transport/local: configurable dial_timeout for connect, default 2s

* Add Close() in closeState to identify the first closer
* Non-first closers get an error
* Reads and Writes from the Conn get an error if the conn was closed
  during the Read / Write was running
* The first closer starts _separate_ goroutine draining the c.frameReads channel
* The first closer then waits for the goroutine that fills c.frameReads
  to exit

refs 3bfe0c16d0
fixes #174

readFrames would block on `reads <-`
   but only after that would stream.Conn.readFrames close c.waitReadFramesDone
   which was too late because stream.Conn.Close would wait for c.waitReadFramesDone to be closed before draining the channel
                              ^^^^^^ (not frameconn.Conn, that closed successfully)

   195 @ 0x1032ae0 0x1006cab 0x1006c81 0x1006a65 0x15505be 0x155163e 0x1060bc1
           0x15505bd       github.com/zrepl/zrepl/rpc/dataconn/stream.readFrames+0x16d             github.com/zrepl/zrepl/rpc/dataconn/stream/stream.go:220
           0x155163d       github.com/zrepl/zrepl/rpc/dataconn/stream.(*Conn).readFrames+0xbd      github.com/zrepl/zrepl/rpc/dataconn/stream/stream_conn.go:71

   195 @ 0x1032ae0 0x10078c8 0x100789e 0x100758b 0x1552678 0x1557a4b 0x1556aec 0x1060bc1
           0x1552677       github.com/zrepl/zrepl/rpc/dataconn/stream.(*Conn).Close+0x77           github.com/zrepl/zrepl/rpc/dataconn/stream/stream_conn.go:191
           0x1557a4a       github.com/zrepl/zrepl/rpc/dataconn.(*Server).serveConn.func1+0x5a      github.com/zrepl/zrepl/rpc/dataconn/dataconn_server.go:93
           0x1556aeb       github.com/zrepl/zrepl/rpc/dataconn.(*Server).serveConn+0x87b           github.com/zrepl/zrepl/rpc/dataconn/dataconn_server.go:176

.circleci/config.yml

@@ -5,6 +5,7 @@ workflows:
     jobs:
       - build-1.11
       - build-1.12
+      - build-1.13
       - build-latest
 jobs:
   # build-latest serves as the template
@@ -27,11 +28,12 @@ jobs:
           command: |
             # used by pip (for docs)
             echo 'export PATH="$HOME/.local/bin:$PATH"' >> $BASH_ENV
+            # we use modules
+            echo 'export GO111MODULE=on' >> $BASH_ENV
 
       - restore_cache:
           keys:
             - source
-            - vendor
             - protobuf
 
       - checkout
@@ -50,22 +52,9 @@ jobs:
           paths:
             - "/usr/include/google/protobuf"
             
-      - run: sudo apt install python3 python3-pip libgirepository1.0-dev
+      - run: sudo apt update && sudo apt install python3 python3-pip libgirepository1.0-dev
       - run: ./lazy.sh devsetup
 
-      - run:
-          name: Install minio client, prepare minio
-          command: |
-            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}
-
-      - run: make vendordeps
-      - save_cache:
-          key: vendor
-          paths:
-            - "./vendor"
-            
       - run: make
       - run: make vet
       - run: make test
@@ -78,11 +67,25 @@ jobs:
           when: always
 
       - run:
-          shell: /bin/bash -euo pipefail
+          shell: /bin/bash -eo pipefail
           when: always
           command: |
+            if [ -n "$CIRCLE_PR_NUMBER" ]; then # CIRCLE_PR_NUMBER is guaranteed to be only present in forked PRs (external)
+              echo "Forked PR detected. Sry, can't trust you with credentials to external artifact store, use CircleCI's instead."
+              exit 0
+            fi
+            set -u # from now on
+            
+            # Download and install minio
+            curl -sSL https://dl.minio.io/client/mc/release/linux-amd64/mc -o ${GOPATH}/bin/mc
+            chmod +x ${GOPATH}/bin/mc
+            mc config host add --api s3v4 zrepl-minio https://minio.cschwarz.com ${MINIO_ACCESS_KEY} ${MINIO_SECRET_KEY}
+            
+            # Upload artifacts
             echo "$CIRCLE_BUILD_URL" > ./artifacts/release/cirlceci_build_url
             mc cp -r artifacts/release "zrepl-minio/zrepl-ci-artifacts/${CIRCLE_SHA1}/${CIRCLE_JOB}/"
+
+            # Push Artifact Link to GitHub
             REPO="zrepl/zrepl"
             COMMIT="${CIRCLE_SHA1}"
             JOB_NAME="${CIRCLE_JOB}"
@@ -91,12 +94,17 @@ jobs:
               -X POST \
               -d '{"context":"zrepl/publish-ci-artifacts", "state": "success", "description":"CI Build Artifacts for '"$JOB_NAME"'", "target_url":"https://minio.cschwarz.com/minio/zrepl-ci-artifacts/'"$COMMIT"'/"}'
 
-
   build-1.11:
     <<: *build-latest
     docker:
-      - image: circleci/golang:1.11
+    - image: circleci/golang:1.11
+
   build-1.12:
     <<: *build-latest
     docker:
     - image: circleci/golang:1.12
+
+  build-1.13:
+    <<: *build-latest
+    docker:
+    - image: circleci/golang:1.13

.github/FUNDING.yml

@@ -0,0 +1,4 @@
+patreon: zrepl
+liberapay: zrepl
+custom: https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96
+ko_fi: problame

Gopkg.toml

@@ -1,67 +0,0 @@
-ignored = [
-    "github.com/inconshreveable/mousetrap",
-]
-
-required = [
-    "golang.org/x/tools/cmd/stringer",
-    "github.com/alvaroloes/enumer",
-    "github.com/golangci/golangci-lint/cmd/golangci-lint",
-    "golang.org/x/tools/cmd/goimports",
-]
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/jinzhu/copier"
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/kr/pretty"
-
-[[constraint]]
-  name = "github.com/pkg/errors"
-  version = "0.8.0"
-
-[[constraint]]
-  branch = "master"
-  name = "github.com/spf13/cobra"
-
-[[constraint]]
-  name = "github.com/stretchr/testify"
-  version = "1.1.4"
-
-[[constraint]]
-  branch = "v2"
-  name = "github.com/zrepl/yaml-config"
-
-[[constraint]]
-  name = "github.com/go-logfmt/logfmt"
-  version = "*"
-
-[[constraint]]
-  name = "github.com/problame/go-netssh"
-  branch = "master"
-
-[[constraint]]
-  name = "github.com/prometheus/client_golang"
-  branch = "master"
-
-[[constraint]]
-  name = "github.com/golang/protobuf"
-  version = "1"
-
-[[constraint]]
-  name = "github.com/fatih/color"
-  version = "1.7.0"
-
-
-[[constraint]]
-  name = "github.com/gdamore/tcell"
-  version = "1.0.0"
-
-[[constraint]]
-  name = "google.golang.org/grpc"
-  version = "1"
-  
-[[constraint]]
-  version = "1.1.0"
-  name = "github.com/google/uuid"

Makefile

@@ -1,4 +1,4 @@
-.PHONY: generate build test vet cover release docs docs-clean clean vendordeps format lint
+.PHONY: generate build test vet cover release docs docs-clean clean format lint platformtest
 .DEFAULT_GOAL := build
 
 ARTIFACTDIR := artifacts
@@ -12,23 +12,24 @@ ifndef _ZREPL_VERSION
         $(error cannot infer variable ZREPL_VERSION using git and variable is not overriden by make invocation)
     endif
 endif
+GO := go
 GO_LDFLAGS := "-X github.com/zrepl/zrepl/version.zreplVersion=$(_ZREPL_VERSION)"
+GO_MOD_READONLY := -mod=readonly
+GO_BUILDFLAGS := $(GO_MOD_READONLY)
+GO_BUILD := GO111MODULE=on $(GO) build $(GO_BUILDFLAGS) -v -ldflags $(GO_LDFLAGS)
 
-GO_BUILD := go build -ldflags $(GO_LDFLAGS)
-
+# keep in sync with vet target
 RELEASE_BINS := $(ARTIFACTDIR)/zrepl-freebsd-amd64
 RELEASE_BINS += $(ARTIFACTDIR)/zrepl-linux-amd64
+RELEASE_BINS += $(ARTIFACTDIR)/zrepl-linux-arm64
 RELEASE_BINS += $(ARTIFACTDIR)/zrepl-darwin-amd64
 
 RELEASE_NOARCH := $(ARTIFACTDIR)/zrepl-noarch.tar
-THIS_PLATFORM_RELEASE_BIN := $(shell bash -c 'source <(go env) && echo "zrepl-$${GOOS}-$${GOARCH}"' )
-
-vendordeps:
-	dep ensure -v -vendor-only
+THIS_PLATFORM_RELEASE_BIN := $(shell bash -c 'source <($(GO) env) && echo "zrepl-$${GOOS}-$${GOARCH}"' )
 
 generate: #not part of the build, must do that manually
 	protoc -I=replication/logic/pdu --go_out=plugins=grpc:replication/logic/pdu replication/logic/pdu/pdu.proto
-	go generate -x ./...
+	$(GO) generate $(GO_BUILDFLAGS) -x ./...
 
 format:
 	goimports -srcdir . -local 'github.com/zrepl/zrepl' -w $(shell find . -type f -name '*.go' -not -path "./vendor/*" -not -name '*.pb.go' -not -name '*_enumer.go')
@@ -37,17 +38,28 @@ lint:
 	golangci-lint run ./...
 
 build:
-	@echo "INFO: In case of missing dependencies, run 'make vendordeps'"
 	$(GO_BUILD) -o "$(ARTIFACTDIR)/zrepl"
 
 test:
-	go test ./...
+	$(GO) test $(GO_BUILDFLAGS) ./...
+	# TODO compile the tests for each supported platform
+	# but `go test -c ./...` is not supported
 
 vet:
+	$(GO) vet $(GO_BUILDFLAGS) ./...
 	# for each supported platform to cover conditional compilation
-	GOOS=linux   go vet ./...
-	GOOS=darwin  go vet ./...
-	GOOS=freebsd go vet ./...
+	# (keep in sync with RELEASE_BINS)
+	GOOS=freebsd	GOARCH=amd64 	$(GO) vet $(GO_BUILDFLAGS) ./...
+	GOOS=linux		GOARCH=amd64 	$(GO) vet $(GO_BUILDFLAGS) ./...
+	GOOS=linux		GOARCH=arm64 	$(GO) vet $(GO_BUILDFLAGS) ./...
+	GOOS=darwin		GOARCH=amd64 	$(GO) vet $(GO_BUILDFLAGS) ./...
+
+ZREPL_PLATFORMTEST_POOLNAME := zreplplatformtest
+ZREPL_PLATFORMTEST_IMAGEPATH := /tmp/zreplplatformtest.pool.img
+$(ARTIFACTDIR)/zrepl_platformtest:
+	$(GO_BUILD) -o "$(ARTIFACTDIR)/zrepl_platformtest" ./platformtest/harness
+platformtest: $(ARTIFACTDIR)/zrepl_platformtest
+	"$(ARTIFACTDIR)/zrepl_platformtest" -poolname "$(ZREPL_PLATFORMTEST_POOLNAME)" -imagepath "$(ZREPL_PLATFORMTEST_IMAGEPATH)"
 
 $(ARTIFACTDIR):
 	mkdir -p "$@"
@@ -58,8 +70,9 @@ $(ARTIFACTDIR)/docs: $(ARTIFACTDIR)
 $(ARTIFACTDIR)/bash_completion: $(RELEASE_BINS)
 	artifacts/$(THIS_PLATFORM_RELEASE_BIN) bashcomp "$@"
 
+.PHONY: $(ARTIFACTDIR)/go_version.txt
 $(ARTIFACTDIR)/go_version.txt:
-	go version > $@
+	$(GO) version > $@
 
 docs: $(ARTIFACTDIR)/docs
 	make -C docs \
@@ -71,12 +84,11 @@ docs-clean:
 		clean \
 		BUILDDIR=../artifacts/docs
 
-
 .PHONY: $(RELEASE_BINS)
 # TODO: two wildcards possible
-$(RELEASE_BINS): $(ARTIFACTDIR)/zrepl-%-amd64: generate $(ARTIFACTDIR) vet test lint
-	@echo "INFO: In case of missing dependencies, run 'make vendordeps'"
-	GOOS=$* GOARCH=amd64   $(GO_BUILD) -o "$(ARTIFACTDIR)/zrepl-$*-amd64"
+$(RELEASE_BINS): $(ARTIFACTDIR)/zrepl-%: generate $(ARTIFACTDIR) vet test lint
+	STEM=$*; GOOS="$${STEM%%-*}"; GOARCH="$${STEM##*-}"; export GOOS GOARCH; \
+		$(GO_BUILD) -o "$(ARTIFACTDIR)/zrepl-$$GOOS-$$GOARCH"
 
 $(RELEASE_NOARCH): docs $(ARTIFACTDIR)/bash_completion $(ARTIFACTDIR)/go_version.txt
 	tar --mtime='1970-01-01' --sort=name \
@@ -96,6 +108,9 @@ release: $(RELEASE_BINS) $(RELEASE_NOARCH)
         echo '[INFO] either git reports checkout is dirty or git is not installed or this is not a git checkout'; \
 		if [ "$(ZREPL_VERSION)" = "" ]; then \
 			echo '[WARN] git checkout is dirty and make variable ZREPL_VERSION was not used to override'; \
+			git status; \
+			echo "git diff:";  \
+			git diff | cat; \
 			exit 1; \
 		fi; \
 	fi;

README.md

@@ -2,7 +2,7 @@
 [![Language: Go](https://img.shields.io/badge/language-Go-6ad7e5.svg)](https://golang.org/)
 [![User Docs](https://img.shields.io/badge/docs-web-blue.svg)](https://zrepl.github.io)
 [![Donate via PayPal](https://img.shields.io/badge/donate-paypal-yellow.svg)](https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96)
-[![Donate via Liberapay](https://img.shields.io/badge/donate-liberapay-yellow.svg)](https://liberapay.com/zrepl/donate)
+[![Donate via Liberapay](https://img.shields.io/liberapay/receives/zrepl.svg?logo=liberapay)](https://liberapay.com/zrepl/donate)
 [![Twitter](https://img.shields.io/twitter/url/https/github.com/zrepl/zrepl.svg?style=social)](https://twitter.com/intent/tweet?text=Wow:&url=https%3A%2F%2Fgithub.com%2Fzrepl%2Fzrepl)
 
 # zrepl
@@ -37,11 +37,12 @@ Check out the *Coding Workflow* section below for details.
 * Ship other material provided in `./dist`, e.g. in `/usr/share/zrepl/`.
 * Use `make release ZREPL_VERSION='mydistro-1.2.3_1'`
     * Your distro's name and any versioning supplemental to zrepl's (e.g. package revision) should be in this string
+* Use `make platformtest` **on a test system** to validate that zrepl's abstractions on top of ZFS work with the system ZFS.
 * Make sure you are informed about new zrepl versions, e.g. by subscribing to GitHub's release RSS feed.
 
 ## Developer Documentation
 
-zrepl is written in [Go](https://golang.org) and uses [`dep`](https://github.com/golang/dep) to manage dependencies.
+zrepl is written in [Go](https://golang.org) and uses [Go modules](https://github.com/golang/go/wiki/Modules) to manage dependencies.
 The documentation is written in [ReStructured Text](http://docutils.sourceforge.net/rst.html) using the [Sphinx](https://www.sphinx-doc.org) framework.
 
 To get started, run `./lazy.sh devsetup` to easily install build dependencies and read `docs/installation.rst -> Compiling from Source`.

build.Dockerfile

@@ -4,9 +4,8 @@ RUN apt-get update && apt-get install -y \
     python3-pip \
     unzip
 
-RUN wget https://github.com/protocolbuffers/protobuf/releases/download/v3.6.1/protoc-3.6.1-linux-x86_64.zip
-RUN echo "6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip" | sha256sum -c
-RUN unzip -d /usr protoc-3.6.1-linux-x86_64.zip
+ADD build.installprotoc.bash ./
+RUN bash build.installprotoc.bash
 
 ADD lazy.sh /tmp/lazy.sh
 ADD docs/requirements.txt /tmp/requirements.txt
@@ -14,18 +13,17 @@ ENV ZREPL_LAZY_DOCS_REQPATH=/tmp/requirements.txt
 RUN /tmp/lazy.sh docdep
 
 # prepare volume mount of git checkout to /zrepl
-RUN mkdir -p /go/src/github.com/zrepl/zrepl
+RUN mkdir -p /src/github.com/zrepl/zrepl
 RUN mkdir -p /.cache && chmod -R 0777 /.cache
-WORKDIR /go/src/github.com/zrepl/zrepl
 
-ADD Gopkg.toml Gopkg.lock  ./
+# $GOPATH is /go
+# Go 1.12 doesn't use modules within GOPATH, but 1.13 and later do
+# => store source outside of GOPATH
+WORKDIR /src
 
-# godep will install the Go dependencies to vendor in order to then build and install
-# build dependencies like stringer to $GOPATH/bin.
-# However, since users volume-mount their Git checkout into /go/src/github.com/zrepl/zrepl
-# the vendor directory will be empty at build time, allowing them to experiment with
-# new checkouts, etc.
-# Thus, we only use the vendored deps for building dependencies.
+# Install build tools (e.g. protobuf generator, stringer) into $GOPATH/bin
+ADD go.mod go.sum ./
 RUN /tmp/lazy.sh godep
+
 RUN chmod -R 0777 /go
 

build.installprotoc.bash

@@ -0,0 +1,25 @@
+#!/usr/bin/env bash
+set -euo pipefail
+set -x
+
+MACH=$(uname -m)
+MACH="${MACH/aarch64/aarch_64}"
+
+VERSION=3.6.1
+FILENAME=protoc-"$VERSION"-linux-"$MACH".zip
+
+if [ -e "$FILENAME" ]; then
+    echo "$FILENAME" already exists 1>&2
+    exit 1
+fi
+
+wget https://github.com/protocolbuffers/protobuf/releases/download/v"$VERSION"/"$FILENAME"
+
+stat "$FILENAME"
+
+sha256sum -c --ignore-missing <<EOF
+6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip
+af8e5aaaf39ddec62ec8dd2be1b8d9602c6da66564883a16393ade5f71170922  protoc-3.6.1-linux-aarch_64.zip
+EOF
+
+unzip -d /usr "$FILENAME"

client/status.go

@@ -11,7 +11,14 @@ import (
 	"sync"
 	"time"
 
+	// tcell is the termbox-compatbile library for abstracting away escape sequences, etc.
+	// as of tcell#252, the number of default distributed terminals is relatively limited
+	// additional terminal definitions can be included via side-effect import
+	// See https://github.com/gdamore/tcell/blob/master/terminfo/base/base.go
+	// See https://github.com/gdamore/tcell/issues/252#issuecomment-533836078
 	"github.com/gdamore/tcell/termbox"
+	_ "github.com/gdamore/tcell/terminfo/s/screen" // tmux on FreeBSD 11 & 12 without ncurses
+
 	"github.com/pkg/errors"
 	"github.com/spf13/pflag"
 	"github.com/zrepl/yaml-config"
@@ -20,6 +27,7 @@ import (
 	"github.com/zrepl/zrepl/daemon"
 	"github.com/zrepl/zrepl/daemon/job"
 	"github.com/zrepl/zrepl/daemon/pruner"
+	"github.com/zrepl/zrepl/daemon/snapper"
 	"github.com/zrepl/zrepl/replication/report"
 )
 
@@ -76,6 +84,8 @@ type tui struct {
 	report map[string]job.Status
 	err    error
 
+	jobFilter string
+
 	replicationProgress map[string]*bytesProgressHistory // by job name
 }
 
@@ -157,6 +167,7 @@ func (t *tui) addIndent(indent int) {
 
 var statusFlags struct {
 	Raw bool
+	Job string
 }
 
 var StatusCmd = &cli.Subcommand{
@@ -164,6 +175,7 @@ var StatusCmd = &cli.Subcommand{
 	Short: "show job activity or dump as JSON for monitoring",
 	SetupFlags: func(f *pflag.FlagSet) {
 		f.BoolVar(&statusFlags.Raw, "raw", false, "dump raw status description from zrepl daemon")
+		f.StringVar(&statusFlags.Job, "job", "", "only dump specified job")
 	},
 	Run: runStatus,
 }
@@ -198,6 +210,7 @@ func runStatus(s *cli.Subcommand, args []string) error {
 	t.lock.Lock()
 	t.err = errors.New("Got no report yet")
 	t.lock.Unlock()
+	t.jobFilter = statusFlags.Job
 
 	err = termbox.Init()
 	if err != nil {
@@ -273,19 +286,29 @@ func (t *tui) draw() {
 		t.write(t.err.Error())
 	} else {
 		//Iterate over map in alphabetical order
-		keys := make([]string, len(t.report))
-		i := 0
+		keys := make([]string, 0, len(t.report))
 		for k := range t.report {
-			keys[i] = k
-			i++
-		}
-		sort.Strings(keys)
-
-		for _, k := range keys {
-			v := t.report[k]
 			if len(k) == 0 || daemon.IsInternalJobName(k) { //Internal job
 				continue
 			}
+			if t.jobFilter != "" && k != t.jobFilter {
+				continue
+			}
+			keys = append(keys, k)
+		}
+		sort.Strings(keys)
+
+		if len(keys) == 0 {
+			t.setIndent(0)
+			t.printf("no jobs to display")
+			t.newline()
+			termbox.Flush()
+			return
+		}
+
+		for _, k := range keys {
+			v := t.report[k]
+
 			t.setIndent(0)
 
 			t.printf("Job: %s", k)
@@ -320,6 +343,15 @@ func (t *tui) draw() {
 				t.addIndent(1)
 				t.renderPrunerReport(activeStatus.PruningReceiver)
 				t.addIndent(-1)
+
+				if v.Type == job.TypePush {
+					t.printf("Snapshotting:")
+					t.newline()
+					t.addIndent(1)
+					t.renderSnapperReport(activeStatus.Snapshotting)
+					t.addIndent(-1)
+				}
+
 			} else if v.Type == job.TypeSnap {
 				snapStatus, ok := v.JobSpecific.(*job.SnapJobStatus)
 				if !ok || snapStatus == nil {
@@ -332,6 +364,19 @@ func (t *tui) draw() {
 				t.addIndent(1)
 				t.renderPrunerReport(snapStatus.Pruning)
 				t.addIndent(-1)
+				t.printf("Snapshotting:")
+				t.newline()
+				t.addIndent(1)
+				t.renderSnapperReport(snapStatus.Snapshotting)
+				t.addIndent(-1)
+			} else if v.Type == job.TypeSource {
+
+				st := v.JobSpecific.(*job.PassiveStatus)
+				t.printf("Snapshotting:\n")
+				t.addIndent(1)
+				t.renderSnapperReport(st.Snapper)
+				t.addIndent(-1)
+
 			} else {
 				t.printf("No status representation for job type '%s', dumping as YAML", v.Type)
 				t.newline()
@@ -384,7 +429,7 @@ func (t *tui) renderReplicationReport(rep *report.Report, history *bytesProgress
 	} else {
 		t.printf("Attempt #%d", len(rep.Attempts))
 		if len(rep.Attempts) > 1 {
-			t.printf(". Previous attempts failed with the follwing statuses:")
+			t.printf(". Previous attempts failed with the following statuses:")
 			t.newline()
 			t.addIndent(1)
 			for i, a := range rep.Attempts[:len(rep.Attempts)-1] {
@@ -545,6 +590,86 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 
 }
 
