }
+
+var _ Side = (*SideSender)(nil)
+
+func (s *SideSender) GetReplicationPosition(ctx context.Context, fs string) (*zfs.FilesystemVersion, error) {
+	if fs == "" {
+		panic("must not pass zero value for fs")
+	}
+	return endpoint.GetMostRecentReplicationCursorOfJob(ctx, fs, s.jobID)
+}
@@ -0,0 +1,70 @@
+// Code generated by "enumer -type=SnapState -json"; DO NOT EDIT.
+
+//
+package pruner
+
+import (
+	"encoding/json"
+	"fmt"
+)
+
+const _SnapStateName = "SnapStateInitializedSnapStateKeepingSnapStateDeletePendingSnapStateDeleteAttempted"
+
+var _SnapStateIndex = [...]uint8{0, 20, 36, 58, 82}
+
+func (i SnapState) String() string {
+	if i < 0 || i >= SnapState(len(_SnapStateIndex)-1) {
+		return fmt.Sprintf("SnapState(%d)", i)
+	}
+	return _SnapStateName[_SnapStateIndex[i]:_SnapStateIndex[i+1]]
+}
+
+var _SnapStateValues = []SnapState{0, 1, 2, 3}
+
+var _SnapStateNameToValueMap = map[string]SnapState{
+	_SnapStateName[0:20]:  0,
+	_SnapStateName[20:36]: 1,
+	_SnapStateName[36:58]: 2,
+	_SnapStateName[58:82]: 3,
+}
+
+// SnapStateString retrieves an enum value from the enum constants string name.
+// Throws an error if the param is not part of the enum.
+func SnapStateString(s string) (SnapState, error) {
+	if val, ok := _SnapStateNameToValueMap[s]; ok {
+		return val, nil
+	}
+	return 0, fmt.Errorf("%s does not belong to SnapState values", s)
+}
+
+// SnapStateValues returns all values of the enum
+func SnapStateValues() []SnapState {
+	return _SnapStateValues
+}
+
+// IsASnapState returns "true" if the value is listed in the enum definition. "false" otherwise
+func (i SnapState) IsASnapState() bool {
+	for _, v := range _SnapStateValues {
+		if i == v {
+			return true
+		}
+	}
+	return false
+}
+
+// MarshalJSON implements the json.Marshaler interface for SnapState
+func (i SnapState) MarshalJSON() ([]byte, error) {
+	return json.Marshal(i.String())
+}
+
+// UnmarshalJSON implements the json.Unmarshaler interface for SnapState
+func (i *SnapState) UnmarshalJSON(data []byte) error {
+	var s string
+	if err := json.Unmarshal(data, &s); err != nil {
+		return fmt.Errorf("SnapState should be a string, got %s", data)
+	}
+
+	var err error
+	*i, err = SnapStateString(s)
+	return err
+}
@@ -0,0 +1,71 @@
+// Code generated by "enumer -type=State -json"; DO NOT EDIT.
+
+//
+package pruner
+
+import (
+	"encoding/json"
+	"fmt"
+)
+
+const _StateName = "StateInitializedStateListFilesystemsStateListFilesystemsErrorStateFanOutFilesystemsStateDone"
+
+var _StateIndex = [...]uint8{0, 16, 36, 61, 83, 92}
+
+func (i State) String() string {
+	if i < 0 || i >= State(len(_StateIndex)-1) {
+		return fmt.Sprintf("State(%d)", i)
+	}
+	return _StateName[_StateIndex[i]:_StateIndex[i+1]]
+}
+
+var _StateValues = []State{0, 1, 2, 3, 4}
+
+var _StateNameToValueMap = map[string]State{
+	_StateName[0:16]:  0,
+	_StateName[16:36]: 1,
+	_StateName[36:61]: 2,
+	_StateName[61:83]: 3,
+	_StateName[83:92]: 4,
+}
+
+// StateString retrieves an enum value from the enum constants string name.
+// Throws an error if the param is not part of the enum.
+func StateString(s string) (State, error) {
+	if val, ok := _StateNameToValueMap[s]; ok {
+		return val, nil
+	}
+	return 0, fmt.Errorf("%s does not belong to State values", s)
+}
+
+// StateValues returns all values of the enum
+func StateValues() []State {
+	return _StateValues
+}
+
+// IsAState returns "true" if the value is listed in the enum definition. "false" otherwise
+func (i State) IsAState() bool {
+	for _, v := range _StateValues {
+		if i == v {
+			return true
+		}
+	}
+	return false
+}
+
+// MarshalJSON implements the json.Marshaler interface for State
+func (i State) MarshalJSON() ([]byte, error) {
+	return json.Marshal(i.String())
+}
+
+// UnmarshalJSON implements the json.Unmarshaler interface for State
+func (i *State) UnmarshalJSON(data []byte) error {
+	var s string
+	if err := json.Unmarshal(data, &s); err != nil {
+		return fmt.Errorf("State should be a string, got %s", data)
+	}
+
+	var err error
+	*i, err = StateString(s)
+	return err
+}
@@ -20,15 +20,14 @@ import (
 )

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

 // Try to keep it compatible with github.com/zrepl/zrepl/endpoint.Endpoint
 type Target interface {
-	ListFilesystems(ctx context.Context, req *pdu.ListFilesystemReq) (*pdu.ListFilesystemRes, error)
-	ListFilesystemVersions(ctx context.Context, req *pdu.ListFilesystemVersionsReq) (*pdu.ListFilesystemVersionsRes, error)
+	Endpoint
 	DestroySnapshots(ctx context.Context, req *pdu.DestroySnapshotsReq) (*pdu.DestroySnapshotsRes, error)
 }

@@ -46,13 +45,14 @@ func GetLogger(ctx context.Context) Logger {
 }

 type args struct {
-	ctx                            context.Context
-	target                         Target
-	receiver                       History
-	rules                          []pruning.KeepRule
-	retryWait                      time.Duration
-	considerSnapAtCursorReplicated bool
-	promPruneSecs                  prometheus.Observer
+	ctx                              context.Context
+	target                           Target
+	sender, receiver                 Endpoint
+	rules                            []pruning.KeepRule
+	retryWait                        time.Duration
+	considerSnapAtCursorReplicated   bool
+	convertAnyStepHoldToStepBookmark bool
+	promPruneSecs                    prometheus.Observer
 }

 type Pruner struct {
@@ -70,11 +70,12 @@ type Pruner struct {
 }

 type PrunerFactory struct {
-	senderRules                    []pruning.KeepRule
-	receiverRules                  []pruning.KeepRule
-	retryWait                      time.Duration
-	considerSnapAtCursorReplicated bool
-	promPruneSecs                  *prometheus.HistogramVec
+	senderRules                      []pruning.KeepRule
+	receiverRules                    []pruning.KeepRule
+	retryWait                        time.Duration
+	considerSnapAtCursorReplicated   bool
+	convertAnyStepHoldToStepBookmark bool
+	promPruneSecs                    *prometheus.HistogramVec
 }

 type LocalPrunerFactory struct {
@@ -122,25 +123,35 @@ func NewPrunerFactory(in config.PruningSenderReceiver, promPruneSecs *prometheus
 		}
 		considerSnapAtCursorReplicated = considerSnapAtCursorReplicated || !knr.KeepSnapshotAtCursor
 	}
+
+	convertAnyStepHoldToStepBookmark := false
+	for _, r := range in.KeepSender {
+		_, ok := r.Ret.(*config.PruneKeepStepHolds)
+		convertAnyStepHoldToStepBookmark = convertAnyStepHoldToStepBookmark || ok
+	}
+
 	f := &PrunerFactory{
-		senderRules:                    keepRulesSender,
-		receiverRules:                  keepRulesReceiver,
-		retryWait:                      envconst.Duration("ZREPL_PRUNER_RETRY_INTERVAL", 10*time.Second),
-		considerSnapAtCursorReplicated: considerSnapAtCursorReplicated,
-		promPruneSecs:                  promPruneSecs,
+		senderRules:                      keepRulesSender,
+		receiverRules:                    keepRulesReceiver,
+		retryWait:                        envconst.Duration("ZREPL_PRUNER_RETRY_INTERVAL", 10*time.Second),
+		considerSnapAtCursorReplicated:   considerSnapAtCursorReplicated,
+		convertAnyStepHoldToStepBookmark: convertAnyStepHoldToStepBookmark,
+		promPruneSecs:                    promPruneSecs,
 	}
 	return f, nil
 }

-func (f *PrunerFactory) BuildSenderPruner(ctx context.Context, target Target, receiver History) *Pruner {
+func (f *PrunerFactory) BuildSenderPruner(ctx context.Context, sender Target, receiver Endpoint) *Pruner {
 	p := &Pruner{
 		args: args{
 			context.WithValue(ctx, contextKeyPruneSide, "sender"),
-			target,
+			sender,
+			sender,
 			receiver,
 			f.senderRules,
 			f.retryWait,
 			f.considerSnapAtCursorReplicated,
+			f.convertAnyStepHoldToStepBookmark,
 			f.promPruneSecs.WithLabelValues("sender"),
 		},
 		state: Plan,
@@ -148,15 +159,17 @@ func (f *PrunerFactory) BuildSenderPruner(ctx context.Context, target Target, re
 	return p
 }

-func (f *PrunerFactory) BuildReceiverPruner(ctx context.Context, target Target, receiver History) *Pruner {
+func (f *PrunerFactory) BuildReceiverPruner(ctx context.Context, receiver Target, sender Endpoint) *Pruner {
 	p := &Pruner{
 		args: args{
 			context.WithValue(ctx, contextKeyPruneSide, "receiver"),
-			target,
+			receiver,
+			sender,
 			receiver,
 			f.receiverRules,
 			f.retryWait,
 			false, // senseless here anyways
+			false, // senseless here anyways
 			f.promPruneSecs.WithLabelValues("receiver"),
 		},
 		state: Plan,
@@ -164,15 +177,17 @@ func (f *PrunerFactory) BuildReceiverPruner(ctx context.Context, target Target,
 	return p
 }

-func (f *LocalPrunerFactory) BuildLocalPruner(ctx context.Context, target Target, receiver History) *Pruner {
+func (f *LocalPrunerFactory) BuildLocalPruner(ctx context.Context, target Target) *Pruner {
 	p := &Pruner{
 		args: args{
 			context.WithValue(ctx, contextKeyPruneSide, "local"),
 			target,
-			receiver,
+			target,
+			target,
 			f.keepRules,
 			f.retryWait,
 			false, // considerSnapAtCursorReplicated is not relevant for local pruning
+			false, // convertAnyStepHoldToStepBookmark is not relevant for local pruning
 			f.promPruneSecs.WithLabelValues("local"),
 		},
 		state: Plan,
@@ -341,11 +356,13 @@ func (s snapshot) Replicated() bool { return s.replicated }

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

+func (s snapshot) CreateTXG() uint64 { return s.fsv.GetCreateTXG() }
+
 func doOneAttempt(a *args, u updater) {

-	ctx, target, receiver := a.ctx, a.target, a.receiver
+	ctx, sender, receiver, target := a.ctx, a.sender, a.receiver, a.target

-	sfssres, err := receiver.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
+	sfssres, err := sender.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
 	if err != nil {
 		u(func(p *Pruner) {
 			p.state = PlanErr
@@ -407,6 +424,10 @@ tfss_loop:

 		pfs.snaps = make([]pruning.Snapshot, 0, len(tfsvs))

+		receiver.ListFilesystemVersions(ctx, &pdu.ListFilesystemVersionsReq{
+			Filesystem: tfs.Path,
+		})
+
 		rcReq := &pdu.ReplicationCursorReq{
 			Filesystem: tfs.Path,
 		}
@@ -415,6 +436,7 @@ tfss_loop:
 			pfsPlanErrAndLog(err, "cannot get replication cursor bookmark")
 			continue tfss_loop
 		}
+
 		if rc.GetNotexist() {
 			err := errors.New("replication cursor bookmark does not exist (one successful replication is required before pruning works)")
 			pfsPlanErrAndLog(err, "")
@@ -8,10 +8,10 @@ import (
 	"time"

 	"github.com/pkg/errors"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"

 	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
 	"github.com/zrepl/zrepl/daemon/hooks"
 	"github.com/zrepl/zrepl/daemon/logging"
 	"github.com/zrepl/zrepl/logger"
@@ -49,7 +49,7 @@ type args struct {
 	ctx            context.Context
 	prefix         string
 	interval       time.Duration
-	fsf            zfs.DatasetFilter
+	fsf            *filters.DatasetMapFilter
 	snapshotsTaken chan<- struct{}
 	hooks          *hooks.List
 	dryRun         bool
@@ -109,7 +109,7 @@ func getLogger(ctx context.Context) Logger {
 	return logging.GetLogger(ctx, logging.SubsysSnapshot)
 }

-func PeriodicFromConfig(g *config.Global, fsf zfs.DatasetFilter, in *config.SnapshottingPeriodic) (*Snapper, error) {
+func PeriodicFromConfig(g *config.Global, fsf *filters.DatasetMapFilter, in *config.SnapshottingPeriodic) (*Snapper, error) {
 	if in.Prefix == "" {
 		return nil, errors.New("prefix must not be empty")
 	}
@@ -383,7 +383,7 @@ func wait(a args, u updater) state {
 	}
 }

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

@@ -5,7 +5,7 @@ import (
 	"fmt"

 	"github.com/zrepl/zrepl/config"
-	"github.com/zrepl/zrepl/zfs"
+	"github.com/zrepl/zrepl/daemon/filters"
 )

 // FIXME: properly abstract snapshotting:
@@ -32,7 +32,7 @@ func (s *PeriodicOrManual) Report() *Report {
 	return nil
 }

-func FromConfig(g *config.Global, fsf zfs.DatasetFilter, in config.SnapshottingEnum) (*PeriodicOrManual, error) {
+func FromConfig(g *config.Global, fsf *filters.DatasetMapFilter, in config.SnapshottingEnum) (*PeriodicOrManual, error) {
 	switch v := in.Ret.(type) {
 	case *config.SnapshottingPeriodic:
 		snapper, err := PeriodicFromConfig(g, fsf, v)
@@ -1 +0,0 @@
-packaging/deb/debian
@@ -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
+}
@@ -27,33 +27,6 @@ We use the following annotations for classifying changes:
 * |bugfix| Change that fixes a bug, no regressions or incompatibilities expected.
 * |docs| Change to the documentation.

-0.3.1
-----
-
-Mostly a bugfix release for :ref:`zrepl 0.3 <release-0.3>`.
-
-* |feature| pruning: add optional ``regex`` field to ``last_n`` rule
-* |docs| pruning: ``grid`` : improve documentation and add an example
-* |bugfix| pruning: ``grid``:  add all snapshots that do not match the regex to the rule's destroy list.
-  This brings the implementation in line with the docs.
-* |bugfix| ``easyrsa`` script in docs
-* |bugfix| platformtest: fix skipping encryption-only tests on systems that don't support encryption
-* |bugfix| replication: report AttemptDone if no filesystems are replicated
-* |feature| status + replication: warning if replication succeeeded without any filesystem being replicated
-* |docs| update multi-job & multi-host setup section
-* RPM Packaging
-* CI infrastructure rework
-* Continuous deployment of that new `stable` branch to zrepl.github.io.
-
-.. 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 GitHub Sponsors| |Donate via Liberapay| |Donate via PayPal|
-   |  Note that PayPal processing fees are relatively high for small donations.
-   |  For SEPA wire transfer and **commercial support**, please `contact Christian directly <https://cschwarz.com>`_.
-
-.. _release-0.3:
-
 0.3
 ---

@@ -61,58 +34,35 @@ This is a big one! Headlining features:

 * **Resumable Send & Recv Support**
  No knobs required, automatically used where supported.
-* **Encrypted Send & Recv Support** for OpenZFS native encryption,
-  :ref:`configurable <job-send-options>` at the job level, i.e., for all filesystems a job is responsible for.
-* **Replication Guarantees**
-  Automatic use of ZFS holds and bookmarks to protect a replicated filesystem from losing synchronization between sender and receiver.
-  By default, zrepl guarantees that incremental replication will always be possible and interrupted steps will always be resumable.
+* **Hold-Protected Send & Recv**
+  Automatic ZFS holds to ensure that we can always resume a replication step.
+* **Encrypted Send & Recv Support** for OpenZFS native encryption.
+  :ref:`Configurable <job-send-options>` at the job level, i.e., for all filesystems a job is responsible for.
+* **Receive-side hold on last received dataset**
+  The counterpart to the replication cursor bookmark on the send-side.
+  Ensures that incremental replication will always be possible between a sender and receiver.

-.. TIP::
+Actual changelog:

-   We highly recommend studying the updated :ref:`overview section of the configuration chapter <overview-how-replication-works>` to understand how replication works.
+* |break_config| **more restrictive job names than in prior zrepl versions**
+  Starting with this version, job names are going to be embedded into ZFS holds and bookmark names.

-.. TIP::
-
-   Go 1.15 changed the default TLS validation policy to **require Subject Alternative Names (SAN) in certificates**.
-   The openssl commands we provided in the quick-start guides up to and including the zrepl 0.3 docs seem not to work properly.
-   If you encounter certificate validation errors regarding SAN and wish to continue to use your old certificates, start the zrepl daemon with env var ``GODEBUG=x509ignoreCN=0``.
-   Alternatively, generate new certificates with SANs (see :ref:`both options int the TLS transport docs <transport-tcp+tlsclientauth-certgen>` ).
-
-Quick-start guides:
-
-* We have added :ref:`another quick-start guide for a typical workstation use case for zrepl <quickstart-backup-to-external-disk>`.
-  Check it out to learn how you can use zrepl to back up your workstation's OpenZFS natively-encrypted root filesystem to an external disk.
-
-Additional changelog:
-
-* |break| Go 1.15 TLS changes mentioned above.
-* |break| |break_config| **more restrictive job names than in prior zrepl versions**
-  Starting with this version, job names are going to be embedded into ZFS holds and bookmark names (see :ref:`this section for details <zrepl-zfs-abstractions>`).
-  Therefore you might need to adjust your job names.
-  **Note that jobs** cannot be renamed easily **once you start using zrepl 0.3.**
 * |break| |mig| replication cursor representation changed

-  * zrepl now manages the :ref:`replication cursor bookmark <zrepl-zfs-abstractions>` per job-filesystem tuple instead of a single replication cursor per filesystem.
+  * zrepl now manages the :ref:`replication cursor bookmark <replication-cursor-and-last-received-hold>` per job-filesystem tuple instead of a single replication cursor per filesystem.
    In the future, this will permit multiple sending jobs to send from the same filesystems.
  * ZFS does not allow bookmark renaming, thus we cannot migrate the old replication cursors.
  * zrepl 0.3 will automatically create cursors in the new format for new replications, and warn if it still finds ones in the old format.
  * Run ``zrepl migrate replication-cursor:v1-v2`` to safely destroy old-format cursors.
    The migration will ensure that only those old-format cursors are destroyed that have been superseeded by new-format cursors.

-* |feature| New option ``listen_freebind`` (tcp, tls, prometheus listener)
-* |feature| :issue:`341` Prometheus metric for failing replications + corresponding Grafana panel
-* |feature| :issue:`265` transport/tcp: support for CIDR masks in client IP whitelist
-* |feature| documented subcommand to generate ``bash`` and ``zsh`` completions
-* |feature| :issue:`307` ``chrome://trace`` -compatible activity tracing of zrepl daemon activity
-* |feature| logging: trace IDs for better log entry correlation with concurrent replication jobs
-* |feature| experimental environment variable for parallel replication (see :issue:`306` )
 * |bugfix| missing logger context vars in control connection handlers
 * |bugfix| improved error messages on ``zfs send`` errors
+* |feature| New option ``listen_freebind`` (tcp, tls, prometheus listener)
 * |bugfix| |docs| snapshotting: clarify sync-up behavior and warn about filesystems
-* |bugfix| transport/ssh: do not leak zombie ssh process on connection failures
  that will not be snapshotted until the sync-up phase is over
-* |docs| Installation: :ref:`FreeBSD jail with iocage <installation-freebsd-jail-with-iocage>`
 * |docs| Document new replication features in the :ref:`config overview <overview-how-replication-works>` and :repomasterlink:`replication/design.md`.
+* |feature| documented subcommand to generate ``bash`` and ``zsh`` completions
 * **[MAINTAINER NOTICE]** New platform tests in this version, please make sure you run them for your distro!
 * **[MAINTAINER NOTICE]** Please add the shell completions to the zrepl packages.

@@ -152,6 +102,14 @@ Additional changelog:
  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 **commercial support**, please `contact Christian directly <https://cschwarz.com>`_.
+
+
 0.1.1
 -----

@@ -12,7 +12,6 @@ Configuration
    configuration/transports
    configuration/filter_syntax
    configuration/sendrecvoptions
-    configuration/replication
    configuration/snapshotting
    configuration/prune
    configuration/logging
@@ -19,7 +19,7 @@ Job Type ``push``
    * - ``type``
      - = ``push``
    * - ``name``
-      - unique name of the job :issue:`(must not change)<327>`
+      - unique name of the job
    * - ``connect``
      - |connect-transport|
    * - ``filesystems``
@@ -47,7 +47,7 @@ Job Type ``sink``
    * - ``type``
      - = ``sink``
    * - ``name``
-      - unique name of the job :issue:`(must not change)<327>`
+      - unique name of the job
    * - ``serve``
      - |serve-transport|
    * - ``root_fs``
@@ -70,7 +70,7 @@ Job Type ``pull``
    * - ``type``
      - = ``pull``
    * - ``name``
-      - unique name of the job :issue:`(must not change)<327>`
+      - unique name of the job
    * - ``connect``
      - |connect-transport|
    * - ``root_fs``
@@ -98,7 +98,7 @@ Job Type ``source``
    * - ``type``
      - = ``source``
    * - ``name``
-      - unique name of the job :issue:`(must not change)<327>`
+      - unique name of the job
    * - ``serve``
      - |serve-transport|
    * - ``filesystems``
@@ -137,7 +137,7 @@ Job type that only takes snapshots and performs pruning on the local machine.
    * - ``type``
      - = ``snap``
    * - ``name``
-      - unique name of the job :issue:`(must not change)<327>`
+      - unique name of the job
    * - ``filesystems``
      - |filter-spec| for filesystems to be snapshotted
    * - ``snapshotting``
@@ -12,7 +12,7 @@ The following paths are considered:
 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:`quick-start guides <quickstart-toc>` or the :sampleconf:`/` directory might also be helpful.
+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
@@ -39,10 +39,6 @@ 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.

-.. NOTE::
-   The job name is persisted in several places on disk and thus :issue:`cannot be changed easily<327>`.
-
-
 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 permissions.
@@ -59,8 +55,8 @@ Note that snapshot-creation denoted by "(snap)" is orthogonal to whether a job i
 | 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 to :ref:`locally attached disk <quickstart-backup-to-external-disk>`      |
-|                       | | with :ref:`local transport <transport-local>` | * Backup FreeBSD boot pool                                                         |
+| 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                                                                       |
@@ -80,19 +76,16 @@ The active side (:ref:`push <job-push>` and :ref:`pull <job-pull>` job) executes
 .. TIP::
  The progress of the active side can be watched live using the ``zrepl status`` subcommand.

-.. _overview-passive-side--client-identity:
-
 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.
-When a client connects, the transport listener performS listener-specific access control (cert validation, IP ACLs, etc)
-and determines the *client identity*.
-The passive side job then uses this client identity as follows:
+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`` might, in the future, embed the client identity in :ref:`zrepl's ZFS abstraction names <zrepl-zfs-abstractions>` in order to support multi-host replication.
+* 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.
@@ -106,7 +99,7 @@ One of the major design goals of the replication module is to avoid any duplicat
 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 high-level steps take place during replication and can be monitored using the ``zrepl status`` subcommand:
+The following steps take place during replication and can be monitored using the ``zrepl status`` subcommand:

 * Plan the replication:

@@ -114,9 +107,9 @@ The following high-level steps take place during replication and can be monitore
  * Build the **replication plan**

    * Per filesystem, compute a diff between sender and receiver snapshots
-    * Build a list of **replication steps**
+    * Build a list of replication steps

-      * If possible, use incremental and resumable sends
+      * If possible, use incremental and resumable sends (``zfs send -i``)
      * Otherwise, use full send of most recent snapshot on sender

  * Retry on errors that are likely temporary (i.e. network failures).
@@ -135,39 +128,19 @@ The following high-level steps take place during replication and can be monitore
   
 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.

-ZFS Background Knowledge
-^^^^^^^^^^^^^^^^^^^^^^^^
-This section gives some background knowledge about ZFS features that zrepl uses to provide guarantees for a replication filesystem.
-Specifically, zrepl guarantees by default that **incremental replication is always possible and that started replication steps can always be resumed if they are interrupted.**
+.. _replication-cursor-and-last-received-hold:

-**ZFS Send Modes & Bookmarks**
-ZFS supports full sends (``zfs send fs@to``) and incremental sends (``zfs send -i @from fs@to``).
-Full sends are used to create a new filesystem on the receiver with the send-side state of ``fs@to``.
-Incremental sends only transfer the delta between ``@from`` and ``@to``.
-Incremental sends require that ``@from`` be present on the receiving side when receiving the incremental stream.
-Incremental sends can also use a ZFS bookmark as *from* on the sending side (``zfs send -i #bm_from fs@to``), where ``#bm_from`` was created using ``zfs bookmark fs@from fs#bm_from``.
-The receiving side must always have the actual snapshot ``@from``, regardless of whether the sending side uses ``@from`` or a bookmark of it.
-
-**Resumable Send & Recv**
-The ``-s`` flag for ``zfs recv`` tells zfs to save the partially received send stream in case it is interrupted.
-To resume the replication, the receiving side filesystem's ``receive_resume_token`` must be passed to a new ``zfs send -t <value> | zfs recv`` command.
-A full send can only be resumed if ``@to`` still exists.
-An incremental send can only be resumed if ``@to`` still exists *and* either ``@from`` still exists *or* a bookmark ``#fbm`` of ``@from`` still exists.
-
-**ZFS Holds**
-ZFS holds prevent a snapshot from being deleted through ``zfs destroy``, letting the destroy fail with a ``datset is busy`` error.
-Holds are created and referred to by a *tag*. They can be thought of as a named, persistent lock on the snapshot.
-
-
-.. _zrepl-zfs-abstractions:
-
-ZFS Abstractions Managed By zrepl
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-With the background knowledge from the previous paragraph, we now summarize the different on-disk ZFS objects that zrepl manages to provide its functionality.
+**Replication cursor** bookmark and **last-received-hold** are managed by zrepl to ensure that future replications can always be done incrementally:
+the replication cursor is a send-side bookmark of the most recent successfully replicated snapshot,
+and the last-received-hold is a hold of that snapshot on the receiving side.
+The replication cursor has the format ``#zrepl_CUSOR_G_<GUID>_J_<JOBNAME>``.
+The last-received-hold tag has the format ``#zrepl_last_received_J_<JOBNAME>``.
+Encoding the job name in the names ensures that multiple sending jobs can replicate the same filesystem to different receivers without interference.
+The ``zrepl holds list`` provides a listing of all bookmarks and holds managed by zrepl.

 .. _replication-placeholder-property:

-**Placeholder filesystems** on the receiving side are regular ZFS filesystems with the ZFS property ``zrepl:placeholder=on``.
+**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 hierarchy 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``.
@@ -175,110 +148,55 @@ 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.

-.. _replication-cursor-and-last-received-hold:
-
-The **replication cursor** bookmark and **last-received-hold** are managed by zrepl to ensure that future replications can always be done incrementally.
-The replication cursor is a send-side bookmark of the most recent successfully replicated snapshot,
-and the last-received-hold is a hold of that snapshot on the receiving side.
-Both are moved atomically after the receiving side has confirmed that a replication step is complete.
-
-The replication cursor has the format ``#zrepl_CUSOR_G_<GUID>_J_<JOBNAME>``.
-The last-received-hold tag has the format ``zrepl_last_received_J_<JOBNAME>``.
-Encoding the job name in the names ensures that multiple sending jobs can replicate the same filesystem to different receivers without interference.
-
-.. _tentative-replication-cursor-bookmarks:
-
-**Tentative replication cursor bookmarks** are short-lived boomkarks that protect the atomic moving-forward of the replication cursor and last-received-hold (see :issue:`this issue <340>`).
-They are only necessary if step holds are not used as per the :ref:`replication.protection <replication-option-protection>` setting.
-The tentative replication cursor has the format ``#zrepl_CUSORTENTATIVE_G_<GUID>_J_<JOBNAME>``.
-The ``zrepl zfs-abstraction list`` command provides a listing of all bookmarks and holds managed by zrepl.
-
-.. _step-holds:
-
-**Step holds** are zfs holds managed by zrepl to ensure that a replication step can always be resumed if it is interrupted, e.g., due to network outage.
-zrepl creates step holds before it attempts a replication step and releases them after the receiver confirms that the replication step is complete.
-For an initial replication ``full @initial_snap``, zrepl puts a zfs hold on ``@initial_snap``.
-For an incremental send ``@from -> @to``, zrepl puts a zfs hold on both ``@from`` and ``@to``.
-Note that ``@from`` is not strictly necessary for resumability -- a bookmark on the sending side would be sufficient --, but size-estimation in currently used OpenZFS versions only works if ``@from`` is a snapshot.
-The hold tag has the format ``zrepl_STEP_J_<JOBNAME>``.
-A job only ever has one active send per filesystem.
-Thus, there are never more than two step holds for a given pair of ``(job,filesystem)``.
-
-**Step bookmarks** are zrepl's equivalent for holds on bookmarks (ZFS does not support putting holds on bookmarks).
-They are intended for a situation where a replication step uses a bookmark ``#bm`` as incremental ``from`` where ``#bm`` is not managed by zrepl.
-To ensure resumability, zrepl copies ``#bm`` to step bookmark ``#zrepl_STEP_G_<GUID>_J_<JOBNAME>``.
-If the replication is interrupted and ``#bm`` is deleted by the user, the step bookmark remains as an incremental source for the resumable send.
-Note that zrepl does not yet support creating step bookmarks because the `corresponding ZFS feature for copying bookmarks <https://github.com/openzfs/zfs/pull/9571>`_ is not yet widely available .
-Subscribe to zrepl :issue:`326` for details.
-
-The ``zrepl zfs-abstraction list`` command provides a listing of all bookmarks and holds managed by zrepl.
-
-.. NOTE::
-
-    More details can be found in the design document :repomasterlink:`replication/design.md`.
-
-Limitations
-^^^^^^^^^^^
-
 .. ATTENTION::

    Currently, zrepl does not replicate filesystem properties.
    When receiving a filesystem, it is never mounted (`-u` flag)  and `mountpoint=none` is set.
    This is temporary and being worked on :issue:`24`.

+.. NOTE::
+
+    More details can be found in the design document :repomasterlink:`replication/design.md`.
+

 .. _jobs-multiple-jobs:

 Multiple Jobs & More than 2 Machines
 ------------------------------------

-The quick-start guides focus on simple setups with a single sender and a single receiver.
-This section documents considerations for more complex setups.
-
 .. ATTENTION::

-   Before you continue, make sure you have a working understanding of :ref:`how zrepl works <overview-how-replication-works>`
-   and :ref:`what zrepl does to ensure <zrepl-zfs-abstractions>` that replication between sender and receiver is always
-   possible without conflicts.
-   This will help you understand why certain kinds of multi-machine setups do not (yet) work.
+  When using multiple jobs across single or multiple machines, the following rules are critical to avoid race conditions & data loss:

-.. NOTE::
+  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.

-   If you can't find your desired configuration, have questions or would like to see improvements to multi-job setups, please `open an issue on GitHub <https://github.com/zrepl/zrepl/issues/new>`_.
+  Explanations & exceptions to above rules are detailed below.

-Multiple Jobs on one Machine
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-As a general rule, multiple jobs configured on one machine **must operate on disjoint sets of filesystems**.
-Otherwise, concurrently running jobs might interfere when operating on the same filesystem.
+If you would like to see improvements to multi-job setups, please `open an issue on GitHub <https://github.com/zrepl/zrepl/issues/new>`_.

-On your setup, ensure that
+No Overlapping
+~~~~~~~~~~~~~~

-* all ``filesystems`` filter specifications are disjoint
-* no ``root_fs`` is a prefix or equal to another ``root_fs``
-* no ``filesystems`` filter maches any ``root_fs``
+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 assumed the snapshot to still be present.
+However, the next replication attempt will re-examine the situation from scratch and should work.

-**Exceptions to the rule**:
+N push jobs to 1 sink
+~~~~~~~~~~~~~~~~~~~~~

-* A ``snap`` and ``push`` job on the same machine can match the same ``filesystems``.
-  To avoid interference, only one of the jobs should be pruning snapshots on the sender, the other one should keep all snapshots.
-  Since the jobs won't coordinate, errors in the log are to be expected, but :ref:`zrepl's ZFS abstractions <zrepl-zfs-abstractions>` ensure that ``push`` and ``sink`` can always replicate incrementally.
-  This scenario is detailed in one of the :ref:`quick-start guides <quickstart-backup-to-external-disk>`.
+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
+~~~~~~~~~~~~~~~~~~~~~~~~~

-More Than 2 Machines
-^^^^^^^^^^^^^^^^^^^^
+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.

-This section might be relevant to users who wish to *fan-in* (N machines replicate to 1) or *fan-out* (replicate 1 machine to N machines).
-
-**Working setups**:
-
-* N ``push`` identities, 1 ``sink`` (as long as the different push jobs have a different :ref:`client identity <overview-passive-side--client-identity>`)
-
-  * ``sink`` constrains each client to a disjoint sub-tree of the sink-side dataset hierarchy ``${root_fs}/${client_identity}``.
-    Therefore, the different clients cannot interfere.
-
-
-**Setups that do not work**:
-
-* N ``pull`` identities, 1 ``source`` job. Tracking :issue:`380`.
+There is currently no way for a pull job to filter which snapshots it should attempt to replicate.
+Thus, it is not possible to just manually assert that the prune rules of all pull jobs are disjoint to avoid replication-prune and prune-prune races.

@@ -46,8 +46,6 @@ Example Configuration:
             grid: 1x1h(keep=all) | 24x1h | 35x1d | 6x30d
             regex: "^zrepl_.*"
           # manually created snapshots will be kept forever on receiver
-           - type: regex
-             regex: "^manual_.*"

 .. DANGER::
    You might have **existing snapshots** of filesystems affected by pruning which you want to keep, i.e. not be destroyed by zrepl.
@@ -84,82 +82,34 @@ Policy ``grid``
        - type: grid
          regex: "^zrepl_.*"
          grid: 1x1h(keep=all) | 24x1h | 35x1d | 6x30d
-                │                │               │
-                └─ 1 repetition of a one-hour interval with keep=all
-                                 │               │
-                                 └─ 24 repetitions of a one-hour interval with keep=1
-                                                 │
-                                                 └─ 6 repetitions of a 30-day interval with keep=1
+                │                │
+                └─ one hour interval
+                                 │
+                                 └─ 24 adjacent one-hour intervals
      ...

-The retention grid can be thought of as a time-based sieve that thins out snapshots as they get older.
+The retention grid can be thought of as a time-based sieve:
+The ``grid`` field specifies a list of adjacent time intervals:
+the left edge of the leftmost (first) interval is the ``creation`` date of the youngest snapshot.
+All intervals to its right describe time intervals further in the past.
+
+Each interval carries a maximum number of snapshots to keep.
+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 ``grid`` field specifies a list of adjacent time intervals.
-Each interval is a bucket with a maximum capacity of ``keep`` snapshots.
 The following procedure happens during pruning:

 #. The list of snapshots is filtered by the regular expression in ``regex``.
-   Only snapshots names that match the regex are considered for this rule, all others will be pruned unless another rule keeps them.
-#. The snapshots that match ``regex`` are placed onto a time axis according to their ``creation`` date.
-   The youngest snapshot is on the left, the oldest on the right.
-#. The first buckets are placed "under" that axis so that the ``grid`` spec's first bucket's left edge aligns with youngest snapshot.
-#. All subsequent buckets are placed adjacent to their predecessor bucket.
-#. Now each snapshot on the axis either falls into one bucket or it is older than our rightmost bucket.
-   Buckets are left-inclusive and right-exclusive which means that a snapshot on the edge of bucket will always 'fall into the right one'.
-#. Snapshots older than the rightmost bucket **not kept** by this gridspec.
-#. For each bucket, we only keep the ``keep`` oldest snapshots.
+   Only snapshots names that match the regex are considered for this rule, all others are not affected.
+#. The filtered list of snapshots is sorted by ``creation``
+#. The left edge of the first interval is aligned to the ``creation`` date of the youngest snapshot
+#. A list of buckets is created, one for each interval
+#. The list of snapshots is split up into the buckets.
+#. For each bucket

-The syntax to describe the bucket list is as follows:
-
-::
-
-     Repeat x Duration (keep=all)
-
-* The **duration** specifies the length of the interval.
-* The **keep** count specifies the number of snapshots that fit into the bucket.
-  It can be either a positive integer or ``all`` (all snapshots are kept).
-* The **repeat** count repeats the bucket definition for the specified number of times.
-
-**Example**:
-
-::
-
-          This grid spec produces the following list of adjacent buckets. For the sake of simplicity,
-          we subject all snapshots to the grid pruning policy by settings `regex: .*`.
-
-          `
-           grid: 1x1h(keep=all) | 2x2h | 1x3h
-           regex: .*
-          `
-
-          0h        1h        2h        3h        4h        5h        6h        7h        8h        9h
-          |         |         |         |         |         |         |         |         |         |
-          |-Bucket1-|-----Bucket 2------|------Bucket 3-----|-----------Bucket 4----------|
-          | keep=all|      keep=1       |       keep=1      |            keep=1           |
-
-
-
-          Let us consider the following set of snapshots @a-zA-C:
-
-
-          |  a  b  c  d  e  f  g  h  i  j  k  l  m  n  o  p  q  r  s  t  u  v  w  x  y  z  A  B  C  D        |
-
-          The `grid` algorithm maps them to their respective buckets:
-
-          Bucket 1: a, b, c
-          Bucket 2: d,e,f,g,h,i,j
-          Bucket 3: k,l,m,n,o,p
-          Bucket 4: q,r, q,r,s,t,u,v,w,x,y,z
-          None:     A,B,C,D
-
-          It then applies the per-bucket pruning logic described above which resulting in the
-          following list of remaining snapshots.
-
-          |  a  b  c                    j                 p                             z                    |
-
-          Note that it only makes sense to grow (not shorten) the interval duration for buckets
-          further in the past since each bucket acts like a low-pass filter for incoming snapshots
-          and adding a less-low-pass-filter after a low-pass one has no effect.
+   #. the contained snapshot list is sorted by creation.
+   #. snapshots from the list, oldest first, are destroyed until the specified ``keep`` count is reached.
+   #. all remaining snapshots on the list are kept.


 .. _prune-keep-last-n:
@@ -175,11 +125,9 @@ Policy ``last_n``
         keep_receiver:
         - type: last_n
           count: 10
-           regex: ^zrepl_.*$ # optional
     ...

-``last_n`` filters the snapshot list by ``regex``, then keeps the last ``count`` snapshots in that list (last = youngest = most recent creation date)
-All snapshots that don't match ``regex`` or exceed ``count`` in the filtered list are destroyed unless matched by other rules.
+``last_n`` keeps the last ``count`` snapshots (last = youngest = most recent creation date).

 .. _prune-keep-regex:

@@ -1,47 +0,0 @@
-.. include:: ../global.rst.inc
-
-
-Replication Options
-===================
-
-
-::
-   
-   jobs:
-   - type: push
-     filesystems: ...
-     replication:
-       protection:
-         initial:     guarantee_resumability # guarantee_{resumability,incremental,nothing}
-         incremental: guarantee_resumability # guarantee_{resumability,incremental,nothing}
-     ...
-
-.. _replication-option-protection:
-
-``protection`` option
--------------------------
-
-The ``protection`` variable controls the degree to which a replicated filesystem is protected from getting out of sync through a zrepl pruner or external tools that destroy snapshots.
-zrepl can guarantee :ref:`resumability <step-holds>` or just :ref:`incremental replication <replication-cursor-and-last-received-hold>`.
-
-``guarantee_resumability`` is the **default** value and guarantees that a replication step is always resumable and that incremental replication will always be possible.
-The implementation uses replication cursors, last-received-hold and step holds.
-
-``guarantee_incremental`` only guarantees that incremental replication will always be possible.
-If a step ``from -> to`` is interrupted and its `to` snapshot is destroyed, zrepl will remove the half-received ``to``'s resume state and start a new step ``from -> to2``.
-The implementation uses replication cursors, tentative replication cursors and last-received-hold.
-
-``guarantee_nothing`` does not make any guarantees with regards to keeping sending and receiving side in sync.
-No bookmarks or holds are created to protect sender and receiver from diverging.
-
-**Tradeoffs**
-
-Using ``guarantee_incremental`` instead of ``guarantee_resumability`` obviously removes the resumability guarantee.
-This means that replication progress is no longer monotonic which might lead to a replication setup that never makes progress if mid-step interruptions are too frequent (e.g. frequent network outages).
-However, the advantage and :issue:`reason for existence <288>` of the ``incremental`` mode is that it allows the pruner to delete snapshots of interrupted replication steps
-which is useful if replication happens so rarely (or fails so frequently) that the amount of disk space exclusively referenced by the step's snapshots becomes intolerable.
-
-.. NOTE::
-
-   When changing this flag, obsoleted zrepl-managed bookmarks and holds will be destroyed on the next replication step that is attempted for each filesystem.
-
@@ -131,7 +131,7 @@ The following environment variables are set:
 * ``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`.
+An empty template hook can be found in :sampleconf:`hooks/template.sh`.

 .. _job-hook-type-postgres-checkpoint:

@@ -91,8 +91,7 @@ The ``tls`` transport uses TCP + TLS with client authentication using client cer
 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, mutual TLS might also be sufficient.
-We provide :ref:`copy-pastable instructions to generate the certificates below <transport-tcp+tlsclientauth-certgen>`.
+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.
@@ -104,16 +103,6 @@ If intermediate CAs are used, the **full chain** must be present in either in th
 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.

-.. NOTE::
-
-   As of Go 1.15 (zrepl 0.3.0 and newer), the Go TLS / x509 library **requrires Subject Alternative Names**
-   be present in certificates. You might need to re-generate your certificates using one of the :ref:`two alternatives
-   provided below<transport-tcp+tlsclientauth-certgen>`.
-
-   Note further that zrepl continues to use the CommonName field to assign client identities.
-   Hence, we recommend to keep the Subject Alternative Name and the CommonName in sync.
-
-
 Serve
 ~~~~~

@@ -158,25 +147,45 @@ 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-certgen:
-
 .. _transport-tcp+tlsclientauth-2machineopenssl:

-Mutual-TLS between Two Machines
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+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

-   (name=HOSTNAME; openssl req -x509 -sha256 -nodes \
-    -newkey rsa:4096 \
-    -days 365 \
-    -keyout $name.key \
-    -out $name.crt -addext "subjectAltName = DNS:$name" -subj "/CN=$name")
+   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:
@@ -207,36 +216,6 @@ The serve & connect configuration will thus look like the following:
     ...


-Certificate Authority using EasyRSA
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-For more than two machines, it might make sense to set up a CA infrastructure.
-Tools like `EasyRSA <https://github.com/OpenVPN/easy-rsa>`_ make this very easy:
-
-::
-
-     #!/usr/bin/env bash
-     set -euo pipefail
-
-     HOSTS=(backupserver prod1 prod2 prod3)
-
-     curl -L https://github.com/OpenVPN/easy-rsa/releases/download/v3.0.7/EasyRSA-3.0.7.tgz > EasyRSA-3.0.7.tgz
-     echo "157d2e8c115c3ad070c1b2641a4c9191e06a32a8e50971847a718251eeb510a8  EasyRSA-3.0.7.tgz" | sha256sum -c
-     rm -rf EasyRSA-3.0.7
-     tar -xf EasyRSA-3.0.7.tgz
-     cd EasyRSA-3.0.7
-     ./easyrsa
-     ./easyrsa init-pki
-     ./easyrsa build-ca nopass
-
-     for host in "${HOSTS[@]}"; do
-         ./easyrsa build-serverClient-full $host nopass
-         echo cert for host $host available at pki/issued/$host.crt
-         echo key for host $host available at pki/private/$host.key
-     done
-     echo ca cert available at pki/ca.crt
-
-
 .. _transport-ssh+stdinserver:

 ``ssh+stdinserver`` Transport
@@ -1,74 +0,0 @@
-#!/usr/bin/env python3
-
-import subprocess
-import re
-
-output = subprocess.run(["git", "tag", "-l"], capture_output=True, check=True, text=True)
-tagRE = re.compile(r"^v(?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(-rc(?P<rc>\d+))?$")
-class Tag:
-    orig: str
-    major: int
-    minor: int
-    patch: int
-    rc: int
-
-    def __str__(self):
-        return self.orig
-
-    def __repr__(self):
-        return self.orig
-
-tags = []
-for line in output.stdout.split("\n"):
-    m = tagRE.match(line)
-    if not m:
-        continue
-    m = m.groupdict()
-
-    t = Tag()
-    t.orig = line
-    t.major = int(m["major"])   
-    t.minor = int(m["minor"])   
-    t.patch = int(m["patch"])
-    t.rc = int(m["rc"] if m["rc"] is not None else 0)
-
-    tags.append(t)
-
-by_major_minor = {}
-
-for tag in tags:
-    key = (tag.major, tag.minor)
-    l = by_major_minor.get(key, [])
-    l.append(tag)
-    by_major_minor[key] = l
-
-latest_by_major_minor = []
-for (mm, l) in by_major_minor.items():
-    # sort ascending by patch-level (rc's weigh less)
-    l.sort(key=lambda tag: (tag.patch, int(tag.rc == 0), tag.rc))
-    latest_by_major_minor.append(l[-1])
-latest_by_major_minor.sort(key=lambda tag: (tag.major, tag.minor))
-
-# print(by_major_minor)
-# print(latest_by_major_minor)
-
-cmdline = []
-
-for latest_patch in latest_by_major_minor:
-    cmdline.append("--whitelist-tags")
-    cmdline.append(f"^{re.escape(latest_patch.orig)}$")
-
-# we want flexibility to update docs for the latest stable release
-# => we have a branch for that, called `stable` which we move manually
-# TODO: in the future, have f"stable-{latest_by_major_minor[-1]}"
-default_version = "stable"
-cmdline.extend(["--whitelist-branches", default_version])
-
-cmdline.extend(["--root-ref", f"{default_version}"])
-cmdline.extend(["--banner-main-ref", f"{default_version}"])
-cmdline.extend(["--show-banner"])
-cmdline.extend(["--sort", "semver"])
-
-cmdline.extend(["--whitelist-branches", "master"])
-
-print(" ".join(cmdline))
@@ -11,12 +11,10 @@
   :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/patrons/zrepl.svg?logo=liberapay
+.. |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.vercel.app%2Fapi%3Fusername%3Dzrepl%26type%3Dpatrons&style=flat&color=yellow
+.. |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
-.. |Donate via GitHub Sponsors| image:: https://img.shields.io/static/v1?label=Sponsor&message=%E2%9D%A4&logo=GitHub&style=flat&color=yellow
-   :target: https://github.com/sponsors/problame
 .. |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

@@ -27,8 +25,4 @@
 .. |recv-options| replace:: :ref:`recv options<job-recv-options>`
 .. |snapshotting-spec| replace:: :ref:`snapshotting specification <job-snapshotting-spec>`
 .. |pruning-spec| replace:: :ref:`pruning specification <prune>`
-.. |filter-spec| replace:: :ref:`filter specification<pattern-filter>`
-
-.. |br| raw:: html
-
-   <br/>
+.. |filter-spec| replace:: :ref:`filter specification<pattern-filter>`
@@ -5,7 +5,7 @@

 .. include:: global.rst.inc

-|GitHub license| |Language: Go| |Twitter| |Donate via Patreon| |Donate via GitHub Sponsors| |Donate via Liberapay| |Donate via PayPal|
+|GitHub license| |Language: Go| |Twitter| |Donate via Patreon| |Donate via Liberapay| |Donate via PayPal| 


 zrepl - ZFS replication
@@ -44,7 +44,7 @@ zrepl - ZFS replication
 Getting started
 ~~~~~~~~~~~~~~~

-The :ref:`10 minute quick-start guides <quickstart-toc>` give you a first impression.
+The :ref:`10 minutes tutorial setup <tutorial>` gives you a first impression.