+func (t *tui) renderSnapperReport(r *snapper.Report) {
+	if r == nil {
+		t.printf("<snapshot type does not have a report>\n")
+		return
+	}
+
+	t.printf("Status: %s", r.State)
+	t.newline()
+
+	if r.Error != "" {
+		t.printf("Error: %s\n", r.Error)
+	}
+	if !r.SleepUntil.IsZero() {
+		t.printf("Sleep until: %s\n", r.SleepUntil)
+	}
+
+	sort.Slice(r.Progress, func(i, j int) bool {
+		return strings.Compare(r.Progress[i].Path, r.Progress[j].Path) == -1
+	})
+
+	t.addIndent(1)
+	defer t.addIndent(-1)
+	dur := func(d time.Duration) string {
+		return d.Round(100 * time.Millisecond).String()
+	}
+
+	type row struct {
+		path, state, duration, remainder, hookReport string
+	}
+	var widths struct {
+		path, state, duration int
+	}
+	rows := make([]*row, len(r.Progress))
+	for i, fs := range r.Progress {
+		r := &row{
+			path:  fs.Path,
+			state: fs.State.String(),
+		}
+		if fs.HooksHadError {
+			r.hookReport = fs.Hooks // FIXME render here, not in daemon
+		}
+		switch fs.State {
+		case snapper.SnapPending:
+			r.duration = "..."
+			r.remainder = ""
+		case snapper.SnapStarted:
+			r.duration = dur(time.Since(fs.StartAt))
+			r.remainder = fmt.Sprintf("snap name: %q", fs.SnapName)
+		case snapper.SnapDone:
+			fallthrough
+		case snapper.SnapError:
+			r.duration = dur(fs.DoneAt.Sub(fs.StartAt))
+			r.remainder = fmt.Sprintf("snap name: %q", fs.SnapName)
+		}
+		rows[i] = r
+		if len(r.path) > widths.path {
+			widths.path = len(r.path)
+		}
+		if len(r.state) > widths.state {
+			widths.state = len(r.state)
+		}
+		if len(r.duration) > widths.duration {
+			widths.duration = len(r.duration)
+		}
+	}
+
+	for _, r := range rows {
+		path := rightPad(r.path, widths.path, " ")
+		state := rightPad(r.state, widths.state, " ")
+		duration := rightPad(r.duration, widths.duration, " ")
+		t.printf("%s %s %s", path, state, duration)
+		t.printfDrawIndentedAndWrappedIfMultiline(" %s", r.remainder)
+		if r.hookReport != "" {
+			t.printfDrawIndentedAndWrappedIfMultiline("%s", r.hookReport)
+		}
+		t.newline()
+	}
+
+}
+
 func times(str string, n int) (out string) {
 	for i := 0; i < n; i++ {
 		out += str
@@ -556,7 +681,7 @@ func rightPad(str string, length int, pad string) string {
 	if len(str) > length {
 		return str[:length]
 	}
-	return str + times(pad, length-len(str))
+	return str + strings.Repeat(pad, length-len(str))
 }
 
 var arrowPositions = `>\|/`

config/config.go

@@ -139,6 +139,7 @@ type SnapshottingPeriodic struct {
 	Type     string        `yaml:"type"`
 	Prefix   string        `yaml:"prefix"`
 	Interval time.Duration `yaml:"interval,positive"`
+	Hooks    HookList      `yaml:"hooks,optional"`
 }
 
 type SnapshottingManual struct {
@@ -202,13 +203,13 @@ type ConnectCommon struct {
 
 type TCPConnect struct {
 	ConnectCommon `yaml:",inline"`
-	Address       string        `yaml:"address"`
+	Address       string        `yaml:"address,hostport"`
 	DialTimeout   time.Duration `yaml:"dial_timeout,zeropositive,default=10s"`
 }
 
 type TLSConnect struct {
 	ConnectCommon `yaml:",inline"`
-	Address       string        `yaml:"address"`
+	Address       string        `yaml:"address,hostport"`
 	Ca            string        `yaml:"ca"`
 	Cert          string        `yaml:"cert"`
 	Key           string        `yaml:"key"`
@@ -230,8 +231,9 @@ type SSHStdinserverConnect struct {
 
 type LocalConnect struct {
 	ConnectCommon  `yaml:",inline"`
-	ListenerName   string `yaml:"listener_name"`
-	ClientIdentity string `yaml:"client_identity"`
+	ListenerName   string        `yaml:"listener_name"`
+	ClientIdentity string        `yaml:"client_identity"`
+	DialTimeout    time.Duration `yaml:"dial_timeout,zeropositive,default=2s"`
 }
 
 type ServeEnum struct {
@@ -244,13 +246,13 @@ type ServeCommon struct {
 
 type TCPServe struct {
 	ServeCommon `yaml:",inline"`
-	Listen      string            `yaml:"listen"`
+	Listen      string            `yaml:"listen,hostport"`
 	Clients     map[string]string `yaml:"clients"`
 }
 
 type TLSServe struct {
 	ServeCommon      `yaml:",inline"`
-	Listen           string        `yaml:"listen"`
+	Listen           string        `yaml:"listen,hostport"`
 	Ca               string        `yaml:"ca"`
 	Cert             string        `yaml:"cert"`
 	Key              string        `yaml:"key"`
@@ -312,7 +314,7 @@ type SyslogLoggingOutlet struct {
 
 type TCPLoggingOutlet struct {
 	LoggingOutletCommon `yaml:",inline"`
-	Address             string               `yaml:"address"`
+	Address             string               `yaml:"address,hostport"`
 	Net                 string               `yaml:"net,default=tcp"`
 	RetryInterval       time.Duration        `yaml:"retry_interval,positive,default=10s"`
 	TLS                 *TCPLoggingOutletTLS `yaml:"tls,optional"`
@@ -330,7 +332,7 @@ type MonitoringEnum struct {
 
 type PrometheusMonitoring struct {
 	Type   string `yaml:"type"`
-	Listen string `yaml:"listen"`
+	Listen string `yaml:"listen,hostport"`
 }
 
 type SyslogFacility syslog.Priority
@@ -361,6 +363,38 @@ type JobDebugSettings struct {
 	RPCLog bool `yaml:"rpc_log,optional,default=false"`
 }
 
+type HookList []HookEnum
+
+type HookEnum struct {
+	Ret interface{}
+}
+
+type HookCommand struct {
+	Path               string            `yaml:"path"`
+	Timeout            time.Duration     `yaml:"timeout,optional,positive,default=30s"`
+	Filesystems        FilesystemsFilter `yaml:"filesystems,optional,default={'<': true}"`
+	HookSettingsCommon `yaml:",inline"`
+}
+
+type HookPostgresCheckpoint struct {
+	HookSettingsCommon `yaml:",inline"`
+	DSN                string            `yaml:"dsn"`
+	Timeout            time.Duration     `yaml:"timeout,optional,positive,default=30s"`
+	Filesystems        FilesystemsFilter `yaml:"filesystems"` // required, user should not CHECKPOINT for every FS
+}
+
+type HookMySQLLockTables struct {
+	HookSettingsCommon `yaml:",inline"`
+	DSN                string            `yaml:"dsn"`
+	Timeout            time.Duration     `yaml:"timeout,optional,positive,default=30s"`
+	Filesystems        FilesystemsFilter `yaml:"filesystems"`
+}
+
+type HookSettingsCommon struct {
+	Type       string `yaml:"type"`
+	ErrIsFatal bool   `yaml:"err_is_fatal,optional,default=false"`
+}
+
 func enumUnmarshal(u func(interface{}, bool) error, types map[string]interface{}) (interface{}, error) {
 	var in struct {
 		Type string
@@ -501,6 +535,15 @@ func (t *SyslogFacility) UnmarshalYAML(u func(interface{}, bool) error) (err err
 	return nil
 }
 
+func (t *HookEnum) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
+	t.Ret, err = enumUnmarshal(u, map[string]interface{}{
+		"command":             &HookCommand{},
+		"postgres-checkpoint": &HookPostgresCheckpoint{},
+		"mysql-lock-tables":   &HookMySQLLockTables{},
+	})
+	return
+}
+
 var ConfigFileDefaultLocations = []string{
 	"/etc/zrepl/zrepl.yml",
 	"/usr/local/etc/zrepl/zrepl.yml",

config/config_snapshotting_test.go

@@ -38,6 +38,29 @@ jobs:
     interval: 10m
 `
 
+	hooks := `
+  snapshotting:
+    type: periodic
+    prefix: zrepl_
+    interval: 10m
+    hooks:
+    - type: command
+      path: /tmp/path/to/command
+    - type: command
+      path: /tmp/path/to/command
+      filesystems: { "zroot<": true, "<": false }
+    - type: postgres-checkpoint
+      dsn: "host=localhost port=5432 user=postgres sslmode=disable"
+      filesystems: {
+          "tank/postgres/data11": true
+      }
+    - type: mysql-lock-tables
+      dsn: "root@tcp(localhost)/"
+      filesystems: {
+        "tank/mysql": true
+      }
+`
+
 	fillSnapshotting := func(s string) string { return fmt.Sprintf(tmpl, s) }
 	var c *Config
 
@@ -55,4 +78,13 @@ jobs:
 		assert.Equal(t, "zrepl_", snp.Prefix)
 	})
 
+	t.Run("hooks", func(t *testing.T) {
+		c = testValidConfig(t, fillSnapshotting(hooks))
+		hs := c.Jobs[0].Ret.(*PushJob).Snapshotting.Ret.(*SnapshottingPeriodic).Hooks
+		assert.Equal(t, hs[0].Ret.(*HookCommand).Filesystems["<"], true)
+		assert.Equal(t, hs[1].Ret.(*HookCommand).Filesystems["zroot<"], true)
+		assert.Equal(t, hs[2].Ret.(*HookPostgresCheckpoint).Filesystems["tank/postgres/data11"], true)
+		assert.Equal(t, hs[3].Ret.(*HookMySQLLockTables).Filesystems["tank/mysql"], true)
+	})
+
 }

config/config_test.go

@@ -1,13 +1,17 @@
 package config
 
 import (
+	"bufio"
 	"bytes"
+	"fmt"
 	"path"
 	"path/filepath"
+	"strings"
 	"testing"
 	"text/template"
 
 	"github.com/kr/pretty"
+	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
 
@@ -65,3 +69,20 @@ func testConfig(t *testing.T, input string) (*Config, error) {
 	t.Helper()
 	return ParseConfigBytes([]byte(input))
 }
+
+func trimSpaceEachLineAndPad(s, pad string) string {
+	var out strings.Builder
+	scan := bufio.NewScanner(strings.NewReader(s))
+	for scan.Scan() {
+		fmt.Fprintf(&out, "%s%s\n", pad, bytes.TrimSpace(scan.Bytes()))
+	}
+	return out.String()
+}
+
+func TestTrimSpaceEachLineAndPad(t *testing.T) {
+	foo := `
+	foo
+	bar baz 
+	`
+	assert.Equal(t, "  \n  foo\n  bar baz\n  \n", trimSpaceEachLineAndPad(foo, "  "))
+}

config/config_transport_test.go

@@ -0,0 +1,99 @@
+package config
+
+import (
+	"fmt"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func TestTransportConnect(t *testing.T) {
+	tmpl := `
+jobs:
+- name: foo
+  type: push
+  connect:
+%s
+  filesystems: {"<": true}
+  snapshotting:
+    type: manual
+  pruning:
+    keep_sender:
+    - type: last_n
+      count: 10
+    keep_receiver:
+    - type: last_n
+      count: 10
+`
+
+	mconf := func(s string) string { return fmt.Sprintf(tmpl, s) }
+
+	type test struct {
+		Name        string
+		ExpectError bool
+		Connect     string
+	}
+
+	testTable := []test{
+		{
+			Name:        "tcp_with_address_and_port",
+			ExpectError: false,
+			Connect: `
+			type: tcp
+			address: 10.0.0.23:42
+			`,
+		},
+		{
+			Name:        "tls_with_host_and_port",
+			ExpectError: false,
+			Connect: `
+			type: tls
+			address: "server1.foo.bar:8888"
+			ca:   /etc/zrepl/ca.crt
+			cert: /etc/zrepl/backupserver.fullchain
+			key:  /etc/zrepl/backupserver.key
+			server_cn: "server1"
+			`,
+		},
+		{
+			Name:        "tcp_without_port",
+			ExpectError: true,
+			Connect: `
+			type: tcp
+			address: 10.0.0.23
+			`,
+		},
+		{
+			Name:        "tls_without_port",
+			ExpectError: true,
+			Connect: `
+			type: tls
+			address: 10.0.0.23
+			ca:   /etc/zrepl/ca.crt
+			cert: /etc/zrepl/backupserver.fullchain
+			key:  /etc/zrepl/backupserver.key
+			server_cn: "server1"
+			`,
+		},
+	}
+
+	for _, tc := range testTable {
+		t.Run(tc.Name, func(t *testing.T) {
+			require.NotEmpty(t, tc.Connect)
+			connect := trimSpaceEachLineAndPad(tc.Connect, "    ")
+			conf := mconf(connect)
+			config, err := testConfig(t, conf)
+			if tc.ExpectError && err == nil {
+				t.Errorf("expected test failure, but got valid config %v", config)
+				return
+			}
+			if !tc.ExpectError && err != nil {
+				t.Errorf("not expecint test failure but got error: %s", err)
+				return
+			}
+			t.Logf("error=%v", err)
+			t.Logf("config=%v", config)
+		})
+	}
+
+}

config/samples/hooks/template.sh

@@ -0,0 +1,22 @@
+#!/usr/bin/env bash
+set -e
+
+[ "$ZREPL_DRYRUN" = "true" ] && DRYRUN="echo DRYRUN: "
+
+pre_snapshot() {
+    $DRYRUN date
+}
+
+post_snapshot() {
+    $DRYRUN date
+}
+
+case "$ZREPL_HOOKTYPE" in
+    pre_snapshot|post_snapshot)
+        "$ZREPL_HOOKTYPE"
+        ;;
+    *)
+        printf 'Unrecognized hook type: %s\n' "$ZREPL_HOOKTYPE"
+        exit 255
+        ;;
+esac

daemon/control.go

@@ -16,6 +16,7 @@ import (
 	"github.com/zrepl/zrepl/daemon/job"
 	"github.com/zrepl/zrepl/daemon/nethelpers"
 	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/util/envconst"
 	"github.com/zrepl/zrepl/version"
 	"github.com/zrepl/zrepl/zfs"
 )
@@ -86,6 +87,12 @@ func (j *controlJob) Run(ctx context.Context) {
 	}
 
 	pprofServer := NewPProfServer(ctx)
+	if listen := envconst.String("ZREPL_DAEMON_AUTOSTART_PPROF_SERVER", ""); listen != "" {
+		pprofServer.Control(PprofServerControlMsg{
+			Run:               true,
+			HttpListenAddress: listen,
+		})
+	}
 
 	mux := http.NewServeMux()
 	mux.Handle(ControlJobEndpointPProf,

daemon/filters/fsmapfilter_test.go

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

daemon/hooks/edge_string.go

@@ -0,0 +1,35 @@
+// Code generated by "stringer -type=Edge"; DO NOT EDIT.
+
+package hooks
+
+import "strconv"
+
+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[Pre-1]
+	_ = x[Callback-2]
+	_ = x[Post-4]
+}
+
+const (
+	_Edge_name_0 = "PreCallback"
+	_Edge_name_1 = "Post"
+)
+
+var (
+	_Edge_index_0 = [...]uint8{0, 3, 11}
+)
+
+func (i Edge) String() string {
+	switch {
+	case 1 <= i && i <= 2:
+		i -= 1
+		return _Edge_name_0[_Edge_index_0[i]:_Edge_index_0[i+1]]
+	case i == 4:
+		return _Edge_name_1
+	default:
+		return "Edge(" + strconv.FormatInt(int64(i), 10) + ")"
+	}
+}

daemon/hooks/hook_config.go

@@ -0,0 +1,52 @@
+package hooks
+
+import (
+	"fmt"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type List []Hook
+
+func HookFromConfig(in config.HookEnum) (Hook, error) {
+	switch v := in.Ret.(type) {
+	case *config.HookCommand:
+		return NewCommandHook(v)
+	case *config.HookPostgresCheckpoint:
+		return PgChkptHookFromConfig(v)
+	case *config.HookMySQLLockTables:
+		return MyLockTablesFromConfig(v)
+	default:
+		return nil, fmt.Errorf("unknown hook type %T", v)
+	}
+}
+
+func ListFromConfig(in *config.HookList) (r *List, err error) {
+	hl := make(List, len(*in))
+
+	for i, h := range *in {
+		hl[i], err = HookFromConfig(h)
+		if err != nil {
+			return nil, fmt.Errorf("create hook #%d: %s", i+1, err)
+		}
+	}
+
+	return &hl, nil
+}
+
+func (l List) CopyFilteredForFilesystem(fs *zfs.DatasetPath) (ret List, err error) {
+	ret = make(List, 0, len(l))
+
+	for _, h := range l {
+		var passFilesystem bool
+		if passFilesystem, err = h.Filesystems().Filter(fs); err != nil {
+			return nil, err
+		}
+		if passFilesystem {
+			ret = append(ret, h)
+		}
+	}
+
+	return ret, nil
+}

daemon/hooks/hook_docs.go

@@ -0,0 +1,34 @@
+// Package hooks implements pre- and post snapshot hooks.
+//
+// Plan is a generic executor for ExpectStepReports before and after an activity specified in a callback.
+// It provides a reporting facility that can be polled while the plan is executing to gather progress information.
+//
+// This package also provides all supported hook type implementations and abstractions around them.
+//
+// Use For Other Kinds Of ExpectStepReports
+//
+// This package REQUIRES REFACTORING before it can be used for other activities than snapshots, e.g. pre- and post-replication:
+//
+// The Hook interface requires a hook to provide a Filesystems() filter, which doesn't make sense for
+// all kinds of activities.
+//
+// The hook implementations should move out of this package.
+// However, there is a lot of tight coupling which to untangle isn't worth it ATM.
+//
+// How This Package Is Used By Package Snapper
+//
+// Deserialize a config.List using ListFromConfig().
+// Then it MUST filter the list to only contain hooks for a particular filesystem using
+// hooksList.CopyFilteredForFilesystem(fs).
+//
+// Then create a CallbackHook using NewCallbackHookForFilesystem().
+//
+// Pass all of the above to NewPlan() which provides a Report() and Run() method:
+//
+// Plan.Run(ctx context.Context,dryRun bool) executes the plan and take a context as argument that should contain a logger added using hooks.WithLogger()).
+// The value of dryRun is passed through to the hooks' Run() method.
+// Command hooks make it available in the environment variable ZREPL_DRYRUN.
+//
+// Plan.Report() can be called while Plan.Run() is executing to give an overview of plan execution progress (future use in "zrepl status").
+//
+package hooks

daemon/hooks/hook_exec.go

@@ -0,0 +1,292 @@
+package hooks
+
+import (
+	"context"
+	"fmt"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/zrepl/zrepl/zfs"
+)
+
+// Re-export type here so that
+// every file in package hooks doesn't
+// have to import github.com/zrepl/zrepl/zfs
+type Filter zfs.DatasetFilter
+
+type Hook interface {
+	Filesystems() Filter
+
+	// If true and the Pre edge invocation of Run fails, Post edge will not run and other Pre edges will not run.
+	ErrIsFatal() bool
+
+	// Run is invoked by HookPlan for a Pre edge.
+	// If HookReport.HadError() == false, the Post edge will be invoked, too.
+	Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport
+
+	String() string
+}
+
+type Phase string
+
+const (
+	PhaseSnapshot = Phase("snapshot")
+	PhaseTesting  = Phase("testing")
+)
+
+func (p Phase) String() string {
+	return string(p)
+}
+
+//go:generate stringer -type=Edge
+type Edge uint
+
+const (
+	Pre = Edge(1 << iota)
+	Callback
+	Post
+)
+
+func (e Edge) StringForPhase(phase Phase) string {
+	return fmt.Sprintf("%s_%s", e.String(), phase.String())
+}
+
+//go:generate enumer -type=StepStatus -trimprefix=Step
+type StepStatus int
+
+const (
+	StepPending StepStatus = 1 << iota
+	StepExec
+	StepOk
+	StepErr
+	StepSkippedDueToFatalErr
+	StepSkippedDueToPreErr
+)
+
+type HookReport interface {
+	String() string
+	HadError() bool
+	Error() string
+}
+
+type Step struct {
+	Hook       Hook
+	Edge       Edge
+	Status     StepStatus
+	Begin, End time.Time
+	// Report may be nil
+	// FIXME cannot serialize this for client status, but contains interesting info (like what error happened)
+	Report HookReport
+	state  map[interface{}]interface{}
+}
+
+func (s Step) String() (out string) {
+	fatal := "~"
+	if s.Hook.ErrIsFatal() && s.Edge == Pre {
+		fatal = "!"
+	}
+	runTime := "..."
+	if s.Status != StepPending {
+		t := s.End.Sub(s.Begin)
+		runTime = t.Round(time.Millisecond).String()
+	}
+	return fmt.Sprintf("[%s] [%5s] %s [%s]  %s", s.Status, runTime, fatal, s.Edge, s.Hook)
+}
+
+type Plan struct {
+	mtx sync.RWMutex
+
+	steps []*Step
+	pre   []*Step // protected by mtx
+	cb    *Step
+	post  []*Step // not reversed, i.e. entry at index i corresponds to pre-edge in pre[i]
+
+	phase Phase
+	env   Env
+}
+
+func NewPlan(hooks *List, phase Phase, cb *CallbackHook, extra Env) (*Plan, error) {
+
+	var pre, post []*Step
+	// TODO sanity check unique name of hook?
+	for _, hook := range *hooks {
+		state := make(map[interface{}]interface{})
+		preE := &Step{
+			Hook:   hook,
+			Edge:   Pre,
+			Status: StepPending,
+			state:  state,
+		}
+		pre = append(pre, preE)
+		postE := &Step{
+			Hook:   hook,
+			Edge:   Post,
+			Status: StepPending,
+			state:  state,
+		}
+		post = append(post, postE)
+	}
+
+	cbE := &Step{
+		Hook:   cb,
+		Edge:   Callback,
+		Status: StepPending,
+	}
+
+	steps := make([]*Step, 0, len(pre)+len(post)+1)
+	steps = append(steps, pre...)
+	steps = append(steps, cbE)
+	for i := len(post) - 1; i >= 0; i-- {
+		steps = append(steps, post[i])
+	}
+
+	plan := &Plan{
+		phase: phase,
+		env:   extra,
+		steps: steps,
+		pre:   pre,
+		post:  post,
+		cb:    cbE,
+	}
+
+	return plan, nil
+}
+
+type PlanReport []Step
+
+func (p *Plan) Report() PlanReport {
+	p.mtx.RLock()
+	defer p.mtx.RUnlock()
+	rep := make([]Step, len(p.steps))
+	for i := range rep {
+		rep[i] = *p.steps[i]
+	}
+	return rep
+}
+
+func (r PlanReport) HadError() bool {
+	for _, e := range r {
+		if e.Status == StepErr {
+			return true
+		}
+	}
+	return false
+}
+
+func (r PlanReport) HadFatalError() bool {
+	for _, e := range r {
+		if e.Status == StepSkippedDueToFatalErr {
+			return true
+		}
+	}
+	return false
+}
+
+func (r PlanReport) String() string {
+	stepStrings := make([]string, len(r))
+	for i, e := range r {
+		stepStrings[i] = fmt.Sprintf("%02d %s", i+1, e)
+	}
+	return strings.Join(stepStrings, "\n")
+}
+
+func (p *Plan) Run(ctx context.Context, dryRun bool) {
+	p.mtx.RLock()
+	defer p.mtx.RUnlock()
+	w := func(f func()) {
+		p.mtx.RUnlock()
+		defer p.mtx.RLock()
+		p.mtx.Lock()
+		defer p.mtx.Unlock()
+		f()
+	}
+	runHook := func(s *Step, ctx context.Context, edge Edge) HookReport {
+		w(func() { s.Status = StepExec })
+		begin := time.Now()
+		r := s.Hook.Run(ctx, edge, p.phase, dryRun, p.env, s.state)
+		end := time.Now()
+		w(func() {
+			s.Report = r
+			s.Status = StepOk
+			if r.HadError() {
+				s.Status = StepErr
+			}
+			s.Begin, s.End = begin, end
+		})
+		return r
+	}
+
+	l := getLogger(ctx)
+
+	// it's a stack, execute until we reach the end of the list (last item in)
+	// or fail inbetween
+	l.Info("run pre-edges in configuration order")
+	next := 0
+	for ; next < len(p.pre); next++ {
+		e := p.pre[next]
+		l := l.WithField("hook", e.Hook)
+		r := runHook(e, ctx, Pre)
+		if r.HadError() {
+			l.WithError(r).Error("hook invocation failed for pre-edge")
+			if e.Hook.ErrIsFatal() {
+				l.Error("the hook run was aborted due to a fatal error in this hook")
+				break
+			}
+		}
+	}
+
+	hadFatalErr := next != len(p.pre)
+	if hadFatalErr {
+		l.Error("fatal error in a pre-snapshot hook invocation")
+		l.Error("no snapshot will be taken")
+		l.Error("only running post-edges for successful pre-edges")
+		w(func() {
+			p.post[next].Status = StepSkippedDueToFatalErr
+			for i := next + 1; i < len(p.pre); i++ {
+				p.pre[i].Status = StepSkippedDueToFatalErr
+				p.post[i].Status = StepSkippedDueToFatalErr
+			}
+			p.cb.Status = StepSkippedDueToFatalErr
+		})
+		return
+	}
+
+	l.Info("running callback")
+	cbR := runHook(p.cb, ctx, Callback)
+	if cbR.HadError() {
+		l.WithError(cbR).Error("callback failed")
+	}
+
+	l.Info("run post-edges for successful pre-edges in reverse configuration order")
+
+	// the constructor produces pre and post entries
+	// post is NOT reversed
+	next-- // now at index of last executed pre-edge
+	for ; next >= 0; next-- {
+		e := p.post[next]
+		l := l.WithField("hook", e.Hook)
+
+		if p.pre[next].Status != StepOk {
+			if p.pre[next].Status != StepErr {
+				panic(fmt.Sprintf("expecting a pre-edge hook report to be either Ok or Err, got %s", p.pre[next].Status))
+			}
+			l.Info("skip post-edge because pre-edge failed")
+			w(func() {
+				e.Status = StepSkippedDueToPreErr
+			})
+			continue
+		}
+
+		report := runHook(e, ctx, Post)
+
+		if report.HadError() {
+			l.WithError(report).Error("hook invocation failed for post-edge")
+			l.Error("subsequent post-edges run regardless of this post-edge failure")
+		}
+
+		// ErrIsFatal is only relevant for Pre
+
+	}
+
+}

daemon/hooks/hook_logging.go

@@ -0,0 +1,103 @@
+package hooks
+
+import (
+	"bufio"
+	"bytes"
+	"context"
+	"sync"
+
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/util/envconst"
+)
+
+type contextKey int
+
+const (
+	contextKeyLog contextKey = 0
+)
+
+type Logger = logger.Logger
+
+func WithLogger(ctx context.Context, log Logger) context.Context {
+	return context.WithValue(ctx, contextKeyLog, log)
+}
+
+func GetLogger(ctx context.Context) Logger { return getLogger(ctx) }
+
+func getLogger(ctx context.Context) Logger {
+	if log, ok := ctx.Value(contextKeyLog).(Logger); ok {
+		return log
+	}
+	return logger.NewNullLogger()
+}
+
+const MAX_HOOK_LOG_SIZE_DEFAULT int = 1 << 20
+
+type logWriter struct {
+	/*
+		Mutex prevents:
+			concurrent writes to buf, scanner in Write([]byte)
+			data race on scanner vs Write([]byte)
+				and concurrent write to buf (call to buf.Reset())
+				in Close()
+
+		(Also, Close() should generally block until any Write() call completes.)
+	*/
+	mtx     *sync.Mutex
+	buf     bytes.Buffer
+	scanner *bufio.Scanner
+	logger  Logger
+	level   logger.Level
+	field   string
+}
+
+func NewLogWriter(mtx *sync.Mutex, logger Logger, level logger.Level, field string) *logWriter {
+	w := new(logWriter)
+	w.mtx = mtx
+	w.scanner = bufio.NewScanner(&w.buf)
+	w.logger = logger
+	w.level = level
+	w.field = field
+	return w
+}
+
+func (w *logWriter) log(line string) {
+	w.logger.WithField(w.field, line).Log(w.level, "hook output")
+}
+
+func (w *logWriter) logUnreadBytes() error {
+	for w.scanner.Scan() {
+		w.log(w.scanner.Text())
+	}
+	if w.buf.Cap() > envconst.Int("ZREPL_MAX_HOOK_LOG_SIZE", MAX_HOOK_LOG_SIZE_DEFAULT) {
+		w.buf.Reset()
+	}
+
+	return nil
+}
+
+func (w *logWriter) Write(in []byte) (int, error) {
+	w.mtx.Lock()
+	defer w.mtx.Unlock()
+
+	n, bufErr := w.buf.Write(in)
+	if bufErr != nil {
+		return n, bufErr
+	}
+
+	err := w.logUnreadBytes()
+	if err != nil {
+		return n, err
+	}
+	// Always reset the scanner for the next Write
+	w.scanner = bufio.NewScanner(&w.buf)
+
+	return n, nil
+}
+
+func (w *logWriter) Close() (err error) {
+	w.mtx.Lock()
+	defer w.mtx.Unlock()
+
+	return w.logUnreadBytes()
+}