 Main Features
 ~~~~~~~~~~~~~
@@ -125,7 +125,7 @@ Table of Contents
   :maxdepth: 2
   :caption: Contents:

-   quickstart
+   tutorial
   installation
   configuration
   usage
@@ -7,14 +7,13 @@ Installation

 .. TIP::

-    Note: check out the :ref:`quick-start guides <quickstart-toc>` if you want a first impression of zrepl.
+    Note: check out the :ref:`tutorial` if you want a first impression of zrepl.

 .. toctree::

    installation/user-privileges
    installation/packages
    installation/apt-repos
-    installation/rpm-repos
    installation/compile-from-source
    installation/freebsd-jail-with-iocage
    installation/what-next
@@ -7,7 +7,7 @@ Debian / Ubuntu APT repositories
 We maintain APT repositories for Debian, Ubuntu and derivatives.
 The fingerprint of the signing key is ``E101 418F D3D6 FBCB 9D65  A62D 7086 99FC 5F2E BF16``.
 It is available at `<https://zrepl.cschwarz.com/apt/apt-key.asc>`_ .
-Please open an issue in on GitHub if you encounter any issues with the repository.
+Please open an issue `in the packaging repository <https://github.com/zrepl/debian-binary-packaging>`_ if you encounter any issues with the repository.

 The following snippet configure the repository for your Debian or Ubuntu release:

@@ -24,6 +24,6 @@ The following snippet configure the repository for your Debian or Ubuntu release

 .. NOTE::

-   Until zrepl reaches 1.0, the repositories will be updated to the latest zrepl release immediately.
+   Until zrepl reaches 1.0, all APT repositories will be updated to the latest zrepl release immediately.
   This includes breaking changes between zrepl versions.
   Use ``apt-mark hold zrepl`` to prevent upgrades of zrepl.
@@ -9,35 +9,37 @@ Producing a release requires **Go 1.11** or newer and **Python 3** + **pip3** +
 A tutorial to install Go is available over at `golang.org <https://golang.org/doc/install>`_.
 Python and pip3 should probably be installed via your distro's package manager.

-::
-   cd to/your/zrepl/checkout
-   python3 -m venv3
-   source venv3/bin/activate
-   ./lazy.sh devsetup
-   make release
-   # build artifacts are available in ./artifacts/release
-
-The Python venv is used for the documentation build dependencies.
-If you just want to build the zrepl binary, leave it out and use `./lazy.sh godep` instead.
-
 Alternatively, you can use the Docker build process:
 it is used to produce the official zrepl `binary releases`_
 and serves as a reference for build dependencies and procedure:

 ::

-    cd to/your/zrepl/checkout
-    # make sure your user has access to the docker socket
-    make release-docker
-    # if you want .deb or .rpm packages, invoke the follwoing
-    # targets _after_ you invoked release-docker
-    make deb-docker
-    make rpm-docker
-    # build artifacts are available in ./artifacts/release
-    # packages are available in ./artifacts
+    git clone https://github.com/zrepl/zrepl.git && \
+    cd zrepl && \
+    sudo docker build -t zrepl_build -f build.Dockerfile . && \
+    sudo docker run -it --rm \
+        -v "${PWD}:/src" \
+        --user "$(id -u):$(id -g)" \
+        zrepl_build make release

+Alternatively, you can install build dependencies on your local system and then build in your ``$GOPATH``:
+
+::
+
+    mkdir -p "${GOPATH}/src/github.com/zrepl/zrepl"
+    git clone https://github.com/zrepl/zrepl.git "${GOPATH}/src/github.com/zrepl/zrepl"
+    cd "${GOPATH}/src/github.com/zrepl/zrepl"
+    python3 -m venv3
+    source venv3/bin/activate
+    ./lazy.sh devsetup
+    make release
+
+The Python venv is used for the documentation build dependencies.
+If you just want to build the zrepl binary, leave it out and use `./lazy.sh godep` instead.
+Either way, all build results are located in the ``artifacts/`` directory.

 .. NOTE::

-    It is your job to install the built binary in the zrepl users's ``$PATH``, e.g. ``/usr/local/bin/zrepl``.
-    Otherwise, the examples in the :ref:`quick-start guides <quickstart-toc>` may need to be adjusted.
+    It is your job to install the appropriate binary in the zrepl users's ``$PATH``, e.g. ``/usr/local/bin/zrepl``.
+    Otherwise, the examples in the :ref:`tutorial` may need to be adjusted.
@@ -126,7 +126,7 @@ Modify the ``/usr/local/etc/zrepl/zrepl.yml`` configuration file.

 .. TIP::

-    Note: check out the :ref:`quick-start guides <quickstart-toc>` for examples of a ``sink`` job.
+    Note: check out the :ref:`tutorial` for examples of a ``sink`` job.

 Now ``zrepl`` can be started.

@@ -30,12 +30,15 @@ The following list may be incomplete, feel free to submit a PR with an update:
    * - Arch Linux
      - ``yay install zrepl``
      - Available on `AUR <https://aur.archlinux.org/packages/zrepl>`_
-    * - Fedora, CentOS, RHEL, OpenSUSE
+    * - Fedora
      - ``dnf install zrepl``
-      - :ref:`RPM repository config <installation-rpm-repos>`
+      - Available on `COPR <https://copr.fedorainfracloud.org/coprs/poettlerric/zrepl/>`_
+    * - CentOS/RHEL
+      - ``yum install zrepl``
+      - Available on `COPR <https://copr.fedorainfracloud.org/coprs/poettlerric/zrepl/>`_
    * - Debian + Ubuntu
      - ``apt install zrepl``
-      - :ref:`APT repository config <installation-apt-repos>`
+      - APT repository config :ref:`see below <installation-apt-repos>`
    * - OmniOS
      - ``pkg install zrepl``
      - Available since `r151030 <https://pkg.omniosce.org/r151030/extra/en/search.shtml?token=zrepl&action=Search>`_
@@ -1,30 +0,0 @@
-.. _installation-rpm-repos:
-
-RPM repositories
-~~~~~~~~~~~~~~~~
-
-We provide a single RPM repository for all RPM-based Linux distros.
-The zrepl binary in the repo is the same as the one published to GitHub.
-Since Go binaries are statically linked, the RPM should work about everywhere.
-
-The fingerprint of the signing key is ``F6F6 E8EA 6F2F 1462 2878 B5DE 50E3 4417 826E 2CE6``.
-It is available at `<https://zrepl.cschwarz.com/rpm/rpm-key.asc>`_ .
-Please open an issue on GitHub if you encounter any issues with the repository.
-
-Copy-paste the following snippet into your shell to set up the zrepl repository.
-Then ``dnf install zrepl`` and make sure to confirm that the signing key matches the one shown above.
-
-::
-
-    cat > /etc/yum.repos.d/zrepl.repo <<EOF
-    [zrepl]
-    name = zrepl
-    baseurl = https://zrepl.cschwarz.com/rpm/repo
-    gpgkey = https://zrepl.cschwarz.com/rpm/rpm-key.asc
-    EOF
-
-.. NOTE::
-
-   Until zrepl reaches 1.0, the repository will be updated to the latest zrepl release immediately.
-   This includes breaking changes between zrepl versions.
-   If that bothers you, use the `dnf versionlock plugin <https://dnf-plugins-core.readthedocs.io/en/latest/versionlock.html>`_ to pin the version of zrepl on your system.
@@ -5,4 +5,4 @@ What next?

 Read the :ref:`configuration chapter<configuration_toc>` and then continue with the :ref:`usage chapter<usage>`.

-**Reminder**: If you want a quick introduction, please read the :ref:`quick-start guides <quickstart-toc>`.
+**Reminder**: If you want a quick introduction, please read the :ref:`tutorial`.
@@ -1,30 +1,14 @@
 #!/bin/bash
 set -eo pipefail

-
-NON_INTERACTIVE=false
-DO_CLONE=false
-while getopts "ca" arg; do
-    case "$arg" in
-        "a")
-            NON_INTERACTIVE=true
-            ;;
-        "c")
-            DO_CLONE=true
-            ;;
-        *)
-            echo invalid option
-            exit 1
-            ;;
-    esac
-done
-
 GHPAGESREPO="git@github.com:zrepl/zrepl.github.io.git"
 SCRIPTDIR=$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )
 PUBLICDIR="${SCRIPTDIR}/public_git"

 checkout_repo_msg() {
    echo "clone ${GHPAGESREPO} to ${PUBLICDIR}:"
+    echo "	git clone ${GHPAGESREPO} ${PUBLICDIR}"
+    git clone "${GHPAGESREPO}" "${PUBLICDIR}"
 }

 exit_msg() {
@@ -41,19 +25,11 @@ cd "$SCRIPTDIR"

 if [ ! -d "$PUBLICDIR" ]; then
    checkout_repo_msg
-    if $DO_CLONE; then
-        git clone "${GHPAGESREPO}" "${PUBLICDIR}"
-    else
-        exit 1
-    fi
+    exit 1
 fi

-if $NON_INTERACTIVE; then
-    echo "non-interactive mode"
-else
-    echo -n "PRESS ENTER to confirm you commited and pushed docs changes and tags to the zrepl repo"
-    read -r
-fi
+echo -n "PRESS ENTER to confirm you commited and pushed docs changes and tags to the zrepl repo"
+read

 pushd "$PUBLICDIR" 

@@ -75,11 +51,11 @@ git rm -rf .
 popd

 echo "building site"
-
-flags="$(python3 gen-sphinx-versioning-flags.py)"
 set -e
 sphinx-versioning build \
-   $flags \
+   --show-banner \
+   --whitelist-branches '^master$' \
+   --whitelist-tags '(v)?\d+\.[1-9][0-9]*.\d+$' \
   docs ./public_git \
   -- -c sphinxconf # older conf.py throw errors because they used
                    # version = subprocess.show_output(["git", "describe"])
@@ -91,7 +67,7 @@ git status --porcelain
 if [[ "$(git status --porcelain)" != "" ]]; then
    CURRENT_COMMIT="${CURRENT_COMMIT}(dirty)" 
 fi
-COMMIT_MSG="sphinx-versioning render from publish.sh - $(date -u) - ${CURRENT_COMMIT}"
+COMMIT_MSG="sphinx-versioning render from publish.sh - `date -u` - ${CURRENT_COMMIT}"

 pushd "$PUBLICDIR"

@@ -1,87 +0,0 @@
-.. include:: global.rst.inc
-
-.. _quickstart-toc:
-
-***********************
-Quick Start by Use Case
-***********************
-
-The goal of this quick-start guide is to give you an impression of how zrepl can accomodate your use case.
-
-Install zrepl
-=============
-
-Follow the :ref:`OS-specific installation instructions <installation_toc>` and come back here.
-
-Overview Of How zrepl Works
-============================
-
-Check out the :ref:`overview section <job-overview>` to get a rough idea of what you are going to configure in the next step, then come back here.
-
-Configuration Examples
-======================
-
-zrepl is configured through a YAML configuration file in ``/etc/zrepl/zrepl.yml``.
-We have prepared example use cases that show-case typical deployments and different functionality of zrepl.
-We encourage you to read through all of the examples to get an idea of what zrepl has to offer, and how you can mix-and-match configurations for your use case.
-Keep the :ref:`full config documentation <configuration_toc>` handy if a config snippet is unclear.
-
-**Example Use Cases**
-
-.. toctree::
-   :titlesonly:
-
-   quickstart/continuous_server_backup
-   quickstart/backup_to_external_disk
-
-Use ``zrepl configcheck`` to validate your configuration.
-No output indicates that everything is fine.
-
-.. NOTE::
-
-   Please open an issue on GitHub if your use case for zrepl is significantly different from those listed above.
-   Or even better, write it up in the same style as above and open a PR!
-
-.. _quickstart-apply-config:
-
-Apply Configuration Changes
-===========================
-
-We hope that you have found a configuration that fits your use case.
-Use ``zrepl configcheck`` once again to make sure the config is correct (output indicates that everything is fine).
-Then restart the zrepl daemon on all systems involved in the replication, likely using ``service zrepl restart`` or ``systemctl restart zrepl``.
-
-Watch it Work
-=============
-
-Run ``zrepl status`` on the active side of the replication setup to monitor snaphotting, replication and pruning activity.
-To re-trigger replication (snapshots are separate!), use ``zrepl signal wakeup JOBNAME``.
-(refer to the example use case document if you are uncertain which job you want to wake up).
-
-You can also use basic UNIX tools to inspect see what's going on.
-If you like tmux, here is a handy script that works on FreeBSD: ::
-
-    pkg install gnu-watch tmux
-    tmux new -s zrepl -d
-    tmux split-window -t zrepl "tail -f /var/log/messages"
-    tmux split-window -t zrepl "gnu-watch 'zfs list -t snapshot -o name,creation -s creation'"
-    tmux split-window -t zrepl "zrepl status"
-    tmux select-layout -t zrepl tiled
-    tmux attach -t zrepl
-
-The Linux equivalent might look like this: ::
-
-    # make sure tmux is installed & let's assume you use systemd + journald
-    tmux new -s zrepl -d
-    tmux split-window -t zrepl  "journalctl -f -u zrepl.service"
-    tmux split-window -t zrepl "watch 'zfs list -t snapshot -o name,creation -s creation'"
-    tmux split-window -t zrepl "zrepl status"
-    tmux select-layout -t zrepl tiled
-    tmux attach -t zrepl
-
-What Next?
-==========
-
-* Read more about :ref:`configuration format, options & job types <configuration_toc>`
-* Configure :ref:`logging <logging>` \& :ref:`monitoring <monitoring>`.
-
@@ -1,36 +0,0 @@
-.. include:: ../global.rst.inc
-
-.. _quickstart-backup-to-external-disk:
-
-Local Snapshots + Offline Backup to an External Disk
-====================================================
-
-This config example shows how we can use zrepl to make periodic snapshots of our local workstation and back it up to a zpool on an external disk which we occassionally connect.
-
-The local snapshots should be taken every 15 minutes for pain-free recovery from CLI disasters (``rm -rf /`` and the like).
-However, we do not want to keep the snapshots around for very long because our workstation is a little tight on disk space.
-Thus, we only keep one hour worth of high-resolution snapshots, then fade them out to one per hour for a day (24 hours), then one per day for 14 days.
-
-At the end of each work day, we connect our external disk that serves as our workstation's local offline backup.
-We want zrepl to inspect the filesystems and snapshots on the external pool, figure out which snapshots were created since the last time we connected the external disk, and use incremental replication to efficiently mirror our workstation to our backup disk.
-Afterwards, we want to clean up old snapshots on the backup pool: we want to keep all snapshots younger than one hour, 24 for each hor of the first day, then 360 daily backups.
-
-A few additional requirements:
-
-* Snapshot creation and pruning on our workstation should happen in the background, without interaction from our side.
-* However, we want to explicitly trigger replication via the command line.
-* We want to use OpenZFS native encryption to protect our data on the external disk.
-  It is absolutely critical that only encrypted data leaves our workstation.
-  **zrepl should provide an easy config knob for this and prevent replication of unencrypted datasets to the external disk.**
-* We want to be able to put off the backups for more than three weeks, i.e., longer than the lifetime of the automatically created snapshots on our workstation.
-  **zrepl should use bookmarks and holds to achieve this goal**.
-* When we yank out the drive during replication and go on a long vacation, we do *not* want the partially replicated snapshot to stick around as it would hold on to too much disk space over time.
-  Therefore, we want zrepl to deviate from its :ref:`default behavior <replication-option-protection>` and sacrifice resumability, but nonetheless retain the ability to do incremental replication once we return from our vacation.
-  **zrepl should provide an easy config knob to disable step holds for incremental replication**.
-
-The following config snippet implements the setup described above.
-You will likely want to customize some aspects mentioned in the top comment in the file.
-
-.. literalinclude:: ../../config/samples/quickstart_backup_to_external_disk.yml
-
-:ref:`Click here <quickstart-apply-config>` to go back to the quickstart guide.
@@ -1,94 +0,0 @@
-.. include:: ../global.rst.inc
-
-.. _quickstart-continuous-replication:
-
-Continuous Backup of a Server
-=============================
-
-This config example shows how we can backup our ZFS-based server to another machine using a zrepl push job.
-
-* Production server ``prod`` with filesystems to back up:
-
-  * ``zroot/var/db``
-  * ``zroot/usr/home`` and all its child filesystems
-  * **except** ``zroot/usr/home/paranoid`` belonging to a user doing backups themselves
-
-* Backup server ``backups`` with
-
-  * Filesystem ``storage/zrepl/sink/prod`` + children dedicated to backups of ``prod``
-
-Our backup solution should fulfill the following requirements:
-
-* Periodically snapshot the filesystems on ``prod`` *every 10 minutes*
-* Incrementally replicate these snapshots to ``storage/zrepl/sink/prod/*`` on ``backups``
-* Keep only very few snapshots on ``prod`` to save disk space
-* Keep a fading history (24 hourly, 30 daily, 6 monthly) of snapshots on ``backups``
-* The network is untrusted - zrepl should use TLS to protect its communication and our data.
-
-Analysis
--------
-
-We can model this situation as two jobs:
-
-* A **push job** on ``prod``
-
-  * Creates the snapshots
-  * Keeps a short history of local snapshots to enable incremental replication to ``backups``
-  * Connects to the ``zrepl daemon`` process on ``backups``
-  * Pushes snapshots ``backups``
-  * Prunes snapshots on ``backups`` after replication is complete
-
-* A **sink job** on ``backups``
-
-  * Accepts connections & responds to requests from ``prod``
-  * Limits client ``prod`` access to filesystem sub-tree ``storage/zrepl/sink/prod``
-
-Generate TLS Certificates
-------------------------
-
-We use the :ref:`TLS client authentication transport <transport-tcp+tlsclientauth>` to protect our data on the wire.
-To get things going quickly, we skip setting up a CA and generate two self-signed certificates as described :ref:`here <transport-tcp+tlsclientauth-2machineopenssl>`.
-For convenience, we generate the key pairs on our local machine and distribute them using ssh:
-
-.. code-block:: bash
-
-   (name=backups; openssl req -x509 -sha256 -nodes \
-    -newkey rsa:4096 \
-    -days 365 \
-    -keyout $name.key \
-    -out $name.crt -addext "subjectAltName = DNS:$name" -subj "/CN=$name")
-
-   (name=prod; openssl req -x509 -sha256 -nodes \
-    -newkey rsa:4096 \
-    -days 365 \
-    -keyout $name.key \
-    -out $name.crt -addext "subjectAltName = DNS:$name" -subj "/CN=$name")
-
-   ssh root@backups "mkdir /etc/zrepl"
-   scp  backups.key backups.crt prod.crt root@backups:/etc/zrepl
-
-   ssh root@prod "mkdir /etc/zrepl"
-   scp  prod.key prod.crt backups.crt root@prod:/etc/zrepl
-
-Note that alternative transports exist, e.g. via :ref:`TCP without TLS <transport-tcp>` or :ref:`ssh <transport-ssh+stdinserver>`.
-
-Configure server ``prod``
-------------------------
-
-We define a **push job** named ``prod_to_backups`` in ``/etc/zrepl/zrepl.yml`` on host ``prod`` :
-
-.. literalinclude:: ../../config/samples/quickstart_continuous_server_backup_sender.yml
-
-.. _tutorial-configure-prod:
-
-Configure server ``backups``
----------------------------
-
-We define a corresponding **sink job** named ``sink`` in ``/etc/zrepl/zrepl.yml`` on host ``backups`` :
-
-.. literalinclude:: ../../config/samples/quickstart_continuous_server_backup_receiver.yml
-
-Go Back To Quickstart Guide
---------------------------
-
-:ref:`Click here <quickstart-apply-config>` to go back to the quickstart guide.
@@ -2,7 +2,7 @@

 .. _supporters:

-|Donate via Patreon| |Donate via GitHub Sponsors| |Donate via Liberapay| |Donate via PayPal|
+|Donate via Patreon| |Donate via Liberapay| |Donate via PayPal|

 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 services listed above.
@@ -32,10 +32,6 @@ We would like to thank the following people and organizations for supporting zre

   <div class="fa fa-code" style="width: 1em;"></div>

-* |supporter-std| `Gordon Schulz <https://github.com/azmodude>`_
-* |supporter-std| `@jwittlincohen <https://github.com/jwittlincohen>`_
-* |supporter-std| `Michael D. Schmitt <https://waterbendingscroll.dancingdragons.org>`_
-* |supporter-std| `Hans Schulz <https://github.com/schulzh>`_
 * |supporter-std| Henning Kessler
 * |supporter-std| `John Ramsden <https://github.com/johnramsden>`_
 * |supporter-std| `DrLuke <https://github.com/drluke>`_
@@ -47,7 +43,7 @@ We would like to thank the following people and organizations for supporting zre
 * |supporter-gold| `MNX.io <https://mnx.io>`_
 * |supporter-std| `Marshall Clyburn <https://github.com/mdclyburn>`_
 * |supporter-code| `Ross Williams <https://github.com/overhacked>`_
-* |supporter-gold| Mike T.
+* |supporter-std| Mike T.
 * |supporter-code| `Justin Scholz <https://github.com/JMoVS>`_
 * |supporter-code| `InsanePrawn <https://github.com/InsanePrawn>`_
 * |supporter-code| `Ben Woods <https://www.freshports.org/sysutils/zrepl/>`_
@@ -0,0 +1,177 @@
+.. include:: global.rst.inc
+
+.. _tutorial:
+
+Tutorial
+========
+
+
+This tutorial shows how zrepl can be used to implement a ZFS-based push backup.
+We assume the following scenario:
+
+* Production server ``prod`` with filesystems to back up:
+
+  * ``zroot/var/db``
+  * ``zroot/usr/home`` and all its child filesystems
+  * **except** ``zroot/usr/home/paranoid`` belonging to a user doing backups themselves
+
+* Backup server ``backups`` with
+
+  * Filesystem ``storage/zrepl/sink/prod`` + children dedicated to backups of ``prod``
+
+Our backup solution should fulfill the following requirements:
+
+* Periodically snapshot the filesystems on ``prod`` *every 10 minutes*
+* Incrementally replicate these snapshots to ``storage/zrepl/sink/prod/*`` on ``backups``
+* Keep only very few snapshots on ``prod`` to save disk space
+* Keep a fading history (24 hourly, 30 daily, 6 monthly) of snapshots on ``backups``
+
+Analysis
+--------
+
+We can model this situation as two jobs:
+
+* A **push job** on ``prod``
+
+  * Creates the snapshots
+  * Keeps a short history of local snapshots to enable incremental replication to ``backups``
+  * Connects to the ``zrepl daemon`` process on ``backups``
+  * Pushes snapshots ``backups``
+  * Prunes snapshots on ``backups`` after replication is complete
+
+* A **sink job** on ``backups``
+
+  * Accepts connections & responds to requests from ``prod``
+  * Limits client ``prod`` access to filesystem sub-tree ``storage/zrepl/sink/prod``
+
+Install zrepl
+-------------
+
+Follow the :ref:`OS-specific installation instructions <installation_toc>` and come back here.
+
+Generate TLS Certificates
+-------------------------
+
+We use the :ref:`TLS client authentication transport <transport-tcp+tlsclientauth>` to protect our data on the wire.
+To get things going quickly, we skip setting up a CA and generate two self-signed certificates as described :ref:`here <transport-tcp+tlsclientauth-2machineopenssl>`.
+For convenience, we generate the key pairs on our local machine and distribute them using ssh:
+
+.. code-block:: bash
+   :emphasize-lines: 6,13
+
+   openssl req -x509 -sha256 -nodes \
+      -newkey rsa:4096 \
+      -days 365 \
+      -keyout backups.key \
+      -out backups.crt
+   # ... and use "backups" as Common Name (CN)
+
+   openssl req -x509 -sha256 -nodes \
+      -newkey rsa:4096 \
+      -days 365 \
+      -keyout prod.key \
+      -out prod.crt
+   # ... and use "prod" as Common Name (CN)