daemon/hooks/hook_type_callback.go

@@ -0,0 +1,65 @@
+package hooks
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type HookJobCallback func(ctx context.Context) error
+
+type CallbackHook struct {
+	cb            HookJobCallback
+	filter        Filter
+	displayString string
+}
+
+func NewCallbackHookForFilesystem(displayString string, fs *zfs.DatasetPath, cb HookJobCallback) *CallbackHook {
+	filter, _ := filters.DatasetMapFilterFromConfig(map[string]bool{fs.ToString(): true})
+	return NewCallbackHook(displayString, cb, filter)
+}
+
+func NewCallbackHook(displayString string, cb HookJobCallback, filter Filter) *CallbackHook {
+	return &CallbackHook{
+		cb:            cb,
+		filter:        filter,
+		displayString: displayString,
+	}
+}
+
+func (h *CallbackHook) Filesystems() Filter {
+	return h.filter
+}
+
+func (h *CallbackHook) ErrIsFatal() bool {
+	return false // callback is by definition
+}
+
+func (h *CallbackHook) String() string {
+	return h.displayString
+}
+
+type CallbackHookReport struct {
+	Name string
+	Err  error
+}
+
+func (r *CallbackHookReport) String() string {
+	if r.HadError() {
+		return r.Error()
+	}
+	return r.Name
+}
+
+func (r *CallbackHookReport) HadError() bool { return r.Err != nil }
+
+func (r *CallbackHookReport) Error() string {
+	return fmt.Sprintf("%s error: %s", r.Name, r.Err)
+}
+
+func (h *CallbackHook) Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport {
+	err := h.cb(ctx)
+	return &CallbackHookReport{h.displayString, err}
+}

daemon/hooks/hook_type_command.go

@@ -0,0 +1,192 @@
+package hooks
+
+import (
+	"context"
+	"fmt"
+	"io"
+	"math"
+	"os"
+	"os/exec"
+	"sort"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/util/circlog"
+	"github.com/zrepl/zrepl/util/envconst"
+)
+
+type HookEnvVar string
+
+const (
+	EnvType     HookEnvVar = "ZREPL_HOOKTYPE"
+	EnvDryRun   HookEnvVar = "ZREPL_DRYRUN"
+	EnvFS       HookEnvVar = "ZREPL_FS"
+	EnvSnapshot HookEnvVar = "ZREPL_SNAPNAME"
+	EnvTimeout  HookEnvVar = "ZREPL_TIMEOUT"
+)
+
+type Env map[HookEnvVar]string
+
+func NewHookEnv(edge Edge, phase Phase, dryRun bool, timeout time.Duration, extra Env) Env {
+	r := Env{
+		EnvTimeout: fmt.Sprintf("%.f", math.Floor(timeout.Seconds())),
+	}
+
+	edgeString := edge.StringForPhase(phase)
+	r[EnvType] = strings.ToLower(edgeString)
+
+	var dryRunString string
+	if dryRun {
+		dryRunString = "true"
+	} else {
+		dryRunString = ""
+	}
+	r[EnvDryRun] = dryRunString
+
+	for k, v := range extra {
+		r[k] = v
+	}
+
+	return r
+}
+
+type CommandHook struct {
+	edge       Edge
+	filter     Filter
+	errIsFatal bool
+	command    string
+	timeout    time.Duration
+}
+
+type CommandHookReport struct {
+	Command                      string
+	Args                         []string // currently always empty
+	Env                          Env
+	Err                          error
+	CapturedStdoutStderrCombined []byte
+}
+
+func (r *CommandHookReport) String() string {
+	// Reproduces a POSIX shell-compatible command line
+	var cmdLine strings.Builder
+	sep := ""
+
+	// Make sure environment variables are always
+	// printed in the same order
+	var hookEnvKeys []HookEnvVar
+	for k := range r.Env {
+		hookEnvKeys = append(hookEnvKeys, k)
+	}
+	sort.Slice(hookEnvKeys, func(i, j int) bool { return string(hookEnvKeys[i]) < string(hookEnvKeys[j]) })
+
+	for _, k := range hookEnvKeys {
+		cmdLine.WriteString(fmt.Sprintf("%s%s='%s'", sep, k, r.Env[k]))
+		sep = " "
+	}
+
+	cmdLine.WriteString(fmt.Sprintf("%s%s", sep, r.Command))
+	for _, a := range r.Args {
+		cmdLine.WriteString(fmt.Sprintf("%s'%s'", sep, a))
+	}
+
+	return fmt.Sprintf("command hook invocation: \"%s\"", cmdLine.String()) // no %q to make copy-pastable
+}
+func (r *CommandHookReport) Error() string {
+	if r.Err == nil {
+		return ""
+	}
+	return fmt.Sprintf("%s FAILED with error: %s", r.String(), r.Err)
+}
+
+func (r *CommandHookReport) HadError() bool {
+	return r.Err != nil
+}
+
+func NewCommandHook(in *config.HookCommand) (r *CommandHook, err error) {
+	r = &CommandHook{
+		errIsFatal: in.ErrIsFatal,
+		command:    in.Path,
+		timeout:    in.Timeout,
+	}
+
+	r.filter, err = filters.DatasetMapFilterFromConfig(in.Filesystems)
+	if err != nil {
+		return nil, fmt.Errorf("cannot parse filesystem filter: %s", err)
+	}
+
+	r.edge = Pre | Post
+
+	return r, nil
+}
+
+func (h *CommandHook) Filesystems() Filter {
+	return h.filter
+}
+
+func (h *CommandHook) ErrIsFatal() bool {
+	return h.errIsFatal
+}
+
+func (h *CommandHook) String() string {
+	return h.command
+}
+
+func (h *CommandHook) Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport {
+	l := getLogger(ctx).WithField("command", h.command)
+
+	cmdCtx, cancel := context.WithTimeout(ctx, h.timeout)
+	defer cancel()
+
+	cmdExec := exec.CommandContext(cmdCtx, h.command)
+
+	hookEnv := NewHookEnv(edge, phase, dryRun, h.timeout, extra)
+	cmdEnv := os.Environ()
+	for k, v := range hookEnv {
+		cmdEnv = append(cmdEnv, fmt.Sprintf("%s=%s", k, v))
+	}
+	cmdExec.Env = cmdEnv
+
+	var scanMutex sync.Mutex
+	combinedOutput, err := circlog.NewCircularLog(envconst.Int("ZREPL_MAX_HOOK_LOG_SIZE", MAX_HOOK_LOG_SIZE_DEFAULT))
+	if err != nil {
+		return &CommandHookReport{Err: err}
+	}
+	logErrWriter := NewLogWriter(&scanMutex, l, logger.Warn, "stderr")
+	logOutWriter := NewLogWriter(&scanMutex, l, logger.Info, "stdout")
+	defer logErrWriter.Close()
+	defer logOutWriter.Close()
+
+	cmdExec.Stderr = io.MultiWriter(logErrWriter, combinedOutput)
+	cmdExec.Stdout = io.MultiWriter(logOutWriter, combinedOutput)
+
+	report := &CommandHookReport{
+		Command: h.command,
+		Env:     hookEnv,
+		// no report.Args
+	}
+
+	err = cmdExec.Start()
+	if err != nil {
+		report.Err = err
+		return report
+	}
+
+	err = cmdExec.Wait()
+	combinedOutputBytes := combinedOutput.Bytes()
+	report.CapturedStdoutStderrCombined = make([]byte, len(combinedOutputBytes))
+	copy(report.CapturedStdoutStderrCombined, combinedOutputBytes)
+	if err != nil {
+		if cmdCtx.Err() == context.DeadlineExceeded {
+			report.Err = fmt.Errorf("timed out after %s: %s", h.timeout, err)
+			return report
+		}
+		report.Err = err
+		return report
+	}
+
+	return report
+}

daemon/hooks/hook_type_mysql_lock_tables.go

@@ -0,0 +1,145 @@
+package hooks
+
+import (
+	"context"
+	"database/sql"
+	"fmt"
+	"strings"
+
+	sqldriver "database/sql/driver"
+
+	"github.com/go-sql-driver/mysql"
+
+	"github.com/pkg/errors"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+// https://dev.mysql.com/doc/mysql-backup-excerpt/5.7/en/backup-methods.html
+//
+// Making Backups Using a File System Snapshot:
+//
+//  If you are using a Veritas file system, you can make a backup like this:
+//
+//    	From a client program, execute FLUSH TABLES WITH READ LOCK.
+//    	From another shell, execute mount vxfs snapshot.
+//    	From the first client, execute UNLOCK TABLES.
+//    	Copy files from the snapshot.
+//    	Unmount the snapshot.
+//
+//	Similar snapshot capabilities may be available in other file systems, such as LVM or ZFS.
+type MySQLLockTables struct {
+	errIsFatal  bool
+	connector   sqldriver.Connector
+	filesystems Filter
+}
+
+type myLockTablesStateKey int
+
+const (
+	myLockTablesConnection myLockTablesStateKey = 1 + iota
+)
+
+func MyLockTablesFromConfig(in *config.HookMySQLLockTables) (*MySQLLockTables, error) {
+	conf, err := mysql.ParseDSN(in.DSN)
+	if err != nil {
+		return nil, errors.Wrap(err, "`dsn` invalid")
+	}
+	cn, err := mysql.NewConnector(conf)
+	if err != nil {
+		return nil, errors.Wrap(err, "`connect` invalid")
+	}
+
+	filesystems, err := filters.DatasetMapFilterFromConfig(in.Filesystems)
+	if err != nil {
+		return nil, errors.Wrap(err, "`filesystems` invalid")
+	}
+
+	return &MySQLLockTables{
+		in.ErrIsFatal,
+		cn,
+		filesystems,
+	}, nil
+}
+
+func (h *MySQLLockTables) ErrIsFatal() bool    { return h.errIsFatal }
+func (h *MySQLLockTables) Filesystems() Filter { return h.filesystems }
+func (h *MySQLLockTables) String() string      { return "MySQL FLUSH TABLES WITH READ LOCK" }
+
+type MyLockTablesReport struct {
+	What string
+	Err  error
+}
+
+func (r *MyLockTablesReport) HadError() bool { return r.Err != nil }
+func (r *MyLockTablesReport) Error() string  { return r.String() }
+func (r *MyLockTablesReport) String() string {
+	var s strings.Builder
+	s.WriteString(r.What)
+	if r.Err != nil {
+		fmt.Fprintf(&s, ": %s", r.Err)
+	}
+	return s.String()
+}
+
+func (h *MySQLLockTables) Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport {
+	fs, ok := extra[EnvFS]
+	if !ok {
+		panic(extra)
+	}
+	dp, err := zfs.NewDatasetPath(fs)
+	if err != nil {
+		panic(err)
+	}
+	if pass, err := h.filesystems.Filter(dp); err != nil {
+		return &MyLockTablesReport{What: "filesystem filter", Err: err}
+	} else if !pass {
+		getLogger(ctx).Debug("filesystem does not match filter, skipping")
+		return &MyLockTablesReport{What: "filesystem filter skipped this filesystem", Err: nil}
+	}
+
+	switch edge {
+	case Pre:
+		err := h.doRunPre(ctx, dp, dryRun, state)
+		return &MyLockTablesReport{"FLUSH TABLES WITH READ LOCK", err}
+	case Post:
+		err := h.doRunPost(ctx, dp, dryRun, state)
+		return &MyLockTablesReport{"UNLOCK TABLES", err}
+	}
+	return &MyLockTablesReport{What: "skipped this edge", Err: nil}
+}
+
+func (h *MySQLLockTables) doRunPre(ctx context.Context, fs *zfs.DatasetPath, dry bool, state map[interface{}]interface{}) (err error) {
+	db := sql.OpenDB(h.connector)
+	defer func(err *error) {
+		if *err != nil {
+			db.Close()
+		}
+	}(&err)
+
+	getLogger(ctx).Debug("do FLUSH TABLES WITH READ LOCK")
+	_, err = db.ExecContext(ctx, "FLUSH TABLES WITH READ LOCK")
+	if err != nil {
+		return
+	}
+
+	state[myLockTablesConnection] = db
+
+	return nil
+}
+
+func (h *MySQLLockTables) doRunPost(ctx context.Context, fs *zfs.DatasetPath, dry bool, state map[interface{}]interface{}) error {
+
+	db := state[myLockTablesConnection].(*sql.DB)
+	defer db.Close()
+
+	getLogger(ctx).Debug("do UNLOCK TABLES")
+	_, err := db.ExecContext(ctx, "UNLOCK TABLES")
+	if err != nil {
+		return err
+	}
+
+	return nil
+}

daemon/hooks/hook_type_postgres_checkpoint.go

@@ -0,0 +1,98 @@
+package hooks
+
+import (
+	"context"
+	"database/sql"
+	"fmt"
+	"math"
+	"time"
+
+	"github.com/lib/pq"
+	"github.com/pkg/errors"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type PgChkptHook struct {
+	errIsFatal  bool
+	connector   *pq.Connector
+	filesystems Filter
+}
+
+func PgChkptHookFromConfig(in *config.HookPostgresCheckpoint) (*PgChkptHook, error) {
+	filesystems, err := filters.DatasetMapFilterFromConfig(in.Filesystems)
+	if err != nil {
+		return nil, errors.Wrap(err, "`filesystems` invalid")
+	}
+	cn, err := pq.NewConnector(in.DSN)
+	if err != nil {
+		return nil, errors.Wrap(err, "`dsn` invalid")
+	}
+
+	return &PgChkptHook{
+		in.ErrIsFatal,
+		cn,
+		filesystems,
+	}, nil
+}
+
+func (h *PgChkptHook) ErrIsFatal() bool    { return h.errIsFatal }
+func (h *PgChkptHook) Filesystems() Filter { return h.filesystems }
+func (h *PgChkptHook) String() string      { return "postgres checkpoint" }
+
+type PgChkptHookReport struct{ Err error }
+
+func (r *PgChkptHookReport) HadError() bool { return r.Err != nil }
+func (r *PgChkptHookReport) Error() string  { return r.Err.Error() }
+func (r *PgChkptHookReport) String() string {
+	if r.Err != nil {
+		return fmt.Sprintf("postgres CHECKPOINT failed: %s", r.Err)
+	} else {
+		return "postgres CHECKPOINT completed"
+	}
+}
+
+func (h *PgChkptHook) Run(ctx context.Context, edge Edge, phase Phase, dryRun bool, extra Env, state map[interface{}]interface{}) HookReport {
+	if edge != Pre {
+		return &PgChkptHookReport{nil}
+	}
+	fs, ok := extra[EnvFS]
+	if !ok {
+		panic(extra)
+	}
+	dp, err := zfs.NewDatasetPath(fs)
+	if err != nil {
+		panic(err)
+	}
+	err = h.doRunPre(ctx, dp, dryRun)
+	return &PgChkptHookReport{err}
+}
+
+func (h *PgChkptHook) doRunPre(ctx context.Context, fs *zfs.DatasetPath, dry bool) error {
+
+	if pass, err := h.filesystems.Filter(fs); err != nil || !pass {
+		getLogger(ctx).Debug("filesystem does not match filter, skipping")
+		return err
+	}
+
+	db := sql.OpenDB(h.connector)
+	defer db.Close()
+	dl, ok := ctx.Deadline()
+	if ok {
+		timeout := uint64(math.Floor(time.Until(dl).Seconds() * 1000)) // TODO go1.13 milliseconds
+		getLogger(ctx).WithField("statement_timeout", timeout).Debug("setting statement timeout for CHECKPOINT")
+		_, err := db.ExecContext(ctx, "SET statement_timeout TO ?", timeout)
+		if err != nil {
+			return err
+		}
+	}
+	if dry {
+		getLogger(ctx).Debug("dry-run - use ping instead of CHECKPOINT")
+		return db.PingContext(ctx)
+	}
+	getLogger(ctx).Info("execute CHECKPOINT command")
+	_, err := db.ExecContext(ctx, "CHECKPOINT")
+	return err
+}

daemon/hooks/hooks_test.go

@@ -0,0 +1,484 @@
+package hooks_test
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"os"
+	"regexp"
+	"testing"
+	"text/template"
+
+	"github.com/stretchr/testify/require"
+
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/hooks"
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type comparisonAssertionFunc func(require.TestingT, interface{}, interface{}, ...interface{})
+type valueAssertionFunc func(require.TestingT, interface{}, ...interface{})
+
+type expectStep struct {
+	ExpectedEdge hooks.Edge
+	ExpectStatus hooks.StepStatus
+	OutputTest   valueAssertionFunc
+	ErrorTest    valueAssertionFunc
+}
+
+type testCase struct {
+	Name           string
+	Config         []string
+	IsSlow         bool
+	SuppressOutput bool
+
+	ExpectCallbackSkipped bool
+	ExpectHadFatalErr     bool
+	ExpectHadError        bool
+	ExpectStepReports     []expectStep
+}
+
+func curryLeft(f comparisonAssertionFunc, expected interface{}) valueAssertionFunc {
+	return curry(f, expected, false)
+}
+
+func curryRight(f comparisonAssertionFunc, expected interface{}) valueAssertionFunc {
+	return curry(f, expected, true)
+}
+
+func curry(f comparisonAssertionFunc, expected interface{}, right bool) (ret valueAssertionFunc) {
+	ret = func(t require.TestingT, s interface{}, v ...interface{}) {
+		var x interface{}
+		var y interface{}
+		if right {
+			x = s
+			y = expected
+		} else {
+			x = expected
+			y = s
+		}
+
+		if len(v) > 0 {
+			f(t, x, y, v)
+		} else {
+			f(t, x, y)
+		}
+	}
+	return
+}
+
+func TestHooks(t *testing.T) {
+	testFSName := "testpool/testdataset"
+	testSnapshotName := "testsnap"
+
+	tmpl, err := template.New("TestHooks").Parse(`
+jobs:
+- name: TestHooks
+  type: snap
+  filesystems: {"<": true}
+  snapshotting:
+    type: periodic
+    interval: 1m
+    prefix: zrepl_snapjob_
+    hooks:
+    {{- template "List" . }}
+  pruning:
+    keep:
+      - type: last_n
+        count: 10
+`)
+	if err != nil {
+		panic(err)
+	}
+
+	regexpTest := func(s string) valueAssertionFunc {
+		return curryLeft(require.Regexp, regexp.MustCompile(s))
+	}
+
+	containsTest := func(s string) valueAssertionFunc {
+		return curryRight(require.Contains, s)
+	}
+
+	testTable := []testCase{
+		testCase{
+			Name: "no_hooks",
+			ExpectStepReports: []expectStep{
+				{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+			},
+		},
+		testCase{
+			Name:           "timeout",
+			IsSlow:         true,
+			ExpectHadError: true,
+			Config:         []string{`{type: command, path: {{.WorkDir}}/test/test-timeout.sh, timeout: 2s}`},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s ZREPL_TIMEOUT=2", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest(`timed out after 2(.\d+)?s`),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToPreErr,
+				},
+			},
+		},
+		testCase{
+			Name:           "check_env",
+			Config:         []string{`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`},
+			ExpectHadError: false,
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST post_testing %s@%s", testFSName, testSnapshotName)),
+				},
+			},
+		},
+
+		testCase{
+			Name:                  "nonfatal_pre_error_continues",
+			ExpectCallbackSkipped: false,
+			ExpectHadError:        true,
+			Config:                []string{`{type: command, path: {{.WorkDir}}/test/test-error.sh}`},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR pre_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToPreErr, // post-edge is not executed for failing pre-edge
+				},
+			},
+		},
+
+		testCase{
+			Name:                  "pre_error_fatal_skips_subsequent_pre_edges_and_callback_and_its_post_edge_and_post_edges",
+			ExpectCallbackSkipped: true,
+			ExpectHadFatalErr:     true,
+			ExpectHadError:        true,
+			Config: []string{
+				`{type: command, path: {{.WorkDir}}/test/test-error.sh, err_is_fatal: true}`,
+				`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`,
+			},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR pre_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepSkippedDueToFatalErr,
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepSkippedDueToFatalErr},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToFatalErr,
+				},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToFatalErr,
+				},
+			},
+		},
+
+		testCase{
+			Name:              "post_error_fails_are_ignored_even_if_fatal",
+			ExpectHadFatalErr: false, // only occurs during Post, so it's not a fatal error
+			ExpectHadError:    true,
+			Config: []string{
+				`{type: command, path: {{.WorkDir}}/test/test-post-error.sh, err_is_fatal: true}`,
+				`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`,
+			},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					// No-action run of test-post-error.sh
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   require.Empty,
+				},
+				expectStep{
+					// Pre run of test-report-env.sh
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST post_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR post_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+			},
+		},
+
+		testCase{
+			Name:   "cleanup_check_env",
+			Config: []string{`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST post_testing %s@%s", testFSName, testSnapshotName)),
+				},
+			},
+		},
+
+		testCase{
+			Name:              "pre_error_cancels_post",
+			Config:            []string{`{type: command, path: {{.WorkDir}}/test/test-pre-error-post-ok.sh}`},
+			ExpectHadError:    true,
+			ExpectHadFatalErr: false,
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR pre_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToPreErr,
+				},
+			},
+		},
+
+		testCase{
+			Name: "pre_error_does_not_cancel_other_posts_but_itself",
+			Config: []string{
+				`{type: command, path: {{.WorkDir}}/test/test-report-env.sh}`,
+				`{type: command, path: {{.WorkDir}}/test/test-pre-error-post-ok.sh}`,
+			},
+			ExpectHadError:    true,
+			ExpectHadFatalErr: false,
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST pre_testing %s@%s", testFSName, testSnapshotName)),
+				},
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepErr,
+					OutputTest:   containsTest(fmt.Sprintf("TEST ERROR pre_testing %s@%s", testFSName, testSnapshotName)),
+					ErrorTest:    regexpTest("^command hook invocation.*exit status 1$"),
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepSkippedDueToPreErr,
+				},
+				expectStep{
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   containsTest(fmt.Sprintf("TEST post_testing %s@%s", testFSName, testSnapshotName)),
+				},
+			},
+		},
+
+		testCase{
+			Name:              "exceed_buffer_limit",
+			SuppressOutput:    true,
+			Config:            []string{`{type: command, path: {{.WorkDir}}/test/test-large-stdout.sh}`},
+			ExpectHadError:    false,
+			ExpectHadFatalErr: false,
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Pre,
+					ExpectStatus: hooks.StepOk,
+					OutputTest: func(t require.TestingT, s interface{}, v ...interface{}) {
+						require.Len(t, s, 1<<20)
+					},
+				},
+				expectStep{ExpectedEdge: hooks.Callback, ExpectStatus: hooks.StepOk},
+				expectStep{
+					// No-action run of above hook
+					ExpectedEdge: hooks.Post,
+					ExpectStatus: hooks.StepOk,
+					OutputTest:   require.Empty,
+				},
+			},
+		},
+
+		/*
+			Following not intended to test functionality of
+			filter package. Only to demonstrate that hook
+			filters are being applied. The following should
+			result in NO hooks running. If it does run, a
+			fatal hooks.RunReport will be returned.
+		*/
+		testCase{
+			Name: "exclude_all_filesystems",
+			Config: []string{
+				`{type: command, path: {{.WorkDir}}/test/test-error.sh, err_is_fatal: true, filesystems: {"<": false}}`,
+			},
+			ExpectStepReports: []expectStep{
+				expectStep{
+					ExpectedEdge: hooks.Callback,
+					ExpectStatus: hooks.StepOk,
+				},
+			},
+		},
+	}
+
+	parseHookConfig := func(t *testing.T, in string) *config.Config {
+		t.Helper()
+		conf, err := config.ParseConfigBytes([]byte(in))
+		require.NoError(t, err)
+		require.NotNil(t, conf)
+		return conf
+	}
+
+	fillHooks := func(tt *testCase) string {
+		// make hook path absolute
+		cwd, err := os.Getwd()
+		if err != nil {
+			panic("os.Getwd() failed")
+		}
+		var hooksTmpl string = "\n"
+		for _, l := range tt.Config {
+			hooksTmpl += fmt.Sprintf("    - %s\n", l)
+		}
+		tmpl.New("List").Parse(hooksTmpl)
+
+		var outBytes bytes.Buffer
+		data := struct {
+			WorkDir string
+		}{
+			WorkDir: cwd,
+		}
+		if err := tmpl.Execute(&outBytes, data); err != nil {
+			panic(err)
+		}
+
+		return outBytes.String()
+	}
+
+	var c *config.Config
+	fs, err := zfs.NewDatasetPath(testFSName)
+	require.NoError(t, err)
+
+	log := logger.NewTestLogger(t)
+
+	var cbReached bool
+	cb := hooks.NewCallbackHookForFilesystem("testcallback", fs, func(_ context.Context) error {
+		cbReached = true
+		return nil
+	})
+
+	hookEnvExtra := hooks.Env{
+		hooks.EnvFS:       fs.ToString(),
+		hooks.EnvSnapshot: testSnapshotName,
+	}
+
+	for _, tt := range testTable {
+		if testing.Short() && tt.IsSlow {
+			continue
+		}
+
+		t.Run(tt.Name, func(t *testing.T) {
+			c = parseHookConfig(t, fillHooks(&tt))
+			snp := c.Jobs[0].Ret.(*config.SnapJob).Snapshotting.Ret.(*config.SnapshottingPeriodic)
+			hookList, err := hooks.ListFromConfig(&snp.Hooks)
+			require.NoError(t, err)
+
+			filteredHooks, err := hookList.CopyFilteredForFilesystem(fs)
+			require.NoError(t, err)
+			plan, err := hooks.NewPlan(&filteredHooks, hooks.PhaseTesting, cb, hookEnvExtra)
+			require.NoError(t, err)
+			t.Logf("REPORT PRE EXECUTION:\n%s", plan.Report())
+
+			cbReached = false
+
+			ctx := context.Background()
+			if testing.Verbose() && !tt.SuppressOutput {
+				ctx = hooks.WithLogger(ctx, log)
+			}
+			plan.Run(ctx, false)
+			report := plan.Report()
+
+			t.Logf("REPORT POST EXECUTION:\n%s", report)
+
+			/*
+			 * TEST ASSERTIONS
+			 */
+
+			t.Logf("len(runReports)=%v", len(report))
+			t.Logf("len(tt.ExpectStepReports)=%v", len(tt.ExpectStepReports))
+			require.Equal(t, len(tt.ExpectStepReports), len(report), "ExpectStepReports must be same length as expected number of hook runs, excluding possible Callback")
+
+			// Check if callback ran, when required
+			if tt.ExpectCallbackSkipped {
+				require.False(t, cbReached, "callback ran but should not have run")
+			} else {
+				require.True(t, cbReached, "callback should have run but did not")
+			}
+
+			// Check if a fatal run error occurred and was expected
+			require.Equal(t, tt.ExpectHadFatalErr, report.HadFatalError(), "non-matching HadFatalError")
+			require.Equal(t, tt.ExpectHadError, report.HadError(), "non-matching HadError")
+
+			if tt.ExpectHadFatalErr {
+				require.True(t, tt.ExpectHadError, "ExpectHadFatalErr implies ExpectHadError")
+			}
+			if !tt.ExpectHadError {
+				require.False(t, tt.ExpectHadFatalErr, "!ExpectHadError implies !ExpectHadFatalErr")
+			}
+
+			// Iterate through each expected hook run
+			for i, hook := range tt.ExpectStepReports {
+				t.Logf("expecting report conforming to %v", hook)
+
+				exp, act := hook.ExpectStatus, report[i].Status
+				require.Equal(t, exp, act, "%s != %s", exp, act)
+
+				// Check for required ExpectedEdge
+				require.NotZero(t, hook.ExpectedEdge, "each hook must have an ExpectedEdge")
+				require.Equal(t, hook.ExpectedEdge, report[i].Edge,
+					"incorrect edge: expected %q, actual %q", hook.ExpectedEdge.String(), report[i].Edge.String(),
+				)
+
+				// Check for expected output
+				if hook.OutputTest != nil {
+					require.IsType(t, (*hooks.CommandHookReport)(nil), report[i].Report)
+					chr := report[i].Report.(*hooks.CommandHookReport)
+					hook.OutputTest(t, string(chr.CapturedStdoutStderrCombined))
+				}
+
+				// Check for expected errors
+				if hook.ErrorTest != nil {
+					hook.ErrorTest(t, string(report[i].Report.Error()))
+				}
+			}
+
+		})
+	}
+}

daemon/hooks/stepstatus_enumer.go

@@ -0,0 +1,77 @@
+// Code generated by "enumer -type=StepStatus -trimprefix=Step"; DO NOT EDIT.
+
+//
+package hooks
+
+import (
+	"fmt"
+)
+
+const (
+	_StepStatusName_0 = "PendingExec"
+	_StepStatusName_1 = "Ok"
+	_StepStatusName_2 = "Err"
+	_StepStatusName_3 = "SkippedDueToFatalErr"
+	_StepStatusName_4 = "SkippedDueToPreErr"
+)
+
+var (
+	_StepStatusIndex_0 = [...]uint8{0, 7, 11}
+	_StepStatusIndex_1 = [...]uint8{0, 2}
+	_StepStatusIndex_2 = [...]uint8{0, 3}
+	_StepStatusIndex_3 = [...]uint8{0, 20}
+	_StepStatusIndex_4 = [...]uint8{0, 18}
+)
+
+func (i StepStatus) String() string {
+	switch {
+	case 1 <= i && i <= 2:
+		i -= 1
+		return _StepStatusName_0[_StepStatusIndex_0[i]:_StepStatusIndex_0[i+1]]
+	case i == 4:
+		return _StepStatusName_1
+	case i == 8:
+		return _StepStatusName_2
+	case i == 16:
+		return _StepStatusName_3
+	case i == 32:
+		return _StepStatusName_4
+	default:
+		return fmt.Sprintf("StepStatus(%d)", i)
+	}
+}
+
+var _StepStatusValues = []StepStatus{1, 2, 4, 8, 16, 32}
+
+var _StepStatusNameToValueMap = map[string]StepStatus{
+	_StepStatusName_0[0:7]:  1,
+	_StepStatusName_0[7:11]: 2,
+	_StepStatusName_1[0:2]:  4,
+	_StepStatusName_2[0:3]:  8,
+	_StepStatusName_3[0:20]: 16,
+	_StepStatusName_4[0:18]: 32,
+}
+
+// StepStatusString retrieves an enum value from the enum constants string name.
+// Throws an error if the param is not part of the enum.
+func StepStatusString(s string) (StepStatus, error) {
+	if val, ok := _StepStatusNameToValueMap[s]; ok {
+		return val, nil
+	}
+	return 0, fmt.Errorf("%s does not belong to StepStatus values", s)
+}
+
+// StepStatusValues returns all values of the enum
+func StepStatusValues() []StepStatus {
+	return _StepStatusValues
+}
+
+// IsAStepStatus returns "true" if the value is listed in the enum definition. "false" otherwise
+func (i StepStatus) IsAStepStatus() bool {
+	for _, v := range _StepStatusValues {
+		if i == v {
+			return true
+		}
+	}
+	return false
+}

daemon/hooks/test/test-error.sh

@@ -0,0 +1,5 @@
+#!/bin/sh -eu
+
+>&2 echo "TEST ERROR $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME"
+
+exit 1

daemon/hooks/test/test-large-stdout.sh

@@ -0,0 +1,18 @@
+#!/bin/sh -eu
+
+# Exceed default hook log size of 1<<20 bytes by double
+# The Go test should fail if buffer size exceeds 1MB
+
+pre_testing() {
+	ln=0
+	while :; do printf '%06d: 012345678901234567890123456789012345678901234567890123456789012345678901\n' "$ln"; ln="$(($ln + 1))"; done \
+	| head -c$(( 2 << 20 ))
+}
+
+case "$ZREPL_HOOKTYPE" in
+	pre_testing)
+		"$ZREPL_HOOKTYPE";;
+	*)
+		# Not handled by this script
+		exit 0;;
+esac

daemon/hooks/test/test-post-error.sh

@@ -0,0 +1,15 @@
+#!/bin/sh -eu
+
+post_testing() {
+	>&2 echo "TEST ERROR $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME"
+
+	exit 1
+}
+
+case "$ZREPL_HOOKTYPE" in
+	post_testing)
+		"$ZREPL_HOOKTYPE";;
+	*)
+		# Not handled by this script
+		exit 0;;
+esac

daemon/hooks/test/test-pre-error-post-ok.sh

@@ -0,0 +1,11 @@
+#!/bin/sh -eu
+
+if [ "$ZREPL_HOOKTYPE" = "pre_testing" ]; then
+	>&2 echo "TEST ERROR $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME"
+	exit 1
+elif [ "$ZREPL_HOOKTYPE" = "post_testing" ]; then
+	echo "TEST $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME"
+else
+	printf "Unknown hook type: %s" "$ZREPL_HOOKTYPE" 
+	exit 255
+fi

daemon/hooks/test/test-report-env.sh

@@ -0,0 +1,3 @@
+#!/bin/sh -eu
+
+echo "TEST $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME ${ZREPL_DRYRUN:+(dry run)}"

daemon/hooks/test/test-timeout.sh

@@ -0,0 +1,5 @@
+#!/bin/sh -eu
+
+echo "TEST $ZREPL_HOOKTYPE $ZREPL_FS@$ZREPL_SNAPNAME ZREPL_TIMEOUT=$ZREPL_TIMEOUT"
+
+exec sleep $(($ZREPL_TIMEOUT + 1))

daemon/job/active.go

@@ -83,6 +83,7 @@ type activeMode interface {
 	SenderReceiver() (logic.Sender, logic.Receiver)
 	Type() Type
 	RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{})
+	SnapperReport() *snapper.Report
 	ResetConnectBackoff()
 }
 
@@ -124,6 +125,10 @@ func (m *modePush) RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{}
 	m.snapper.Run(ctx, wakeUpCommon)
 }
 
+func (m *modePush) SnapperReport() *snapper.Report {
+	return m.snapper.Report()
+}
+
 func (m *modePush) ResetConnectBackoff() {
 	m.setupMtx.Lock()
 	defer m.setupMtx.Unlock()
@@ -206,6 +211,10 @@ func (m *modePull) RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{}
 	}
 }
 
+func (m *modePull) SnapperReport() *snapper.Report {
+	return nil
+}
+
 func (m *modePull) ResetConnectBackoff() {
 	m.setupMtx.Lock()
 	defer m.setupMtx.Unlock()
@@ -279,6 +288,7 @@ func (j *ActiveSide) Name() string { return j.name }
 type ActiveSideStatus struct {
 	Replication                    *report.Report
 	PruningSender, PruningReceiver *pruner.Report
+	Snapshotting                   *snapper.Report
 }
 
 func (j *ActiveSide) Status() *Status {
@@ -295,6 +305,7 @@ func (j *ActiveSide) Status() *Status {
 	if tasks.prunerReceiver != nil {
 		s.PruningReceiver = tasks.prunerReceiver.Report()
 	}
+	s.Snapshotting = j.mode.SnapperReport()
 	return &Status{Type: t, JobSpecific: s}
 }
 

daemon/job/activesidestate_enumer.go

@@ -1,5 +1,6 @@
 // Code generated by "enumer -type=ActiveSideState"; DO NOT EDIT.
 
+//
 package job
 
 import (

daemon/job/passive.go

@@ -27,6 +27,7 @@ type PassiveSide struct {
 type passiveMode interface {
 	Handler() rpc.Handler
 	RunPeriodic(ctx context.Context)
+	SnapperReport() *snapper.Report // may be nil
 	Type() Type
 }
 
@@ -40,7 +41,8 @@ func (m *modeSink) Handler() rpc.Handler {
 	return endpoint.NewReceiver(m.rootDataset, true)
 }
 
-func (m *modeSink) RunPeriodic(_ context.Context) {}
+func (m *modeSink) RunPeriodic(_ context.Context)  {}
+func (m *modeSink) SnapperReport() *snapper.Report { return nil }
 
 func modeSinkFromConfig(g *config.Global, in *config.SinkJob) (m *modeSink, err error) {
 	m = &modeSink{}
@@ -85,6 +87,10 @@ func (m *modeSource) RunPeriodic(ctx context.Context) {
 	m.snapper.Run(ctx, nil)
 }
 
+func (m *modeSource) SnapperReport() *snapper.Report {
+	return m.snapper.Report()
+}
+
 func passiveSideFromConfig(g *config.Global, in *config.PassiveJob, mode passiveMode) (s *PassiveSide, err error) {
 
 	s = &PassiveSide{mode: mode, name: in.Name}
@@ -97,10 +103,15 @@ func passiveSideFromConfig(g *config.Global, in *config.PassiveJob, mode passive
 
 func (j *PassiveSide) Name() string { return j.name }
 
-type PassiveStatus struct{}
+type PassiveStatus struct {
+	Snapper *snapper.Report
+}
 
 func (s *PassiveSide) Status() *Status {
-	return &Status{Type: s.mode.Type()} // FIXME PassiveStatus
+	st := &PassiveStatus{
+		Snapper: s.mode.SnapperReport(),
+	}
+	return &Status{Type: s.mode.Type(), JobSpecific: st}
 }
 
 func (j *PassiveSide) OwnedDatasetSubtreeRoot() (rfs *zfs.DatasetPath, ok bool) {

daemon/job/snapjob.go

@@ -66,7 +66,8 @@ func (j *SnapJob) RegisterMetrics(registerer prometheus.Registerer) {
 }
 
 type SnapJobStatus struct {
-	Pruning *pruner.Report
+	Pruning      *pruner.Report
+	Snapshotting *snapper.Report // may be nil
 }
 
 func (j *SnapJob) Status() *Status {
@@ -75,6 +76,7 @@ func (j *SnapJob) Status() *Status {
 	if j.pruner != nil {
 		s.Pruning = j.pruner.Report()
 	}
+	s.Snapshotting = j.snapper.Report()
 	return &Status{Type: t, JobSpecific: s}
 }
 

daemon/logging/build_logging.go

@@ -11,6 +11,7 @@ import (
 	"github.com/pkg/errors"
 
 	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/hooks"
 	"github.com/zrepl/zrepl/daemon/pruner"
 	"github.com/zrepl/zrepl/daemon/snapper"
 	"github.com/zrepl/zrepl/endpoint"
@@ -37,7 +38,7 @@ func OutletsFromConfig(in config.LoggingOutletEnumList) (*logger.Outlets, error)
 	var syslogOutlets, stdoutOutlets int
 	for lei, le := range in {
 
-		outlet, minLevel, err := parseOutlet(le)
+		outlet, minLevel, err := ParseOutlet(le)
 		if err != nil {
 			return nil, errors.Wrapf(err, "cannot parse outlet #%d", lei)
 		}
@@ -72,6 +73,7 @@ const (
 	SubsyEndpoint      Subsystem = "endpoint"
 	SubsysPruning      Subsystem = "pruning"
 	SubsysSnapshot     Subsystem = "snapshot"
+	SubsysHooks        Subsystem = "hook"
 	SubsysTransport    Subsystem = "transport"
 	SubsysTransportMux Subsystem = "transportmux"
 	SubsysRPC          Subsystem = "rpc"
@@ -85,6 +87,7 @@ func WithSubsystemLoggers(ctx context.Context, log logger.Logger) context.Contex
 	ctx = endpoint.WithLogger(ctx, log.WithField(SubsysField, SubsyEndpoint))
 	ctx = pruner.WithLogger(ctx, log.WithField(SubsysField, SubsysPruning))
 	ctx = snapper.WithLogger(ctx, log.WithField(SubsysField, SubsysSnapshot))
+	ctx = hooks.WithLogger(ctx, log.WithField(SubsysField, SubsysHooks))
 	ctx = transport.WithLogger(ctx, log.WithField(SubsysField, SubsysTransport))
 	ctx = transportmux.WithLogger(ctx, log.WithField(SubsysField, SubsysTransportMux))
 	ctx = rpc.WithLoggers(ctx,
@@ -123,7 +126,7 @@ func parseLogFormat(i interface{}) (f EntryFormatter, err error) {
 
 }
 
-func parseOutlet(in config.LoggingOutletEnum) (o logger.Outlet, level logger.Level, err error) {
+func ParseOutlet(in config.LoggingOutletEnum) (o logger.Outlet, level logger.Level, err error) {
 
 	parseCommon := func(common config.LoggingOutletCommon) (logger.Level, EntryFormatter, error) {
 		if common.Level == "" || common.Format == "" {

daemon/logging/logging_formatters.go

@@ -88,7 +88,7 @@ func (f *HumanFormatter) Format(e *logger.Entry) (out []byte, err error) {
 	prefixFields := []string{JobField, SubsysField}
 	prefixed := make(map[string]bool, len(prefixFields)+2)
 	for _, field := range prefixFields {
-		val, ok := e.Fields[field].(string)
+		val, ok := e.Fields[field]
 		if !ok {
 			continue
 		}

daemon/pruner/state_enumer.go

@@ -1,5 +1,6 @@
 // Code generated by "enumer -type=State"; DO NOT EDIT.
 
+//
 package pruner
 
 import (

daemon/snapper/snapper.go

@@ -11,6 +11,7 @@ import (
 
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/daemon/filters"
+	"github.com/zrepl/zrepl/daemon/hooks"
 	"github.com/zrepl/zrepl/logger"
 	"github.com/zrepl/zrepl/zfs"
 )
@@ -25,18 +26,20 @@ const (
 	SnapError
 )
 
+// All fields protected by Snapper.mtx
 type snapProgress struct {
 	state SnapState
 
 	// SnapStarted, SnapDone, SnapError
-	name    string
-	startAt time.Time
+	name     string
+	startAt  time.Time
+	hookPlan *hooks.Plan
 
 	// SnapDone
 	doneAt time.Time
 
-	// SnapErr
-	err error
+	// SnapErr TODO disambiguate state
+	runResults hooks.PlanReport
 }
 
 type args struct {
@@ -46,6 +49,8 @@ type args struct {
 	interval       time.Duration
 	fsf            *filters.DatasetMapFilter
 	snapshotsTaken chan<- struct{}
+	hooks          *hooks.List
+	dryRun         bool
 }
 
 type Snapper struct {
@@ -58,7 +63,7 @@ type Snapper struct {
 	lastInvocation time.Time
 
 	// valid for state Snapshotting
-	plan map[*zfs.DatasetPath]snapProgress
+	plan map[*zfs.DatasetPath]*snapProgress
 
 	// valid for state SyncUp and Waiting
 	sleepUntil time.Time
@@ -123,10 +128,16 @@ func PeriodicFromConfig(g *config.Global, fsf *filters.DatasetMapFilter, in *con
 		return nil, errors.New("interval must be positive")
 	}
 
+	hookList, err := hooks.ListFromConfig(&in.Hooks)
+	if err != nil {
+		return nil, errors.Wrap(err, "hook config error")
+	}
+
 	args := args{
 		prefix:   in.Prefix,
 		interval: in.Interval,
 		fsf:      fsf,
+		hooks:    hookList,
 		// ctx and log is set in Run()
 	}
 
@@ -141,6 +152,7 @@ func (s *Snapper) Run(ctx context.Context, snapshotsTaken chan<- struct{}) {
 	s.args.snapshotsTaken = snapshotsTaken
 	s.args.ctx = ctx
 	s.args.log = getLogger(ctx)
+	s.args.dryRun = false // for future expansion
 
 	u := func(u func(*Snapper)) State {
 		s.mtx.Lock()
@@ -224,24 +236,30 @@ func plan(a args, u updater) state {
 		return onErr(err, u)
 	}
 
-	plan := make(map[*zfs.DatasetPath]snapProgress, len(fss))
+	plan := make(map[*zfs.DatasetPath]*snapProgress, len(fss))
 	for _, fs := range fss {
-		plan[fs] = snapProgress{state: SnapPending}
+		plan[fs] = &snapProgress{state: SnapPending}
 	}
 	return u(func(s *Snapper) {
 		s.state = Snapshotting
 		s.plan = plan
+		s.err = nil
 	}).sf()
 }
 
 func snapshot(a args, u updater) state {
 
-	var plan map[*zfs.DatasetPath]snapProgress
+	var plan map[*zfs.DatasetPath]*snapProgress
 	u(func(snapper *Snapper) {
 		plan = snapper.plan
 	})
 
-	hadErr := false
+	hookMatchCount := make(map[hooks.Hook]int, len(*a.hooks))
+	for _, h := range *a.hooks {
+		hookMatchCount[h] = 0
+	}
+
+	anyFsHadErr := false
 	// TODO channel programs -> allow a little jitter?
 	for fs, progress := range plan {
 		suffix := time.Now().In(time.UTC).Format("20060102_150405_000")
@@ -251,27 +269,71 @@ func snapshot(a args, u updater) state {
 			WithField("fs", fs.ToString()).
 			WithField("snap", snapname)
 
+		hookEnvExtra := hooks.Env{
+			hooks.EnvFS:       fs.ToString(),
+			hooks.EnvSnapshot: snapname,
+		}
+
+		jobCallback := hooks.NewCallbackHookForFilesystem("snapshot", fs, func(_ context.Context) (err error) {
+			l.Debug("create snapshot")
+			err = zfs.ZFSSnapshot(fs, snapname, false) // TODO propagagte context to ZFSSnapshot
+			if err != nil {
+				l.WithError(err).Error("cannot create snapshot")
+			}
+			return
+		})
+
+		fsHadErr := false
+		var planReport hooks.PlanReport
+		var plan *hooks.Plan
+		{
+			filteredHooks, err := a.hooks.CopyFilteredForFilesystem(fs)
+			if err != nil {
+				l.WithError(err).Error("unexpected filter error")
+				fsHadErr = true
+				goto updateFSState
+			}
+			// account for running hooks
+			for _, h := range filteredHooks {
+				hookMatchCount[h] = hookMatchCount[h] + 1
+			}
+
+			var planErr error
+			plan, planErr = hooks.NewPlan(&filteredHooks, hooks.PhaseSnapshot, jobCallback, hookEnvExtra)
+			if planErr != nil {
+				fsHadErr = true
+				l.WithError(planErr).Error("cannot create job hook plan")
+				goto updateFSState
+			}
+		}
 		u(func(snapper *Snapper) {
 			progress.name = snapname
 			progress.startAt = time.Now()
+			progress.hookPlan = plan
 			progress.state = SnapStarted
 		})
-
-		l.Debug("create snapshot")
-		err := zfs.ZFSSnapshot(fs, snapname, false)
-		if err != nil {
-			hadErr = true
-			l.WithError(err).Error("cannot create snapshot")
-		}
-		doneAt := time.Now()
-
-		u(func(snapper *Snapper) {
-			progress.doneAt = doneAt
-			progress.state = SnapDone
-			if err != nil {
-				progress.state = SnapError
-				progress.err = err
+		{
+			l := hooks.GetLogger(a.ctx).WithField("fs", fs.ToString()).WithField("snap", snapname)
+			l.WithField("report", plan.Report().String()).Debug("begin run job plan")
+			plan.Run(hooks.WithLogger(a.ctx, l), a.dryRun)
+			planReport = plan.Report()
+			fsHadErr = planReport.HadError() // not just fatal errors
+			if fsHadErr {
+				l.WithField("report", planReport.String()).Error("end run job plan with error")
+			} else {
+				l.WithField("report", planReport.String()).Info("end run job plan successful")
 			}
+		}
+
+	updateFSState:
+		anyFsHadErr = anyFsHadErr || fsHadErr
+		u(func(snapper *Snapper) {
+			progress.doneAt = time.Now()
+			progress.state = SnapDone
+			if fsHadErr {
+				progress.state = SnapError
+			}
+			progress.runResults = planReport
 		})
 	}
 
@@ -283,12 +345,26 @@ func snapshot(a args, u updater) state {
 		}
 	}
 
+	for h, mc := range hookMatchCount {
+		if mc == 0 {
+			hookIdx := -1
+			for idx, ah := range *a.hooks {
+				if ah == h {
+					hookIdx = idx
+					break
+				}
+			}
+			a.log.WithField("hook", h.String()).WithField("hook_number", hookIdx+1).Warn("hook did not match any snapshotted filesystems")
+		}
+	}
+
 	return u(func(snapper *Snapper) {
-		if hadErr {
+		if anyFsHadErr {
 			snapper.state = ErrorWait
 			snapper.err = errors.New("one or more snapshots could not be created, check logs for details")
 		} else {
 			snapper.state = Waiting
+			snapper.err = nil
 		}
 	}).sf()
 }

daemon/snapper/snapper_all.go

@@ -24,6 +24,14 @@ func (s *PeriodicOrManual) Run(ctx context.Context, wakeUpCommon chan<- struct{}
 	}
 }
 
+// Returns nil if manual
+func (s *PeriodicOrManual) Report() *Report {
+	if s.s != nil {
+		return s.s.Report()
+	}
+	return nil
+}
+
 func FromConfig(g *config.Global, fsf *filters.DatasetMapFilter, in config.SnapshottingEnum) (*PeriodicOrManual, error) {
 	switch v := in.Ret.(type) {
 	case *config.SnapshottingPeriodic:

daemon/snapper/snapper_report.go

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

daemon/snapper/snapstate_string.go

@@ -4,6 +4,16 @@ package snapper
 
 import "strconv"
 
+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[SnapPending-1]
+	_ = x[SnapStarted-2]
+	_ = x[SnapDone-4]
+	_ = x[SnapError-8]
+}
+
 const (
 	_SnapState_name_0 = "SnapPendingSnapStarted"
 	_SnapState_name_1 = "SnapDone"

daemon/snapper/state_string.go

@@ -4,6 +4,19 @@ package snapper
 
 import "strconv"
 
+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[SyncUp-1]
+	_ = x[SyncUpErrWait-2]
+	_ = x[Planning-4]
+	_ = x[Snapshotting-8]
+	_ = x[Waiting-16]
+	_ = x[ErrorWait-32]
+	_ = x[Stopped-64]
+}
+
 const (
 	_State_name_0 = "SyncUpSyncUpErrWait"
 	_State_name_1 = "Planning"

docs/changelog.rst

@@ -1,3 +1,5 @@
+.. include:: global.rst.inc
+
 .. |break_config| replace:: **[CONFIG]**
 .. |break| replace:: **[BREAK]**
 .. |bugfix| replace:: [BUG]
@@ -25,6 +27,46 @@ We use the following annotations for classifying changes:
 * |bugfix| Change that fixes a bug, no regressions or incompatibilities expected.
 * |docs| Change to the documentation.
 