+
+   ssh root@backups "mkdir /etc/zrepl"
+   scp  backups.key backups.crt prod.crt root@backups:/etc/zrepl
+
+   ssh root@prod "mkdir /etc/zrepl"
+   scp  prod.key prod.crt backups.crt root@prod:/etc/zrepl
+
+
+Configure server ``prod``
+-------------------------
+
+We define a **push job** named ``prod_to_backups`` in ``/etc/zrepl/zrepl.yml`` on host ``prod`` : ::
+
+    jobs:
+    - name: prod_to_backups
+      type: push 
+      connect:
+        type: tls
+        address: "backups.example.com:8888"
+        ca: /etc/zrepl/backups.crt
+        cert: /etc/zrepl/prod.crt
+        key:  /etc/zrepl/prod.key
+        server_cn: "backups"
+      filesystems: {
+        "zroot/var/db": true,
+        "zroot/usr/home<": true,
+        "zroot/usr/home/paranoid": false
+      }
+      snapshotting:
+        type: periodic
+        prefix: zrepl_
+        interval: 10m
+      pruning:
+        keep_sender:
+        - type: not_replicated
+        - type: last_n
+          count: 10
+        keep_receiver:
+        - type: grid
+          grid: 1x1h(keep=all) | 24x1h | 30x1d | 6x30d
+          regex: "^zrepl_"
+
+.. _tutorial-configure-prod:
+
+Configure server ``backups``
+----------------------------
+
+We define a corresponding **sink job** named ``sink`` in ``/etc/zrepl/zrepl.yml`` on host ``backups`` : ::
+
+    jobs:
+    - name: sink
+      type: sink 
+      serve:
+         type: tls
+         listen: ":8888"
+         ca: "/etc/zrepl/prod.crt"
+         cert: "/etc/zrepl/backups.crt"
+         key: "/etc/zrepl/backups.key"
+         client_cns:
+           - "prod"
+      root_fs: "storage/zrepl/sink"
+
+
+Apply Configuration Changes
+---------------------------
+
+We use ``zrepl configcheck`` to catch any configuration errors: no output indicates that everything is fine.
+If that is the case, restart the zrepl daemon on **both** ``prod`` and ``backups`` using ``service zrepl restart`` or ``systemctl restart zrepl``.
+
+
+Watch it Work
+-------------
+
+Run ``zrepl status`` on ``prod`` to monitor the replication and pruning activity.
+To re-trigger replication (snapshots are separate!), use ``zrepl signal wakeup prod_to_backups`` on ``prod``.
+
+If you like tmux, here is a handy script that works on FreeBSD: ::
+
+    pkg install gnu-watch tmux
+    tmux new -s zrepl -d
+    tmux split-window -t zrepl "tail -f /var/log/messages"
+    tmux split-window -t zrepl "gnu-watch 'zfs list -t snapshot -o name,creation -s creation | grep zrepl_'"
+    tmux split-window -t zrepl "zrepl status"
+    tmux select-layout -t zrepl tiled
+    tmux attach -t zrepl
+
+The Linux equivalent might look like this: ::
+
+    # make sure tmux is installed & let's assume you use systemd + journald
+    tmux new -s zrepl -d
+    tmux split-window -t zrepl  "journalctl -f -u zrepl.service"
+    tmux split-window -t zrepl "watch 'zfs list -t snapshot -o name,creation -s creation | grep zrepl_'"
+    tmux split-window -t zrepl "zrepl status"
+    tmux select-layout -t zrepl tiled
+    tmux attach -t zrepl
+
+Summary
+-------
+
+Congratulations, you have a working push backup. Where to go next?
+
+* Read more about :ref:`configuration format, options & job types <configuration_toc>`
+* Configure :ref:`logging <logging>` \& :ref:`monitoring <monitoring>`.
@@ -38,7 +38,7 @@ CLI Overview
    * - ``zrepl migrate``
      - | perform on-disk state / ZFS property migrations
        | (see :ref:`changelog <changelog>` for details)
-    * - ``zrepl zfs-abstraction``
+    * - ``zrepl zfs-abstractions``
      - list and remove zrepl's abstractions on top of ZFS, e.g. holds and step bookmarks (see :ref:`overview <replication-cursor-and-last-received-hold>` )

 .. _usage-zrepl-daemon:
@@ -10,7 +10,6 @@ import (

 	"github.com/kr/pretty"
 	"github.com/pkg/errors"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"

 	"github.com/zrepl/zrepl/replication/logic/pdu"
@@ -113,6 +112,13 @@ func (s *Sender) ListFilesystemVersions(ctx context.Context, r *pdu.ListFilesyst
 	for i := range fsvs {
 		rfsvs[i] = pdu.FilesystemVersionFromZFS(&fsvs[i])
 	}
+
+	sendAbstractions := sendAbstractionsCacheSingleton.GetByFS(ctx, lp.ToString())
+	rSabsInfo := make([]*pdu.SendAbstraction, len(sendAbstractions))
+	for i := range rSabsInfo {
+		rSabsInfo[i] = SendAbstractionToPDU(sendAbstractions[i])
+	}
+
 	res := &pdu.ListFilesystemVersionsRes{Versions: rfsvs}
 	return res, nil

@@ -209,30 +215,36 @@ func (s *Sender) Send(ctx context.Context, r *pdu.SendReq) (*pdu.SendRes, io.Rea
 		return res, nil, nil
 	}

-	// create holds or bookmarks of `From` and `To` to guarantee one of the following:
-	// - that the replication step can always be resumed (`holds`),
-	// - that the replication step can be interrupted and a future replication
-	//   step with same or different `To` but same `From` is still possible (`bookmarks`)
-	// - nothing (`none`)
-	//
-	// ...
-	//
-	// ... actually create the abstractions
-	replicationGuaranteeOptions, err := replicationGuaranteeOptionsFromPDU(r.GetReplicationConfig().Protection)
-	if err != nil {
-		return nil, nil, err
-	}
-	replicationGuaranteeStrategy := replicationGuaranteeOptions.Strategy(sendArgs.From != nil)
-	liveAbs, err := replicationGuaranteeStrategy.SenderPreSend(ctx, s.jobId, &sendArgs)
-	if err != nil {
-		return nil, nil, err
-	}
-	for _, a := range liveAbs {
-		if a != nil {
-			abstractionsCacheSingleton.Put(a)
+	// create a replication cursor for `From` (usually an idempotent no-op because SendCompleted already created it before)
+	var fromReplicationCursor Abstraction
+	if sendArgs.From != nil {
+		// For all but the first replication, this should always be a no-op because SendCompleted already moved the cursor
+		fromReplicationCursor, err = CreateReplicationCursor(ctx, sendArgs.FS, *sendArgs.FromVersion, s.jobId) // no shadow
+		if err == zfs.ErrBookmarkCloningNotSupported {
+			getLogger(ctx).Debug("not creating replication cursor from bookmark because ZFS does not support it")
+			// fallthrough
+		} else if err != nil {
+			return nil, nil, errors.Wrap(err, "cannot set replication cursor to `from` version before starting send")
 		}
 	}

+	var fromHold, toHold Abstraction
+	// make sure `From` doesn't go away in order to make this step resumable
+	if sendArgs.From != nil {
+		fromHold, err = HoldStep(ctx, sendArgs.FS, *sendArgs.FromVersion, s.jobId) // no shadow
+		if err == zfs.ErrBookmarkCloningNotSupported {
+			getLogger(ctx).Debug("not creating step bookmark because ZFS does not support it")
+			// fallthrough
+		} else if err != nil {
+			return nil, nil, errors.Wrapf(err, "cannot hold `from` version %q before starting send", *sendArgs.FromVersion)
+		}
+	}
+	// make sure `To` doesn't go away in order to make this step resumable
+	toHold, err = HoldStep(ctx, sendArgs.FS, sendArgs.ToVersion, s.jobId)
+	if err != nil {
+		return nil, nil, errors.Wrapf(err, "cannot hold `to` version %q before starting send", sendArgs.ToVersion)
+	}
+
 	// cleanup the mess that _this function_ might have created in prior failed attempts:
 	//
 	// In summary, we delete every endpoint ZFS abstraction created on this filesystem for this job id,
@@ -253,6 +265,7 @@ func (s *Sender) Send(ctx context.Context, r *pdu.SendReq) (*pdu.SendRes, io.Rea
 		ctx, endSpan := trace.WithSpan(ctx, "cleanup-stale-abstractions")
 		defer endSpan()

+		liveAbs := []Abstraction{fromHold, toHold, fromReplicationCursor}
 		keep := func(a Abstraction) (keep bool) {
 			keep = false
 			for _, k := range liveAbs {
@@ -269,31 +282,23 @@ func (s *Sender) Send(ctx context.Context, r *pdu.SendReq) (*pdu.SendRes, io.Rea
 			}
 			for _, staleVersion := range obsoleteAbs {
 				for _, mustLiveVersion := range mustLiveVersions {
-					isSendArg := zfs.FilesystemVersionEqualIdentity(mustLiveVersion, staleVersion.GetFilesystemVersion())
-					stepHoldBasedGuaranteeStrategy := false
-					k := replicationGuaranteeStrategy.Kind()
-					switch k {
-					case ReplicationGuaranteeKindResumability:
-						stepHoldBasedGuaranteeStrategy = true
-					case ReplicationGuaranteeKindIncremental:
-					case ReplicationGuaranteeKindNone:
-					default:
-						panic(fmt.Sprintf("this is supposed to be an exhaustive match, got %v", k))
-					}
-					isSnapshot := mustLiveVersion.IsSnapshot()
-					if isSendArg && (!isSnapshot || stepHoldBasedGuaranteeStrategy) {
+					if zfs.FilesystemVersionEqualIdentity(mustLiveVersion, staleVersion.GetFilesystemVersion()) {
 						panic(fmt.Sprintf("impl error: %q would be destroyed because it is considered stale but it is part of of sendArgs=%s", mustLiveVersion.String(), pretty.Sprint(sendArgs)))
 					}
 				}
 			}
 		}
-		destroyTypes := AbstractionTypeSet{
-			AbstractionStepHold:                           true,
-			AbstractionTentativeReplicationCursorBookmark: true,
-		}
-		abstractionsCacheSingleton.TryBatchDestroy(ctx, s.jobId, sendArgs.FS, destroyTypes, keep, check)
+		sendAbstractionsCacheSingleton.TryBatchDestroy(ctx, s.jobId, sendArgs.FS, keep, check)
 	}()

+	if fromHold != nil {
+		sendAbstractionsCacheSingleton.Put(fromHold)
+	}
+	sendAbstractionsCacheSingleton.Put(toHold)
+	if fromReplicationCursor != nil {
+		sendAbstractionsCacheSingleton.Put(fromReplicationCursor)
+	}
+
 	sendStream, err := zfs.ZFSSend(ctx, sendArgs)
 	if err != nil {
 		// it's ok to not destroy the abstractions we just created here, a new send attempt will take care of it
@@ -326,32 +331,35 @@ func (p *Sender) SendCompleted(ctx context.Context, r *pdu.SendCompletedReq) (*p
 		return nil, errors.Wrap(err, "validate `to` exists")
 	}

-	replicationGuaranteeOptions, err := replicationGuaranteeOptionsFromPDU(orig.GetReplicationConfig().Protection)
-	if err != nil {
-		return nil, err
-	}
-	liveAbs, err := replicationGuaranteeOptions.Strategy(from != nil).SenderPostRecvConfirmed(ctx, p.jobId, fs, to)
-	if err != nil {
-		return nil, err
-	}
-	for _, a := range liveAbs {
-		if a != nil {
-			abstractionsCacheSingleton.Put(a)
+	log := func(ctx context.Context) Logger {
+		log := getLogger(ctx).WithField("to_guid", to.Guid).
+			WithField("fs", fs).
+			WithField("to", to.RelName)
+		if from != nil {
+			log = log.WithField("from", from.RelName).WithField("from_guid", from.Guid)
 		}
+		return log
 	}
-	keep := func(a Abstraction) (keep bool) {
-		keep = false
-		for _, k := range liveAbs {
-			keep = keep || AbstractionEquals(a, k)
+
+	toReplicationCursor, err := CreateReplicationCursor(ctx, fs, to, p.jobId)
+	if err != nil {
+		if err == zfs.ErrBookmarkCloningNotSupported {
+			log(ctx).Debug("not setting replication cursor, bookmark cloning not supported")
+		} else {
+			msg := "cannot move replication cursor, keeping hold on `to` until successful"
+			log(ctx).WithError(err).Error(msg)
+			err = errors.Wrap(err, msg)
+			// it is correct to not destroy from and to step holds if we can't move the cursor!
+			return &pdu.SendCompletedRes{}, err
 		}
-		return keep
+	} else {
+		log(ctx).WithField("to_cursor", toReplicationCursor.String()).Info("successfully created `to` replication cursor")
 	}
-	destroyTypes := AbstractionTypeSet{
-		AbstractionStepHold:                           true,
-		AbstractionTentativeReplicationCursorBookmark: true,
-		AbstractionReplicationCursorBookmarkV2:        true,
+
+	keep := func(a Abstraction) bool {
+		return AbstractionEquals(a, toReplicationCursor)
 	}
-	abstractionsCacheSingleton.TryBatchDestroy(ctx, p.jobId, fs, destroyTypes, keep, nil)
+	sendAbstractionsCacheSingleton.TryBatchDestroy(ctx, p.jobId, fs, keep, nil)

 	return &pdu.SendCompletedRes{}, nil

@@ -427,6 +435,8 @@ type ReceiverConfig struct {

 	RootWithoutClientComponent *zfs.DatasetPath // TODO use
 	AppendClientIdentity       bool
+
+	UpdateLastReceivedHold bool
 }

 func (c *ReceiverConfig) copyIn() {
@@ -667,7 +677,7 @@ func (s *Receiver) Receive(ctx context.Context, req *pdu.ReceiveReq, receive io.
 		f := zfs.NewDatasetPathForest()
 		f.Add(lp)
 		getLogger(ctx).Debug("begin tree-walk")
-		f.WalkTopDown(func(v *zfs.DatasetPathVisit) (visitChildTree bool) {
+		f.WalkTopDown(func(v zfs.DatasetPathVisit) (visitChildTree bool) {
 			if v.Path.Equal(lp) {
 				return false
 			}
@@ -695,7 +705,7 @@ func (s *Receiver) Receive(ctx context.Context, req *pdu.ReceiveReq, receive io.
 				}
 				l := getLogger(ctx).WithField("placeholder_fs", v.Path)
 				l.Debug("create placeholder filesystem")
-				err := zfs.ZFSCreatePlaceholderFilesystem(ctx, v.Path, v.Parent.Path)
+				err := zfs.ZFSCreatePlaceholderFilesystem(ctx, v.Path)
 				if err != nil {
 					l.WithError(err).Error("cannot create placeholder filesystem")
 					visitErr = err
@@ -851,38 +861,12 @@ func (s *Receiver) Receive(ctx context.Context, req *pdu.ReceiveReq, receive io.
 		return nil, errors.Wrap(err, msg)
 	}

-	replicationGuaranteeOptions, err := replicationGuaranteeOptionsFromPDU(req.GetReplicationConfig().Protection)
-	if err != nil {
-		return nil, err
-	}
-	replicationGuaranteeStrategy := replicationGuaranteeOptions.Strategy(ph.FSExists)
-	liveAbs, err := replicationGuaranteeStrategy.ReceiverPostRecv(ctx, s.conf.JobID, lp.ToString(), toRecvd)
-	if err != nil {
-		return nil, err
-	}
-	for _, a := range liveAbs {
-		if a != nil {
-			abstractionsCacheSingleton.Put(a)
+	if s.conf.UpdateLastReceivedHold {
+		log.Debug("move last-received-hold")
+		if err := MoveLastReceivedHold(ctx, lp.ToString(), toRecvd, s.conf.JobID); err != nil {
+			return nil, errors.Wrap(err, "cannot move last-received-hold")
 		}
 	}
-	keep := func(a Abstraction) (keep bool) {
-		keep = false
-		for _, k := range liveAbs {
-			keep = keep || AbstractionEquals(a, k)
-		}
-		return keep
-	}
-	check := func(obsoleteAbs []Abstraction) {
-		for _, abs := range obsoleteAbs {
-			if zfs.FilesystemVersionEqualIdentity(abs.GetFilesystemVersion(), toRecvd) {
-				panic(fmt.Sprintf("would destroy endpoint abstraction around the filesystem version we just received %s", abs))
-			}
-		}
-	}
-	destroyTypes := AbstractionTypeSet{
-		AbstractionLastReceivedHold: true,
-	}
-	abstractionsCacheSingleton.TryBatchDestroy(ctx, s.conf.JobID, lp.ToString(), destroyTypes, keep, check)

 	return &pdu.ReceiveRes{}, nil
 }
@@ -1,209 +0,0 @@
-package endpoint
-
-import (
-	"context"
-	"fmt"
-	"sync"
-
-	"github.com/prometheus/client_golang/prometheus"
-
-	"github.com/zrepl/zrepl/daemon/logging/trace"
-	"github.com/zrepl/zrepl/util/chainlock"
-)
-
-var abstractionsCacheMetrics struct {
-	count prometheus.Gauge
-}
-
-func init() {
-	abstractionsCacheMetrics.count = prometheus.NewGauge(prometheus.GaugeOpts{
-		Namespace: "zrepl",
-		Subsystem: "endpoint",
-		Name:      "abstractions_cache_entry_count",
-		Help:      "number of abstractions tracked in the abstractionsCache data structure",
-	})
-}
-
-var abstractionsCacheSingleton = newAbstractionsCache()
-
-func AbstractionsCacheInvalidate(fs string) {
-	abstractionsCacheSingleton.InvalidateFSCache(fs)
-}
-
-type abstractionsCacheDidLoadFSState int
-
-const (
-	abstractionsCacheDidLoadFSStateNo abstractionsCacheDidLoadFSState = iota // 0-value has meaning
-	abstractionsCacheDidLoadFSStateInProgress
-	abstractionsCacheDidLoadFSStateDone
-)
-
-type abstractionsCache struct {
-	mtx              chainlock.L
-	abstractions     []Abstraction
-	didLoadFS        map[string]abstractionsCacheDidLoadFSState
-	didLoadFSChanged *sync.Cond
-}
-
-func newAbstractionsCache() *abstractionsCache {
-	c := &abstractionsCache{
-		didLoadFS: make(map[string]abstractionsCacheDidLoadFSState),
-	}
-	c.didLoadFSChanged = c.mtx.NewCond()
-	return c
-}
-
-func (s *abstractionsCache) Put(a Abstraction) {
-	defer s.mtx.Lock().Unlock()
-
-	var zeroJobId JobID
-	if a.GetJobID() == nil {
-		panic("abstraction must not have nil job id")
-	} else if *a.GetJobID() == zeroJobId {
-		panic(fmt.Sprintf("abstraction must not have zero-value job id: %s", a))
-	}
-
-	s.abstractions = append(s.abstractions, a)
-	abstractionsCacheMetrics.count.Set(float64(len(s.abstractions)))
-}
-
-func (s *abstractionsCache) InvalidateFSCache(fs string) {
-	// FIXME: O(n)
-	newAbs := make([]Abstraction, 0, len(s.abstractions))
-	for _, a := range s.abstractions {
-		if a.GetFS() != fs {
-			newAbs = append(newAbs, a)
-		}
-	}
-	s.abstractions = newAbs
-	abstractionsCacheMetrics.count.Set(float64(len(s.abstractions)))
-
-	s.didLoadFS[fs] = abstractionsCacheDidLoadFSStateNo
-	s.didLoadFSChanged.Broadcast()
-
-}
-
-// - logs errors in getting on-disk abstractions
-// - only fetches on-disk abstractions once, but every time from the in-memory store
-//
-// That means that for precise results, all abstractions created by the endpoint must be .Put into this cache.
-func (s *abstractionsCache) GetAndDeleteByJobIDAndFS(ctx context.Context, jobID JobID, fs string, types AbstractionTypeSet, keep func(a Abstraction) bool) (ret []Abstraction) {
-	defer s.mtx.Lock().Unlock()
-	defer trace.WithSpanFromStackUpdateCtx(&ctx)()
-	var zeroJobId JobID
-	if jobID == zeroJobId {
-		panic("must not pass zero-value job id")
-	}
-	if fs == "" {
-		panic("must not pass zero-value fs")
-	}
-
-	s.tryLoadOnDiskAbstractions(ctx, fs)
-
-	// FIXME O(n)
-	var remaining []Abstraction
-	for _, a := range s.abstractions {
-		aJobId := *a.GetJobID()
-		aFS := a.GetFS()
-		if aJobId == jobID && aFS == fs && types[a.GetType()] && !keep(a) {
-			ret = append(ret, a)
-		} else {
-			remaining = append(remaining, a)
-		}
-	}
-	s.abstractions = remaining
-	abstractionsCacheMetrics.count.Set(float64(len(s.abstractions)))
-
-	return ret
-}
-
-// caller must hold s.mtx
-func (s *abstractionsCache) tryLoadOnDiskAbstractions(ctx context.Context, fs string) {
-	for s.didLoadFS[fs] != abstractionsCacheDidLoadFSStateDone {
-		if s.didLoadFS[fs] == abstractionsCacheDidLoadFSStateInProgress {
-			s.didLoadFSChanged.Wait()
-			continue
-		}
-		if s.didLoadFS[fs] != abstractionsCacheDidLoadFSStateNo {
-			panic(fmt.Sprintf("unreachable: %v", s.didLoadFS[fs]))
-		}
-
-		s.didLoadFS[fs] = abstractionsCacheDidLoadFSStateInProgress
-		defer s.didLoadFSChanged.Broadcast()
-
-		var onDiskAbs []Abstraction
-		var err error
-		s.mtx.DropWhile(func() {
-			onDiskAbs, err = s.tryLoadOnDiskAbstractionsImpl(ctx, fs) // no shadow
-		})
-
-		if err != nil {
-			s.didLoadFS[fs] = abstractionsCacheDidLoadFSStateNo
-			getLogger(ctx).WithField("fs", fs).WithError(err).Error("cannot list abstractions for filesystem")
-		} else {
-			s.didLoadFS[fs] = abstractionsCacheDidLoadFSStateDone
-			s.abstractions = append(s.abstractions, onDiskAbs...)
-			getLogger(ctx).WithField("fs", fs).WithField("abstractions", onDiskAbs).Debug("loaded step abstractions for filesystem")
-		}
-		return
-	}
-}
-
-// caller should _not hold s.mtx
-func (s *abstractionsCache) tryLoadOnDiskAbstractionsImpl(ctx context.Context, fs string) ([]Abstraction, error) {
-	defer trace.WithSpanFromStackUpdateCtx(&ctx)()
-
-	q := ListZFSHoldsAndBookmarksQuery{
-		FS: ListZFSHoldsAndBookmarksQueryFilesystemFilter{
-			FS: &fs,
-		},
-		JobID: nil,
-		What: AbstractionTypeSet{
-			AbstractionStepHold:                           true,
-			AbstractionTentativeReplicationCursorBookmark: true,
-			AbstractionReplicationCursorBookmarkV2:        true,
-			AbstractionLastReceivedHold:                   true,
-		},
-		Concurrency: 1,
-	}
-	abs, absErrs, err := ListAbstractions(ctx, q)
-	if err != nil {
-		return nil, err
-	}
-	// safe to ignore absErrs here, this is best-effort cleanup
-	if len(absErrs) > 0 {
-		return nil, ListAbstractionsErrors(absErrs)
-	}
-	return abs, nil
-}
-
-func (s *abstractionsCache) TryBatchDestroy(ctx context.Context, jobId JobID, fs string, types AbstractionTypeSet, keep func(a Abstraction) bool, check func(willDestroy []Abstraction)) {
-	// no s.mtx, we only use the public interface in this function
-
-	defer trace.WithSpanFromStackUpdateCtx(&ctx)()
-
-	obsoleteAbs := s.GetAndDeleteByJobIDAndFS(ctx, jobId, fs, types, keep)
-
-	if check != nil {
-		check(obsoleteAbs)
-	}
-
-	hadErr := false
-	for res := range BatchDestroy(ctx, obsoleteAbs) {
-		if res.DestroyErr != nil {
-			hadErr = true
-			getLogger(ctx).
-				WithField("abstraction", res.Abstraction).
-				WithError(res.DestroyErr).
-				Error("cannot destroy abstraction")
-		} else {
-			getLogger(ctx).
-				WithField("abstraction", res.Abstraction).
-				Info("destroyed abstraction")
-		}
-	}
-	if hadErr {
-		s.InvalidateFSCache(fs)
-	}
-
-}
@@ -1,214 +0,0 @@
-package endpoint
-
-import (
-	"context"
-	"fmt"
-
-	"github.com/pkg/errors"
-
-	"github.com/zrepl/zrepl/replication/logic/pdu"
-	"github.com/zrepl/zrepl/zfs"
-)
-
-type ReplicationGuaranteeOptions struct {
-	Initial     ReplicationGuaranteeKind
-	Incremental ReplicationGuaranteeKind
-}
-
-func replicationGuaranteeOptionsFromPDU(in *pdu.ReplicationConfigProtection) (o ReplicationGuaranteeOptions, _ error) {
-	if in == nil {
-		return o, errors.New("pdu.ReplicationConfigProtection must not be nil")
-	}
-	initial, err := replicationGuaranteeKindFromPDU(in.GetInitial())
-	if err != nil {
-		return o, errors.Wrap(err, "pdu.ReplicationConfigProtection: field Initial")
-	}
-	incremental, err := replicationGuaranteeKindFromPDU(in.GetIncremental())
-	if err != nil {
-		return o, errors.Wrap(err, "pdu.ReplicationConfigProtection: field Incremental")
-	}
-	o = ReplicationGuaranteeOptions{
-		Initial:     initial,
-		Incremental: incremental,
-	}
-	return o, nil
-}
-
-func replicationGuaranteeKindFromPDU(in pdu.ReplicationGuaranteeKind) (k ReplicationGuaranteeKind, _ error) {
-	switch in {
-	case pdu.ReplicationGuaranteeKind_GuaranteeNothing:
-		return ReplicationGuaranteeKindNone, nil
-	case pdu.ReplicationGuaranteeKind_GuaranteeIncrementalReplication:
-		return ReplicationGuaranteeKindIncremental, nil
-	case pdu.ReplicationGuaranteeKind_GuaranteeResumability:
-		return ReplicationGuaranteeKindResumability, nil
-
-	case pdu.ReplicationGuaranteeKind_GuaranteeInvalid:
-		fallthrough
-	default:
-		return k, errors.Errorf("%q", in.String())
-	}
-}
-
-func (o ReplicationGuaranteeOptions) Strategy(incremental bool) ReplicationGuaranteeStrategy {
-	g := o.Initial
-	if incremental {
-		g = o.Incremental
-	}
-	return ReplicationGuaranteeFromKind(g)
-}
-
-//go:generate enumer -type=ReplicationGuaranteeKind -json -transform=snake -trimprefix=ReplicationGuaranteeKind
-type ReplicationGuaranteeKind int
-
-const (
-	ReplicationGuaranteeKindResumability ReplicationGuaranteeKind = 1 << iota
-	ReplicationGuaranteeKindIncremental
-	ReplicationGuaranteeKindNone
-)
-
-type ReplicationGuaranteeStrategy interface {
-	Kind() ReplicationGuaranteeKind
-	SenderPreSend(ctx context.Context, jid JobID, sendArgs *zfs.ZFSSendArgsValidated) (keep []Abstraction, err error)
-	ReceiverPostRecv(ctx context.Context, jid JobID, fs string, toRecvd zfs.FilesystemVersion) (keep []Abstraction, err error)
-	SenderPostRecvConfirmed(ctx context.Context, jid JobID, fs string, to zfs.FilesystemVersion) (keep []Abstraction, err error)
-}
-