+0.2
+---
+
+* |feature| :ref:`Pre- and Post-Snapshot Hooks <job-snapshotting-hooks>`
+  with built-in support for MySQL and Postgres checkpointing
+  as well as custom scripts (thanks, `@overhacked <https://github.com/overhacked>`_!)
+* |feature| Use ``zfs destroy pool/fs@snap1,snap2,...`` CLI feature if available
+* |feature| Linux ARM64 Docker build support & binary builds
+* |feature| ``zrepl status`` now displays snapshotting reports
+* |feature| ``zrepl status --job <JOBNAME>`` filter flag
+* |bugfix| i386 build
+* |bugfix| early validation of host:port tuples in config
+* |bugfix| ``zrepl status`` now supports ``TERM=screen`` (tmux on FreeBSD / FreeNAS)
+* |bugfix| ignore *connection reset by peer* errors when shutting down connections
+* |bugfix| correct error messages when receive-side pool or ``root_fs`` dataset is not imported
+* |bugfix| fail fast for misconfigured local transport
+* |bugfix| race condition in replication report generation would crash the daemon when running ``zrepl status``
+* |bugfix| rpc goroutine leak in ``push`` mode if zfs recv fails on the ``sink`` side
+* [MAINTAINER NOTICE] Go modules for dependency management both inside and outside of GOPATH
+  (``lazy.sh`` and ``Makefile`` force ``GO111MODULE=on``)
+* [MAINTAINER NOTICE] ``make platformtest`` target to check zrepl's ZFS abstractions (screen scraping, etc.).
+  These tests only work on a system with ZFS installed, and must be run as root because they create a file-backed pool for each test case.
+  The pool name ``zreplplatformtest`` is reserved for this use case.
+  Only run ``make platformtest`` on test systems, e.g. a FreeBSD VM image.
+
+.. NOTE::
+   |  zrepl is a spare-time project primarily developed by `Christian Schwarz <https://cschwarz.com>`_.
+   |  You can support maintenance and feature development through one of the following services:
+   |  |Donate via Patreon| |Donate via Liberapay| |Donate via PayPal|
+   |  For commerical support, please `contact Christian directly <https://cschwarz.com>`_.
+
+
+0.1.1
+-----
+
+* |bugfix| :issue:`162` :commit:`d6304f4` : fix I/O timeout errors on variable receive rate
+
+  * A significant reduction or sudden stall of the receive rate (e.g. recv pool has other I/O to do)
+    would cause a ``writev I/O timeout`` error after approximately ten seconds.
+
 0.1
 ---
 

docs/configuration.rst

@@ -7,10 +7,11 @@ Configuration
 
 .. toctree::
 
-    configuration/preface
+    configuration/overview
     configuration/jobs
     configuration/transports
     configuration/filter_syntax
+    configuration/snapshotting
     configuration/prune
     configuration/logging
     configuration/monitoring

docs/configuration/jobs.rst

@@ -1,222 +1,9 @@
 .. include:: ../global.rst.inc
 
-.. |serve-transport| replace:: :ref:`serve specification<transport>`
-.. |connect-transport| replace:: :ref:`connect specification<transport>`
-.. |snapshotting-spec| replace:: :ref:`snapshotting specification <job-snapshotting-spec>`
-.. |pruning-spec| replace:: :ref:`pruning specification <prune>`
-.. |filter-spec| replace:: :ref:`filter specification<pattern-filter>`
-
 .. _job:
 
-Job Types & Replication
-=======================
-
-.. _job-overview:
-
-Overview & Terminology
-----------------------
-
-A *job* is the unit of activity tracked by the zrepl daemon.
-Every job has a unique ``name``, a ``type`` and type-dependent fields which are documented on this page.
-
-Replication always happens between a pair of jobs: one is the **active side**, and one the **passive side**.
-The active side executes the replication logic whereas the passive side responds to requests after checking the active side's permissions.
-For communication, the active side connects to the passive side using a :ref:`transport <transport>` and starts issuing remote procedure calls (RPCs).
-
-The following table shows how different job types can be combined to achieve both push and pull mode setups:
-
-+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
-| Setup name            | active side  | passive side                     | use case                                                                           |
-+=======================+==============+==================================+====================================================================================+
-| Push mode             | ``push``     | ``sink``                         | * Laptop backup                                                                    |
-|                       |              |                                  | * NAS behind NAT to offsite                                                        |
-+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
-| Pull mode             | ``pull``     | ``source``                       | * Central backup-server for many nodes                                             |
-|                       |              |                                  | * Remote server to NAS behind NAT                                                  |
-+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
-| Local replication     | | ``push`` + ``sink`` in one config             | * Backup FreeBSD boot pool                                                         |
-|                       | | with :ref:`local transport <transport-local>` |                                                                                    |
-+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
-| Snap & prune-only     | ``snap``     | N/A                              | * | Snapshots & pruning but no replication                                         |
-|                       |              |                                  |   | required                                                                       |
-|                       |              |                                  | * Workaround for :ref:`source-side pruning <prune-workaround-source-side-pruning>` |
-+-----------------------+--------------+----------------------------------+------------------------------------------------------------------------------------+
-
-How the Active Side Works
-~~~~~~~~~~~~~~~~~~~~~~~~~
-
-The active side (:ref:`push <job-push>` and :ref:`pull <job-pull>` job) executes the replication and pruning logic:
-
-* Wakeup because of finished snapshotting (``push`` job) or pull interval ticker (``pull`` job).
-* Connect to the corresponding passive side using a :ref:`transport <transport>` and instantiate an RPC client.
-* Replicate data from the sending to the receiving side.
-* Prune on sender & receiver.
-
-.. TIP::
-  The progress of the active side can be watched live using the ``zrepl status`` subcommand.
-
-How the Passive Side Works
-~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-The passive side (:ref:`sink <job-sink>` and :ref:`source <job-source>`) waits for connections from the corresponding active side,
-using the transport listener type specified in the ``serve`` field of the job configuration.
-Each transport listener provides a client's identity to the passive side job.
-It uses the client identity for access control:
-
-* The ``sink`` job maps requests from different client identities to their respective sub-filesystem tree ``root_fs/${client_identity}``.
-* The ``source`` job has a whitelist of client identities that are allowed pull access.
-
-.. TIP::
-   The implementation of the ``sink`` job requires that the connecting client identities be a valid ZFS filesystem name components.
-
-How Replication Works
-~~~~~~~~~~~~~~~~~~~~~
-
-One of the major design goals of the replication module is to avoid any duplication of the nontrivial logic.
-As such, the code works on abstract senders and receiver **endpoints**, where typically one will be implemented by a local program object and the other is an RPC client instance.
-Regardless of push- or pull-style setup, the logic executes on the active side, i.e. in the ``push`` or ``pull`` job.
-
-The following steps take place during replication and can be monitored using the ``zrepl status`` subcommand:
-
-* Plan the replication:
-
-  * Compare sender and receiver filesystem snapshots
-  * Build the **replication plan**
-
-    * Per filesystem, compute a diff between sender and receiver snapshots
-    * Build a list of replication steps
-
-      * If possible, use incremental sends (``zfs send -i``)
-      * Otherwise, use full send of most recent snapshot on sender
-      * Give up on filesystems that cannot be replicated without data loss
-
-  * Retry on errors that are likely temporary (i.e. network failures).
-  * Give up on filesystems where a permanent error was received over RPC.
-
-* Execute the plan
-
-  * Perform replication steps in the following order:
-    Among all filesystems with pending replication steps, pick the filesystem whose next replication step's snapshot is the oldest.
-  * Create placeholder filesystems on the receiving side to mirror the dataset paths on the sender to ``root_fs/${client_identity}``.
-  * After a successful replication step, update the replication cursor bookmark (see below).
-   
-The idea behind the execution order of replication steps is that if the sender snapshots all filesystems simultaneously at fixed intervals, the receiver will have all filesystems snapshotted at time ``T1`` before the first snapshot at ``T2 = T1 + $interval`` is replicated.
-
-.. _replication-cursor-bookmark:
-
-The **replication cursor bookmark** ``#zrepl_replication_cursor`` is kept per filesystem on the sending side of a replication setup:
-It is a bookmark of the most recent successfully replicated snapshot to the receiving side.
-It is is used by the :ref:`not_replicated <prune-keep-not-replicated>` keep rule to identify all snapshots that have not yet been replicated to the receiving side.
-Regardless of whether that keep rule is used, the bookmark ensures that replication can always continue incrementally.
-Note that there is only one cursor bookmark per filesystem, which prohibits multiple jobs to replicate the same filesystem (:ref:`see below<jobs-multiple-jobs>`).
-
-.. _replication-placeholder-property:
-
-**Placeholder filesystems** on the receiving side are regular ZFS filesystems with the placeholder property ``zrepl:placeholder=on``.
-Placeholders allow the receiving side to mirror the sender's ZFS dataset hierachy without replicating every filesystem at every intermediary dataset path component.
-Consider the following example: ``S/H/J`` shall be replicated to ``R/sink/job/S/H/J``, but neither ``S/H`` nor ``S`` shall be replicated.
-ZFS requires the existence of ``R/sink/job/S`` and ``R/sink/job/S/H`` in order to receive into ``R/sink/job/S/H/J``.
-Thus, zrepl creates the parent filesystems as placeholders on the receiving side.
-If at some point ``S/H`` and ``S`` shall be replicated, the receiving side invalidates the placeholder flag automatically.
-The ``zrepl test placeholder`` command can be used to check whether a filesystem is a placeholder.
-
-.. ATTENTION::
-
-    Currently, zrepl does not replicate filesystem properties.
-    Whe receiving a filesystem, it is never mounted (`-u` flag)  and `mountpoint=none` is set.
-    This is temporary and being worked on :issue:`24`.
-
-
-.. _job-snapshotting-spec:
-
-Taking Snaphots
----------------
-
-The ``push``, ``source`` and ``snap`` jobs can automatically take periodic snapshots of the filesystems matched by the ``filesystems`` filter field.
-The snapshot names are composed of a user-defined prefix followed by a UTC date formatted like ``20060102_150405_000``.
-We use UTC because it will avoid name conflicts when switching time zones or between summer and winter time.
-
-For ``push`` jobs, replication is automatically triggered after all filesystems have been snapshotted.
-
-::
-
-    jobs:
-    - type: push
-      filesystems: {
-        "<": true,
-        "tmp": false
-      }
-      snapshotting:
-        type: periodic
-        prefix: zrepl_
-        interval: 10m
-      ...
-
-There is also a ``manual`` snapshotting type, which covers the following use cases:
-
-* Existing infrastructure for automatic snapshots: you only want to use this zrepl job for replication.
-* Run scripts before and after taking snapshots (like locking database tables).
-  We are working on better integration for this use case: see :issue:`74`.
-* Handling snapshotting through a separate ``snap`` job.
-
-Note that you will have to trigger replication manually using the ``zrepl signal wakeup JOB`` subcommand in that case.
-
-::
-
-   jobs:
-   - type: push
-     filesystems: {
-       "<": true,
-       "tmp": false
-     }
-     snapshotting:
-       type: manual
-     ...
-
-.. _jobs-multiple-jobs:
-
-Multiple Jobs & More than 2 Machines
-------------------------------------
-
-.. ATTENTION::
-
-  When using multiple jobs across single or multiple machines, the following rules are critical to avoid race conditions & data loss:
-
-  1. The sets of ZFS filesystems matched by the ``filesystems`` filter fields must be disjoint across all jobs configured on a machine.
-  2. The ZFS filesystem subtrees of jobs with ``root_fs`` must be disjoint.
-  3. Across all zrepl instances on all machines in the replication domain, there must be a 1:1 correspondence between active and passive jobs.
-
-  Explanations & exceptions to above rules are detailed below.
-
-If you would like to see improvements to multi-job setups, please `open an issue on GitHub <https://github.com/zrepl/zrepl/issues/new>`_.
-
-No Overlapping
-~~~~~~~~~~~~~~
-
-Jobs run independently of each other.
-If two jobs match the same filesystem with their ``filesystems`` filter, they will operate on that filesystem independently and potentially in parallel.
-For example, if job A prunes snapshots that job B is planning to replicate, the replication will fail because B asssumed the snapshot to still be present.
-More subtle race conditions can occur with the :ref:`replication cursor bookmark <replication-cursor-bookmark>`, which currently only exists once per filesystem.
-
-N push jobs to 1 sink
-~~~~~~~~~~~~~~~~~~~~~
-
-The :ref:`sink job <job-sink>` namespaces by client identity.
-It is thus safe to push to one sink job with different client identities.
-If the push jobs have the same client identity, the filesystems matched by the push jobs must be disjoint to avoid races.
-
-N pull jobs from 1 source
-~~~~~~~~~~~~~~~~~~~~~~~~~
-
-Multiple pull jobs pulling from the same source have potential for race conditions during pruning:
-each pull job prunes the source side independently, causing replication-prune and prune-prune races.
-
-There is currently no way for a pull job to filter which snapshots it should attempt to replicate.
-Thus, it is not possibe to just manually assert that the prune rules of all pull jobs are disjoint to avoid replication-prune and prune-prune races.
-
-
-------------------------------------------------------------------------------
-
+Job Types in Detail
+===================
 
 .. _job-push:
 

docs/configuration/misc.rst

@@ -6,16 +6,16 @@ Miscellaneous
 Runtime Directories & UNIX Sockets
 ----------------------------------
 
-zrepl daemon creates various UNIX sockets to allow communicating with it:
+The zrepl daemon needs to open various UNIX sockets in a runtime directory:
 
-* the :ref:`transport-ssh+stdinserver` transport connects to a socket named after ``client_identity`` parameter
-* the ``control`` CLI subcommand connects to a defined control socket
+* a ``control`` socket that the CLI commands use to interact with the daemon
+* the :ref:`transport-ssh+stdinserver` listener opens one socket per configured client, named after ``client_identity`` parameter
 
-There is no further authentication on these sockets.
-Therefore we have to make sure they can only be created and accessed by ``zrepl daemon``.
-In fact, ``zrepl daemon`` will not bind a socket to a path in a directory that is world-accessible.
+There is no authentication on these sockets except the UNIX permissions.
+The zrepl daemon will refuse to bind any of the above sockets in a directory that is world-accessible.
 
-The directories can be configured in the main configuration file, the defaults are provided below:
+The following sections of the ``global`` config shows the default paths.
+The shell script below shows how the default runtime directory can be created.
 
 ::
 
@@ -27,6 +27,12 @@ The directories can be configured in the main configuration file, the defaults a
           sockdir: /var/run/zrepl/stdinserver
 
 
+::
+
+    mkdir -p /var/run/zrepl/stdinserver
+    chmod -R 0700 /var/run/zrepl
+
+
 Durations & Intervals
 ---------------------
 

docs/configuration/overview.rst

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

docs/configuration/preface.rst

@@ -1,39 +0,0 @@
-
-.. _configuration_preface:
-
-=======
-Preface
-=======
-
------------------------
-Configuration File Path
------------------------
-
-zrepl searches for its main configuration file in the following locations (in that order):
-
-* If set, the location specified via the global ``--config`` flag
-* ``/etc/zrepl/zrepl.yml``
-* ``/usr/local/etc/zrepl/zrepl.yml``
-
-The examples in the :ref:`tutorial` or the :sampleconf:`/` directory should provide a good starting point.
-
--------------------
-Runtime Directories
--------------------
-
-zrepl requires runtime directories for various UNIX sockets --- they are documented in the :ref:`config file<conf-runtime-directories>`.
-Your package maintainer / init script should take care of creating them.
-Alternatively, for default settings, the following should to the trick.
-
-::
-
-    mkdir -p /var/run/zrepl/stdinserver
-    chmod -R 0700 /var/run/zrepl
-
-
-----------
-Validating
-----------
-
-The config can be validated using the ``zrepl configcheck`` subcommand.
-

docs/configuration/snapshotting.rst

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

docs/configuration/transports.rst

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

docs/global.rst.inc

@@ -2,3 +2,25 @@
 
 .. 
   .. |macroname| replace:: macro expansion
+
+.. |GitHub license| image:: https://img.shields.io/github/license/zrepl/zrepl.svg
+   :target: https://github.com/zrepl/zrepl/blob/master/LICENSE
+.. |Language: Go| image:: https://img.shields.io/badge/language-Go-6ad7e5.svg
+   :target: https://golang.org/
+.. |User Docs| image:: https://img.shields.io/badge/docs-web-blue.svg
+   :target: https://zrepl.github.io
+.. |Donate via PayPal| image:: https://img.shields.io/badge/donate-paypal-yellow.svg
+   :target: https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96
+.. |Donate via Liberapay| image:: https://img.shields.io/liberapay/receives/zrepl.svg?logo=liberapay
+   :target: https://liberapay.com/zrepl/donate
+.. |Donate via Patreon| image:: https://img.shields.io/endpoint.svg?url=https%3A%2F%2Fshieldsio-patreon.herokuapp.com%2Fzrepl%2Fpledges&style=flat&color=yellow
+   :target: https://www.patreon.com/zrepl
+.. |Twitter| image:: https://img.shields.io/twitter/url/https/github.com/zrepl/zrepl.svg?style=social
+   :target: https://twitter.com/intent/tweet?text=Wow:&url=https%3A%2F%2Fgithub.com%2Fzrepl%2Fzrepl
+
+
+.. |serve-transport| replace:: :ref:`serve specification<transport>`
+.. |connect-transport| replace:: :ref:`connect specification<transport>`
+.. |snapshotting-spec| replace:: :ref:`snapshotting specification <job-snapshotting-spec>`
+.. |pruning-spec| replace:: :ref:`pruning specification <prune>`
+.. |filter-spec| replace:: :ref:`filter specification<pattern-filter>`

docs/index.rst

@@ -5,21 +5,8 @@
 
 .. include:: global.rst.inc
 
-|GitHub license| |Language: Go| |User Docs| |Donate via PayPal| |Donate
-via Liberapay| |Twitter|
+|GitHub license| |Language: Go| |Twitter| |Donate via Patreon| |Donate via Liberapay| |Donate via PayPal| 
 
-.. |GitHub license| image:: https://img.shields.io/github/license/zrepl/zrepl.svg
-   :target: https://github.com/zrepl/zrepl/blob/master/LICENSE
-.. |Language: Go| image:: https://img.shields.io/badge/language-Go-6ad7e5.svg
-   :target: https://golang.org/
-.. |User Docs| image:: https://img.shields.io/badge/docs-web-blue.svg
-   :target: https://zrepl.github.io
-.. |Donate via PayPal| image:: https://img.shields.io/badge/donate-paypal-yellow.svg
-   :target: https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96
-.. |Donate via Liberapay| image:: https://img.shields.io/badge/donate-liberapay-yellow.svg
-   :target: https://liberapay.com/zrepl/donate
-.. |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
 
 zrepl - ZFS replication
 -----------------------
@@ -76,7 +63,8 @@ Main Features
 
 * **Automatic snapshot management**
 
-  * [x] Periodic filesystem snapshots
+  * [x] Periodic :ref:`filesystem snapshots <job-snapshotting-spec>`
+  * [x] Support for :ref:`pre- and post-snapshot hooks <job-snapshotting-hooks>` with builtins for MySQL & Postgres
   * [x] Flexible :ref:`pruning rule system <prune>`
 
     * [x] Age-based fading (grandfathering scheme)
@@ -109,6 +97,7 @@ Contributing
 
 We are happy about any help we can get!
 
+* :ref:`Financial Support <supporters>`
 * Explore the codebase
 
   * These docs live in the ``docs/`` subdirectory
@@ -124,6 +113,8 @@ We are happy about any help we can get!
     Make sure to read the `Developer Documentation section <https://github.com/zrepl/zrepl>`_ and open new issues or pull requests there.
 
 
+
+
 Table of Contents
 ~~~~~~~~~~~~~~~~~
 
@@ -139,4 +130,5 @@ Table of Contents
    pr
    changelog
    GitHub Repository & Issue Tracker <https://github.com/zrepl/zrepl>
+   supporters
 

docs/installation.rst

@@ -34,6 +34,9 @@ The following list may be incomplete, feel free to submit a PR with an update:
     * - FreeBSD
       - ``pkg install zrepl``
       - `<https://www.freshports.org/sysutils/zrepl/>`_
+    * - MacOS
+      - ``brew install zrepl``
+      - Available on `homebrew <https://brew.sh>`_
     * - Others
       -
       - Use `binary releases`_ or build from source.
@@ -41,7 +44,7 @@ The following list may be incomplete, feel free to submit a PR with an update:
 Compile From Source
 ~~~~~~~~~~~~~~~~~~~
 
-Producing a release requires **Go 1.9** or newer and **Python 3** + **pip3** + ``docs/requirements.txt`` for the Sphinx documentation.
+Producing a release requires **Go 1.11** or newer and **Python 3** + **pip3** + ``docs/requirements.txt`` for the Sphinx documentation.
 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.
 
@@ -55,9 +58,9 @@ and serves as a reference for build dependencies and procedure:
     cd zrepl
     sudo docker build -t zrepl_build -f build.Dockerfile .
     sudo docker run -it --rm \
-        -v "${PWD}:/go/src/github.com/zrepl/zrepl" \
+        -v "${PWD}:/src" \
         --user "$(id -u):$(id -g)" \
-        zrepl_build make vendordeps release
+        zrepl_build make release
 
 Alternatively, you can install build dependencies on your local system and then build in your ``$GOPATH``:
 
@@ -69,7 +72,7 @@ Alternatively, you can install build dependencies on your local system and then
     python3 -m venv3
     source venv3/bin/activate
     ./lazy.sh devsetup
-    make vendordeps release
+    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.
@@ -83,8 +86,7 @@ Either way, all build results are located in the ``artifacts/`` directory.
 What next?
 ----------
 
-Read the :ref:`configuration chapter<configuration_toc>` and optionally create the :ref:`runtime directories <configuration_preface>`.
-Afterwards, continue with the :ref:`usage chapter<usage>`.
+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:`tutorial`.
 

docs/requirements.txt

@@ -1,20 +1,30 @@
 alabaster==0.7.12
-Babel==2.6.0
-certifi==2018.10.15
+attrs==19.1.0
+Babel==2.7.0
+certifi==2019.6.16
 chardet==3.0.4
-docutils==0.14
-idna==2.7
+Click==7.0
+colorclass==2.2.0
+docutils==0.15.2
+idna==2.8
 imagesize==1.1.0
-Jinja2==2.10
-MarkupSafe==1.1.0
-packaging==18.0
-Pygments==2.2.0
-pyparsing==2.3.0
-pytz==2018.7
-requests==2.20.1
-six==1.11.0
-snowballstemmer==1.2.1
-Sphinx==1.8.2
-sphinx-rtd-theme==0.4.2
-sphinxcontrib-websupport==1.1.0
-urllib3==1.24.1
+Jinja2==2.10.1
+MarkupSafe==1.1.1
+packaging==19.1
+Pygments==2.4.2
+pyparsing==2.4.2
+pytz==2019.2
+requests==2.22.0
+six==1.12.0
+snowballstemmer==1.9.1
+Sphinx==1.8.5
+sphinx-rtd-theme==0.4.3
+sphinxcontrib-applehelp==1.0.1
+sphinxcontrib-devhelp==1.0.1
+sphinxcontrib-htmlhelp==1.0.2
+sphinxcontrib-jsmath==1.0.1
+sphinxcontrib-qthelp==1.0.2
+sphinxcontrib-serializinghtml==1.1.3
+-e git://github.com/rwblair/sphinxcontrib-versioning.git@7e3885a389a809e17ea55261316b7b0e98dbf98f#egg=sphinxcontrib-versioning
+sphinxcontrib-websupport==1.1.2
+urllib3==1.25.3

docs/supporters.rst

@@ -0,0 +1,28 @@
+.. include:: global.rst.inc
+
+.. _supporters:
+
+|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.
+For commerical support, please `contact Christian directly <https://cschwarz.com>`_.
+
+**Thanks for your support!**
+
+.. NOTE ::
+
+   PayPal takes a relatively high fixed processing fee plus percentage of the donation.
+   Larger less-frequent donations make more sense there.
+
+Supporters
+==========
+
+We would like to thank the following people / organizations for supporting zrepl through monetary and other means:
+
+* `Ross Williams <https://github.com/overhacked>`_
+* Mike T.
+* `Justin Scholz <https://github.com/JMoVS>`_
+* `InsanePrawn <https://github.com/InsanePrawn>`_
+* `Ben Woods <https://www.freshports.org/sysutils/zrepl/>`_
+* `Anton Schirg <https://github.com/antonxy>`_

endpoint/endpoint.go

@@ -380,6 +380,15 @@ func (s *Receiver) Receive(ctx context.Context, req *pdu.ReceiveReq, receive zfs
 			}
 
 			if !ph.FSExists {
+				if s.rootWithoutClientComponent.HasPrefix(v.Path) {
+					if v.Path.Length() == 1 {
+						visitErr = fmt.Errorf("pool %q not imported", v.Path.ToString())
+					} else {
+						visitErr = fmt.Errorf("root_fs %q does not exist", s.rootWithoutClientComponent.ToString())
+					}
+					getLogger(ctx).WithError(visitErr).Error("placeholders are only created automatically below root_fs")
+					return false
+				}
 				l := getLogger(ctx).WithField("placeholder_fs", v.Path)
 				l.Debug("create placeholder filesystem")
 				err := zfs.ZFSCreatePlaceholderFilesystem(v.Path)
@@ -445,30 +454,34 @@ func (s *Receiver) DestroySnapshots(ctx context.Context, req *pdu.DestroySnapsho
 }
 
 func doDestroySnapshots(ctx context.Context, lp *zfs.DatasetPath, snaps []*pdu.FilesystemVersion) (*pdu.DestroySnapshotsRes, error) {
-	fsvs := make([]*zfs.FilesystemVersion, len(snaps))
+	reqs := make([]*zfs.DestroySnapOp, len(snaps))
+	ress := make([]*pdu.DestroySnapshotRes, len(snaps))
+	errs := make([]error, len(snaps))
 	for i, fsv := range snaps {
 		if fsv.Type != pdu.FilesystemVersion_Snapshot {
 			return nil, fmt.Errorf("version %q is not a snapshot", fsv.Name)
 		}
-		var err error
-		fsvs[i], err = fsv.ZFSFilesystemVersion()
-		if err != nil {
-			return nil, err
+		ress[i] = &pdu.DestroySnapshotRes{
+			Snapshot: fsv,
+			// Error set after batch operation
+		}
+		reqs[i] = &zfs.DestroySnapOp{
+			Filesystem: lp.ToString(),
+			Name:       fsv.Name,
+			ErrOut:     &errs[i],
 		}
 	}
-	res := &pdu.DestroySnapshotsRes{
-		Results: make([]*pdu.DestroySnapshotRes, len(fsvs)),
-	}
-	for i, fsv := range fsvs {
-		err := zfs.ZFSDestroyFilesystemVersion(lp, fsv)
-		errMsg := ""
-		if err != nil {
-			errMsg = err.Error()
-		}
-		res.Results[i] = &pdu.DestroySnapshotRes{
-			Snapshot: pdu.FilesystemVersionFromZFS(fsv),
-			Error:    errMsg,
+	zfs.ZFSDestroyFilesystemVersions(reqs)
+	for i := range reqs {
+		if errs[i] != nil {
+			if de, ok := errs[i].(*zfs.DestroySnapshotsError); ok && len(de.Reason) == 1 {
+				ress[i].Error = de.Reason[0]
+			} else {
+				ress[i].Error = errs[i].Error()
+			}
 		}
 	}
-	return res, nil
+	return &pdu.DestroySnapshotsRes{
+		Results: ress,
+	}, nil
 }

go.mod

@@ -0,0 +1,82 @@
+module github.com/zrepl/zrepl
+
+go 1.12
+
+require (
+	github.com/OpenPeeDeeP/depguard v0.0.0-20181229194401-1f388ab2d810 // indirect
+	github.com/alvaroloes/enumer v1.1.1
+	github.com/beorn7/perks v0.0.0-20180321164747-3a771d992973 // indirect
+	github.com/fatih/color v1.7.0
+	github.com/ftrvxmtrx/fd v0.0.0-20150925145434-c6d800382fff // indirect
+	github.com/gdamore/tcell v1.2.0
+	github.com/go-critic/go-critic v0.3.4 // indirect
+	github.com/go-logfmt/logfmt v0.3.0
+	github.com/go-sql-driver/mysql v1.4.1-0.20190907122137-b2c03bcae3d4
+	github.com/gogo/protobuf v1.2.1 // indirect
+	github.com/golang/mock v1.2.0 // indirect
+	github.com/golang/protobuf v1.2.0
+	github.com/golangci/errcheck v0.0.0-20181223084120-ef45e06d44b6 // indirect
+	github.com/golangci/go-tools v0.0.0-20190124090046-35a9f45a5db0 // indirect
+	github.com/golangci/gocyclo v0.0.0-20180528144436-0a533e8fa43d // indirect
+	github.com/golangci/gofmt v0.0.0-20181222123516-0b8337e80d98 // indirect
+	github.com/golangci/golangci-lint v1.17.1
+	github.com/golangci/gosec v0.0.0-20180901114220-8afd9cbb6cfb // indirect
+	github.com/golangci/lint-1 v0.0.0-20181222135242-d2cdd8c08219 // indirect
+	github.com/golangci/misspell v0.3.4 // indirect
+	github.com/golangci/revgrep v0.0.0-20180812185044-276a5c0a1039 // indirect
+	github.com/google/uuid v1.1.1
+	github.com/jinzhu/copier v0.0.0-20170922082739-db4671f3a9b8
+	github.com/k0kubun/colorstring v0.0.0-20150214042306-9440f1994b88 // indirect
+	github.com/konsorten/go-windows-terminal-sequences v1.0.2 // indirect
+	github.com/kr/logfmt v0.0.0-20140226030751-b84e30acd515 // indirect
+	github.com/kr/pretty v0.1.0
+	github.com/lib/pq v1.2.0
+	github.com/mattn/go-isatty v0.0.3
+	github.com/matttproud/golang_protobuf_extensions v1.0.0 // indirect
+	github.com/mitchellh/go-homedir v1.1.0 // indirect
+	github.com/montanaflynn/stats v0.5.0
+	github.com/pkg/errors v0.8.0
+	github.com/pkg/profile v1.2.1
+	github.com/problame/go-netssh v0.0.0-20190110232351-09d6bc45d284
+	github.com/prometheus/client_golang v0.0.0-20180410130117-e11c6ff8170b
+	github.com/prometheus/client_model v0.0.0-20171117100541-99fa1f4be8e5 // indirect
+	github.com/prometheus/common v0.0.0-20180413074202-d0f7cd64bda4 // indirect
+	github.com/prometheus/procfs v0.0.0-20180408092902-8b1c2da0d56d // indirect
+	github.com/sergi/go-diff v1.0.0 // indirect
+	github.com/sirupsen/logrus v1.4.0 // indirect
+	github.com/spf13/afero v1.2.2 // indirect
+	github.com/spf13/cobra v0.0.2
+	github.com/spf13/jwalterweatherman v1.1.0 // indirect
+	github.com/spf13/pflag v1.0.3
+	github.com/spf13/viper v1.3.2 // indirect
+	github.com/stretchr/testify v1.4.0
+	github.com/theckman/goconstraint v1.11.0 // indirect
+	github.com/yudai/gojsondiff v0.0.0-20170107030110-7b1b7adf999d
+	github.com/yudai/golcs v0.0.0-20170316035057-ecda9a501e82 // indirect
+	github.com/yudai/pp v2.0.1+incompatible // indirect
+	github.com/zrepl/yaml-config v0.0.0-20190928121844-af7ca3f8448f
+	golang.org/x/crypto v0.0.0-20190325154230-a5d413f7728c // indirect
+	golang.org/x/net v0.0.0-20190313220215-9f648a60d977
+	golang.org/x/sync v0.0.0-20190423024810-112230192c58
+	golang.org/x/sys v0.0.0-20190626150813-e07cf5db2756
+	golang.org/x/tools v0.0.0-20190524210228-3d17549cdc6b
+	google.golang.org/genproto v0.0.0-20181202183823-bd91e49a0898 // indirect
+	google.golang.org/grpc v1.17.0
+	mvdan.cc/unparam v0.0.0-20190310220240-1b9ccfa71afe // indirect
+	sourcegraph.com/sqs/pbtypes v1.0.0 // indirect
+)
+
+// invalid dates in transitive dependencies (first validated in Go 1.13, didn't fail in earlier Go versions)
+replace (
+	github.com/go-critic/go-critic v0.0.0-20181204210945-1df300866540 => github.com/go-critic/go-critic v0.3.5-0.20190526074819-1df300866540
+	github.com/go-critic/go-critic v0.0.0-20181204210945-ee9bf5809ead => github.com/go-critic/go-critic v0.3.5-0.20190210220443-ee9bf5809ead
+	github.com/golangci/errcheck v0.0.0-20181003203344-ef45e06d44b6 => github.com/golangci/errcheck v0.0.0-20181223084120-ef45e06d44b6
+	github.com/golangci/go-tools v0.0.0-20180109140146-35a9f45a5db0 => github.com/golangci/go-tools v0.0.0-20190124090046-35a9f45a5db0
+	github.com/golangci/go-tools v0.0.0-20180109140146-af6baa5dc196 => github.com/golangci/go-tools v0.0.0-20190318060251-af6baa5dc196
+	github.com/golangci/gofmt v0.0.0-20181105071733-0b8337e80d98 => github.com/golangci/gofmt v0.0.0-20181222123516-0b8337e80d98
+	github.com/golangci/gosec v0.0.0-20180901114220-66fb7fc33547 => github.com/golangci/gosec v0.0.0-20190211064107-66fb7fc33547
+	github.com/golangci/ineffassign v0.0.0-20180808204949-42439a7714cc => github.com/golangci/ineffassign v0.0.0-20190609212857-42439a7714cc
+	github.com/golangci/lint-1 v0.0.0-20180610141402-ee948d087217 => github.com/golangci/lint-1 v0.0.0-20190420132249-ee948d087217
+	golang.org/x/tools v0.0.0-20190125232054-379209517ffe => golang.org/x/tools v0.0.0-20190205201329-379209517ffe
+	mvdan.cc/unparam v0.0.0-20190124213536-fbb59629db34 => mvdan.cc/unparam v0.0.0-20190209190245-fbb59629db34
+)