-func ReplicationGuaranteeFromKind(k ReplicationGuaranteeKind) ReplicationGuaranteeStrategy {
-	switch k {
-	case ReplicationGuaranteeKindNone:
-		return ReplicationGuaranteeNone{}
-	case ReplicationGuaranteeKindIncremental:
-		return ReplicationGuaranteeIncremental{}
-	case ReplicationGuaranteeKindResumability:
-		return ReplicationGuaranteeResumability{}
-	default:
-		panic(fmt.Sprintf("unreachable: %q %T", k, k))
-	}
-}
-
-type ReplicationGuaranteeNone struct{}
-
-func (g ReplicationGuaranteeNone) Kind() ReplicationGuaranteeKind {
-	return ReplicationGuaranteeKindNone
-}
-
-func (g ReplicationGuaranteeNone) SenderPreSend(ctx context.Context, jid JobID, sendArgs *zfs.ZFSSendArgsValidated) (keep []Abstraction, err error) {
-	return nil, nil
-}
-
-func (g ReplicationGuaranteeNone) ReceiverPostRecv(ctx context.Context, jid JobID, fs string, toRecvd zfs.FilesystemVersion) (keep []Abstraction, err error) {
-	return nil, nil
-}
-
-func (g ReplicationGuaranteeNone) SenderPostRecvConfirmed(ctx context.Context, jid JobID, fs string, to zfs.FilesystemVersion) (keep []Abstraction, err error) {
-	return nil, nil
-}
-
-type ReplicationGuaranteeIncremental struct{}
-
-func (g ReplicationGuaranteeIncremental) Kind() ReplicationGuaranteeKind {
-	return ReplicationGuaranteeKindIncremental
-}
-
-func (g ReplicationGuaranteeIncremental) SenderPreSend(ctx context.Context, jid JobID, sendArgs *zfs.ZFSSendArgsValidated) (keep []Abstraction, err error) {
-	if sendArgs.FromVersion != nil {
-		from, err := CreateTentativeReplicationCursor(ctx, sendArgs.FS, *sendArgs.FromVersion, jid)
-		if err != nil {
-			if err == zfs.ErrBookmarkCloningNotSupported {
-				getLogger(ctx).WithField("replication_guarantee", g).
-					WithField("bookmark", sendArgs.From.FullPath(sendArgs.FS)).
-					Info("bookmark cloning is not supported, speculating that `from` will not be destroyed until step is done")
-			} else {
-				return nil, err
-			}
-		}
-		keep = append(keep, from)
-	}
-	to, err := CreateTentativeReplicationCursor(ctx, sendArgs.FS, sendArgs.ToVersion, jid)
-	if err != nil {
-		return nil, err
-	}
-	keep = append(keep, to)
-
-	return keep, nil
-}
-
-func (g ReplicationGuaranteeIncremental) ReceiverPostRecv(ctx context.Context, jid JobID, fs string, toRecvd zfs.FilesystemVersion) (keep []Abstraction, err error) {
-	return receiverPostRecvCommon(ctx, jid, fs, toRecvd)
-}
-
-func (g ReplicationGuaranteeIncremental) SenderPostRecvConfirmed(ctx context.Context, jid JobID, fs string, to zfs.FilesystemVersion) (keep []Abstraction, err error) {
-	return senderPostRecvConfirmedCommon(ctx, jid, fs, to)
-}
-
-type ReplicationGuaranteeResumability struct{}
-
-func (g ReplicationGuaranteeResumability) Kind() ReplicationGuaranteeKind {
-	return ReplicationGuaranteeKindResumability
-}
-
-func (g ReplicationGuaranteeResumability) SenderPreSend(ctx context.Context, jid JobID, sendArgs *zfs.ZFSSendArgsValidated) (keep []Abstraction, err error) {
-	// try to hold the FromVersion
-	if sendArgs.FromVersion != nil {
-		if sendArgs.FromVersion.Type == zfs.Bookmark {
-			getLogger(ctx).WithField("replication_guarantee", g).WithField("fromVersion", sendArgs.FromVersion.FullPath(sendArgs.FS)).
-				Debug("cannot hold a bookmark, speculating that `from` will not be destroyed until step is done")
-		} else {
-			from, err := HoldStep(ctx, sendArgs.FS, *sendArgs.FromVersion, jid)
-			if err != nil {
-				return nil, err
-			}
-			keep = append(keep, from)
-		}
-		// fallthrough
-	}
-
-	to, err := HoldStep(ctx, sendArgs.FS, sendArgs.ToVersion, jid)
-	if err != nil {
-		return nil, err
-	}
-	keep = append(keep, to)
-
-	return keep, nil
-}
-
-func (g ReplicationGuaranteeResumability) ReceiverPostRecv(ctx context.Context, jid JobID, fs string, toRecvd zfs.FilesystemVersion) (keep []Abstraction, err error) {
-	return receiverPostRecvCommon(ctx, jid, fs, toRecvd)
-}
-
-func (g ReplicationGuaranteeResumability) SenderPostRecvConfirmed(ctx context.Context, jid JobID, fs string, to zfs.FilesystemVersion) (keep []Abstraction, err error) {
-	return senderPostRecvConfirmedCommon(ctx, jid, fs, to)
-}
-
-// helper function used by multiple strategies
-func senderPostRecvConfirmedCommon(ctx context.Context, jid JobID, fs string, to zfs.FilesystemVersion) (keep []Abstraction, err error) {
-
-	log := getLogger(ctx).WithField("toVersion", to.FullPath(fs))
-
-	toReplicationCursor, err := CreateReplicationCursor(ctx, fs, to, jid)
-	if err != nil {
-		if err == zfs.ErrBookmarkCloningNotSupported {
-			log.Debug("not setting replication cursor, bookmark cloning not supported")
-		} else {
-			msg := "cannot move replication cursor, keeping hold on `to` until successful"
-			log.WithError(err).Error(msg)
-			err = errors.Wrap(err, msg)
-			return nil, err
-		}
-	} else {
-		log.WithField("to_cursor", toReplicationCursor.String()).Info("successfully created `to` replication cursor")
-	}
-
-	return []Abstraction{toReplicationCursor}, nil
-}
-
-// helper function used by multiple strategies
-func receiverPostRecvCommon(ctx context.Context, jid JobID, fs string, toRecvd zfs.FilesystemVersion) (keep []Abstraction, err error) {
-	getLogger(ctx).Debug("create new last-received-hold")
-	lrh, err := CreateLastReceivedHold(ctx, fs, toRecvd, jid)
-	if err != nil {
-		return nil, err
-	}
-	return []Abstraction{lrh}, nil
-}
@@ -3,5 +3,5 @@ package endpoint
 import "github.com/prometheus/client_golang/prometheus"

 func RegisterMetrics(r prometheus.Registerer) {
-	r.MustRegister(abstractionsCacheMetrics.count)
+	r.MustRegister(sendAbstractionsCacheMetrics.count)
 }
@@ -0,0 +1,232 @@
+package endpoint
+
+import (
+	"context"
+	"fmt"
+	"sync"
+
+	"github.com/prometheus/client_golang/prometheus"
+	"github.com/zrepl/zrepl/daemon/logging/trace"
+	"github.com/zrepl/zrepl/util/chainlock"
+)
+
+var sendAbstractionsCacheMetrics struct {
+	count prometheus.Gauge
+}
+
+func init() {
+	sendAbstractionsCacheMetrics.count = prometheus.NewGauge(prometheus.GaugeOpts{
+		Namespace: "zrepl",
+		Subsystem: "endpoint",
+		Name:      "send_abstractions_cache_entry_count",
+		Help:      "number of send abstractions tracked in the sendAbstractionsCache data structure",
+	})
+}
+
+var sendAbstractionsCacheSingleton = newSendAbstractionsCache()
+
+func SendAbstractionsCacheInvalidate(fs string) {
+	sendAbstractionsCacheSingleton.InvalidateFSCache(fs)
+}
+
+type sendAbstractionsCacheDidLoadFSState int
+
+const (
+	sendAbstractionsCacheDidLoadFSStateNo sendAbstractionsCacheDidLoadFSState = iota // 0-value has meaning
+	sendAbstractionsCacheDidLoadFSStateInProgress
+	sendAbstractionsCacheDidLoadFSStateDone
+)
+
+type sendAbstractionsCache struct {
+	mtx              chainlock.L
+	abstractions     []Abstraction
+	didLoadFS        map[string]sendAbstractionsCacheDidLoadFSState
+	didLoadFSChanged *sync.Cond
+}
+
+func newSendAbstractionsCache() *sendAbstractionsCache {
+	c := &sendAbstractionsCache{
+		didLoadFS: make(map[string]sendAbstractionsCacheDidLoadFSState),
+	}
+	c.didLoadFSChanged = c.mtx.NewCond()
+	return c
+}
+
+func (s *sendAbstractionsCache) Put(a Abstraction) {
+	defer s.mtx.Lock().Unlock()
+
+	var zeroJobId JobID
+	if a.GetJobID() == nil {
+		panic("abstraction must not have nil job id")
+	} else if *a.GetJobID() == zeroJobId {
+		panic(fmt.Sprintf("abstraction must not have zero-value job id: %s", a))
+	}
+
+	s.abstractions = append(s.abstractions, a)
+	sendAbstractionsCacheMetrics.count.Set(float64(len(s.abstractions)))
+}
+
+func (s *sendAbstractionsCache) InvalidateFSCache(fs string) {
+	// FIXME: O(n)
+	newAbs := make([]Abstraction, 0, len(s.abstractions))
+	for _, a := range s.abstractions {
+		if a.GetFS() != fs {
+			newAbs = append(newAbs, a)
+		}
+	}
+	s.abstractions = newAbs
+	sendAbstractionsCacheMetrics.count.Set(float64(len(s.abstractions)))
+
+	s.didLoadFS[fs] = sendAbstractionsCacheDidLoadFSStateNo
+	s.didLoadFSChanged.Broadcast()
+
+}
+
+func (s *sendAbstractionsCache) GetByFS(ctx context.Context, fs string) (ret []Abstraction) {
+	defer s.mtx.Lock().Unlock()
+	defer trace.WithSpanFromStackUpdateCtx(&ctx)()
+	if fs == "" {
+		panic("must not pass zero-value fs")
+	}
+
+	s.tryLoadOnDiskSendAbstractions(ctx, fs)
+
+	for _, a := range s.abstractions {
+		if a.GetFS() == fs {
+			ret = append(ret, a)
+		}
+	}
+
+	return ret
+}
+
+// - logs errors in getting on-disk abstractions
+// - only fetches on-disk abstractions once, but every time from the in-memory store
+//
+// That means that for precise results, all abstractions created by the endpoint must be .Put into this cache.
+func (s *sendAbstractionsCache) GetAndDeleteByJobIDAndFS(ctx context.Context, jobID JobID, fs string) (ret []Abstraction) {
+	defer s.mtx.Lock().Unlock()
+	defer trace.WithSpanFromStackUpdateCtx(&ctx)()
+	var zeroJobId JobID
+	if jobID == zeroJobId {
+		panic("must not pass zero-value job id")
+	}
+	if fs == "" {
+		panic("must not pass zero-value fs")
+	}
+
+	s.tryLoadOnDiskSendAbstractions(ctx, fs)
+
+	// FIXME O(n)
+	var remaining []Abstraction
+	for _, a := range s.abstractions {
+		aJobId := *a.GetJobID()
+		aFS := a.GetFS()
+		if aJobId == jobID && aFS == fs {
+			ret = append(ret, a)
+		} else {
+			remaining = append(remaining, a)
+		}
+	}
+	s.abstractions = remaining
+	sendAbstractionsCacheMetrics.count.Set(float64(len(s.abstractions)))
+
+	return ret
+}
+
+// caller must hold s.mtx
+func (s *sendAbstractionsCache) tryLoadOnDiskSendAbstractions(ctx context.Context, fs string) {
+	for s.didLoadFS[fs] != sendAbstractionsCacheDidLoadFSStateDone {
+		if s.didLoadFS[fs] == sendAbstractionsCacheDidLoadFSStateInProgress {
+			s.didLoadFSChanged.Wait()
+			continue
+		}
+		if s.didLoadFS[fs] != sendAbstractionsCacheDidLoadFSStateNo {
+			panic(fmt.Sprintf("unreachable: %v", s.didLoadFS[fs]))
+		}
+
+		s.didLoadFS[fs] = sendAbstractionsCacheDidLoadFSStateInProgress
+		defer s.didLoadFSChanged.Broadcast()
+
+		var onDiskAbs []Abstraction
+		var err error
+		s.mtx.DropWhile(func() {
+			onDiskAbs, err = s.tryLoadOnDiskSendAbstractionsImpl(ctx, fs) // no shadow
+		})
+
+		if err != nil {
+			s.didLoadFS[fs] = sendAbstractionsCacheDidLoadFSStateNo
+			getLogger(ctx).WithField("fs", fs).WithError(err).Error("cannot list send step abstractions for filesystem")
+		} else {
+			s.didLoadFS[fs] = sendAbstractionsCacheDidLoadFSStateDone
+			s.abstractions = append(s.abstractions, onDiskAbs...)
+			getLogger(ctx).WithField("fs", fs).WithField("abstractions", onDiskAbs).Debug("loaded step abstractions for filesystem")
+		}
+		return
+	}
+}
+
+// caller should _not hold s.mtx
+func (s *sendAbstractionsCache) tryLoadOnDiskSendAbstractionsImpl(ctx context.Context, fs string) ([]Abstraction, error) {
+	defer trace.WithSpanFromStackUpdateCtx(&ctx)()
+
+	q := ListZFSHoldsAndBookmarksQuery{
+		FS: ListZFSHoldsAndBookmarksQueryFilesystemFilter{
+			FS: &fs,
+		},
+		JobID: nil,
+		What: AbstractionTypeSet{
+			AbstractionStepHold:                    true,
+			AbstractionStepBookmark:                true,
+			AbstractionReplicationCursorBookmarkV2: true,
+		},
+		Concurrency: 1,
+	}
+	abs, absErrs, err := ListAbstractions(ctx, q)
+	if err != nil {
+		return nil, err
+	}
+	// safe to ignore absErrs here, this is best-effort cleanup
+	if len(absErrs) > 0 {
+		return nil, ListAbstractionsErrors(absErrs)
+	}
+	return abs, nil
+}
+
+func (s *sendAbstractionsCache) TryBatchDestroy(ctx context.Context, jobId JobID, fs string, keep func(a Abstraction) bool, check func(willDestroy []Abstraction)) {
+	// no s.mtx, we only use the public interface in this function
+
+	defer trace.WithSpanFromStackUpdateCtx(&ctx)()
+
+	allSendStepAbstractions := s.GetAndDeleteByJobIDAndFS(ctx, jobId, fs)
+
+	var obsoleteAbs []Abstraction
+	for _, a := range allSendStepAbstractions {
+		if !keep(a) {
+			obsoleteAbs = append(obsoleteAbs, a)
+		}
+	}
+
+	if check != nil {
+		check(obsoleteAbs)
+	}
+
+	hadErr := false
+	for res := range BatchDestroy(ctx, obsoleteAbs) {
+		if res.DestroyErr != nil {
+			hadErr = true
+			getLogger(ctx).
+				WithField("abstraction", res.Abstraction).
+				WithError(res.DestroyErr).
+				Error("cannot destroy stale send step abstraction")
+		} else {
+			getLogger(ctx).
+				WithField("abstraction", res.Abstraction).
+				Info("destroyed stale send step abstraction")
+		}
+	}
+	if hadErr {
+		s.InvalidateFSCache(fs)
+	}
+
+}
@@ -23,19 +23,19 @@ type AbstractionType string
 // There are a lot of exhaustive switches on AbstractionType in the code base.
 // When adding a new abstraction type, make sure to search and update them!
 const (
-	AbstractionStepHold                           AbstractionType = "step-hold"
-	AbstractionLastReceivedHold                   AbstractionType = "last-received-hold"
-	AbstractionTentativeReplicationCursorBookmark AbstractionType = "tentative-replication-cursor-bookmark-v2"
-	AbstractionReplicationCursorBookmarkV1        AbstractionType = "replication-cursor-bookmark-v1"
-	AbstractionReplicationCursorBookmarkV2        AbstractionType = "replication-cursor-bookmark-v2"
+	AbstractionStepBookmark                AbstractionType = "step-bookmark"
+	AbstractionStepHold                    AbstractionType = "step-hold"
+	AbstractionLastReceivedHold            AbstractionType = "last-received-hold"
+	AbstractionReplicationCursorBookmarkV1 AbstractionType = "replication-cursor-bookmark-v1"
+	AbstractionReplicationCursorBookmarkV2 AbstractionType = "replication-cursor-bookmark-v2"
 )

 var AbstractionTypesAll = map[AbstractionType]bool{
-	AbstractionStepHold:                           true,
-	AbstractionLastReceivedHold:                   true,
-	AbstractionTentativeReplicationCursorBookmark: true,
-	AbstractionReplicationCursorBookmarkV1:        true,
-	AbstractionReplicationCursorBookmarkV2:        true,
+	AbstractionStepBookmark:                true,
+	AbstractionStepHold:                    true,
+	AbstractionLastReceivedHold:            true,
+	AbstractionReplicationCursorBookmarkV1: true,
+	AbstractionReplicationCursorBookmarkV2: true,
 }

 // Implementation Note:
@@ -80,12 +80,12 @@ func AbstractionEquals(a, b Abstraction) bool {

 func (t AbstractionType) Validate() error {
 	switch t {
+	case AbstractionStepBookmark:
+		return nil
 	case AbstractionStepHold:
 		return nil
 	case AbstractionLastReceivedHold:
 		return nil
-	case AbstractionTentativeReplicationCursorBookmark:
-		return nil
 	case AbstractionReplicationCursorBookmarkV1:
 		return nil
 	case AbstractionReplicationCursorBookmarkV2:
@@ -185,8 +185,8 @@ type BookmarkExtractor func(fs *zfs.DatasetPath, v zfs.FilesystemVersion) Abstra
 // returns nil if the abstraction type is not bookmark-based
 func (t AbstractionType) BookmarkExtractor() BookmarkExtractor {
 	switch t {
-	case AbstractionTentativeReplicationCursorBookmark:
-		return TentativeReplicationCursorExtractor
+	case AbstractionStepBookmark:
+		return StepBookmarkExtractor
 	case AbstractionReplicationCursorBookmarkV1:
 		return ReplicationCursorV1Extractor
 	case AbstractionReplicationCursorBookmarkV2:
@@ -205,7 +205,7 @@ type HoldExtractor = func(fs *zfs.DatasetPath, v zfs.FilesystemVersion, tag stri
 // returns nil if the abstraction type is not hold-based
 func (t AbstractionType) HoldExtractor() HoldExtractor {
 	switch t {
-	case AbstractionTentativeReplicationCursorBookmark:
+	case AbstractionStepBookmark:
 		return nil
 	case AbstractionReplicationCursorBookmarkV1:
 		return nil
@@ -220,23 +220,6 @@ func (t AbstractionType) HoldExtractor() HoldExtractor {
 	}
 }

-func (t AbstractionType) BookmarkNamer() func(fs string, guid uint64, jobId JobID) (string, error) {
-	switch t {
-	case AbstractionTentativeReplicationCursorBookmark:
-		return TentativeReplicationCursorBookmarkName
-	case AbstractionReplicationCursorBookmarkV1:
-		panic("shouldn't be creating new ones")
-	case AbstractionReplicationCursorBookmarkV2:
-		return ReplicationCursorBookmarkName
-	case AbstractionStepHold:
-		return nil
-	case AbstractionLastReceivedHold:
-		return nil
-	default:
-		panic(fmt.Sprintf("unimpl: %q", t))
-	}
-}
-
 type ListZFSHoldsAndBookmarksQuery struct {
 	FS ListZFSHoldsAndBookmarksQueryFilesystemFilter
 	// What abstraction types should match (any contained in the set)
@@ -714,9 +697,11 @@ func ListStale(ctx context.Context, q ListZFSHoldsAndBookmarksQuery) (*Staleness
 		return nil, &ListStaleQueryError{errors.New("ListStale cannot have Until != nil set on query")}
 	}

-	// if asking for step holds must also ask for replication cursor bookmarks (for firstNotStale)
+	// if asking for step holds, must also as for step bookmarks (same kind of abstraction)
+	// as well as replication cursor bookmarks (for firstNotStale)
 	ifAnyThenAll := AbstractionTypeSet{
 		AbstractionStepHold:                    true,
+		AbstractionStepBookmark:                true,
 		AbstractionReplicationCursorBookmarkV2: true,
 	}
 	if q.What.ContainsAnyOf(ifAnyThenAll) && !q.What.ContainsAll(ifAnyThenAll) {
@@ -745,7 +730,7 @@ type fsAjobAtype struct {
 }

 // For step holds and bookmarks, only those older than the most recent replication cursor
-// of their (filesystem,job) are considered because younger ones cannot be stale by definition
+// of their (filesystem,job) is considered because younger ones cannot be stale by definition
 // (if we destroy them, we might actually lose the hold on the `To` for an ongoing incremental replication)
 //
 // For replication cursors and last-received-holds, only the most recent one is kept.
@@ -787,6 +772,8 @@ func listStaleFiltering(abs []Abstraction, sinceBound *CreateTXGRangeBound) *Sta
 			}

 		// stepFirstNotStaleCandidate.step
+		case AbstractionStepBookmark:
+			fallthrough
 		case AbstractionStepHold:
 			if c.step == nil || (*c.step).GetCreateTXG() < a.GetCreateTXG() {
 				a := a
@@ -810,7 +797,7 @@ func listStaleFiltering(abs []Abstraction, sinceBound *CreateTXGRangeBound) *Sta
 	for k := range by {
 		l := by[k]

-		if k.Type == AbstractionStepHold {
+		if k.Type == AbstractionStepHold || k.Type == AbstractionStepBookmark {
 			// all older than the most recent cursor are stale, others are always live

 			// if we don't have a replication cursor yet, use untilBound = nil
@@ -4,10 +4,12 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
+	"regexp"
 	"sort"

 	"github.com/pkg/errors"

+	"github.com/zrepl/zrepl/util/errorarray"
 	"github.com/zrepl/zrepl/zfs"
 )

@@ -51,28 +53,6 @@ func ParseReplicationCursorBookmarkName(fullname string) (uint64, JobID, error)
 	return guid, jobID, err
 }

-const tentativeReplicationCursorBookmarkNamePrefix = "zrepl_CURSORTENTATIVE_"
-
-// v must be validated by caller
-func TentativeReplicationCursorBookmarkName(fs string, guid uint64, id JobID) (string, error) {
-	return tentativeReplicationCursorBookmarkNameImpl(fs, guid, id.String())
-}
-
-func tentativeReplicationCursorBookmarkNameImpl(fs string, guid uint64, jobid string) (string, error) {
-	return makeJobAndGuidBookmarkName(tentativeReplicationCursorBookmarkNamePrefix, fs, guid, jobid)
-}
-
-// name is the full bookmark name, including dataset path
-//
-// err != nil always means that the bookmark is not a step bookmark
-func ParseTentativeReplicationCursorBookmarkName(fullname string) (guid uint64, jobID JobID, err error) {
-	guid, jobID, err = parseJobAndGuidBookmarkName(fullname, tentativeReplicationCursorBookmarkNamePrefix)
-	if err != nil {
-		err = errors.Wrap(err, "parse step bookmark name") // no shadow!
-	}
-	return guid, jobID, err
-}
-
 // may return nil for both values, indicating there is no cursor
 func GetMostRecentReplicationCursorOfJob(ctx context.Context, fs string, jobID JobID) (*zfs.FilesystemVersion, error) {
 	fsp, err := zfs.NewDatasetPath(fs)
@@ -142,23 +122,23 @@ func GetReplicationCursors(ctx context.Context, dp *zfs.DatasetPath, jobID JobID
 //
 // returns ErrBookmarkCloningNotSupported if version is a bookmark and bookmarking bookmarks is not supported by ZFS
 func CreateReplicationCursor(ctx context.Context, fs string, target zfs.FilesystemVersion, jobID JobID) (a Abstraction, err error) {
-	return createBookmarkAbstraction(ctx, AbstractionReplicationCursorBookmarkV2, fs, target, jobID)
-}

-func CreateTentativeReplicationCursor(ctx context.Context, fs string, target zfs.FilesystemVersion, jobID JobID) (a Abstraction, err error) {
-	return createBookmarkAbstraction(ctx, AbstractionTentativeReplicationCursorBookmark, fs, target, jobID)
-}
-
-func createBookmarkAbstraction(ctx context.Context, abstractionType AbstractionType, fs string, target zfs.FilesystemVersion, jobID JobID) (a Abstraction, err error) {
-
-	bookmarkNamer := abstractionType.BookmarkNamer()
-	if bookmarkNamer == nil {
-		panic(abstractionType)
+	bookmarkname, err := ReplicationCursorBookmarkName(fs, target.GetGuid(), jobID)
+	if err != nil {
+		return nil, errors.Wrap(err, "determine replication cursor name")
 	}

-	bookmarkname, err := bookmarkNamer(fs, target.GetGuid(), jobID)
-	if err != nil {
-		return nil, errors.Wrapf(err, "determine %s name", abstractionType)
+	if target.IsBookmark() && target.GetName() == bookmarkname {
+		return &bookmarkBasedAbstraction{
+			Type:              AbstractionReplicationCursorBookmarkV2,
+			FS:                fs,
+			FilesystemVersion: target,
+			JobID:             jobID,
+		}, nil
+	}
+
+	if !target.IsSnapshot() {
+		return nil, zfs.ErrBookmarkCloningNotSupported
 	}

 	// idempotently create bookmark (guid is encoded in it)
@@ -172,13 +152,125 @@ func createBookmarkAbstraction(ctx context.Context, abstractionType AbstractionT
 	}

 	return &bookmarkBasedAbstraction{
-		Type:              abstractionType,
+		Type:              AbstractionReplicationCursorBookmarkV2,
 		FS:                fs,
 		FilesystemVersion: cursorBookmark,
 		JobID:             jobID,
 	}, nil
 }

+const (
+	ReplicationCursorBookmarkNamePrefix = "zrepl_last_received_J_"
+)
+
+var lastReceivedHoldTagRE = regexp.MustCompile("^zrepl_last_received_J_(.+)$")
+
+// err != nil always means that the bookmark is not a step bookmark
+func ParseLastReceivedHoldTag(tag string) (JobID, error) {
+	match := lastReceivedHoldTagRE.FindStringSubmatch(tag)
+	if match == nil {
+		return JobID{}, errors.Errorf("parse last-received-hold tag: does not match regex %s", lastReceivedHoldTagRE.String())
+	}
+	jobId, err := MakeJobID(match[1])
+	if err != nil {
+		return JobID{}, errors.Wrap(err, "parse last-received-hold tag: invalid job id field")
+	}
+	return jobId, nil
+}
+
+func LastReceivedHoldTag(jobID JobID) (string, error) {
+	return lastReceivedHoldImpl(jobID.String())
+}
+
+func lastReceivedHoldImpl(jobid string) (string, error) {
+	tag := fmt.Sprintf("%s%s", ReplicationCursorBookmarkNamePrefix, jobid)
+	if err := zfs.ValidHoldTag(tag); err != nil {
+		return "", err
+	}
+	return tag, nil
+}
+
+func CreateLastReceivedHold(ctx context.Context, fs string, to zfs.FilesystemVersion, jobID JobID) (Abstraction, error) {
+
+	if !to.IsSnapshot() {
+		return nil, errors.Errorf("last-received-hold: target must be a snapshot: %s", to.FullPath(fs))
+	}
+
+	tag, err := LastReceivedHoldTag(jobID)
+	if err != nil {
+		return nil, errors.Wrap(err, "last-received-hold: hold tag")
+	}
+
+	// we never want to be without a hold
+	// => hold new one before releasing old hold
+
+	err = zfs.ZFSHold(ctx, fs, to, tag)
+	if err != nil {
+		return nil, errors.Wrap(err, "last-received-hold: hold newly received")
+	}
+
+	return &holdBasedAbstraction{
+		Type:              AbstractionLastReceivedHold,
+		FS:                fs,
+		FilesystemVersion: to,
+		JobID:             jobID,
+		Tag:               tag,
+	}, nil
+}
+
+func MoveLastReceivedHold(ctx context.Context, fs string, to zfs.FilesystemVersion, jobID JobID) error {
+
+	_, err := CreateLastReceivedHold(ctx, fs, to, jobID)
+	if err != nil {
+		return err
+	}
+
+	q := ListZFSHoldsAndBookmarksQuery{
+		What: AbstractionTypeSet{
+			AbstractionLastReceivedHold: true,
+		},
+		FS: ListZFSHoldsAndBookmarksQueryFilesystemFilter{
+			FS: &fs,
+		},
+		JobID: &jobID,
+		CreateTXG: CreateTXGRange{
+			Since: nil,
+			Until: &CreateTXGRangeBound{
+				CreateTXG: to.GetCreateTXG(),
+				Inclusive: &zfs.NilBool{B: false},
+			},
+		},
+		Concurrency: 1,
+	}
+	abs, absErrs, err := ListAbstractions(ctx, q)
+	if err != nil {
+		return errors.Wrap(err, "last-received-hold: list")
+	}
+	if len(absErrs) > 0 {
+		return errors.Wrap(ListAbstractionsErrors(absErrs), "last-received-hold: list")
+	}
+
+	getLogger(ctx).WithField("last-received-holds", fmt.Sprintf("%s", abs)).Debug("releasing last-received-holds")
+
+	var errs []error
+	for res := range BatchDestroy(ctx, abs) {
+		log := getLogger(ctx).
+			WithField("last-received-hold", res.Abstraction)
+		if res.DestroyErr != nil {
+			errs = append(errs, res.DestroyErr)
+			log.WithError(err).
+				Error("cannot release last-received-hold")
+		} else {
+			log.Info("released last-received-hold")
+		}
+	}
+	if len(errs) == 0 {
+		return nil
+	} else {
+		return errorarray.Wrap(errs, "last-received-hold: release")
+	}
+}
+
 func ReplicationCursorV2Extractor(fs *zfs.DatasetPath, v zfs.FilesystemVersion) (_ Abstraction) {
 	if v.Type != zfs.Bookmark {
 		panic("impl error")
@@ -216,28 +308,24 @@ func ReplicationCursorV1Extractor(fs *zfs.DatasetPath, v zfs.FilesystemVersion)
 	return nil
 }

-var _ BookmarkExtractor = TentativeReplicationCursorExtractor
+var _ HoldExtractor = LastReceivedHoldExtractor

-func TentativeReplicationCursorExtractor(fs *zfs.DatasetPath, v zfs.FilesystemVersion) (_ Abstraction) {
-	if v.Type != zfs.Bookmark {
+func LastReceivedHoldExtractor(fs *zfs.DatasetPath, v zfs.FilesystemVersion, holdTag string) Abstraction {
+	var err error
+
+	if v.Type != zfs.Snapshot {
 		panic("impl error")
 	}

-	fullname := v.ToAbsPath(fs)
-
-	guid, jobid, err := ParseTentativeReplicationCursorBookmarkName(fullname)
-	if guid != v.Guid {
-		// TODO log this possibly tinkered-with bookmark
-		return nil
-	}
+	jobID, err := ParseLastReceivedHoldTag(holdTag)
 	if err == nil {
-		bm := &bookmarkBasedAbstraction{
-			Type:              AbstractionTentativeReplicationCursorBookmark,
+		return &holdBasedAbstraction{
+			Type:              AbstractionLastReceivedHold,
 			FS:                fs.ToString(),
 			FilesystemVersion: v,
-			JobID:             jobid,
+			Tag:               holdTag,
+			JobID:             jobID,
 		}
-		return bm
 	}
 	return nil
 }
@@ -1,92 +0,0 @@
-package endpoint
-
-import (
-	"context"
-	"fmt"
-	"regexp"
-
-	"github.com/pkg/errors"
-
-	"github.com/zrepl/zrepl/zfs"
-)
-
-const (
-	LastReceivedHoldTagNamePrefix = "zrepl_last_received_J_"
-)
-
-var lastReceivedHoldTagRE = regexp.MustCompile("^zrepl_last_received_J_(.+)$")
-
-var _ HoldExtractor = LastReceivedHoldExtractor
-
-func LastReceivedHoldExtractor(fs *zfs.DatasetPath, v zfs.FilesystemVersion, holdTag string) Abstraction {
-	var err error
-
-	if v.Type != zfs.Snapshot {
-		panic("impl error")
-	}
-
-	jobID, err := ParseLastReceivedHoldTag(holdTag)
-	if err == nil {
-		return &holdBasedAbstraction{
-			Type:              AbstractionLastReceivedHold,
-			FS:                fs.ToString(),
-			FilesystemVersion: v,
-			Tag:               holdTag,
-			JobID:             jobID,
-		}
-	}
-	return nil
-}
-
-// err != nil always means that the bookmark is not a step bookmark
-func ParseLastReceivedHoldTag(tag string) (JobID, error) {
-	match := lastReceivedHoldTagRE.FindStringSubmatch(tag)
-	if match == nil {
-		return JobID{}, errors.Errorf("parse last-received-hold tag: does not match regex %s", lastReceivedHoldTagRE.String())
-	}
-	jobId, err := MakeJobID(match[1])
-	if err != nil {
-		return JobID{}, errors.Wrap(err, "parse last-received-hold tag: invalid job id field")
-	}
-	return jobId, nil
-}
-
-func LastReceivedHoldTag(jobID JobID) (string, error) {
-	return lastReceivedHoldImpl(jobID.String())
-}
-
-func lastReceivedHoldImpl(jobid string) (string, error) {
-	tag := fmt.Sprintf("%s%s", LastReceivedHoldTagNamePrefix, jobid)
-	if err := zfs.ValidHoldTag(tag); err != nil {
-		return "", err
-	}
-	return tag, nil
-}
-
-func CreateLastReceivedHold(ctx context.Context, fs string, to zfs.FilesystemVersion, jobID JobID) (Abstraction, error) {
-
-	if !to.IsSnapshot() {
-		return nil, errors.Errorf("last-received-hold: target must be a snapshot: %s", to.FullPath(fs))
-	}
-
-	tag, err := LastReceivedHoldTag(jobID)
-	if err != nil {
-		return nil, errors.Wrap(err, "last-received-hold: hold tag")
-	}
-
-	// we never want to be without a hold
-	// => hold new one before releasing old hold
-
-	err = zfs.ZFSHold(ctx, fs, to, tag)
-	if err != nil {
-		return nil, errors.Wrap(err, "last-received-hold: hold newly received")
-	}
-
-	return &holdBasedAbstraction{
-		Type:              AbstractionLastReceivedHold,
-		FS:                fs,
-		FilesystemVersion: to,
-		JobID:             jobID,
-		Tag:               tag,
-	}, nil
-}
@@ -0,0 +1,28 @@
+package endpoint
+
+import "github.com/zrepl/zrepl/replication/logic/pdu"
+
+func SendAbstractionToPDU(a Abstraction) *pdu.SendAbstraction {
+	var ty pdu.SendAbstraction_SendAbstractionType
+	switch a.GetType() {
+	case AbstractionLastReceivedHold:
+		panic(a)
+	case AbstractionReplicationCursorBookmarkV1:
+		panic(a)
+	case AbstractionReplicationCursorBookmarkV2:
+		ty = pdu.SendAbstraction_ReplicationCursorV2
+	case AbstractionStepHold:
+		ty = pdu.SendAbstraction_StepHold
+	case AbstractionStepBookmark:
+		ty = pdu.SendAbstraction_StepBookmark
+	default:
+		panic(a)
+	}
+
+	version := a.GetFilesystemVersion()
+	return &pdu.SendAbstraction{
+		Type:    ty,
+		JobID:   (*a.GetJobID()).String(),
+		Version: pdu.FilesystemVersionFromZFS(&version),
+	}
+}
@@ -0,0 +1,159 @@
+package endpoint
+
+import (
+	"context"
+	"fmt"
+	"regexp"
+
+	"github.com/pkg/errors"
+
+	"github.com/zrepl/zrepl/zfs"
+)
+
+var stepHoldTagRE = regexp.MustCompile("^zrepl_STEP_J_(.+)")
+
+func StepHoldTag(jobid JobID) (string, error) {
+	return stepHoldTagImpl(jobid.String())
+}
+
+func stepHoldTagImpl(jobid string) (string, error) {
+	t := fmt.Sprintf("zrepl_STEP_J_%s", jobid)
+	if err := zfs.ValidHoldTag(t); err != nil {
+		return "", err
+	}
+	return t, nil
+}
+
+// err != nil always means that the bookmark is not a step bookmark
+func ParseStepHoldTag(tag string) (JobID, error) {
+	match := stepHoldTagRE.FindStringSubmatch(tag)
+	if match == nil {
+		return JobID{}, fmt.Errorf("parse hold tag: match regex %q", stepHoldTagRE)
+	}
+	jobID, err := MakeJobID(match[1])
+	if err != nil {
+		return JobID{}, errors.Wrap(err, "parse hold tag: invalid job id field")
+	}
+	return jobID, nil
+}
+
+const stepBookmarkNamePrefix = "zrepl_STEP"
+
+// v must be validated by caller
+func StepBookmarkName(fs string, guid uint64, id JobID) (string, error) {
+	return stepBookmarkNameImpl(fs, guid, id.String())
+}
+
+func stepBookmarkNameImpl(fs string, guid uint64, jobid string) (string, error) {
+	return makeJobAndGuidBookmarkName(stepBookmarkNamePrefix, fs, guid, jobid)
+}
+
+// name is the full bookmark name, including dataset path
+//
+// err != nil always means that the bookmark is not a step bookmark
+func ParseStepBookmarkName(fullname string) (guid uint64, jobID JobID, err error) {
+	guid, jobID, err = parseJobAndGuidBookmarkName(fullname, stepBookmarkNamePrefix)
+	if err != nil {
+		err = errors.Wrap(err, "parse step bookmark name") // no shadow!
+	}
+	return guid, jobID, err
+}
+
+// idempotently hold / step-bookmark `version`
+//
+// returns ErrBookmarkCloningNotSupported if version is a bookmark and bookmarking bookmarks is not supported by ZFS
+func HoldStep(ctx context.Context, fs string, v zfs.FilesystemVersion, jobID JobID) (Abstraction, error) {
+	if v.IsSnapshot() {
+
+		tag, err := StepHoldTag(jobID)
+		if err != nil {
+			return nil, errors.Wrap(err, "step hold tag")
+		}
+
+		if err := zfs.ZFSHold(ctx, fs, v, tag); err != nil {
+			return nil, errors.Wrap(err, "step hold: zfs")
+		}
+
+		return &holdBasedAbstraction{
+			Type:              AbstractionStepHold,
+			FS:                fs,
+			Tag:               tag,
+			JobID:             jobID,
+			FilesystemVersion: v,
+		}, nil
+	}
+
+	if !v.IsBookmark() {
+		panic(fmt.Sprintf("version must bei either snapshot or bookmark, got %#v", v))
+	}
+
+	bmname, err := StepBookmarkName(fs, v.Guid, jobID)
+	if err != nil {
+		return nil, errors.Wrap(err, "create step bookmark: determine bookmark name")
+	}
+	// idempotently create bookmark
+	stepBookmark, err := zfs.ZFSBookmark(ctx, fs, v, bmname)
+	if err != nil {
+		if err == zfs.ErrBookmarkCloningNotSupported {
+			// TODO we could actually try to find a local snapshot that has the requested GUID
+			// 		however, the replication algorithm prefers snapshots anyways, so this quest
+			// 		is most likely not going to be successful. Also, there's the possibility that
+			//      the caller might want to filter what snapshots are eligibile, and this would
+			//      complicate things even further.
+			return nil, err // TODO go1.13 use wrapping
+		}
+		return nil, errors.Wrap(err, "create step bookmark: zfs")
+	}
+	return &bookmarkBasedAbstraction{
+		Type:              AbstractionStepBookmark,
+		FS:                fs,
+		FilesystemVersion: stepBookmark,
+		JobID:             jobID,
+	}, nil
+}
+
+var _ BookmarkExtractor = StepBookmarkExtractor
+
+func StepBookmarkExtractor(fs *zfs.DatasetPath, v zfs.FilesystemVersion) (_ Abstraction) {
+	if v.Type != zfs.Bookmark {
+		panic("impl error")
+	}
+
+	fullname := v.ToAbsPath(fs)
+
+	guid, jobid, err := ParseStepBookmarkName(fullname)
+	if guid != v.Guid {
+		// TODO log this possibly tinkered-with bookmark
+		return nil
+	}
+	if err == nil {
+		bm := &bookmarkBasedAbstraction{
+			Type:              AbstractionStepBookmark,
+			FS:                fs.ToString(),
+			FilesystemVersion: v,
+			JobID:             jobid,
+		}
+		return bm
+	}
+	return nil
+}
+
+var _ HoldExtractor = StepHoldExtractor
+
+func StepHoldExtractor(fs *zfs.DatasetPath, v zfs.FilesystemVersion, holdTag string) Abstraction {
+	if v.Type != zfs.Snapshot {
+		panic("impl error")
+	}
+
+	jobID, err := ParseStepHoldTag(holdTag)
+	if err == nil {
+		return &holdBasedAbstraction{
+			Type:              AbstractionStepHold,
+			FS:                fs.ToString(),
+			Tag:               holdTag,
+			FilesystemVersion: v,
+			JobID:             jobID,
+		}
+	}
+	return nil
+}
@@ -1,83 +0,0 @@
-package endpoint
-
-import (
-	"context"
-	"fmt"
-	"regexp"
-
-	"github.com/pkg/errors"
-
-	"github.com/zrepl/zrepl/zfs"
-)
-
-var stepHoldTagRE = regexp.MustCompile("^zrepl_STEP_J_(.+)")
-
-func StepHoldTag(jobid JobID) (string, error) {
-	return stepHoldTagImpl(jobid.String())
-}
-
-func stepHoldTagImpl(jobid string) (string, error) {
-	t := fmt.Sprintf("zrepl_STEP_J_%s", jobid)
-	if err := zfs.ValidHoldTag(t); err != nil {
-		return "", err
-	}
-	return t, nil
-}
-
-// err != nil always means that the bookmark is not a step bookmark
-func ParseStepHoldTag(tag string) (JobID, error) {
-	match := stepHoldTagRE.FindStringSubmatch(tag)
-	if match == nil {
-		return JobID{}, fmt.Errorf("parse hold tag: match regex %q", stepHoldTagRE)
-	}
-	jobID, err := MakeJobID(match[1])
-	if err != nil {
-		return JobID{}, errors.Wrap(err, "parse hold tag: invalid job id field")
-	}
-	return jobID, nil
-}
-
-// idempotently hold `version`
-func HoldStep(ctx context.Context, fs string, v zfs.FilesystemVersion, jobID JobID) (Abstraction, error) {
-	if !v.IsSnapshot() {
-		panic(fmt.Sprintf("version must be a snapshot got %#v", v))
-	}
-
-	tag, err := StepHoldTag(jobID)
-	if err != nil {
-		return nil, errors.Wrap(err, "step hold tag")
-	}
-
-	if err := zfs.ZFSHold(ctx, fs, v, tag); err != nil {
-		return nil, errors.Wrap(err, "step hold: zfs")
-	}
-
-	return &holdBasedAbstraction{
-		Type:              AbstractionStepHold,
-		FS:                fs,
-		Tag:               tag,
-		JobID:             jobID,
-		FilesystemVersion: v,
-	}, nil
-
-}
-
-var _ HoldExtractor = StepHoldExtractor
-
-func StepHoldExtractor(fs *zfs.DatasetPath, v zfs.FilesystemVersion, holdTag string) Abstraction {
-	if v.Type != zfs.Snapshot {
-		panic("impl error")
-	}
-
-	jobID, err := ParseStepHoldTag(holdTag)
-	if err == nil {
-		return &holdBasedAbstraction{
-			Type:              AbstractionStepHold,
-			FS:                fs.ToString(),
-			Tag:               holdTag,
-			FilesystemVersion: v,
-			JobID:             jobID,
-		}
-	}
-	return nil
-}
@@ -24,9 +24,9 @@ func MakeJobID(s string) (JobID, error) {
 		return JobID{}, errors.Wrap(err, "must be usable as a dataset path component")
 	}

-	if _, err := tentativeReplicationCursorBookmarkNameImpl("pool/ds", 0xface601d, s); err != nil {
+	if _, err := stepBookmarkNameImpl("pool/ds", 0xface601d, s); err != nil {
 		// note that this might still fail due to total maximum name length, but we can't enforce that
-		return JobID{}, errors.Wrap(err, "must be usable for a tentative replication cursor bookmark")
+		return JobID{}, errors.Wrap(err, "must be usable for a step bookmark")
 	}

 	if _, err := stepHoldTagImpl(s); err != nil {
@@ -1,80 +0,0 @@
-// Code generated by "enumer -type=ReplicationGuaranteeKind -json -transform=snake -trimprefix=ReplicationGuaranteeKind"; DO NOT EDIT.
-
-//
-package endpoint
-
-import (
-	"encoding/json"
-	"fmt"
-)
-
-const (
-	_ReplicationGuaranteeKindName_0 = "resumabilityincremental"
-	_ReplicationGuaranteeKindName_1 = "none"
-)
-
-var (
-	_ReplicationGuaranteeKindIndex_0 = [...]uint8{0, 12, 23}
-	_ReplicationGuaranteeKindIndex_1 = [...]uint8{0, 4}
-)
-
-func (i ReplicationGuaranteeKind) String() string {
-	switch {
-	case 1 <= i && i <= 2:
-		i -= 1
-		return _ReplicationGuaranteeKindName_0[_ReplicationGuaranteeKindIndex_0[i]:_ReplicationGuaranteeKindIndex_0[i+1]]
-	case i == 4:
-		return _ReplicationGuaranteeKindName_1
-	default:
-		return fmt.Sprintf("ReplicationGuaranteeKind(%d)", i)
-	}
-}
-
-var _ReplicationGuaranteeKindValues = []ReplicationGuaranteeKind{1, 2, 4}
-
-var _ReplicationGuaranteeKindNameToValueMap = map[string]ReplicationGuaranteeKind{
-	_ReplicationGuaranteeKindName_0[0:12]:  1,
-	_ReplicationGuaranteeKindName_0[12:23]: 2,
-	_ReplicationGuaranteeKindName_1[0:4]:   4,
-}
-
-// ReplicationGuaranteeKindString retrieves an enum value from the enum constants string name.
-// Throws an error if the param is not part of the enum.
-func ReplicationGuaranteeKindString(s string) (ReplicationGuaranteeKind, error) {
-	if val, ok := _ReplicationGuaranteeKindNameToValueMap[s]; ok {
-		return val, nil
-	}
-	return 0, fmt.Errorf("%s does not belong to ReplicationGuaranteeKind values", s)
-}
-
-// ReplicationGuaranteeKindValues returns all values of the enum
-func ReplicationGuaranteeKindValues() []ReplicationGuaranteeKind {
-	return _ReplicationGuaranteeKindValues
-}
-
-// IsAReplicationGuaranteeKind returns "true" if the value is listed in the enum definition. "false" otherwise
-func (i ReplicationGuaranteeKind) IsAReplicationGuaranteeKind() bool {
-	for _, v := range _ReplicationGuaranteeKindValues {
-		if i == v {
-			return true
-		}
-	}
-	return false
-}
-
-// MarshalJSON implements the json.Marshaler interface for ReplicationGuaranteeKind
-func (i ReplicationGuaranteeKind) MarshalJSON() ([]byte, error) {
-	return json.Marshal(i.String())
-}
-
-// UnmarshalJSON implements the json.Unmarshaler interface for ReplicationGuaranteeKind
-func (i *ReplicationGuaranteeKind) UnmarshalJSON(data []byte) error {
-	var s string
-	if err := json.Unmarshal(data, &s); err != nil {
-		return fmt.Errorf("ReplicationGuaranteeKind should be a string, got %s", data)
-	}
-
-	var err error
-	*i, err = ReplicationGuaranteeKindString(s)
-	return err
-}
@@ -22,7 +22,7 @@ require (
 	github.com/onsi/gomega v1.7.0 // indirect
 	github.com/pkg/errors v0.8.1
 	github.com/pkg/profile v1.2.1
-	github.com/problame/go-netssh v0.0.0-20200601114649-26439f9f0dc5
+	github.com/problame/go-netssh v0.0.0-20191209123953-18d8aa6923c7
 	github.com/prometheus/client_golang v1.2.1
 	github.com/prometheus/common v0.7.0
 	github.com/sergi/go-diff v1.0.1-0.20180205163309-da645544ed44 // go1.12 thinks it needs this
@@ -212,8 +212,6 @@ github.com/problame/go-netssh v0.0.0-20191026123024-f34099f4f6b1 h1:HH8yzlaZq/A8
 github.com/problame/go-netssh v0.0.0-20191026123024-f34099f4f6b1/go.mod h1:JRs3nyBjvOv/9YHC+Vlw9z1Jfzl5s5t0BNnTzmclj1c=
 github.com/problame/go-netssh v0.0.0-20191209123953-18d8aa6923c7 h1:Oik8tLrLhoMq4fduDRuNNOAQ+q0M0dXkjAYLUf4+mAc=
 github.com/problame/go-netssh v0.0.0-20191209123953-18d8aa6923c7/go.mod h1:Ba6RaFpK1+IHWs1wLZ6sCBuXkt4iAVLeOG5GinXHusM=
-github.com/problame/go-netssh v0.0.0-20200601114649-26439f9f0dc5 h1:1eSrO2WAeSXjMol8WRKW9LekL1252VIMaKQwWkmEdvs=
-github.com/problame/go-netssh v0.0.0-20200601114649-26439f9f0dc5/go.mod h1:Ba6RaFpK1+IHWs1wLZ6sCBuXkt4iAVLeOG5GinXHusM=
 github.com/prometheus/client_golang v0.0.0-20180410130117-e11c6ff8170b/go.mod h1:7SWBe2y4D6OKWSNQJUaRYU/AaXPKyh/dDVn+NZz0KFw=
 github.com/prometheus/client_golang v0.9.1/go.mod h1:7SWBe2y4D6OKWSNQJUaRYU/AaXPKyh/dDVn+NZz0KFw=
 github.com/prometheus/client_golang v1.0.0/go.mod h1:db9x61etRT2tGnBNRi70OPL5FsnadC4Ky3P0J6CfImo=
@@ -1,4 +1,4 @@
-#!/usr/bin/env bash
+#!/bin/bash
 set -e


@@ -14,31 +14,23 @@ step() {
    echo "${bold}$1${normal}"
 }

+if ! type go >/dev/null; then
+    step "go binary not installed or not in \$PATH" 1>&2
+    exit 1
+fi
+
+if [ -z "$GOPATH" ]; then
+    step "Make sure you have your GOPATH configured correctly" 1>&2
+    exit 1
+fi
+
+CHECKOUTPATH="${GOPATH}/src/github.com/zrepl/zrepl"
+
 godep() {
    step "install build dependencies (versions pinned in build/go.mod and build/tools.go)"
-
-    if ! type go >/dev/null; then
-        step "go binary not installed or not in \$PATH" 1>&2
-        exit 1
-    fi
-
-    if [ -z "$GOPATH" ]; then
-        step "Your GOPATH is not configured correctly" 1>&2
-        exit 1
-    fi
-
-    if ! (echo "$PATH" | grep "${GOPATH}/bin" > /dev/null); then
-        step "GOPATH/bin is not in your PATH (it should be towards the start of it)"
-        exit 1
-    fi
-
    pushd "$(dirname "${BASH_SOURCE[0]}")"/build
    set -x
    export GO111MODULE=on # otherwise, a checkout of this repo in GOPATH will disable modules on Go 1.12 and earlier
-    source <(go env)
-    export GOOS="$GOHOSTOS"
-    export GOARCH="$GOHOSTARCH"
-    # TODO GOARM=$GOHOSTARM?
    go build -v -mod=readonly -o "$GOPATH/bin/stringer"      golang.org/x/tools/cmd/stringer
    go build -v -mod=readonly -o "$GOPATH/bin/protoc-gen-go" github.com/golang/protobuf/protoc-gen-go
    go build -v -mod=readonly -o "$GOPATH/bin/enumer"        github.com/alvaroloes/enumer
@@ -59,9 +51,8 @@ docdep() {
        exit 1
    fi
    step "Installing doc build dependencies"
-    # shellcheck disable=SC2155
-    local reqpath="$(dirname "${BASH_SOURCE[0]}")/docs/requirements.txt"
-    if [ -n "$ZREPL_LAZY_DOCS_REQPATH" ]; then
+    local reqpath="${CHECKOUTPATH}/docs/requirements.txt"
+    if [ ! -z "$ZREPL_LAZY_DOCS_REQPATH" ]; then
        reqpath="$ZREPL_LAZY_DOCS_REQPATH"
    fi
    pip3 install -r "$reqpath"
@@ -72,15 +63,9 @@ release() {
    make release
 }

-# shellcheck disable=SC2198
-if [ -z "$@" ]; then
-    step "No command specified, exiting"
-    exit 1
-fi
-
 for cmd in "$@"; do
    case "$cmd" in
-        godep|docdep|release|docs)
+        godep|docdep|release_bins|docs)
            eval $cmd
            continue
            ;;
@@ -82,7 +82,7 @@ func (l Level) String() string {
 	case Error:
 		return "error"
 	default:
-		return fmt.Sprintf("unknown level %d", l)
+		return string(l)
 	}
 }

@@ -1,11 +0,0 @@
-FROM debian:latest
-
-RUN apt-get update && apt-get install -y \
-	build-essential \
-    devscripts \
-    dh-exec
-
-RUN mkdir -p /build/src && chmod -R 0777 /build
-
-WORKDIR /build/src
-
@@ -1,5 +0,0 @@
-zrepl (VERSION) unstable;
-
-  Check https://zrepl.github.io/changelog.html.
-
- -- zrepl build process <me@cschwarz.com>  DATE_DASH_R_OUTPUT
@@ -1 +0,0 @@
-11
@@ -1,8 +0,0 @@
-Source: zrepl
-Maintainer: Christian Schwarz <me@cschwarz.com>
-Build-Depends: dh-exec
-
-Package: zrepl
-Architecture: arm64 amd64 armhf i386
-Description: ZFS replication
-  zrepl is a one-stop solution for ZFS filesystem replication
@@ -1,50 +0,0 @@
-#!/usr/bin/make -f
-
-%:
-	dh $@
-
-override_dh_strip:
-
-override_dh_auto_build:
-
-override_dh_auto_clean:
-
-	rm -rf debian/renamedir
-
-override_dh_auto_install:
-
-	# install the zrepl.service for
-	# dh_install_systemd and dh_systemd_{start,enable}
-	install dist/systemd/zrepl.service debian/zrepl.service
-	sed -i 's#ExecStart=/usr/local/bin/zrepl #ExecStart=/usr/bin/zrepl #' debian/zrepl.service
-	sed -i 's#ExecStartPre=/usr/local/bin/zrepl #ExecStartPre=/usr/bin/zrepl #' debian/zrepl.service
-
-	mkdir -p debian/renamedir
-
-	# install binary
-	stat artifacts/$(ZREPL_DPKG_ZREPL_BINARY_FILENAME)
-	cp --preserve=all artifacts/$(ZREPL_DPKG_ZREPL_BINARY_FILENAME) debian/renamedir/zrepl
-	dh_install debian/renamedir/zrepl usr/bin
-
-	# install bash completion
-	dh_install artifacts/bash_completion etc/bash_completion.d/zrepl
-	dh_install artifacts/_zrepl.zsh_completion usr/share/zsh/vendor-completions
-
-	# install docs
-	dh_install artifacts/docs/html usr/share/doc/zrepl/docs/
-
-	# install examples
-	dh_install config/samples/* usr/share/doc/zrepl/examples
-
-	# install default config
-	mkdir -p debian/tmp/etc/zrepl
-	chmod 0700 debian/tmp/etc/zrepl
-	sed 's#USR_SHARE_ZREPL#/usr/share/doc/zrepl#' packaging/systemd-default-zrepl.yml > debian/renamedir/zrepl.yml
-	dh_install debian/renamedir/zrepl.yml etc/zrepl
-
-	# save git revision of this packaging repo
-	echo $$(git rev-parse HEAD) > debian/packaging-repo-git-revision
-	dh_install debian/packaging-repo-git-revision usr/share/doc/zrepl
-
-override_dh_auto_test:
-	# don't run tests at this point
@@ -1,7 +0,0 @@
-FROM fedora:latest
-
-RUN dnf install -y git make bash rpm-build 'dnf-command(builddep)'