go.sum

@@ -0,0 +1,358 @@
+cloud.google.com/go v0.26.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
+github.com/BurntSushi/toml v0.3.1 h1:WXkYYl6Yr3qBf1K79EBnL4mak0OimBfB0XUf9Vl28OQ=
+github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
+github.com/DATA-DOG/go-sqlmock v1.3.3 h1:CWUqKXe0s8A2z6qCgkP4Kru7wC11YoAnoupUKFDnH08=
+github.com/DATA-DOG/go-sqlmock v1.3.3/go.mod h1:f/Ixk793poVmq4qj/V1dPUg2JEAKC73Q5eFN3EC/SaM=
+github.com/OpenPeeDeeP/depguard v0.0.0-20180806142446-a69c782687b2/go.mod h1:7/4sitnI9YlQgTLLk734QlzXT8DuHVnAyztLplQjk+o=
+github.com/OpenPeeDeeP/depguard v0.0.0-20181229194401-1f388ab2d810 h1:pFks+oaqVWDFq0KsLQZFB2pQGB5Us0HfMSBENtieXOs=
+github.com/OpenPeeDeeP/depguard v0.0.0-20181229194401-1f388ab2d810/go.mod h1:7/4sitnI9YlQgTLLk734QlzXT8DuHVnAyztLplQjk+o=
+github.com/StackExchange/wmi v0.0.0-20180116203802-5d049714c4a6/go.mod h1:3eOhrUMpNV+6aFIbp5/iudMxNCF27Vw2OZgy4xEx0Fg=
+github.com/alvaroloes/enumer v1.1.1 h1:v+1scNqEhSFAPb9tLwblQegeTXJhnw8hf9O0amTbOvQ=
+github.com/alvaroloes/enumer v1.1.1/go.mod h1:FxrjvuXoDAx9isTJrv4c+T410zFi0DtXIT0m65DJ+Wo=
+github.com/armon/consul-api v0.0.0-20180202201655-eb2c6b5be1b6/go.mod h1:grANhF5doyWs3UAsr3K4I6qtAmlQcZDesFNEHPZAzj8=
+github.com/beorn7/perks v0.0.0-20180321164747-3a771d992973 h1:xJ4a3vCFaGF/jqvzLMYoU8P317H5OQ+Via4RmuPwCS0=
+github.com/beorn7/perks v0.0.0-20180321164747-3a771d992973/go.mod h1:Dwedo/Wpr24TaqPxmxbtue+5NUziq4I4S80YR8gNf3Q=
+github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
+github.com/coreos/etcd v3.3.10+incompatible/go.mod h1:uF7uidLiAD3TWHmW31ZFd/JWoc32PjwdhPthX9715RE=
+github.com/coreos/go-etcd v2.0.0+incompatible/go.mod h1:Jez6KQU2B/sWsbdaef3ED8NzMklzPG4d5KIOhIy30Tk=
+github.com/coreos/go-semver v0.2.0/go.mod h1:nnelYz7RCh+5ahJtPPxZlU+153eP4D4r3EedlOD2RNk=
+github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
+github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/fatih/color v1.6.0/go.mod h1:Zm6kSWBoL9eyXnKyktHP6abPY2pDugNf5KwzbycvMj4=
+github.com/fatih/color v1.7.0 h1:DkWD4oS2D8LGGgTQ6IvwJJXSL5Vp2ffcQg58nFV38Ys=
+github.com/fatih/color v1.7.0/go.mod h1:Zm6kSWBoL9eyXnKyktHP6abPY2pDugNf5KwzbycvMj4=
+github.com/fsnotify/fsnotify v1.4.7 h1:IXs+QLmnXW2CcXuY+8Mzv/fWEsPGWxqefPtCP5CnV9I=
+github.com/fsnotify/fsnotify v1.4.7/go.mod h1:jwhsz4b93w/PPRr/qN1Yymfu8t87LnFCMoQvtojpjFo=
+github.com/ftrvxmtrx/fd v0.0.0-20150925145434-c6d800382fff h1:zk1wwii7uXmI0znwU+lqg+wFL9G5+vm5I+9rv2let60=
+github.com/ftrvxmtrx/fd v0.0.0-20150925145434-c6d800382fff/go.mod h1:yUhRXHewUVJ1k89wHKP68xfzk7kwXUx/DV1nx4EBMbw=
+github.com/gdamore/encoding v1.0.0 h1:+7OoQ1Bc6eTm5niUzBa0Ctsh6JbMW6Ra+YNuAtDBdko=
+github.com/gdamore/encoding v1.0.0/go.mod h1:alR0ol34c49FCSBLjhosxzcPHQbf2trDkoo5dl+VrEg=
+github.com/gdamore/tcell v1.2.0 h1:ikixzsxc8K8o3V2/CEmyoEW8mJZaNYQQ3NP3VIQdUe4=
+github.com/gdamore/tcell v1.2.0/go.mod h1:Hjvr+Ofd+gLglo7RYKxxnzCBmev3BzsS67MebKS4zMM=
+github.com/go-critic/go-critic v0.3.4 h1:FYaiaLjX0Nqei80KPhm4CyFQUBbmJwSrHxQ73taaGBc=
+github.com/go-critic/go-critic v0.3.4/go.mod h1:AHR42Lk/E/aOznsrYdMYeIQS5RH10HZHSqP+rD6AJrc=
+github.com/go-critic/go-critic v0.3.5-0.20190526074819-1df300866540/go.mod h1:+sE8vrLDS2M0pZkBk0wy6+nLdKexVDrl/jBqQOTDThA=
+github.com/go-lintpack/lintpack v0.5.1/go.mod h1:NwZuYi2nUHho8XEIZ6SIxihrnPoqBTDqfpXvXAN0sXM=
+github.com/go-lintpack/lintpack v0.5.2 h1:DI5mA3+eKdWeJ40nU4d6Wc26qmdG8RCi/btYq0TuRN0=
+github.com/go-lintpack/lintpack v0.5.2/go.mod h1:NwZuYi2nUHho8XEIZ6SIxihrnPoqBTDqfpXvXAN0sXM=
+github.com/go-logfmt/logfmt v0.3.0 h1:8HUsc87TaSWLKwrnumgC8/YconD2fJQsRJAsWaPg2ic=
+github.com/go-logfmt/logfmt v0.3.0/go.mod h1:Qt1PoO58o5twSAckw1HlFXLmHsOX5/0LbT9GBnD5lWE=
+github.com/go-ole/go-ole v1.2.1/go.mod h1:7FAglXiTm7HKlQRDeOQ6ZNUHidzCWXuZWq/1dTyBNF8=
+github.com/go-sql-driver/mysql v1.4.1-0.20190907122137-b2c03bcae3d4 h1:0suja/iKSDbEIYLbrS/8C7iArJiWpgCNcR+zwAHu7Ig=
+github.com/go-sql-driver/mysql v1.4.1-0.20190907122137-b2c03bcae3d4/go.mod h1:zAC/RDZ24gD3HViQzih4MyKcchzm+sOG5ZlKdlhCg5w=
+github.com/go-sql-driver/mysql v1.4.1 h1:g24URVg0OFbNUTx9qqY1IRZ9D9z3iPyi5zKhQZpNwpA=
+github.com/go-sql-driver/mysql v1.4.1/go.mod h1:zAC/RDZ24gD3HViQzih4MyKcchzm+sOG5ZlKdlhCg5w=
+github.com/go-toolsmith/astcast v0.0.0-20181028201508-b7a89ed70af1/go.mod h1:TEo3Ghaj7PsZawQHxT/oBvo4HK/sl1RcuUHDKTTju+o=
+github.com/go-toolsmith/astcast v1.0.0 h1:JojxlmI6STnFVG9yOImLeGREv8W2ocNUM+iOhR6jE7g=
+github.com/go-toolsmith/astcast v1.0.0/go.mod h1:mt2OdQTeAQcY4DQgPSArJjHCcOwlX+Wl/kwN+LbLGQ4=
+github.com/go-toolsmith/astcopy v0.0.0-20180903214859-79b422d080c4/go.mod h1:c9CPdq2AzM8oPomdlPniEfPAC6g1s7NqZzODt8y6ib8=
+github.com/go-toolsmith/astcopy v1.0.0 h1:OMgl1b1MEpjFQ1m5ztEO06rz5CUd3oBv9RF7+DyvdG8=
+github.com/go-toolsmith/astcopy v1.0.0/go.mod h1:vrgyG+5Bxrnz4MZWPF+pI4R8h3qKRjjyvV/DSez4WVQ=
+github.com/go-toolsmith/astequal v0.0.0-20180903214952-dcb477bfacd6/go.mod h1:H+xSiq0+LtiDC11+h1G32h7Of5O3CYFJ99GVbS5lDKY=
+github.com/go-toolsmith/astequal v1.0.0 h1:4zxD8j3JRFNyLN46lodQuqz3xdKSrur7U/sr0SDS/gQ=
+github.com/go-toolsmith/astequal v1.0.0/go.mod h1:H+xSiq0+LtiDC11+h1G32h7Of5O3CYFJ99GVbS5lDKY=
+github.com/go-toolsmith/astfmt v0.0.0-20180903215011-8f8ee99c3086/go.mod h1:mP93XdblcopXwlyN4X4uodxXQhldPGZbcEJIimQHrkg=
+github.com/go-toolsmith/astfmt v1.0.0 h1:A0vDDXt+vsvLEdbMFJAUBI/uTbRw1ffOPnxsILnFL6k=
+github.com/go-toolsmith/astfmt v1.0.0/go.mod h1:cnWmsOAuq4jJY6Ct5YWlVLmcmLMn1JUPuQIHCY7CJDw=
+github.com/go-toolsmith/astinfo v0.0.0-20180906194353-9809ff7efb21/go.mod h1:dDStQCHtmZpYOmjRP/8gHHnCCch3Zz3oEgCdZVdtweU=
+github.com/go-toolsmith/astp v0.0.0-20180903215135-0af7e3c24f30/go.mod h1:SV2ur98SGypH1UjcPpCatrV5hPazG6+IfNHbkDXBRrk=
+github.com/go-toolsmith/astp v1.0.0 h1:alXE75TXgcmupDsMK1fRAy0YUzLzqPVvBKoyWV+KPXg=
+github.com/go-toolsmith/astp v1.0.0/go.mod h1:RSyrtpVlfTFGDYRbrjyWP1pYu//tSFcvdYrA8meBmLI=
+github.com/go-toolsmith/pkgload v0.0.0-20181119091011-e9e65178eee8 h1:vVouagbdmqTVlCIAxpyYsNNTbkKZ3V66VpKOLU/s6W4=
+github.com/go-toolsmith/pkgload v0.0.0-20181119091011-e9e65178eee8/go.mod h1:WoMrjiy4zvdS+Bg6z9jZH82QXwkcgCBX6nOfnmdaHks=
+github.com/go-toolsmith/pkgload v1.0.0 h1:4DFWWMXVfbcN5So1sBNW9+yeiMqLFGl1wFLTL5R0Tgg=
+github.com/go-toolsmith/pkgload v1.0.0/go.mod h1:5eFArkbO80v7Z0kdngIxsRXRMTaX4Ilcwuh3clNrQJc=
+github.com/go-toolsmith/strparse v0.0.0-20180903215201-830b6daa1241/go.mod h1:YI2nUKP9YGZnL/L1/DLFBfixrcjslWct4wyljWhSRy8=
+github.com/go-toolsmith/strparse v1.0.0 h1:Vcw78DnpCAKlM20kSbAyO4mPfJn/lyYA4BJUDxe2Jb4=
+github.com/go-toolsmith/strparse v1.0.0/go.mod h1:YI2nUKP9YGZnL/L1/DLFBfixrcjslWct4wyljWhSRy8=
+github.com/go-toolsmith/typep v0.0.0-20181030061450-d63dc7650676/go.mod h1:JSQCQMUPdRlMZFswiq3TGpNp1GMktqkR2Ns5AIQkATU=
+github.com/go-toolsmith/typep v1.0.0 h1:zKymWyA1TRYvqYrYDrfEMZULyrhcnGY3x7LDKU2XQaA=
+github.com/go-toolsmith/typep v1.0.0/go.mod h1:JSQCQMUPdRlMZFswiq3TGpNp1GMktqkR2Ns5AIQkATU=
+github.com/gobwas/glob v0.2.3 h1:A4xDbljILXROh+kObIiy5kIaPYD8e96x1tgBhUI5J+Y=
+github.com/gobwas/glob v0.2.3/go.mod h1:d3Ez4x06l9bZtSvzIay5+Yzi0fmZzPgnTbPcKjJAkT8=
+github.com/gogo/protobuf v1.1.1/go.mod h1:r8qH/GZQm5c6nD/R0oafs1akxWv10x8SbQlK7atdtwQ=
+github.com/gogo/protobuf v1.2.1 h1:/s5zKNz0uPFCZ5hddgPdo2TK2TVrUNMn0OOX8/aZMTE=
+github.com/gogo/protobuf v1.2.1/go.mod h1:hp+jE20tsWTFYpLwKvXlhS1hjn+gTNwPg2I6zVXpSg4=
+github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b h1:VKtxabqXZkF25pY9ekfRL6a582T4P37/31XEstQ5p58=
+github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b/go.mod h1:SBH7ygxi8pfUlaOkMMuAQtPIUF8ecWP5IEl/CR7VP2Q=
+github.com/golang/lint v0.0.0-20180702182130-06c8688daad7/go.mod h1:tluoj9z5200jBnyusfRPU2LqT6J+DAorxEvtC7LHB+E=
+github.com/golang/mock v1.0.0/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/mock v1.1.1/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/mock v1.2.0 h1:28o5sBqPkBsMGnC6b4MvE2TzSr5/AT4c/1fLqVGIwlk=
+github.com/golang/mock v1.2.0/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/protobuf v1.2.0 h1:P3YflyNX/ehuJFLhxviNdFxQPkGK5cDcApsge1SqnvM=
+github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golangci/check v0.0.0-20180506172741-cfe4005ccda2 h1:23T5iq8rbUYlhpt5DB4XJkc6BU31uODLD1o1gKvZmD0=
+github.com/golangci/check v0.0.0-20180506172741-cfe4005ccda2/go.mod h1:k9Qvh+8juN+UKMCS/3jFtGICgW8O96FVaZsaxdzDkR4=
+github.com/golangci/dupl v0.0.0-20180902072040-3e9179ac440a h1:w8hkcTqaFpzKqonE9uMCefW1WDie15eSP/4MssdenaM=
+github.com/golangci/dupl v0.0.0-20180902072040-3e9179ac440a/go.mod h1:ryS0uhF+x9jgbj/N71xsEqODy9BN81/GonCZiOzirOk=
+github.com/golangci/errcheck v0.0.0-20181223084120-ef45e06d44b6 h1:YYWNAGTKWhKpcLLt7aSj/odlKrSrelQwlovBpDuf19w=
+github.com/golangci/errcheck v0.0.0-20181223084120-ef45e06d44b6/go.mod h1:DbHgvLiFKX1Sh2T1w8Q/h4NAI8MHIpzCdnBUDTXU3I0=
+github.com/golangci/go-misc v0.0.0-20180628070357-927a3d87b613 h1:9kfjN3AdxcbsZBf8NjltjWihK2QfBBBZuv91cMFfDHw=
+github.com/golangci/go-misc v0.0.0-20180628070357-927a3d87b613/go.mod h1:SyvUF2NxV+sN8upjjeVYr5W7tyxaT1JVtvhKhOn2ii8=
+github.com/golangci/go-tools v0.0.0-20190124090046-35a9f45a5db0 h1:MRhC9XbUjE6XDOInSJ8pwHuPagqsyO89QDU9IdVhe3o=
+github.com/golangci/go-tools v0.0.0-20190124090046-35a9f45a5db0/go.mod h1:unzUULGw35sjyOYjUt0jMTXqHlZPpPc6e+xfO4cd6mM=
+github.com/golangci/go-tools v0.0.0-20190318060251-af6baa5dc196/go.mod h1:unzUULGw35sjyOYjUt0jMTXqHlZPpPc6e+xfO4cd6mM=
+github.com/golangci/goconst v0.0.0-20180610141641-041c5f2b40f3 h1:pe9JHs3cHHDQgOFXJJdYkK6fLz2PWyYtP4hthoCMvs8=
+github.com/golangci/goconst v0.0.0-20180610141641-041c5f2b40f3/go.mod h1:JXrF4TWy4tXYn62/9x8Wm/K/dm06p8tCKwFRDPZG/1o=
+github.com/golangci/gocyclo v0.0.0-20180528134321-2becd97e67ee/go.mod h1:ozx7R9SIwqmqf5pRP90DhR2Oay2UIjGuKheCBCNwAYU=
+github.com/golangci/gocyclo v0.0.0-20180528144436-0a533e8fa43d h1:pXTK/gkVNs7Zyy7WKgLXmpQ5bHTrq5GDsp8R9Qs67g0=
+github.com/golangci/gocyclo v0.0.0-20180528144436-0a533e8fa43d/go.mod h1:ozx7R9SIwqmqf5pRP90DhR2Oay2UIjGuKheCBCNwAYU=
+github.com/golangci/gofmt v0.0.0-20181222123516-0b8337e80d98 h1:0OkFarm1Zy2CjCiDKfK9XHgmc2wbDlRMD2hD8anAJHU=
+github.com/golangci/gofmt v0.0.0-20181222123516-0b8337e80d98/go.mod h1:9qCChq59u/eW8im404Q2WWTrnBUQKjpNYKMbU4M7EFU=
+github.com/golangci/golangci-lint v1.17.1 h1:lc8Hf9GPCjIr0hg3S/xhvFT1+Hydass8F1xchr8jkME=
+github.com/golangci/golangci-lint v1.17.1/go.mod h1:+5sJSl2h3aly+fpmL2meSP8CaSKua2E4Twi9LPy7b1g=
+github.com/golangci/gosec v0.0.0-20180901114220-8afd9cbb6cfb h1:Bi7BYmZVg4C+mKGi8LeohcP2GGUl2XJD4xCkJoZSaYc=
+github.com/golangci/gosec v0.0.0-20180901114220-8afd9cbb6cfb/go.mod h1:ON/c2UR0VAAv6ZEAFKhjCLplESSmRFfZcDLASbI1GWo=
+github.com/golangci/gosec v0.0.0-20190211064107-66fb7fc33547/go.mod h1:0qUabqiIQgfmlAmulqxyiGkkyF6/tOGSnY2cnPVwrzU=
+github.com/golangci/ineffassign v0.0.0-20190609212857-42439a7714cc h1:gLLhTLMk2/SutryVJ6D4VZCU3CUqr8YloG7FPIBWFpI=
+github.com/golangci/ineffassign v0.0.0-20190609212857-42439a7714cc/go.mod h1:e5tpTHCfVze+7EpLEozzMB3eafxo2KT5veNg1k6byQU=
+github.com/golangci/lint-1 v0.0.0-20181222135242-d2cdd8c08219 h1:utua3L2IbQJmauC5IXdEA547bcoU5dozgQAfc8Onsg4=
+github.com/golangci/lint-1 v0.0.0-20181222135242-d2cdd8c08219/go.mod h1:/X8TswGSh1pIozq4ZwCfxS0WA5JGXguxk94ar/4c87Y=
+github.com/golangci/lint-1 v0.0.0-20190420132249-ee948d087217/go.mod h1:66R6K6P6VWk9I95jvqGxkqJxVWGFy9XlDwLwVz1RCFg=
+github.com/golangci/maligned v0.0.0-20180506175553-b1d89398deca h1:kNY3/svz5T29MYHubXix4aDDuE3RWHkPvopM/EDv/MA=
+github.com/golangci/maligned v0.0.0-20180506175553-b1d89398deca/go.mod h1:tvlJhZqDe4LMs4ZHD0oMUlt9G2LWuDGoisJTBzLMV9o=
+github.com/golangci/misspell v0.0.0-20180809174111-950f5d19e770/go.mod h1:dEbvlSfYbMQDtrpRMQU675gSDLDNa8sCPPChZ7PhiVA=
+github.com/golangci/misspell v0.3.4 h1:kAD5J0611Zo2VirUn9KFP15RjEqkmfEfnFxyIVxZCYg=
+github.com/golangci/misspell v0.3.4/go.mod h1:dEbvlSfYbMQDtrpRMQU675gSDLDNa8sCPPChZ7PhiVA=
+github.com/golangci/prealloc v0.0.0-20180630174525-215b22d4de21 h1:leSNB7iYzLYSSx3J/s5sVf4Drkc68W2wm4Ixh/mr0us=
+github.com/golangci/prealloc v0.0.0-20180630174525-215b22d4de21/go.mod h1:tf5+bzsHdTM0bsB7+8mt0GUMvjCgwLpTapNZHU8AajI=
+github.com/golangci/revgrep v0.0.0-20180526074752-d9c87f5ffaf0/go.mod h1:qOQCunEYvmd/TLamH+7LlVccLvUH5kZNhbCgTHoBbp4=
+github.com/golangci/revgrep v0.0.0-20180812185044-276a5c0a1039 h1:XQKc8IYQOeRwVs36tDrEmTgDgP88d5iEURwpmtiAlOM=
+github.com/golangci/revgrep v0.0.0-20180812185044-276a5c0a1039/go.mod h1:qOQCunEYvmd/TLamH+7LlVccLvUH5kZNhbCgTHoBbp4=
+github.com/golangci/unconvert v0.0.0-20180507085042-28b1c447d1f4 h1:zwtduBRr5SSWhqsYNgcuWO2kFlpdOZbP0+yRjmvPGys=
+github.com/golangci/unconvert v0.0.0-20180507085042-28b1c447d1f4/go.mod h1:Izgrg8RkN3rCIMLGE9CyYmU9pY2Jer6DgANEnZ/L/cQ=
+github.com/google/go-cmp v0.2.0 h1:+dTQ8DZQJz0Mb/HjFlkptS1FeQ4cWSnN941F8aEG4SQ=
+github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
+github.com/google/uuid v1.1.1 h1:Gkbcsh/GbpXz7lPftLA3P6TYMwjCLYm83jiFQZF/3gY=
+github.com/google/uuid v1.1.1/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/gostaticanalysis/analysisutil v0.0.0-20190318220348-4088753ea4d3 h1:JVnpOZS+qxli+rgVl98ILOXVNbW+kb5wcxeGx8ShUIw=
+github.com/gostaticanalysis/analysisutil v0.0.0-20190318220348-4088753ea4d3/go.mod h1:eEOZF4jCKGi+aprrirO9e7WKB3beBRtWgqGunKl6pKE=
+github.com/hashicorp/hcl v0.0.0-20180404174102-ef8a98b0bbce/go.mod h1:oZtUIOe8dh44I2q6ScRibXws4Ajl+d+nod3AaR9vL5w=
+github.com/hashicorp/hcl v1.0.0 h1:0Anlzjpi4vEasTeNFn2mLJgTSwt0+6sfsiTG8qcWGx4=
+github.com/hashicorp/hcl v1.0.0/go.mod h1:E5yfLk+7swimpb2L/Alb/PJmXilQ/rhwaUYs4T20WEQ=
+github.com/hpcloud/tail v1.0.0 h1:nfCOvKYfkgYP8hkirhJocXT2+zOD8yUNjXaWfTlyFKI=
+github.com/hpcloud/tail v1.0.0/go.mod h1:ab1qPbhIpdTxEkNHXyeSf5vhxWSCs/tWer42PpOxQnU=
+github.com/inconshreveable/mousetrap v1.0.0 h1:Z8tu5sraLXCXIcARxBp/8cbvlwVa7Z1NHg9XEKhtSvM=
+github.com/inconshreveable/mousetrap v1.0.0/go.mod h1:PxqpIevigyE2G7u3NXJIT2ANytuPF1OarO4DADm73n8=
+github.com/jinzhu/copier v0.0.0-20170922082739-db4671f3a9b8 h1:+dKzeuiDYbD/Cfi/sEkkNrs6z4/WZ8ZIR8dAbUpjXbU=
+github.com/jinzhu/copier v0.0.0-20170922082739-db4671f3a9b8/go.mod h1:yL958EeXv8Ylng6IfnvG4oflryUi3vgA3xPs9hmII1s=
+github.com/k0kubun/colorstring v0.0.0-20150214042306-9440f1994b88 h1:uC1QfSlInpQF+M0ao65imhwqKnz3Q2z/d8PWZRMQvDM=
+github.com/k0kubun/colorstring v0.0.0-20150214042306-9440f1994b88/go.mod h1:3w7q1U84EfirKl04SVQ/s7nPm1ZPhiXd34z40TNz36k=
+github.com/kisielk/errcheck v1.1.0/go.mod h1:EZBBE59ingxPouuu3KfxchcWSUPOHkagtvWXihfKN4Q=
+github.com/kisielk/gotool v0.0.0-20161130080628-0de1eaf82fa3/go.mod h1:jxZFDH7ILpTPQTk+E2s+z4CUas9lVNjIuKR4c5/zKgM=
+github.com/kisielk/gotool v1.0.0 h1:AV2c/EiW3KqPNT9ZKl07ehoAGi4C5/01Cfbblndcapg=
+github.com/kisielk/gotool v1.0.0/go.mod h1:XhKaO+MFFWcvkIS/tQcRk01m1F5IRFswLeQ+oQHNcck=
+github.com/klauspost/compress v1.4.0/go.mod h1:RyIbtBH6LamlWaDj8nUwkbUhJ87Yi3uG0guNDohfE1A=
+github.com/klauspost/compress v1.4.1/go.mod h1:RyIbtBH6LamlWaDj8nUwkbUhJ87Yi3uG0guNDohfE1A=
+github.com/klauspost/cpuid v0.0.0-20180405133222-e7e905edc00e/go.mod h1:Pj4uuM528wm8OyEC2QMXAi2YiTZ96dNQPGgoMS4s3ek=
+github.com/klauspost/cpuid v1.2.0/go.mod h1:Pj4uuM528wm8OyEC2QMXAi2YiTZ96dNQPGgoMS4s3ek=
+github.com/konsorten/go-windows-terminal-sequences v1.0.1/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
+github.com/konsorten/go-windows-terminal-sequences v1.0.2 h1:DB17ag19krx9CFsz4o3enTrPXyIXCl+2iCXH/aMAp9s=
+github.com/konsorten/go-windows-terminal-sequences v1.0.2/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
+github.com/kr/logfmt v0.0.0-20140226030751-b84e30acd515 h1:T+h1c/A9Gawja4Y9mFVWj2vyii2bbUNDw3kt9VxK2EY=
+github.com/kr/logfmt v0.0.0-20140226030751-b84e30acd515/go.mod h1:+0opPa2QZZtGFBFZlji/RkVcI2GknAs/DXo4wKdlNEc=
+github.com/kr/pretty v0.1.0 h1:L/CwN0zerZDmRFUapSPitk6f+Q3+0za1rQkzVuMiMFI=
+github.com/kr/pretty v0.1.0/go.mod h1:dAy3ld7l9f0ibDNOQOHHMYYIIbhfbHSm3C4ZsoJORNo=
+github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
+github.com/kr/text v0.1.0 h1:45sCR5RtlFHMR4UwH9sdQ5TC8v0qDQCHnXt+kaKSTVE=
+github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
+github.com/lib/pq v1.2.0 h1:LXpIM/LZ5xGFhOpXAQUIMM1HdyqzVYM13zNdjCEEcA0=
+github.com/lib/pq v1.2.0/go.mod h1:5WUZQaWbwv1U+lTReE5YruASi9Al49XbQIvNi/34Woo=
+github.com/logrusorgru/aurora v0.0.0-20181002194514-a7b3b318ed4e/go.mod h1:7rIyQOR62GCctdiQpZ/zOJlFyk6y+94wXzv6RNZgaR4=
+github.com/lucasb-eyer/go-colorful v1.0.2 h1:mCMFu6PgSozg9tDNMMK3g18oJBX7oYGrC09mS6CXfO4=
+github.com/lucasb-eyer/go-colorful v1.0.2/go.mod h1:0MS4r+7BZKSJ5mw4/S5MPN+qHFF1fYclkSPilDOKW0s=
+github.com/magiconair/properties v1.7.6/go.mod h1:PppfXfuXeibc/6YijjN8zIbojt8czPbwD3XqdrwzmxQ=
+github.com/magiconair/properties v1.8.0 h1:LLgXmsheXeRoUOBOjtwPQCWIYqM/LU1ayDtDePerRcY=
+github.com/magiconair/properties v1.8.0/go.mod h1:PppfXfuXeibc/6YijjN8zIbojt8czPbwD3XqdrwzmxQ=
+github.com/mattn/go-colorable v0.0.9 h1:UVL0vNpWh04HeJXV0KLcaT7r06gOH2l4OW6ddYRUIY4=
+github.com/mattn/go-colorable v0.0.9/go.mod h1:9vuHe8Xs5qXnSaW/c/ABM9alt+Vo+STaOChaDxuIBZU=
+github.com/mattn/go-isatty v0.0.3 h1:ns/ykhmWi7G9O+8a448SecJU3nSMBXJfqQkl0upE1jI=
+github.com/mattn/go-isatty v0.0.3/go.mod h1:M+lRXTBqGeGNdLjl/ufCoiOlB5xdOkqRJdNxMWT7Zi4=
+github.com/mattn/go-runewidth v0.0.4 h1:2BvfKmzob6Bmd4YsL0zygOqfdFnK7GR4QL06Do4/p7Y=
+github.com/mattn/go-runewidth v0.0.4/go.mod h1:LwmH8dsx7+W8Uxz3IHJYH5QSwggIsqBzpuz5H//U1FU=
+github.com/mattn/goveralls v0.0.2/go.mod h1:8d1ZMHsd7fW6IRPKQh46F2WRpyib5/X4FOpevwGNQEw=
+github.com/matttproud/golang_protobuf_extensions v1.0.0 h1:YNOwxxSJzSUARoD9KRZLzM9Y858MNGCOACTvCW9TSAc=
+github.com/matttproud/golang_protobuf_extensions v1.0.0/go.mod h1:D8He9yQNgCq6Z5Ld7szi9bcBfOoFv/3dc6xSMkL2PC0=
+github.com/mitchellh/go-homedir v1.0.0/go.mod h1:SfyaCUpYCn1Vlf4IUYiD9fPX4A5wJrkLzIz1N1q0pr0=
+github.com/mitchellh/go-homedir v1.1.0 h1:lukF9ziXFxDFPkA1vsr5zpc1XuPDn/wFntq5mG+4E0Y=
+github.com/mitchellh/go-homedir v1.1.0/go.mod h1:SfyaCUpYCn1Vlf4IUYiD9fPX4A5wJrkLzIz1N1q0pr0=
+github.com/mitchellh/go-ps v0.0.0-20170309133038-4fdf99ab2936/go.mod h1:r1VsdOzOPt1ZSrGZWFoNhsAedKnEd6r9Np1+5blZCWk=
+github.com/mitchellh/mapstructure v0.0.0-20180220230111-00c29f56e238/go.mod h1:FVVH3fgwuzCH5S8UJGiWEs2h04kUh9fWfEaFds41c1Y=
+github.com/mitchellh/mapstructure v1.1.2 h1:fmNYVwqnSfB9mZU6OS2O6GsXM+wcskZDuKQzvN1EDeE=
+github.com/mitchellh/mapstructure v1.1.2/go.mod h1:FVVH3fgwuzCH5S8UJGiWEs2h04kUh9fWfEaFds41c1Y=
+github.com/montanaflynn/stats v0.5.0 h1:2EkzeTSqBB4V4bJwWrt5gIIrZmpJBcoIRGS2kWLgzmk=
+github.com/montanaflynn/stats v0.5.0/go.mod h1:wL8QJuTMNUDYhXwkmfOly8iTdp5TEcJFWZD2D7SIkUc=
+github.com/mozilla/tls-observatory v0.0.0-20180409132520-8791a200eb40/go.mod h1:SrKMQvPiws7F7iqYp8/TX+IhxCYhzr6N/1yb8cwHsGk=
+github.com/nbutton23/zxcvbn-go v0.0.0-20160627004424-a22cb81b2ecd/go.mod h1:o96djdrsSGy3AWPyBgZMAGfxZNfgntdJG+11KU4QvbU=
+github.com/nbutton23/zxcvbn-go v0.0.0-20171102151520-eafdab6b0663 h1:Ri1EhipkbhWsffPJ3IPlrb4SkTOPa2PfRXp3jchBczw=
+github.com/nbutton23/zxcvbn-go v0.0.0-20171102151520-eafdab6b0663/go.mod h1:o96djdrsSGy3AWPyBgZMAGfxZNfgntdJG+11KU4QvbU=
+github.com/onsi/ginkgo v0.0.0-20170829012221-11459a886d9c/go.mod h1:lLunBs/Ym6LB5Z9jYTR76FiuTmxDTDusOGeTQH+WWjE=
+github.com/onsi/ginkgo v1.6.0 h1:Ix8l273rp3QzYgXSR+c8d1fTG7UPgYkOSELPhiY/YGw=
+github.com/onsi/ginkgo v1.6.0/go.mod h1:lLunBs/Ym6LB5Z9jYTR76FiuTmxDTDusOGeTQH+WWjE=
+github.com/onsi/gomega v0.0.0-20170829124025-dcabb60a477c/go.mod h1:C1qb7wdrVGGVU+Z6iS04AVkA3Q65CEZX59MT0QO5uiA=
+github.com/onsi/gomega v1.4.2 h1:3mYCb7aPxS/RU7TI1y4rkEn1oKmPRjNJLNEXgw7MH2I=
+github.com/onsi/gomega v1.4.2/go.mod h1:ex+gbHU/CVuBBDIJjb2X0qEXbFg53c61hWP/1CpauHY=
+github.com/pascaldekloe/name v0.0.0-20180628100202-0fd16699aae1 h1:/I3lTljEEDNYLho3/FUB7iD/oc2cEFgVmbHzV+O0PtU=
+github.com/pascaldekloe/name v0.0.0-20180628100202-0fd16699aae1/go.mod h1:eD5JxqMiuNYyFNmyY9rkJ/slN8y59oEu4Ei7F8OoKWQ=
+github.com/pelletier/go-toml v1.1.0/go.mod h1:5z9KED0ma1S8pY6P1sdut58dfprrGBbd/94hg7ilaic=
+github.com/pelletier/go-toml v1.2.0 h1:T5zMGML61Wp+FlcbWjRDT7yAxhJNAiPPLOFECq181zc=
+github.com/pelletier/go-toml v1.2.0/go.mod h1:5z9KED0ma1S8pY6P1sdut58dfprrGBbd/94hg7ilaic=
+github.com/pkg/errors v0.8.0 h1:WdK/asTD0HN+q6hsWO3/vpuAkAr+tw6aNJNDFFf0+qw=
+github.com/pkg/errors v0.8.0/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
+github.com/pkg/profile v1.2.1 h1:F++O52m40owAmADcojzM+9gyjmMOY/T4oYJkgFDH8RE=
+github.com/pkg/profile v1.2.1/go.mod h1:hJw3o1OdXxsrSjjVksARp5W95eeEaEfptyVZyv6JUPA=
+github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
+github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/problame/go-netssh v0.0.0-20190110232351-09d6bc45d284 h1:GX8YFBfSL6GlKymQZOsQllPbxVamfEYpis7oaJLXYZA=
+github.com/problame/go-netssh v0.0.0-20190110232351-09d6bc45d284/go.mod h1:XMR9fjP5/wMyFK0Ifk9ZHDFem4Rukm+ChvoYr2nxrgw=
+github.com/prometheus/client_golang v0.0.0-20180410130117-e11c6ff8170b h1:ymnAeobXUW0sfmt5n8euMzW4vgdQvZNTagdYMcpK63g=
+github.com/prometheus/client_golang v0.0.0-20180410130117-e11c6ff8170b/go.mod h1:7SWBe2y4D6OKWSNQJUaRYU/AaXPKyh/dDVn+NZz0KFw=
+github.com/prometheus/client_model v0.0.0-20171117100541-99fa1f4be8e5 h1:cLL6NowurKLMfCeQy4tIeph12XNQWgANCNvdyrOYKV4=
+github.com/prometheus/client_model v0.0.0-20171117100541-99fa1f4be8e5/go.mod h1:MbSGuTsp3dbXC40dX6PRTWyKYBIrTGTE9sqQNg2J8bo=
+github.com/prometheus/common v0.0.0-20180413074202-d0f7cd64bda4 h1:rPmL0pCWefiqV7RjMYuBUrLbMJmNVjRKz82caELPTUQ=
+github.com/prometheus/common v0.0.0-20180413074202-d0f7cd64bda4/go.mod h1:daVV7qP5qjZbuso7PdcryaAu0sAZbrN9i7WWcTMWvro=
+github.com/prometheus/procfs v0.0.0-20180408092902-8b1c2da0d56d h1:RCcsxyRr6+/pLg6wr0cUjPovhEhSNOtPh0SOz6u3hGU=
+github.com/prometheus/procfs v0.0.0-20180408092902-8b1c2da0d56d/go.mod h1:c3At6R/oaqEKCNdg8wHV1ftS6bRYblBhIjjI8uT2IGk=
+github.com/quasilyte/go-consistent v0.0.0-20190521200055-c6f3937de18c/go.mod h1:5STLWrekHfjyYwxBRVRXNOSewLJ3PWfDJd1VyTS21fI=
+github.com/rogpeppe/go-internal v1.1.0/go.mod h1:M8bDsm7K2OlrFYOpmOWEs/qY81heoFRclV5y23lUDJ4=
+github.com/rogpeppe/go-internal v1.2.1/go.mod h1:M8bDsm7K2OlrFYOpmOWEs/qY81heoFRclV5y23lUDJ4=
+github.com/ryanuber/go-glob v0.0.0-20170128012129-256dc444b735/go.mod h1:807d1WSdnB0XRJzKNil9Om6lcp/3a0v4qIHxIXzX/Yc=
+github.com/sergi/go-diff v1.0.0 h1:Kpca3qRNrduNnOQeazBd0ysaKrUJiIuISHxogkT9RPQ=
+github.com/sergi/go-diff v1.0.0/go.mod h1:0CfEIISq7TuYL3j771MWULgwwjU+GofnZX9QAmXWZgo=
+github.com/shirou/gopsutil v0.0.0-20180427012116-c95755e4bcd7/go.mod h1:5b4v6he4MtMOwMlS0TUMTu2PcXUg8+E1lC7eC3UO/RA=
+github.com/shirou/w32 v0.0.0-20160930032740-bb4de0191aa4/go.mod h1:qsXQc7+bwAM3Q1u/4XEfrquwF8Lw7D7y5cD8CuHnfIc=
+github.com/shurcooL/go v0.0.0-20180423040247-9e1955d9fb6e h1:MZM7FHLqUHYI0Y/mQAt3d2aYa0SiNms/hFqC9qJYolM=
+github.com/shurcooL/go v0.0.0-20180423040247-9e1955d9fb6e/go.mod h1:TDJrrUr11Vxrven61rcy3hJMUqaf/CLWYhHNPmT14Lk=
+github.com/shurcooL/go-goon v0.0.0-20170922171312-37c2f522c041 h1:llrF3Fs4018ePo4+G/HV/uQUqEI1HMDjCeOf2V6puPc=
+github.com/shurcooL/go-goon v0.0.0-20170922171312-37c2f522c041/go.mod h1:N5mDOmsrJOB+vfqUK+7DmDyjhSLIIBnXo9lvZJj3MWQ=
+github.com/sirupsen/logrus v1.0.5/go.mod h1:pMByvHTf9Beacp5x1UXfOR9xyW/9antXMhjMPG0dEzc=
+github.com/sirupsen/logrus v1.4.0 h1:yKenngtzGh+cUSSh6GWbxW2abRqhYUSR/t/6+2QqNvE=
+github.com/sirupsen/logrus v1.4.0/go.mod h1:LxeOpSwHxABJmUn/MG1IvRgCAasNZTLOkJPxbbu5VWo=
+github.com/sourcegraph/go-diff v0.5.1 h1:gO6i5zugwzo1RVTvgvfwCOSVegNuvnNi6bAD1QCmkHs=
+github.com/sourcegraph/go-diff v0.5.1/go.mod h1:j2dHj3m8aZgQO8lMTcTnBcXkRRRqi34cd2MNlA9u1mE=
+github.com/spf13/afero v1.1.0/go.mod h1:j4pytiNVoe2o6bmDsKpLACNPDBIoEAkihy7loJ1B0CQ=
+github.com/spf13/afero v1.1.2/go.mod h1:j4pytiNVoe2o6bmDsKpLACNPDBIoEAkihy7loJ1B0CQ=
+github.com/spf13/afero v1.2.2 h1:5jhuqJyZCZf2JRofRvN/nIFgIWNzPa3/Vz8mYylgbWc=
+github.com/spf13/afero v1.2.2/go.mod h1:9ZxEEn6pIJ8Rxe320qSDBk6AsU0r9pR7Q4OcevTdifk=
+github.com/spf13/cast v1.2.0/go.mod h1:r2rcYCSwa1IExKTDiTfzaxqT2FNHs8hODu4LnUfgKEg=
+github.com/spf13/cast v1.3.0 h1:oget//CVOEoFewqQxwr0Ej5yjygnqGkvggSE/gB35Q8=
+github.com/spf13/cast v1.3.0/go.mod h1:Qx5cxh0v+4UWYiBimWS+eyWzqEqokIECu5etghLkUJE=
+github.com/spf13/cobra v0.0.2 h1:NfkwRbgViGoyjBKsLI0QMDcuMnhM+SBg3T0cGfpvKDE=
+github.com/spf13/cobra v0.0.2/go.mod h1:1l0Ry5zgKvJasoi3XT1TypsSe7PqH0Sj9dhYf7v3XqQ=
+github.com/spf13/jwalterweatherman v0.0.0-20180109140146-7c0cea34c8ec/go.mod h1:cQK4TGJAtQXfYWX+Ddv3mKDzgVb68N+wFjFa4jdeBTo=
+github.com/spf13/jwalterweatherman v1.0.0/go.mod h1:cQK4TGJAtQXfYWX+Ddv3mKDzgVb68N+wFjFa4jdeBTo=
+github.com/spf13/jwalterweatherman v1.1.0 h1:ue6voC5bR5F8YxI5S67j9i582FU4Qvo2bmqnqMYADFk=
+github.com/spf13/jwalterweatherman v1.1.0/go.mod h1:aNWZUN0dPAAO/Ljvb5BEdw96iTZ0EXowPYD95IqWIGo=
+github.com/spf13/pflag v1.0.1/go.mod h1:DYY7MBk1bdzusC3SYhjObp+wFpr4gzcvqqNjLnInEg4=
+github.com/spf13/pflag v1.0.3 h1:zPAT6CGy6wXeQ7NtTnaTerfKOsV6V6F8agHXFiazDkg=
+github.com/spf13/pflag v1.0.3/go.mod h1:DYY7MBk1bdzusC3SYhjObp+wFpr4gzcvqqNjLnInEg4=
+github.com/spf13/viper v1.0.2/go.mod h1:A8kyI5cUJhb8N+3pkfONlcEcZbueH6nhAm0Fq7SrnBM=
+github.com/spf13/viper v1.3.2 h1:VUFqw5KcqRf7i70GOzW7N+Q7+gxVBkSSqiXB12+JQ4M=
+github.com/spf13/viper v1.3.2/go.mod h1:ZiWeW+zYFKm7srdB9IoDzzZXaJaI5eL9QjNiN/DMA2s=
+github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
+github.com/stretchr/objx v0.1.1/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
+github.com/stretchr/testify v1.2.2 h1:bSDNvY7ZPG5RlJ8otE/7V6gMiyenm9RtJ7IUVIAoJ1w=
+github.com/stretchr/testify v1.2.2/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
+github.com/stretchr/testify v1.4.0 h1:2E4SXV/wtOkTonXsotYi4li6zVWxYlZuYNCXe9XRJyk=
+github.com/stretchr/testify v1.4.0/go.mod h1:j7eGeouHqKxXV5pUuKE4zz7dFj8WfuZ+81PSLYec5m4=
+github.com/theckman/goconstraint v1.11.0 h1:oBUwN5wpE4dwyPhRGraEgJsFTr+JtLWiDnaJZJeeXI0=
+github.com/theckman/goconstraint v1.11.0/go.mod h1:zkCR/f2kOULTk/h1ujgyB9BlCNLaqlQ6GN2Zl4mg81g=
+github.com/timakin/bodyclose v0.0.0-20190407043127-4a873e97b2bb h1:lI9ufgFfvuqRctP9Ny8lDDLbSWCMxBPletcSqrnyFYM=
+github.com/timakin/bodyclose v0.0.0-20190407043127-4a873e97b2bb/go.mod h1:Qimiffbc6q9tBWlVV6x0P9sat/ao1xEkREYPPj9hphk=
+github.com/ugorji/go/codec v0.0.0-20181204163529-d75b2dcb6bc8/go.mod h1:VFNgLljTbGfSG7qAOspJ7OScBnGdDN/yBr0sguwnwf0=
+github.com/valyala/bytebufferpool v1.0.0/go.mod h1:6bBcMArwyJ5K/AmCkWv1jt77kVWyCJ6HpOuEn7z0Csc=
+github.com/valyala/fasthttp v1.2.0/go.mod h1:4vX61m6KN+xDduDNwXrhIAVZaZaZiQ1luJk8LWSxF3s=
+github.com/valyala/quicktemplate v1.1.1/go.mod h1:EH+4AkTd43SvgIbQHYu59/cJyxDoOVRUAfrukLPuGJ4=
+github.com/valyala/tcplisten v0.0.0-20161114210144-ceec8f93295a/go.mod h1:v3UYOV9WzVtRmSR+PDvWpU/qWl4Wa5LApYYX4ZtKbio=
+github.com/xordataexchange/crypt v0.0.3-0.20170626215501-b2862e3d0a77/go.mod h1:aYKd//L2LvnjZzWKhF00oedf4jCCReLcmhLdhm1A27Q=
+github.com/yudai/gojsondiff v0.0.0-20170107030110-7b1b7adf999d h1:yJIizrfO599ot2kQ6Af1enICnwBD3XoxgX3MrMwot2M=
+github.com/yudai/gojsondiff v0.0.0-20170107030110-7b1b7adf999d/go.mod h1:AY32+k2cwILAkW1fbgxQ5mUmMiZFgLIV+FBNExI05xg=
+github.com/yudai/golcs v0.0.0-20170316035057-ecda9a501e82 h1:BHyfKlQyqbsFN5p3IfnEUduWvb9is428/nNb5L3U01M=
+github.com/yudai/golcs v0.0.0-20170316035057-ecda9a501e82/go.mod h1:lgjkn3NuSvDfVJdfcVVdX+jpBxNmX4rDAzaS45IcYoM=
+github.com/yudai/pp v2.0.1+incompatible h1:Q4//iY4pNF6yPLZIigmvcl7k/bPgrcTPIFIcmawg5bI=
+github.com/yudai/pp v2.0.1+incompatible/go.mod h1:PuxR/8QJ7cyCkFp/aUDS+JY727OFEZkTdatxwunjIkc=
+github.com/zrepl/yaml-config v0.0.0-20190903123849-809f4603401a h1:l37btErUqI1yNmt5mf8kzglbeQfo32+TRacnxthiuck=
+github.com/zrepl/yaml-config v0.0.0-20190903123849-809f4603401a/go.mod h1:JmNwisZzOvW4GfpfLvhZ+gtyKLsIiA+WC+wNKJGJaFg=
+github.com/zrepl/yaml-config v0.0.0-20190928121844-af7ca3f8448f h1:3MuiGfgMHCSwKUcsuI7ODbi50j+evTB7SsoOBMNC5Fk=
+github.com/zrepl/yaml-config v0.0.0-20190928121844-af7ca3f8448f/go.mod h1:JmNwisZzOvW4GfpfLvhZ+gtyKLsIiA+WC+wNKJGJaFg=
+golang.org/x/crypto v0.0.0-20180904163835-0709b304e793/go.mod h1:6SG95UA2DQfeDnfUPMdvaQW0Q7yPrPDi9nlGo2tz2b4=
+golang.org/x/crypto v0.0.0-20181203042331-505ab145d0a9/go.mod h1:6SG95UA2DQfeDnfUPMdvaQW0Q7yPrPDi9nlGo2tz2b4=
+golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
+golang.org/x/crypto v0.0.0-20190313024323-a1f597ede03a/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
+golang.org/x/crypto v0.0.0-20190325154230-a5d413f7728c h1:Vj5n4GlwjmQteupaxJ9+0FNOmBrHfq7vN4btdGoDZgI=
+golang.org/x/crypto v0.0.0-20190325154230-a5d413f7728c/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
+golang.org/x/lint v0.0.0-20180702182130-06c8688daad7/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
+golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
+golang.org/x/net v0.0.0-20170915142106-8351a756f30f/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20180826012351-8a410e7b638d/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20180906233101-161cd47e91fd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20180911220305-26e67e76b6c3/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20181106065722-10aee1819953/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190311183353-d8887717615a/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190313220215-9f648a60d977 h1:actzWV6iWn3GLqN8dZjzsB+CLt+gaV2+wsxroxiQI8I=
+golang.org/x/net v0.0.0-20190313220215-9f648a60d977/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
+golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20190423024810-112230192c58 h1:8gQV6CLnAEikrhgkHFbMAEhagSSnXWGV915qUMm9mrU=
+golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sys v0.0.0-20171026204733-164713f0dfce/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20180905080454-ebe1bf3edb33/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20180909124046-d0be0721c37e/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20181205085412-a5c9d58dba9a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190312061237-fead79001313/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190626150813-e07cf5db2756 h1:9nuHUbU8dRnRRfj9KjWUVrJeoexdbeMjttk6Oh1rD10=
+golang.org/x/sys v0.0.0-20190626150813-e07cf5db2756/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/text v0.0.0-20170915090833-1cbadb444a80/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/text v0.3.0 h1:g61tztE5qeGQ89tm6NTjjM9VPIm088od1l6aSorWRWg=
+golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/tools v0.0.0-20170915040203-e531a2a1c15f/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20180221164845-07fd8470d635/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20180828015842-6cd1fcedba52/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20181117154741-2ddaf7f79a09/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20181205014116-22934f0fdb62/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190110163146-51295c7ec13a/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190121143147-24cd39ecf745/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190213192042-740235f6c0d8 h1:b0PLhFjEMqrIqsD4rS5sp0YxBYgdqKzX0KkkGGKLV8I=
+golang.org/x/tools v0.0.0-20190213192042-740235f6c0d8/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190311212946-11955173bddd/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190311215038-5c2858a9cfe5/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190322203728-c1a832b0ad89/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190521203540-521d6ed310dd/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+golang.org/x/tools v0.0.0-20190524210228-3d17549cdc6b h1:iEAPfYPbYbxG/2lNN4cMOHkmgKNsCuUwkxlDCK46UlU=
+golang.org/x/tools v0.0.0-20190524210228-3d17549cdc6b/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+google.golang.org/appengine v1.1.0/go.mod h1:EbEs0AVv82hx2wNQdGPgUI5lhzA/G0D9YwlJXL52JkM=
+google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8/go.mod h1:JiN7NxoALGmiZfu7CAH4rXhgtRTLTxftemlI0sWmxmc=
+google.golang.org/genproto v0.0.0-20181202183823-bd91e49a0898 h1:yvw+zsSmSM02Z5H3ZdEV7B7Ql7eFrjQTnmByJvK+3J8=
+google.golang.org/genproto v0.0.0-20181202183823-bd91e49a0898/go.mod h1:7Ep/1NZk928CDR8SjdVbjWNpdIf6nzjE3BTgJDr2Atg=
+google.golang.org/grpc v1.16.0/go.mod h1:0JHn/cJsOMiMfNA9+DeHDlAU7KAAB5GDlYFpa9MZMio=
+google.golang.org/grpc v1.17.0 h1:TRJYBgMclJvGYn2rIMjj+h9KtMt5r1Ij7ODVRIZkwhk=
+google.golang.org/grpc v1.17.0/go.mod h1:6QZJwpn2B+Zp71q/5VxRsJ6NXXVCE5NRUHRo+f3cWCs=
+gopkg.in/airbrake/gobrake.v2 v2.0.9/go.mod h1:/h5ZAUhDkGaJfjzjKLSjv6zCL6O0LLBxU4K+aSYdM/U=
+gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127 h1:qIbj1fsPNlZgppZ+VLlY7N33q108Sa+fhmuc+sWQYwY=
+gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/errgo.v2 v2.1.0/go.mod h1:hNsd1EY+bozCKY1Ytp96fpM3vjJbqLJn88ws8XvfDNI=
+gopkg.in/fsnotify.v1 v1.4.7 h1:xOHLXZwVvI9hhs+cLKq5+I5onOuwQLhQwiu63xxlHs4=
+gopkg.in/fsnotify.v1 v1.4.7/go.mod h1:Tz8NjZHkW78fSQdbUxIjBTcgA1z1m8ZHf0WmKUhAMys=
+gopkg.in/gemnasium/logrus-airbrake-hook.v2 v2.1.2/go.mod h1:Xk6kEKp8OKb+X14hQBKWaSkCsqBpgog8nAV2xsGOxlo=
+gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 h1:uRGJdciOHaEIrze2W8Q3AKkepLTh2hOroT7a+7czfdQ=
+gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7/go.mod h1:dt/ZhP58zS4L8KSrWDmTeBkI65Dw0HsyUHuEVlX15mw=
+gopkg.in/yaml.v2 v2.0.0-20170812160011-eb3733d160e7/go.mod h1:JAlM8MvJe8wmxCU4Bli9HhUf9+ttbYbLASfIpnQbh74=
+gopkg.in/yaml.v2 v2.2.1/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v2 v2.2.2 h1:ZCJp+EgiOT7lHqUV2J862kp8Qj64Jo6az82+3Td9dZw=
+gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+honnef.co/go/tools v0.0.0-20180728063816-88497007e858/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+mvdan.cc/interfacer v0.0.0-20180901003855-c20040233aed h1:WX1yoOaKQfddO/mLzdV4wptyWgoH/6hwLs7QHTixo0I=
+mvdan.cc/interfacer v0.0.0-20180901003855-c20040233aed/go.mod h1:Xkxe497xwlCKkIaQYRfC7CSLworTXY9RMqwhhCm+8Nc=
+mvdan.cc/lint v0.0.0-20170908181259-adc824a0674b h1:DxJ5nJdkhDlLok9K6qO+5290kphDJbHOQO1DFFFTeBo=
+mvdan.cc/lint v0.0.0-20170908181259-adc824a0674b/go.mod h1:2odslEg/xrtNQqCYg2/jCoyKnw3vv5biOc3JnIcYfL4=
+mvdan.cc/unparam v0.0.0-20190209190245-fbb59629db34/go.mod h1:H6SUd1XjIs+qQCyskXg5OFSrilMRUkD8ePJpHKDPaeY=
+mvdan.cc/unparam v0.0.0-20190310220240-1b9ccfa71afe h1:Ekmnp+NcP2joadI9pbK4Bva87QKZSeY7le//oiMrc9g=
+mvdan.cc/unparam v0.0.0-20190310220240-1b9ccfa71afe/go.mod h1:BnhuWBAqxH3+J5bDybdxgw5ZfS+DsVd4iylsKQePN8o=
+sourcegraph.com/sqs/pbtypes v0.0.0-20180604144634-d3ebe8f20ae4/go.mod h1:ketZ/q3QxT9HOBeFhu6RdvsftgpsbFHBF5Cas6cDKZ0=
+sourcegraph.com/sqs/pbtypes v1.0.0 h1:f7lAwqviDEGvON4kRv0o5V7FT/IQK+tbkF664XMbP3o=
+sourcegraph.com/sqs/pbtypes v1.0.0/go.mod h1:3AciMUv4qUuRHRHhOG4TZOB+72GdPVz5k+c648qsFS4=