-ADD packaging/rpm/zrepl.spec /tmp/zrepl.spec
-RUN dnf builddep -y /tmp/zrepl.spec
-RUN mkdir -p /build/src && chmod -R 0777 /build
-WORKDIR /build/src
@@ -1,78 +0,0 @@
-# This Spec file packages pre-built artifacts in the artifacts directory
-# into an RPM. This means that rpmbuild won't actually build any new artifacts
-#
-# Read the Makefile first, then come back here.
-
-# Global meta data
-Version:          SUBSTITUTED_BY_MAKEFILE
-%global common_description %{expand:
-zrepl is a one-stop, integrated solution for ZFS replication.}
-
-# use gzip to be compatible with centos7
-%define _source_payload w9.gzdio
-%define _binary_payload w9.gzdio
-
-# don't strip pre-built binaries
-%define __strip /usr/bin/true
-
-Name:             zrepl
-Release:          1
-Summary:          One-stop, integrated solution for ZFS replication
-License:          MIT
-URL:              https://zrepl.github.io/
-# Source: we use rpmbuild --build-in-tree => no source
-BuildRequires:    systemd
-Requires(post):   systemd
-Requires(preun):  systemd
-Requires(postun): systemd
-
-%description
-%{common_description}
-
-%prep
-# we use rpmbuild --build-in-tree => no prep or setup
-
-%build
-# we don't actually need to build anything here, that has already been done by the Makefile
-
-# Correct the path in the systemd unit file
-sed s:/usr/local/bin/:%{_bindir}/:g dist/systemd/zrepl.service > artifacts/rpmbuild/zrepl.service
-
-# Generate the default configuration file
-sed 's#USR_SHARE_ZREPL#%{_datadir}/doc/zrepl#' packaging/systemd-default-zrepl.yml > artifacts/rpmbuild/zrepl.yml
-
-%install
-install -Dm 0755 artifacts/%{_zrepl_binary_filename}    %{buildroot}%{_bindir}/zrepl
-install -Dm 0644 artifacts/rpmbuild/zrepl.service       %{buildroot}%{_unitdir}/zrepl.service
-install -Dm 0644 artifacts/_zrepl.zsh_completion        %{buildroot}%{_datadir}/zsh/site-functions/_zrepl
-install -Dm 0644 artifacts/bash_completion              %{buildroot}%{_datadir}/bash-completion/completions/zrepl
-install -Dm 0644 artifacts/rpmbuild/zrepl.yml           %{buildroot}%{_sysconfdir}/zrepl/zrepl.yml
-install -d                                              %{buildroot}%{_datadir}/doc/zrepl
-cp -a   artifacts/docs/html                             %{buildroot}%{_datadir}/doc/zrepl/html
-cp -a   config/samples                                  %{buildroot}%{_datadir}/doc/zrepl/examples
-
-%post
-%systemd_post zrepl.service
-
-
-%preun
-%systemd_preun zrepl.service
-
-
-%postun
-%systemd_postun_with_restart zrepl.service
-
-
-%files
-%defattr(-,root,root)
-%license LICENSE
-%{_bindir}/zrepl
-%config %{_unitdir}/zrepl.service
-%dir %{_sysconfdir}/zrepl
-%config %{_sysconfdir}/zrepl/zrepl.yml
-%{_datadir}/zsh/site-functions/_zrepl
-%{_datadir}/bash-completion/completions/zrepl
-%{_datadir}/doc/zrepl
-
-%changelog
-# TODO: auto-fill changelog from git? -> need same solution for debian
@@ -1,13 +0,0 @@
-global:
-  logging:
-    # use syslog instead of stdout because it makes journald happy
-    - type: syslog
-      format: human
-      level: warn
-
-jobs:
-#  - name: foo
-#    type: bar
-
-# see USR_SHARE_ZREPL/examples
-# or https://zrepl.github.io/configuration/overview.html
@@ -11,7 +11,6 @@ import (

 	"github.com/fatih/color"
 	"github.com/pkg/errors"
-
 	"github.com/zrepl/zrepl/daemon/logging/trace"

 	"github.com/zrepl/zrepl/config"
@@ -2,8 +2,9 @@ package tests

 import (
 	"fmt"
-	"strings"

+	"github.com/kr/pretty"
+	"github.com/stretchr/testify/require"
 	"github.com/zrepl/zrepl/platformtest"
 	"github.com/zrepl/zrepl/zfs"
 )
@@ -17,7 +18,9 @@ func BatchDestroy(ctx *platformtest.Context) {
 		+  "foo bar@1"
 		+  "foo bar@2"
 		+  "foo bar@3"
+		+  "foo bar@4"
 		R  zfs hold zrepl_platformtest "${ROOTDS}/foo bar@2"
+		R  zfs hold zrepl_platformtest "${ROOTDS}/foo bar@4"
 	`)

 	reqs := []*zfs.DestroySnapOp{
@@ -31,24 +34,40 @@ func BatchDestroy(ctx *platformtest.Context) {
 			Filesystem: fmt.Sprintf("%s/foo bar", ctx.RootDataset),
 			Name:       "2",
 		},
+		&zfs.DestroySnapOp{
+			ErrOut:     new(error),
+			Filesystem: fmt.Sprintf("%s/foo bar", ctx.RootDataset),
+			Name:       "non existent",
+		},
+		&zfs.DestroySnapOp{
+			ErrOut:     new(error),
+			Filesystem: fmt.Sprintf("%s/foo bar", ctx.RootDataset),
+			Name:       "4",
+		},
 	}
 	zfs.ZFSDestroyFilesystemVersions(ctx, reqs)
+
+	pretty.Println(reqs)
+
 	if *reqs[0].ErrOut != nil {
 		panic("expecting no error")
 	}
-	err := (*reqs[1].ErrOut).Error()
-	if !strings.Contains(err, fmt.Sprintf("%s/foo bar@2", ctx.RootDataset)) {
-		panic(fmt.Sprintf("expecting error about being unable to destroy @2: %T\n%s", err, err))
-	}
+
+	eBusy, ok := (*reqs[1].ErrOut).(*zfs.ErrDestroySnapshotDatasetIsBusy)
+	require.True(ctx, ok)
+	require.Equal(ctx, reqs[1].Name, eBusy.Name)
+
+	require.Nil(ctx, *reqs[2].ErrOut, "destroying non-existent snap is not an error (idempotence)")
+
+	eBusy, ok = (*reqs[3].ErrOut).(*zfs.ErrDestroySnapshotDatasetIsBusy)
+	require.True(ctx, ok)
+	require.Equal(ctx, reqs[3].Name, eBusy.Name)

 	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
 	!N "foo bar@3"
 	!E "foo bar@1"
 	!E "foo bar@2"
-	R zfs release zrepl_platformtest "${ROOTDS}/foo bar@2"
-	-  "foo bar@2"
-	-  "foo bar@1"
-	-  "foo bar"
+	!E "foo bar@4"
 	`)

 }
@@ -14,23 +14,16 @@ var Cases = []Case{BatchDestroy,
 	ListFilesystemVersionsUserrefs,
 	ListFilesystemVersionsZeroExistIsNotAnError,
 	ListFilesystemsNoFilter,
+	Pruner2NotReplicated,
 	ReceiveForceIntoEncryptedErr,
 	ReceiveForceRollbackWorksUnencrypted,
-	ReplicationFailingInitialParentProhibitsChildReplication,
 	ReplicationIncrementalCleansUpStaleAbstractionsWithCacheOnSecondReplication,
 	ReplicationIncrementalCleansUpStaleAbstractionsWithoutCacheOnSecondReplication,
-	ReplicationIncrementalDestroysStepHoldsIffIncrementalStepHoldsAreDisabledButStepHoldsExist,
 	ReplicationIncrementalIsPossibleIfCommonSnapshotIsDestroyed,
-	ReplicationIsResumableFullSend__both_GuaranteeResumability,
-	ReplicationIsResumableFullSend__initial_GuaranteeIncrementalReplication_incremental_GuaranteeIncrementalReplication,
-	ReplicationIsResumableFullSend__initial_GuaranteeResumability_incremental_GuaranteeIncrementalReplication,
-	ReplicationReceiverErrorWhileStillSending,
-	ReplicationStepCompletedLostBehavior__GuaranteeIncrementalReplication,
-	ReplicationStepCompletedLostBehavior__GuaranteeResumability,
+	ReplicationIsResumableFullSend,
 	ResumableRecvAndTokenHandling,
 	ResumeTokenParsing,
-	SendArgsValidationEncryptedSendOfUnencryptedDatasetForbidden__EncryptionSupported_false,
-	SendArgsValidationEncryptedSendOfUnencryptedDatasetForbidden__EncryptionSupported_true,
+	SendArgsValidationEncryptedSendOfUnencryptedDatasetForbidden,
 	SendArgsValidationResumeTokenDifferentFilesystemForbidden,
 	SendArgsValidationResumeTokenEncryptionMismatchForbidden,
 	UndestroyableSnapshotParsing,
@@ -4,7 +4,6 @@ import (
 	"path"

 	"github.com/stretchr/testify/require"
-
 	"github.com/zrepl/zrepl/platformtest"
 	"github.com/zrepl/zrepl/zfs"
 )
@@ -4,7 +4,6 @@ import (
 	"fmt"

 	"github.com/stretchr/testify/assert"
-
 	"github.com/zrepl/zrepl/platformtest"
 	"github.com/zrepl/zrepl/zfs"
 )
@@ -0,0 +1,331 @@
+package tests
+
+import (
+	"encoding/json"
+	"fmt"
+	"time"
+
+	"github.com/prometheus/client_golang/prometheus"
+	"github.com/stretchr/testify/require"
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/daemon/pruner"
+	"github.com/zrepl/zrepl/endpoint"
+	"github.com/zrepl/zrepl/platformtest"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+func PrunerNotReplicated(ctx *platformtest.Context) {
+
+	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
+        DESTROYROOT
+        CREATEROOT
+        +  "foo bar"
+        +  "foo bar@1"
+        +  "foo bar@2"
+        +  "foo bar@3"
+        +  "foo bar@4"
+        +  "foo bar@5"
+    `)
+
+	c, err := config.ParseConfigBytes([]byte(fmt.Sprintf(`
+jobs:
+- name: prunetest
+  type: push
+  filesystems: {
+      "%s/foo bar<": true
+  }
+  connect:
+    type: tcp
+    address: 255.255.255.255:255
+  snapshotting:
+    type: manual
+  pruning:
+    keep_sender:
+    - type: not_replicated
+    - type: last_n
+      count: 1
+    keep_receiver:
+    - type: last_n
+      count: 2
+    `, ctx.RootDataset)))
+	require.NoError(ctx, err)
+
+	pushJob := c.Jobs[0].Ret.(*config.PushJob)
+
+	dummyHistVec := prometheus.NewHistogramVec(prometheus.HistogramOpts{
+		Namespace: "foo",
+		Subsystem: "foo",
+		Name:      "foo",
+		Help:      "foo",
+	}, []string{"foo"})
+
+	prunerFactory, err := pruner.NewPrunerFactory(pushJob.Pruning, dummyHistVec)
+	require.NoError(ctx, err)
+
+	senderJid := endpoint.MustMakeJobID("sender-job")
+
+	fsfilter, err := filters.DatasetMapFilterFromConfig(pushJob.Filesystems)
+	require.NoError(ctx, err)
+
+	sender := endpoint.NewSender(endpoint.SenderConfig{
+		FSF: fsfilter,
+		Encrypt: &zfs.NilBool{
+			B: false,
+		},
+		JobID: senderJid,
+	})
+
+	fs := ctx.RootDataset + "/foo bar"
+
+	// create a replication cursor to make pruning work at all
+	_, err = endpoint.CreateReplicationCursor(ctx, fs, fsversion(ctx, fs, "@2"), senderJid)
+	require.NoError(ctx, err)
+
+	p := prunerFactory.BuildSenderPruner(ctx, sender, sender)
+
+	p.Prune()
+
+	report := p.Report()
+
+	reportJSON, err := json.MarshalIndent(report, "", "  ")
+	require.NoError(ctx, err)
+	ctx.Logf("%s\n", string(reportJSON))
+
+	require.Equal(ctx, pruner.Done.String(), report.State)
+	require.Len(ctx, report.Completed, 1)
+	fsReport := report.Completed[0]
+	require.Equal(ctx, fs, fsReport.Filesystem)
+	require.Empty(ctx, fsReport.SkipReason)
+	require.Empty(ctx, fsReport.LastError)
+	require.Len(ctx, fsReport.DestroyList, 1)
+	require.Equal(ctx, fsReport.DestroyList[0], pruner.SnapshotReport{
+		Name:       "1",
+		Replicated: true,
+		Date:       fsReport.DestroyList[0].Date,
+	})
+
+}
+
+func PrunerNoKeepNotReplicatedNoKeepStepHoldConvertsAnyStepHoldToBookmark(ctx *platformtest.Context) {
+
+	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
+        DESTROYROOT
+        CREATEROOT
+        +  "foo bar"
+        +  "foo bar@1"
+        +  "foo bar@2"
+        +  "foo bar@3"
+        +  "foo bar@4"
+        +  "foo bar@5"
+    `)
+
+	c, err := config.ParseConfigBytes([]byte(fmt.Sprintf(`
+jobs:
+- name: prunetest
+  type: push
+  filesystems: {
+      "%s/foo bar<": true
+  }
+  connect:
+    type: tcp
+    address: 255.255.255.255:255
+  snapshotting:
+    type: manual
+  pruning:
+    keep_sender:
+    - type: last_n
+      count: 1
+    keep_receiver:
+    - type: last_n
+      count: 2
+    `, ctx.RootDataset)))
+	require.NoError(ctx, err)
+
+	pushJob := c.Jobs[0].Ret.(*config.PushJob)
+
+	dummyHistVec := prometheus.NewHistogramVec(prometheus.HistogramOpts{
+		Namespace: "foo",
+		Subsystem: "foo",
+		Name:      "foo",
+		Help:      "foo",
+	}, []string{"foo"})
+
+	prunerFactory, err := pruner.NewPrunerFactory(pushJob.Pruning, dummyHistVec)
+	require.NoError(ctx, err)
+
+	senderJid := endpoint.MustMakeJobID("sender-job")
+
+	fsfilter, err := filters.DatasetMapFilterFromConfig(pushJob.Filesystems)
+	require.NoError(ctx, err)
+
+	sender := endpoint.NewSender(endpoint.SenderConfig{
+		FSF: fsfilter,
+		Encrypt: &zfs.NilBool{
+			B: false,
+		},
+        JobID: senderJid,
+	})
+
+	fs := ctx.RootDataset + "/foo bar"
+
+	// create a replication cursor to make pruning work at all
+	_, err = endpoint.CreateReplicationCursor(ctx, fs, fsversion(ctx, fs, "@2"), senderJid)
+	require.NoError(ctx, err)
+
+	// create step holds for the incremental @2->@3
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@2"), senderJid)
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@3"), senderJid)
+	// create step holds for another job
+	otherJid := endpoint.MustMakeJobID("other-job")
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@2"), otherJid)
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@3"), otherJid)
+
+	p := prunerFactory.BuildSenderPruner(ctx, sender, sender)
+
+	p.Prune()
+
+	report := p.Report()
+
+	reportJSON, err := json.MarshalIndent(report, "", "  ")
+	require.NoError(ctx, err)
+	ctx.Logf("%s\n", string(reportJSON))
+
+	require.Equal(ctx, pruner.Done.String(), report.State)
+	require.Len(ctx, report.Completed, 1)
+	fsReport := report.Completed[0]
+	require.Equal(ctx, fs, fsReport.Filesystem)
+	require.Empty(ctx, fsReport.SkipReason)
+	require.Empty(ctx, fsReport.LastError)
+	expectDestroyList := []pruner.SnapshotReport{
+		{
+			Name:       "1",
+			Replicated: true,
+		},
+		{
+			Name:       "2",
+			Replicated: true,
+		},
+		{
+			Name:       "3",
+			Replicated: true,
+		},
+		{
+			Name:       "4",
+			Replicated: true,
+		},
+	}
+	for _, d := range fsReport.DestroyList {
+		d.Date = time.Time{}
+	}
+	require.Subset(ctx, fsReport.DestroyList, expectDestroyList)
+}
+
+
+
+func PrunerNoKeepNotReplicatedButKeepStepHold(ctx *platformtest.Context) {
+
+	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
+        DESTROYROOT
+        CREATEROOT
+        +  "foo bar"
+        +  "foo bar@1"
+        +  "foo bar@2"
+        +  "foo bar@3"
+        +  "foo bar@4"
+        +  "foo bar@5"
+    `)
+
+	c, err := config.ParseConfigBytes([]byte(fmt.Sprintf(`
+jobs:
+- name: prunetest
+  type: push
+  filesystems: {
+      "%s/foo bar<": true
+  }
+  connect:
+    type: tcp
+    address: 255.255.255.255:255
+  snapshotting:
+    type: manual
+  pruning:
+    keep_sender:
+    - type: step_holds
+    - type: last_n
+      count: 1
+    keep_receiver:
+    - type: last_n
+      count: 2
+    `, ctx.RootDataset)))
+	require.NoError(ctx, err)
+
+	pushJob := c.Jobs[0].Ret.(*config.PushJob)
+
+	dummyHistVec := prometheus.NewHistogramVec(prometheus.HistogramOpts{
+		Namespace: "foo",
+		Subsystem: "foo",
+		Name:      "foo",
+		Help:      "foo",
+	}, []string{"foo"})
+
+	prunerFactory, err := pruner.NewPrunerFactory(pushJob.Pruning, dummyHistVec)
+	require.NoError(ctx, err)
+
+	senderJid := endpoint.MustMakeJobID("sender-job")
+
+	fsfilter, err := filters.DatasetMapFilterFromConfig(pushJob.Filesystems)
+	require.NoError(ctx, err)
+
+	sender := endpoint.NewSender(endpoint.SenderConfig{
+		FSF: fsfilter,
+		Encrypt: &zfs.NilBool{
+			B: false,
+		},
+        JobID: senderJid,
+	})
+
+	fs := ctx.RootDataset + "/foo bar"
+
+	// create a replication cursor to make pruning work at all
+	_, err = endpoint.CreateReplicationCursor(ctx, fs, fsversion(ctx, fs, "@2"), senderJid)
+	require.NoError(ctx, err)
+
+	// create step holds for the incremental @2->@3
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@2"), senderJid)
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@3"), senderJid)
+	// create step holds for another job
+	otherJid := endpoint.MustMakeJobID("other-job")
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@2"), otherJid)
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@3"), otherJid)
+
+	p := prunerFactory.BuildSenderPruner(ctx, sender, sender)
+
+	p.Prune()
+
+	report := p.Report()
+
+	reportJSON, err := json.MarshalIndent(report, "", "  ")
+	require.NoError(ctx, err)
+	ctx.Logf("%s\n", string(reportJSON))
+
+	require.Equal(ctx, pruner.Done.String(), report.State)
+	require.Len(ctx, report.Completed, 1)
+	fsReport := report.Completed[0]
+	require.Equal(ctx, fs, fsReport.Filesystem)
+	require.Empty(ctx, fsReport.SkipReason)
+	require.Empty(ctx, fsReport.LastError)
+	expectDestroyList := []pruner.SnapshotReport{
+		{
+			Name:       "1",
+			Replicated: true,
+		},
+		{
+			Name:       "4",
+			Replicated: true,
+		},
+	}
+	for _, d := range fsReport.DestroyList {
+		d.Date = time.Time{}
+	}
+	require.Subset(ctx, fsReport.DestroyList, expectDestroyList)
+}
@@ -0,0 +1,137 @@
+package tests
+
+import (
+	"encoding/json"
+	"fmt"
+	"time"
+
+	"github.com/kr/pretty"
+	"github.com/stretchr/testify/require"
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/daemon/pruner.v2"
+	"github.com/zrepl/zrepl/endpoint"
+	"github.com/zrepl/zrepl/platformtest"
+	"github.com/zrepl/zrepl/pruning"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+func Pruner2NotReplicated(ctx *platformtest.Context) {
+
+	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
+        DESTROYROOT
+        CREATEROOT
+        +  "foo bar"
+        +  "foo bar@1"
+        +  "foo bar@2"
+        +  "foo bar@3"
+        +  "foo bar@4"
+        +  "foo bar@5"
+	`)
+
+	fs := ctx.RootDataset + "/foo bar"
+	senderJid := endpoint.MustMakeJobID("sender-job")
+	otherJid1 := endpoint.MustMakeJobID("other-job-1")
+	otherJid2 := endpoint.MustMakeJobID("other-job-2")
+
+
+	// create step holds for the incremental @2->@3
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@2"), senderJid)
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@3"), senderJid)
+	// create step holds for other-job-1 @2 -> @3
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@2"), otherJid1)
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@3"), otherJid1)
+	// create step hold for other-job-2 @1 (will be pruned)
+	endpoint.HoldStep(ctx, fs, fsversion(ctx, fs, "@1"), otherJid2)
+
+	c, err := config.ParseConfigBytes([]byte(fmt.Sprintf(`
+jobs:
+- name: prunetest
+  type: push
+  filesystems: {
+      "%s/foo bar": true
+  }
+  connect:
+    type: tcp
+    address: 255.255.255.255:255
+  snapshotting:
+    type: manual
+  pruning:
+    keep_sender:
+    #- type: not_replicated
+    - type: step_holds
+    - type: last_n
+      count: 1
+    keep_receiver:
+    - type: last_n
+      count: 2
+    `, ctx.RootDataset)))
+	require.NoError(ctx, err)
+
+	pushJob := c.Jobs[0].Ret.(*config.PushJob)
+
+	require.NoError(ctx, err)
+
+	fsfilter, err := filters.DatasetMapFilterFromConfig(pushJob.Filesystems)
+	require.NoError(ctx, err)
+
+	matchedFilesystems, err := zfs.ZFSListMapping(ctx, fsfilter)
+	ctx.Logf("%s", pretty.Sprint(matchedFilesystems))
+	require.NoError(ctx, err)
+	require.Len(ctx, matchedFilesystems, 1)
+
+	sideSender := pruner.NewSideSender(senderJid)
+
+	keepRules, err := pruning.RulesFromConfig(senderJid, pushJob.Pruning.KeepSender)
+	require.NoError(ctx, err)
+	p := pruner.NewPruner(fsfilter, senderJid, sideSender, keepRules)
+
+	runDone := make(chan *pruner.Report)
+	go func() {
+		runDone <- p.Run(ctx)
+	}()
+
+	var report *pruner.Report
+	// concurrency stress
+out:
+	for {
+		select {
+		case <-time.After(10 * time.Millisecond):
+			p.Report(ctx)
+		case report = <-runDone:
+			break out
+		}
+	}
+	ctx.Logf("%s\n", pretty.Sprint(report))
+
+	reportJSON, err := json.MarshalIndent(report, "", "  ")
+	require.NoError(ctx, err)
+	ctx.Logf("%s\n", string(reportJSON))
+
+	ctx.FailNow()
+	// fs := ctx.RootDataset + "/foo bar"
+
+	// // create a replication cursor to make pruning work at all
+	// _, err = endpoint.CreateReplicationCursor(ctx, fs, fsversion(ctx, fs, "@2"), senderJid)
+	// require.NoError(ctx, err)
+
+	// p := prunerFactory.BuildSenderPruner(ctx, sender, sender)
+
+	// p.Prune()
+
+	// report := p.Report()
+
+	// require.Equal(ctx, pruner.Done.String(), report.State)
+	// require.Len(ctx, report.Completed, 1)
+	// fsReport := report.Completed[0]
+	// require.Equal(ctx, fs, fsReport.Filesystem)
+	// require.Empty(ctx, fsReport.SkipReason)
+	// require.Empty(ctx, fsReport.LastError)
+	// require.Len(ctx, fsReport.DestroyList, 1)
+	// require.Equal(ctx, fsReport.DestroyList[0], pruner.SnapshotReport{
+	// 	Name:       "1",
+	// 	Replicated: true,
+	// 	Date:       fsReport.DestroyList[0].Date,
+	// })
+
+}
@@ -4,20 +4,11 @@ import (
 	"fmt"

 	"github.com/stretchr/testify/require"
-
 	"github.com/zrepl/zrepl/platformtest"
 	"github.com/zrepl/zrepl/zfs"
 )

 func ReceiveForceIntoEncryptedErr(ctx *platformtest.Context) {
-
-	supported, err := zfs.EncryptionCLISupported(ctx)
-	require.NoError(ctx, err, "encryption feature test failed")
-	if !supported {
-		ctx.SkipNow()
-		return
-	}
-
 	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
 		DESTROYROOT
 		CREATEROOT
@@ -41,7 +32,6 @@ func ReceiveForceIntoEncryptedErr(ctx *platformtest.Context) {

 	sendStream, err := zfs.ZFSSend(ctx, sendArgs)
 	require.NoError(ctx, err)
-	defer sendStream.Close()

 	recvOpts := zfs.RecvOptions{
 		RollbackAndForceRecv: true,
@@ -5,7 +5,6 @@ import (

 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
-
 	"github.com/zrepl/zrepl/platformtest"
 	"github.com/zrepl/zrepl/zfs"
 )
@@ -4,13 +4,10 @@ import (
 	"context"
 	"fmt"
 	"io"
-	"os"
 	"path"
-	"sort"

 	"github.com/kr/pretty"
 	"github.com/stretchr/testify/require"
-
 	"github.com/zrepl/zrepl/daemon/filters"
 	"github.com/zrepl/zrepl/endpoint"
 	"github.com/zrepl/zrepl/platformtest"
@@ -29,13 +26,10 @@ import (
 // of a new sender and receiver instance and one blocking invocation
 // of the replication engine without encryption
 type replicationInvocation struct {
-	sjid, rjid        endpoint.JobID
-	sfs               string
-	sfilter           *filters.DatasetMapFilter
-	rfsRoot           string
-	interceptSender   func(e *endpoint.Sender) logic.Sender
-	interceptReceiver func(e *endpoint.Receiver) logic.Receiver
-	guarantee         pdu.ReplicationConfigProtection
+	sjid, rjid      endpoint.JobID
+	sfs             string
+	rfsRoot         string
+	interceptSender func(e *endpoint.Sender) logic.Sender
 }

 func (i replicationInvocation) Do(ctx *platformtest.Context) *report.Report {
@@ -43,33 +37,23 @@ func (i replicationInvocation) Do(ctx *platformtest.Context) *report.Report {
 	if i.interceptSender == nil {
 		i.interceptSender = func(e *endpoint.Sender) logic.Sender { return e }
 	}
-	if i.interceptReceiver == nil {
-		i.interceptReceiver = func(e *endpoint.Receiver) logic.Receiver { return e }
-	}

-	if i.sfs != "" && i.sfilter != nil || i.sfs == "" && i.sfilter == nil {
-		panic("either sfs or sfilter must be set")
-	}
-	if i.sfilter == nil {
-		i.sfilter = filters.NewDatasetMapFilter(1, true)
-		err := i.sfilter.Add(i.sfs, "ok")
-		require.NoError(ctx, err)
-	}
+	sfilter := filters.NewDatasetMapFilter(1, true)
+	err := sfilter.Add(i.sfs, "ok")
+	require.NoError(ctx, err)
 	sender := i.interceptSender(endpoint.NewSender(endpoint.SenderConfig{
-		FSF:     i.sfilter.AsFilter(),
+		FSF:     sfilter.AsFilter(),
 		Encrypt: &zfs.NilBool{B: false},
 		JobID:   i.sjid,
 	}))
-	receiver := i.interceptReceiver(endpoint.NewReceiver(endpoint.ReceiverConfig{
+	receiver := endpoint.NewReceiver(endpoint.ReceiverConfig{
 		JobID:                      i.rjid,
 		AppendClientIdentity:       false,
 		RootWithoutClientComponent: mustDatasetPath(i.rfsRoot),
-	}))
+		UpdateLastReceivedHold:     true,
+	})
 	plannerPolicy := logic.PlannerPolicy{
 		EncryptedSend: logic.TriFromBool(false),
-		ReplicationConfig: pdu.ReplicationConfig{
-			Protection: &i.guarantee,
-		},
 	}

 	report, wait := replication.Do(
@@ -102,11 +86,10 @@ func ReplicationIncrementalIsPossibleIfCommonSnapshotIsDestroyed(ctx *platformte
 	snap1 := fsversion(ctx, sfs, "@1")

 	rep := replicationInvocation{
-		sjid:      sjid,
-		rjid:      rjid,
-		sfs:       sfs,
-		rfsRoot:   rfsRoot,
-		guarantee: *pdu.ReplicationConfigProtectionWithKind(pdu.ReplicationGuaranteeKind_GuaranteeResumability),
+		sjid:    sjid,
+		rjid:    rjid,
+		sfs:     sfs,
+		rfsRoot: rfsRoot,
 	}
 	rfs := rep.ReceiveSideFilesystem()

@@ -166,11 +149,10 @@ func implReplicationIncrementalCleansUpStaleAbstractions(ctx *platformtest.Conte
 	rfsRoot := ctx.RootDataset + "/receiver"

 	rep := replicationInvocation{
-		sjid:      sjid,
-		rjid:      rjid,
-		sfs:       sfs,
-		rfsRoot:   rfsRoot,
-		guarantee: *pdu.ReplicationConfigProtectionWithKind(pdu.ReplicationGuaranteeKind_GuaranteeResumability),
+		sjid:    sjid,
+		rjid:    rjid,
+		sfs:     sfs,
+		rfsRoot: rfsRoot,
 	}
 	rfs := rep.ReceiveSideFilesystem()

@@ -196,7 +178,7 @@ func implReplicationIncrementalCleansUpStaleAbstractions(ctx *platformtest.Conte
 	require.NoError(ctx, err)
 	require.Contains(ctx, holds, expectRjidHoldTag)

-	// create artifical stale replication cursors & step holds
+	// create artifical stale replication cursors
 	createArtificalStaleAbstractions := func(jobId endpoint.JobID) []endpoint.Abstraction {
 		snap2Cursor, err := endpoint.CreateReplicationCursor(ctx, sfs, snap2, jobId) // no shadow
 		require.NoError(ctx, err)
@@ -220,7 +202,7 @@ func implReplicationIncrementalCleansUpStaleAbstractions(ctx *platformtest.Conte
 	snap5 := fsversion(ctx, sfs, "@5")

 	if invalidateCacheBeforeSecondReplication {
-		endpoint.AbstractionsCacheInvalidate(sfs)
+		endpoint.SendAbstractionsCacheInvalidate(sfs)
 	}

 	// do another replication
@@ -340,59 +322,7 @@ func (s *PartialSender) Send(ctx context.Context, r *pdu.SendReq) (r1 *pdu.SendR
 	return r1, r2, r3
 }

-func ReplicationIsResumableFullSend__both_GuaranteeResumability(ctx *platformtest.Context) {
-
-	setup := replicationIsResumableFullSendSetup{
-		protection: pdu.ReplicationConfigProtection{
-			Initial:     pdu.ReplicationGuaranteeKind_GuaranteeResumability,
-			Incremental: pdu.ReplicationGuaranteeKind_GuaranteeResumability,
-		},
-		expectDatasetIsBusyErrorWhenDestroySnapshotWhilePartiallyReplicated: true,
-		expectAllThreeSnapshotsToThreeBePresentAfterLoop:                    true,
-		expectNoSnapshotsOnReceiverAfterLoop:                                false,
-	}
-
-	implReplicationIsResumableFullSend(ctx, setup)
-}
-
-func ReplicationIsResumableFullSend__initial_GuaranteeResumability_incremental_GuaranteeIncrementalReplication(ctx *platformtest.Context) {
-
-	setup := replicationIsResumableFullSendSetup{
-		protection: pdu.ReplicationConfigProtection{
-			Initial:     pdu.ReplicationGuaranteeKind_GuaranteeResumability,
-			Incremental: pdu.ReplicationGuaranteeKind_GuaranteeIncrementalReplication,
-		},
-		expectDatasetIsBusyErrorWhenDestroySnapshotWhilePartiallyReplicated: true,
-		expectAllThreeSnapshotsToThreeBePresentAfterLoop:                    true,
-		expectNoSnapshotsOnReceiverAfterLoop:                                false,
-	}
-
-	implReplicationIsResumableFullSend(ctx, setup)
-}
-
-func ReplicationIsResumableFullSend__initial_GuaranteeIncrementalReplication_incremental_GuaranteeIncrementalReplication(ctx *platformtest.Context) {
-
-	setup := replicationIsResumableFullSendSetup{
-		protection: pdu.ReplicationConfigProtection{
-			Initial:     pdu.ReplicationGuaranteeKind_GuaranteeIncrementalReplication,
-			Incremental: pdu.ReplicationGuaranteeKind_GuaranteeIncrementalReplication,
-		},
-		expectDatasetIsBusyErrorWhenDestroySnapshotWhilePartiallyReplicated: false,
-		expectAllThreeSnapshotsToThreeBePresentAfterLoop:                    false,
-		expectNoSnapshotsOnReceiverAfterLoop:                                true,
-	}
-
-	implReplicationIsResumableFullSend(ctx, setup)
-}
-
-type replicationIsResumableFullSendSetup struct {
-	protection                                                          pdu.ReplicationConfigProtection
-	expectDatasetIsBusyErrorWhenDestroySnapshotWhilePartiallyReplicated bool
-	expectAllThreeSnapshotsToThreeBePresentAfterLoop                    bool
-	expectNoSnapshotsOnReceiverAfterLoop                                bool
-}
-
-func implReplicationIsResumableFullSend(ctx *platformtest.Context, setup replicationIsResumableFullSendSetup) {
+func ReplicationIsResumableFullSend(ctx *platformtest.Context) {

 	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
 		CREATEROOT
@@ -423,9 +353,7 @@ func implReplicationIsResumableFullSend(ctx *platformtest.Context, setup replica
 		interceptSender: func(e *endpoint.Sender) logic.Sender {
 			return &PartialSender{Sender: e, failAfterByteCount: 1 << 20}
 		},
-		guarantee: setup.protection,
 	}
-
 	rfs := rep.ReceiveSideFilesystem()

 	for i := 2; i < 10; i++ {
@@ -439,11 +367,8 @@ func implReplicationIsResumableFullSend(ctx *platformtest.Context, setup replica
 			// and we wrote dummy data 1<<22 bytes, thus at least
 			// for the first 4 times this should not be possible
 			// due to step holds
-			if setup.expectDatasetIsBusyErrorWhenDestroySnapshotWhilePartiallyReplicated {
-				ctx.Logf("i=%v", i)
-				require.Error(ctx, err)
-				require.Contains(ctx, err.Error(), "dataset is busy")
-			}
+			require.Error(ctx, err)
+			require.Contains(ctx, err.Error(), "dataset is busy")
 		}

 		// and create some additional snapshots that could
@@ -462,456 +387,10 @@ func implReplicationIsResumableFullSend(ctx *platformtest.Context, setup replica
 		}
 	}

-	if setup.expectAllThreeSnapshotsToThreeBePresentAfterLoop {
-		// make sure all the filesystem versions we created
-		// were replicated by the replication loop
-		_ = fsversion(ctx, rfs, "@1")
-		_ = fsversion(ctx, rfs, "@2")
-		_ = fsversion(ctx, rfs, "@3")
-	}
-
-	if setup.expectNoSnapshotsOnReceiverAfterLoop {
-		versions, err := zfs.ZFSListFilesystemVersions(ctx, mustDatasetPath(rfs), zfs.ListFilesystemVersionsOptions{})
-		require.NoError(ctx, err)
-		require.Empty(ctx, versions)
-	}
-
-}
-
-func ReplicationIncrementalDestroysStepHoldsIffIncrementalStepHoldsAreDisabledButStepHoldsExist(ctx *platformtest.Context) {
-
-	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
-		CREATEROOT
-		+  "sender"
-		+  "receiver"
-		R  zfs create -p "${ROOTDS}/receiver/${ROOTDS}"
-	`)
-
-	sjid := endpoint.MustMakeJobID("sender-job")
-	rjid := endpoint.MustMakeJobID("receiver-job")
-
-	sfs := ctx.RootDataset + "/sender"
-	rfsRoot := ctx.RootDataset + "/receiver"
-
-	// fully replicate snapshots @1
-	{
-		mustSnapshot(ctx, sfs+"@1")
-		rep := replicationInvocation{
-			sjid:      sjid,
-			rjid:      rjid,
-			sfs:       sfs,
-			rfsRoot:   rfsRoot,
-			guarantee: *pdu.ReplicationConfigProtectionWithKind(pdu.ReplicationGuaranteeKind_GuaranteeResumability),
-		}
-		rfs := rep.ReceiveSideFilesystem()
-		report := rep.Do(ctx)
-		ctx.Logf("\n%s", pretty.Sprint(report))
-		// assert this worked (not the main subject of the test)
-		_ = fsversion(ctx, rfs, "@1")
-	}
-
-	// create a large snapshot @2
-	{
-		sfsmp, err := zfs.ZFSGetMountpoint(ctx, sfs)
-		require.NoError(ctx, err)
-		require.True(ctx, sfsmp.Mounted)
-		writeDummyData(path.Join(sfsmp.Mountpoint, "dummy.data"), 1<<22)
-		mustSnapshot(ctx, sfs+"@2")
-	}
-	snap2sfs := fsversion(ctx, sfs, "@2")
-
-	// partially replicate snapshots @2 with step holds enabled
-	// to effect a step-holds situation
-	{
-		rep := replicationInvocation{
-			sjid:      sjid,
-			rjid:      rjid,
-			sfs:       sfs,
-			rfsRoot:   rfsRoot,
-			guarantee: *pdu.ReplicationConfigProtectionWithKind(pdu.ReplicationGuaranteeKind_GuaranteeResumability), // !
-			interceptSender: func(e *endpoint.Sender) logic.Sender {
-				return &PartialSender{Sender: e, failAfterByteCount: 1 << 20}
-			},
-		}
-		rfs := rep.ReceiveSideFilesystem()
-		report := rep.Do(ctx)
-		ctx.Logf("\n%s", pretty.Sprint(report))
-		// assert this partial receive worked
-		_, err := zfs.ZFSGetFilesystemVersion(ctx, rfs+"@2")
-		ctx.Logf("%T %s", err, err)
-		_, notFullyReceived := err.(*zfs.DatasetDoesNotExist)
-		require.True(ctx, notFullyReceived)
-		// assert step holds are in place
-		abs, absErrs, err := endpoint.ListAbstractions(ctx, endpoint.ListZFSHoldsAndBookmarksQuery{
-			FS: endpoint.ListZFSHoldsAndBookmarksQueryFilesystemFilter{
-				FS: &sfs,
-			},
-			Concurrency: 1,
-			JobID:       &sjid,
-			What:        endpoint.AbstractionTypeSet{endpoint.AbstractionStepHold: true},
-		})
-		require.NoError(ctx, err)
-		require.Empty(ctx, absErrs)
-		require.Len(ctx, abs, 2)
-		sort.Slice(abs, func(i, j int) bool {
-			return abs[i].GetCreateTXG() < abs[j].GetCreateTXG()
-		})
-		require.True(ctx, zfs.FilesystemVersionEqualIdentity(abs[0].GetFilesystemVersion(), fsversion(ctx, sfs, "@1")))
-		require.True(ctx, zfs.FilesystemVersionEqualIdentity(abs[1].GetFilesystemVersion(), fsversion(ctx, sfs, "@2")))
-	}
-
-	//
-	// end of test setup
-	//
-
-	// retry replication with incremental step holds disabled (set to bookmarks-only in this case)
-	// - replication should not fail due to holds-related stuff
-	// - replication should fail intermittently due to partial sender being fully read
-	// - the partial sender is 1/4th the length of the stream, thus expect
-	//   successful replication after 5 more attempts
-	rep := replicationInvocation{
-		sjid:      sjid,
-		rjid:      rjid,
-		sfs:       sfs,
-		rfsRoot:   rfsRoot,
-		guarantee: *pdu.ReplicationConfigProtectionWithKind(pdu.ReplicationGuaranteeKind_GuaranteeIncrementalReplication), // !
-		interceptSender: func(e *endpoint.Sender) logic.Sender {
-			return &PartialSender{Sender: e, failAfterByteCount: 1 << 20}
-		},
-	}
-	rfs := rep.ReceiveSideFilesystem()
-	for i := 0; ; i++ {
-		require.True(ctx, i < 5)
-		report := rep.Do(ctx)
-		ctx.Logf("retry run=%v\n%s", i, pretty.Sprint(report))
-		_, err := zfs.ZFSGetFilesystemVersion(ctx, rfs+"@2")
-		if err == nil {
-			break
-		}
-	}
-
-	// assert replication worked
-	fsversion(ctx, rfs, "@2")
-
-	// assert no step holds exist
-	abs, absErrs, err := endpoint.ListAbstractions(ctx, endpoint.ListZFSHoldsAndBookmarksQuery{
-		FS: endpoint.ListZFSHoldsAndBookmarksQueryFilesystemFilter{
-			FS: &sfs,
-		},
-		Concurrency: 1,
-		JobID:       &sjid,
-		What:        endpoint.AbstractionTypeSet{endpoint.AbstractionStepHold: true},
-	})
-	require.NoError(ctx, err)
-	require.Empty(ctx, absErrs)
-	require.Len(ctx, abs, 0)
-
-	// assert that the replication cursor bookmark exists
-	abs, absErrs, err = endpoint.ListAbstractions(ctx, endpoint.ListZFSHoldsAndBookmarksQuery{
-		FS: endpoint.ListZFSHoldsAndBookmarksQueryFilesystemFilter{
-			FS: &sfs,
-		},
-		Concurrency: 1,
-		JobID:       &sjid,
-		What:        endpoint.AbstractionTypeSet{endpoint.AbstractionReplicationCursorBookmarkV2: true},
-	})
-	require.NoError(ctx, err)
-	require.Empty(ctx, absErrs)
-	require.Len(ctx, abs, 1)
-	require.True(ctx, zfs.FilesystemVersionEqualIdentity(abs[0].GetFilesystemVersion(), snap2sfs))
-}
-
-func ReplicationStepCompletedLostBehavior__GuaranteeResumability(ctx *platformtest.Context) {
-	scenario := replicationStepCompletedLostBehavior_impl(ctx, pdu.ReplicationGuaranteeKind_GuaranteeResumability)
-
-	require.Error(ctx, scenario.deleteSfs1Err, "protected by holds")
-	require.Contains(ctx, scenario.deleteSfs1Err.Error(), "dataset is busy")
-
-	require.Error(ctx, scenario.deleteSfs2Err, "protected by holds")
-	require.Contains(ctx, scenario.deleteSfs2Err.Error(), "dataset is busy")
-
-	require.Nil(ctx, scenario.finalReport.Error())
-	_ = fsversion(ctx, scenario.rfs, "@3") // @3 ade it to the other side
-}
-
-func ReplicationStepCompletedLostBehavior__GuaranteeIncrementalReplication(ctx *platformtest.Context) {
-	scenario := replicationStepCompletedLostBehavior_impl(ctx, pdu.ReplicationGuaranteeKind_GuaranteeIncrementalReplication)
-
-	require.NoError(ctx, scenario.deleteSfs1Err, "not protected by holds")
-	require.NoError(ctx, scenario.deleteSfs2Err, "not protected by holds")
-
-	// step bookmarks should protect against loss of StepCompleted message
-	require.Nil(ctx, scenario.finalReport.Error())
-	_ = fsversion(ctx, scenario.rfs, "@3") // @3 ade it to the other side
-}
-
-type FailSendCompletedSender struct {
-	*endpoint.Sender
-}
-
-var _ logic.Sender = (*FailSendCompletedSender)(nil)
-
-func (p *FailSendCompletedSender) SendCompleted(ctx context.Context, r *pdu.SendCompletedReq) (*pdu.SendCompletedRes, error) {
-	return nil, fmt.Errorf("[mock] SendCompleted not delivered to actual endpoint")
-}
-
-type replicationStepCompletedLost_scenario struct {
-	rfs                          string
-	deleteSfs1Err, deleteSfs2Err error
-	finalReport                  *report.FilesystemReport
-}
-
-func replicationStepCompletedLostBehavior_impl(ctx *platformtest.Context, guaranteeKind pdu.ReplicationGuaranteeKind) *replicationStepCompletedLost_scenario {
-
-	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
-		CREATEROOT
-		+  "sender"
-		+  "receiver"
-		R  zfs create -p "${ROOTDS}/receiver/${ROOTDS}"
-	`)
-
-	sjid := endpoint.MustMakeJobID("sender-job")
-	rjid := endpoint.MustMakeJobID("receiver-job")
-
-	sfs := ctx.RootDataset + "/sender"
-	rfsRoot := ctx.RootDataset + "/receiver"
-
-	// fully replicate snapshots @1
-	{
-		mustSnapshot(ctx, sfs+"@1")
-		rep := replicationInvocation{
-			sjid:      sjid,
-			rjid:      rjid,
-			sfs:       sfs,
-			rfsRoot:   rfsRoot,
-			guarantee: *pdu.ReplicationConfigProtectionWithKind(guaranteeKind),
-		}
-		rfs := rep.ReceiveSideFilesystem()
-		report := rep.Do(ctx)
-		ctx.Logf("\n%s", pretty.Sprint(report))
-		// assert this worked (not the main subject of the test)
-		_ = fsversion(ctx, rfs, "@1")
-	}
-
-	// create a second snapshot @2
-	mustSnapshot(ctx, sfs+"@2")
-
-	// fake loss of stepcompleted message
-	rep := replicationInvocation{
-		sjid:      sjid,
-		rjid:      rjid,
-		sfs:       sfs,
-		rfsRoot:   rfsRoot,
-		guarantee: *pdu.ReplicationConfigProtectionWithKind(guaranteeKind),
-		interceptSender: func(e *endpoint.Sender) logic.Sender {
-			return &FailSendCompletedSender{e}
-		},
-	}
-	rfs := rep.ReceiveSideFilesystem()
-	report := rep.Do(ctx)
-	ctx.Logf("\n%s", pretty.Sprint(report))
-
-	// assert the replication worked
+	// make sure all the filesystem versions we created
+	// were replicated by the replication loop
+	_ = fsversion(ctx, rfs, "@1")
 	_ = fsversion(ctx, rfs, "@2")
-	// and that we hold it using a last-received-hold
-	abs, absErrs, err := endpoint.ListAbstractions(ctx, endpoint.ListZFSHoldsAndBookmarksQuery{
-		FS: endpoint.ListZFSHoldsAndBookmarksQueryFilesystemFilter{
-			FS: &rfs,
-		},
-		Concurrency: 1,
-		JobID:       &rjid,
-		What:        endpoint.AbstractionTypeSet{endpoint.AbstractionLastReceivedHold: true},
-	})
-	require.NoError(ctx, err)
-	require.Empty(ctx, absErrs)
-	require.Len(ctx, abs, 1)
-	require.True(ctx, zfs.FilesystemVersionEqualIdentity(abs[0].GetFilesystemVersion(), fsversion(ctx, rfs, "@2")))
-
-	// now try to delete @2 on the sender, this should work because don't have step holds on it
-	deleteSfs2Err := zfs.ZFSDestroy(ctx, sfs+"@2")
-	// defer check to caller
-
-	// and create a new snapshot on the sender
-	mustSnapshot(ctx, sfs+"@3")
-
-	// now we have: sender @1, @3
-	//              recver @1, @2
-
-	// delete @1 on both sides to demonstrate that, if we didn't have bookmarks, we would be out of sync
-	deleteSfs1Err := zfs.ZFSDestroy(ctx, sfs+"@1")
-	// defer check to caller
-	err = zfs.ZFSDestroy(ctx, rfs+"@1")
-	require.NoError(ctx, err)
-
-	// attempt replication and return the filesystem report report
-	{
-		rep := replicationInvocation{
-			sjid:      sjid,
-			rjid:      rjid,
-			sfs:       sfs,
-			rfsRoot:   rfsRoot,
-			guarantee: *pdu.ReplicationConfigProtectionWithKind(guaranteeKind),
-		}
-		report := rep.Do(ctx)
-		ctx.Logf("expecting failure:\n%s", pretty.Sprint(report))
-		require.Len(ctx, report.Attempts, 1)
-		require.Len(ctx, report.Attempts[0].Filesystems, 1)
-		return &replicationStepCompletedLost_scenario{
-			rfs:           rfs,
-			deleteSfs1Err: deleteSfs1Err,
-			deleteSfs2Err: deleteSfs2Err,
-			finalReport:   report.Attempts[0].Filesystems[0],
-		}
-	}
+	_ = fsversion(ctx, rfs, "@3")

 }
-
-type ErroringReceiver struct {
-	recvErr error
-	*endpoint.Receiver
-}
-
-func (r *ErroringReceiver) Receive(ctx context.Context, req *pdu.ReceiveReq, stream io.ReadCloser) (*pdu.ReceiveRes, error) {
-	return nil, r.recvErr
-}
-
-type NeverEndingSender struct {
-	*endpoint.Sender
-}
-
-func (s *NeverEndingSender) Send(ctx context.Context, req *pdu.SendReq) (r *pdu.SendRes, stream io.ReadCloser, _ error) {
-	stream = nil
-	r = &pdu.SendRes{
-		UsedResumeToken: false,
-		ExpectedSize:    1 << 30,
-	}
-	if req.DryRun {
-		return r, stream, nil
-	}
-	dz, err := os.Open("/dev/zero")
-	if err != nil {
-		panic(err)
-	}
-	return r, dz, nil
-}
-
-func ReplicationReceiverErrorWhileStillSending(ctx *platformtest.Context) {
-
-	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
-		CREATEROOT
-		+  "sender"
-		+  "sender@1"
-		+  "receiver"
-		R  zfs create -p "${ROOTDS}/receiver/${ROOTDS}"
-	`)
-
-	sjid := endpoint.MustMakeJobID("sender-job")
-	rjid := endpoint.MustMakeJobID("receiver-job")
-
-	sfs := ctx.RootDataset + "/sender"
-	rfsRoot := ctx.RootDataset + "/receiver"
-
-	mockRecvErr := fmt.Errorf("YiezahK3thie8ahKiel5sah2uugei2ize1yi8feivuu7musoat")
-
-	rep := replicationInvocation{
-		sjid:      sjid,
-		rjid:      rjid,
-		sfs:       sfs,
-		rfsRoot:   rfsRoot,
-		guarantee: *pdu.ReplicationConfigProtectionWithKind(pdu.ReplicationGuaranteeKind_GuaranteeNothing),
-		interceptReceiver: func(r *endpoint.Receiver) logic.Receiver {
-			return &ErroringReceiver{recvErr: mockRecvErr, Receiver: r}
-		},
-		interceptSender: func(s *endpoint.Sender) logic.Sender {
-			return &NeverEndingSender{s}
-		},
-	}
-
-	// first replication
-	report := rep.Do(ctx)
-	ctx.Logf("\n%s", pretty.Sprint(report))
-
-	require.Len(ctx, report.Attempts, 1)
-	attempt := report.Attempts[0]
-	require.Nil(ctx, attempt.PlanError)
-	require.Len(ctx, attempt.Filesystems, 1)
-	afs := attempt.Filesystems[0]
-	require.Nil(ctx, afs.PlanError)
-	require.Len(ctx, afs.Steps, 1)
-	require.Nil(ctx, afs.PlanError)
-	require.NotNil(ctx, afs.StepError)
-	require.Contains(ctx, afs.StepError.Err, mockRecvErr.Error())
-}
-
-func ReplicationFailingInitialParentProhibitsChildReplication(ctx *platformtest.Context) {
-
-	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
-		CREATEROOT
-		+  "sender"
-		+  "sender/a"
-		+  "sender/a/child"
-		+  "sender/aa"
-		+  "receiver"
-		R  zfs create -p "${ROOTDS}/receiver/${ROOTDS}"
-		R  zfs snapshot -r ${ROOTDS}/sender@initial
-	`)
-
-	sjid := endpoint.MustMakeJobID("sender-job")
-	rjid := endpoint.MustMakeJobID("receiver-job")
-
-	fsA := ctx.RootDataset + "/sender/a"
-	fsAChild := ctx.RootDataset + "/sender/a/child"
-	fsAA := ctx.RootDataset + "/sender/aa"
-
-	sfilter := filters.NewDatasetMapFilter(3, true)
-	mustAddToSFilter := func(fs string) {
-		err := sfilter.Add(fs, "ok")
-		require.NoError(ctx, err)
-	}
-	mustAddToSFilter(fsA)
-	mustAddToSFilter(fsAChild)
-	mustAddToSFilter(fsAA)
-	rfsRoot := ctx.RootDataset + "/receiver"
-
-	mockRecvErr := fmt.Errorf("yifae4ohPhaquaes0hohghiep9oufie4roo7quoWooluaj2ee8")
-
-	rep := replicationInvocation{
-		sjid:      sjid,
-		rjid:      rjid,
-		sfilter:   sfilter,
-		rfsRoot:   rfsRoot,
-		guarantee: *pdu.ReplicationConfigProtectionWithKind(pdu.ReplicationGuaranteeKind_GuaranteeNothing),
-		interceptReceiver: func(r *endpoint.Receiver) logic.Receiver {
-			return &ErroringReceiver{recvErr: mockRecvErr, Receiver: r}
-		},
-	}
-
-	r := rep.Do(ctx)
-	ctx.Logf("\n%s", pretty.Sprint(r))
-
-	require.Len(ctx, r.Attempts, 1)
-	attempt := r.Attempts[0]
-	require.Nil(ctx, attempt.PlanError)
-	require.Len(ctx, attempt.Filesystems, 3)
-
-	fsByName := make(map[string]*report.FilesystemReport, len(attempt.Filesystems))
-	for _, fs := range attempt.Filesystems {
-		fsByName[fs.Info.Name] = fs
-	}
-
-	require.Contains(ctx, fsByName, fsA)
-	require.Contains(ctx, fsByName, fsAA)
-	require.Contains(ctx, fsByName, fsAA)
-
-	checkFS := func(fs string, expectErrMsg string) {
-		rep := fsByName[fs]
-		require.Len(ctx, rep.Steps, 1)
-		require.Nil(ctx, rep.PlanError)
-		require.NotNil(ctx, rep.StepError)
-		require.Contains(ctx, rep.StepError.Err, expectErrMsg)
-	}
-
-	checkFS(fsA, mockRecvErr.Error())
-	checkFS(fsAChild, "parent(s) failed during initial replication")
-	checkFS(fsAA, mockRecvErr.Error()) // fsAA is not treated as a child of fsA
-}
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Christian Schwarz	4a0104a44f	WIP endpoint abstractions + pruning integration / pruner rewrite	2020-05-29 00:09:43 +02:00
Christian Schwarz	f0146d03d3	docs: update supporters	2020-05-24 18:39:02 +02:00
Christian Schwarz	742e28d2a3	zfs: ZFSListFilesystemVersions: remove handling of io.ErrUnexpectedEOF ZFSListChan returns (*DatasetDoesNotExist) for the case mentioned in the comment	2020-05-24 18:39:02 +02:00
Christian Schwarz	58f364dc04	zfs: fix error message formatting for send args validation	2020-05-24 18:39:02 +02:00
Christian Schwarz	a5c9450b2d	[#321 ] platformtest: add test for zfs.ZFSHolds	2020-05-24 18:39:02 +02:00
Christian Schwarz	b222574ff3	[#321 ] endpoint: ListAbstractions: acutally emit one Abstraction per matching hold	2020-05-24 18:39:02 +02:00
Christian Schwarz	28e1c486f5	[#321 ] platformtest: minimal integration tests for package replication # Conflicts: # platformtest/tests/generated_cases.go	2020-05-24 18:39:02 +02:00
Christian Schwarz	82370f3555	[#321 ] platformtest: generate test case list + coverage tooling	2020-05-24 18:39:02 +02:00
Christian Schwarz	4f6c57fbba	[#321 ] platformtest: fix test ListFilesystemVersionsZeroExistIsNotAnError	2020-05-24 18:39:02 +02:00
Christian Schwarz	5a558e3b71	[#316 ] endpoint: delete unreachable code	2020-05-24 18:09:33 +02:00
Christian Schwarz	b33f670b9d	[#316 ] endpoint / replication protocol: more robust step-holds and replication cursor management - drop HintMostRecentCommonAncestor rpc call - it is wrong to put faith into the active side of the replication to always make that call (we might not trust it, ref pull setup) - clean up step holds + step bookmarks + replication cursor bookmarks on send RPC instead - this makes it symmetric with Receive RPC - use a cache (endpoint.sendAbstractionsCache) to avoid the cost of listing the on-disk endpoint abstractions state on every step The "create" methods for endpoint abstractions (CreateReplicationCursor, HoldStep) are now fully idempotent and return an Abstraction. Notes about endpoint.sendAbstractionsCache: - fills lazily from disk state on first `Get` operation - fill from disk is generally only attempted once - unless the `ListAbstractions` fails, in which case the fill from disk is retried on next `Get` (the current `Get` will observe a subset of the actual on-disk abstractions) - the `Invalidate` method is called - it is a global (zrepl process-wide) cache fixes #316	2020-05-19 11:30:02 +02:00