lazy.sh

@@ -1,6 +1,7 @@
 #!/bin/bash
 set -e
 
+
 if tty -s; then
     bold=$(tput bold)
     normal=$(tput sgr0)
@@ -26,21 +27,15 @@ fi
 CHECKOUTPATH="${GOPATH}/src/github.com/zrepl/zrepl"
 
 godep() {
-    step "Install go dep using 'go get' to \$GOPATH/bin"
-    go get -u github.com/golang/dep/cmd/dep
-    if ! type dep ; then
-        echo "Unable to install go dep" 1>&2
-        exit 1
-    fi
-    step "Fetching dependencies using 'dep ensure'"
-    dep ensure -v -vendor-only
-    step "go install build dependencies fetched using dep"
-    # these will be in the vendor directory
-    go build -o "$GOPATH/bin/stringer"      ./vendor/golang.org/x/tools/cmd/stringer
-    go build -o "$GOPATH/bin/protoc-gen-go" ./vendor/github.com/golang/protobuf/protoc-gen-go
-    go build -o "$GOPATH/bin/enumer"        ./vendor/github.com/alvaroloes/enumer
-    go build -o "$GOPATH/bin/goimports"     ./vendor/golang.org/x/tools/cmd/goimports
-    go build -o "$GOPATH/bin/golangci-lint" ./vendor/github.com/golangci/golangci-lint/cmd/golangci-lint
+    step "install build dependencies (versions pinned using go.mod and tools.go)"
+    set -x
+    export GO111MODULE=on # otherwise, a checkout of this repo in GOPATH will disable modules on Go 1.12 and earlier
+    go build -v -mod=readonly -o "$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
+    go build -v -mod=readonly -o "$GOPATH/bin/goimports"     golang.org/x/tools/cmd/goimports
+    go build -v -mod=readonly -o "$GOPATH/bin/golangci-lint" github.com/golangci/golangci-lint/cmd/golangci-lint
+    set +x
     if ! type stringer || ! type protoc-gen-go || ! type enumer || ! type goimports || ! type golangci-lint; then
         echo "Installed dependencies but can't find them in \$PATH, adjust it to contain \$GOPATH/bin" 1>&2
         exit 1

logger/logger.go

@@ -20,6 +20,7 @@ type Logger interface {
 	WithField(field string, val interface{}) Logger
 	WithFields(fields Fields) Logger
 	WithError(err error) Logger
+	Log(level Level, msg string)
 	Debug(msg string)
 	Info(msg string)
 	Warn(msg string)
@@ -158,6 +159,10 @@ func (l *loggerImpl) WithError(err error) Logger {
 	return l.WithField(FieldError, val)
 }
 
+func (l *loggerImpl) Log(level Level, msg string) {
+	l.log(level, msg)
+}
+
 func (l *loggerImpl) Debug(msg string) {
 	l.log(Debug, msg)
 }

logger/nulllogger.go

@@ -13,6 +13,7 @@ func (n nullLogger) ReplaceField(field string, val interface{}) Logger { return
 func (n nullLogger) WithField(field string, val interface{}) Logger    { return n }
 func (n nullLogger) WithFields(fields Fields) Logger                   { return n }
 func (n nullLogger) WithError(err error) Logger                        { return n }
+func (nullLogger) Log(level Level, msg string)                         {}
 func (nullLogger) Debug(msg string)                                    {}
 func (nullLogger) Info(msg string)                                     {}
 func (nullLogger) Warn(msg string)                                     {}

platformtest/harness/harness.go

@@ -0,0 +1,79 @@
+package main
+
+import (
+	"context"
+	"flag"
+	"fmt"
+	"path/filepath"
+	"time"
+
+	"github.com/fatih/color"
+	"github.com/pkg/errors"
+	"github.com/zrepl/zrepl/config"
+	"github.com/zrepl/zrepl/daemon/logging"
+	"github.com/zrepl/zrepl/logger"
+	"github.com/zrepl/zrepl/platformtest"
+	"github.com/zrepl/zrepl/platformtest/tests"
+)
+
+func main() {
+
+	var args struct {
+		createArgs platformtest.ZpoolCreateArgs
+	}
+	flag.StringVar(&args.createArgs.PoolName, "poolname", "", "")
+	flag.StringVar(&args.createArgs.ImagePath, "imagepath", "", "")
+	flag.Int64Var(&args.createArgs.ImageSize, "imagesize", 100*(1<<20), "")
+	flag.StringVar(&args.createArgs.Mountpoint, "mountpoint", "none", "")
+	flag.Parse()
+
+	outlets := logger.NewOutlets()
+	outlet, level, err := logging.ParseOutlet(config.LoggingOutletEnum{Ret: &config.StdoutLoggingOutlet{
+		LoggingOutletCommon: config.LoggingOutletCommon{
+			Level:  "debug",
+			Format: "human",
+		},
+	}})
+	if err != nil {
+		panic(err)
+	}
+	outlets.Add(outlet, level)
+	logger := logger.NewLogger(outlets, 1*time.Second)
+
+	if err := args.createArgs.Validate(); err != nil {
+		logger.Error(err.Error())
+		panic(err)
+	}
+	ctx := platformtest.WithLogger(context.Background(), logger)
+	ex := platformtest.NewEx(logger)
+
+	bold := color.New(color.Bold)
+	for _, c := range tests.Cases {
+		// ATTENTION future parallelism must pass c by value into closure!
+		err := func() error {
+			bold.Printf("BEGIN TEST CASE %s\n", c)
+			pool, err := platformtest.CreateOrReplaceZpool(ctx, ex, args.createArgs)
+			if err != nil {
+				return errors.Wrap(err, "create test pool")
+			}
+			defer func() {
+				if err := pool.Destroy(ctx, ex); err != nil {
+					fmt.Printf("error destroying test pool: %s", err)
+				}
+			}()
+			ctx := &platformtest.Context{
+				Context:     ctx,
+				RootDataset: filepath.Join(pool.Name(), "rootds"),
+			}
+			c(ctx)
+			bold.Printf("DONE  TEST CASE %s\n", c)
+			fmt.Println()
+			return nil
+		}()
+		if err != nil {
+			bold.Printf("TEST CASE FAILED WITH ERROR:\n")
+			fmt.Println(err)
+		}
+	}
+
+}

platformtest/harness/harness_test.go

@@ -0,0 +1,44 @@
+package main
+
+import (
+	"os"
+	"strings"
+	"testing"
+)
+
+// Idea taken from
+// https://github.com/openSUSE/umoci/blob/v0.2.1/cmd/umoci/main_test.go
+//
+/* How to generate coverage:
+   go test -c -covermode=atomic -cover -coverpkg github.com/zrepl/zrepl/...
+   sudo ../logmockzfs/logzfsenv /tmp/zrepl_platform_test.log /usr/bin/zfs \
+       ./harness.test -test.coverprofile=/tmp/harness.out \
+       -test.v __DEVEL--i-heard-you-like-tests \
+       -imagepath /tmp/testpool.img -poolname zreplplatformtest
+    go tool cover -html=/tmp/harness.out -o /tmp/harness.html
+*/
+// Merge with existing coverage reports using gocovmerge:
+//  https://github.com/wadey/gocovmerge
+
+func TestMain(t *testing.T) {
+	var (
+		args []string
+		run  bool
+	)
+
+	for _, arg := range os.Args {
+		switch {
+		case arg == "__DEVEL--i-heard-you-like-tests":
+			run = true
+		case strings.HasPrefix(arg, "-test"):
+		case strings.HasPrefix(arg, "__DEVEL"):
+		default:
+			args = append(args, arg)
+		}
+	}
+	os.Args = args
+
+	if run {
+		main()
+	}
+}

platformtest/logmockzfs/logzfsenv

@@ -0,0 +1,9 @@
+#!/usr/bin/env bash
+set -ue
+export ZREPL_MOCK_ZFS_COMMAND_LOG="$1"
+shift
+export ZREPL_MOCK_ZFS_PATH="$1"
+shift
+export PATH="$(dirname "${BASH_SOURCE[0]}" )":"$PATH"
+args=("$@")
+exec "${args[@]}"

platformtest/logmockzfs/zfs

@@ -0,0 +1,12 @@
+#!/usr/bin/env bash
+set -eu
+args=("$@")
+
+if [ -n "$ZREPL_MOCK_ZFS_COMMAND_LOG" ]; then
+    (
+        flock -x 200
+        jq --compact-output -n --argjson args "$(printf '%s\0' "${args[@]}" | jq -Rsc 'split("\u0000")')" '{date: now|strflocaltime("%Y-%m-%dT%H:%M:%S"), command: $args}' >> "$ZREPL_MOCK_ZFS_COMMAND_LOG"
+    ) 200>"$ZREPL_MOCK_ZFS_COMMAND_LOG".lock
+fi
+
+exec "$ZREPL_MOCK_ZFS_PATH" "${args[@]}"

platformtest/platformtest.go

@@ -0,0 +1,24 @@
+package platformtest
+
+import (
+	"context"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+type Context struct {
+	context.Context
+	RootDataset string
+}
+
+var _ assert.TestingT = (*Context)(nil)
+var _ require.TestingT = (*Context)(nil)
+
+func (c *Context) Errorf(format string, args ...interface{}) {
+	getLog(c).Printf(format, args...)
+}
+
+func (c *Context) FailNow() {
+	panic(nil)
+}

platformtest/platformtest_exec.go

@@ -0,0 +1,45 @@
+package platformtest
+
+import (
+	"context"
+	"fmt"
+	"os/exec"
+
+	"github.com/zrepl/zrepl/util/circlog"
+)
+
+type ex struct {
+	log Logger
+}
+
+func NewEx(log Logger) Execer {
+	return &ex{log}
+}
+
+func (e *ex) RunExpectSuccessNoOutput(ctx context.Context, cmd string, args ...string) error {
+	return e.runNoOutput(true, ctx, cmd, args...)
+}
+
+func (e *ex) RunExpectFailureNoOutput(ctx context.Context, cmd string, args ...string) error {
+	return e.runNoOutput(false, ctx, cmd, args...)
+}
+
+func (e *ex) runNoOutput(expectSuccess bool, ctx context.Context, cmd string, args ...string) error {
+	log := e.log.WithField("command", fmt.Sprintf("%q %q", cmd, args))
+	log.Debug("begin executing")
+	defer log.Debug("done executing")
+	ecmd := exec.CommandContext(ctx, cmd, args...)
+	buf, _ := circlog.NewCircularLog(32 << 10)
+	ecmd.Stdout, ecmd.Stderr = buf, buf
+	err := ecmd.Run()
+	log.Printf("command output: %s", buf.String())
+	if _, ok := err.(*exec.ExitError); err != nil && !ok {
+		panic(err)
+	}
+	if expectSuccess && err != nil {
+		return fmt.Errorf("expecting no error, got error: %s\n%s", err, buf.String())
+	} else if !expectSuccess && err == nil {
+		return fmt.Errorf("expecting error, got no error")
+	}
+	return nil
+}

platformtest/platformtest_logging.go

@@ -0,0 +1,24 @@
+package platformtest
+
+import (
+	"context"
+
+	"github.com/zrepl/zrepl/logger"
+)
+
+type Logger = logger.Logger
+
+type contextKey int
+
+const (
+	contextKeyLogger contextKey = iota
+)
+
+func WithLogger(ctx context.Context, logger Logger) context.Context {
+	ctx = context.WithValue(ctx, contextKeyLogger, logger)
+	return ctx
+}
+
+func getLog(ctx context.Context) Logger {
+	return ctx.Value(contextKeyLogger).(Logger)
+}

platformtest/platformtest_ops.go

@@ -0,0 +1,260 @@
+package platformtest
+
+import (
+	"bufio"
+	"bytes"
+	"context"
+	"fmt"
+	"os"
+	"os/exec"
+	"strings"
+	"unicode"
+)
+
+type Execer interface {
+	RunExpectSuccessNoOutput(ctx context.Context, cmd string, args ...string) error
+	RunExpectFailureNoOutput(ctx context.Context, cmd string, args ...string) error
+}
+
+type Stmt interface {
+	Run(context context.Context, e Execer) error
+}
+
+type Op string
+
+const (
+	AssertExists    Op = "!E"
+	AssertNotExists Op = "!N"
+	Add             Op = "+"
+	Del             Op = "-"
+	RunCmd          Op = "R"
+	DestroyRoot     Op = "DESTROYROOT"
+	CreateRoot      Op = "CREATEROOT"
+)
+
+type DestroyRootOp struct {
+	Path string
+}
+
+func (o *DestroyRootOp) Run(ctx context.Context, e Execer) error {
+	// early-exit if it doesn't exist
+	if err := e.RunExpectSuccessNoOutput(ctx, "zfs", "get", "-H", "name", o.Path); err != nil {
+		getLog(ctx).WithField("root_ds", o.Path).Info("assume root ds doesn't exist")
+		return nil
+	}
+	return e.RunExpectSuccessNoOutput(ctx, "zfs", "destroy", "-r", o.Path)
+}
+
+type FSOp struct {
+	Op   Op
+	Path string
+}
+
+func (o *FSOp) Run(ctx context.Context, e Execer) error {
+	switch o.Op {
+	case AssertExists:
+		return e.RunExpectSuccessNoOutput(ctx, "zfs", "get", "-H", "name", o.Path)
+	case AssertNotExists:
+		return e.RunExpectFailureNoOutput(ctx, "zfs", "get", "-H", "name", o.Path)
+	case Add:
+		return e.RunExpectSuccessNoOutput(ctx, "zfs", "create", o.Path)
+	case Del:
+		return e.RunExpectSuccessNoOutput(ctx, "zfs", "destroy", o.Path)
+	default:
+		panic(o.Op)
+	}
+}
+
+type SnapOp struct {
+	Op   Op
+	Path string
+}
+
+func (o *SnapOp) Run(ctx context.Context, e Execer) error {
+	switch o.Op {
+	case AssertExists:
+		return e.RunExpectSuccessNoOutput(ctx, "zfs", "get", "-H", "name", o.Path)
+	case AssertNotExists:
+		return e.RunExpectFailureNoOutput(ctx, "zfs", "get", "-H", "name", o.Path)
+	case Add:
+		return e.RunExpectSuccessNoOutput(ctx, "zfs", "snapshot", o.Path)
+	case Del:
+		return e.RunExpectSuccessNoOutput(ctx, "zfs", "destroy", o.Path)
+	default:
+		panic(o.Op)
+	}
+}
+
+type RunOp struct {
+	RootDS string
+	Script string
+}
+
+func (o *RunOp) Run(ctx context.Context, e Execer) error {
+	cmd := exec.CommandContext(ctx, "/usr/bin/env", "bash", "-c", o.Script)
+	cmd.Env = os.Environ()
+	cmd.Env = append(cmd.Env, fmt.Sprintf("ROOTDS=%s", o.RootDS))
+	log := getLog(ctx).WithField("script", o.Script)
+	log.Info("start script")
+	defer log.Info("script done")
+	output, err := cmd.CombinedOutput()
+	if _, ok := err.(*exec.ExitError); err != nil && !ok {
+		panic(err)
+	}
+	log.Printf("script output:\n%s", output)
+	return err
+}
+
+type LineError struct {
+	Line string
+	What string
+}
+
+func (e LineError) Error() string {
+	return fmt.Sprintf("%q: %s", e.Line, e.What)
+}
+
+type RunKind int
+
+const (
+	PanicErr RunKind = 1 << iota
+	RunAll
+)
+
+func Run(ctx context.Context, rk RunKind, rootds string, stmtsStr string) {
+	stmt, err := parseSequence(rootds, stmtsStr)
+	if err != nil {
+		panic(err)
+	}
+	execer := NewEx(getLog(ctx))
+	for _, s := range stmt {
+		err := s.Run(ctx, execer)
+		if err == nil {
+			continue
+		}
+		if rk == PanicErr {
+			panic(err)
+		} else if rk == RunAll {
+			continue
+		} else {
+			panic(rk)
+		}
+	}
+}
+
+func isNoSpace(r rune) bool {
+	return !unicode.IsSpace(r)
+}
+
+func splitQuotedWords(data []byte, atEOF bool) (advance int, token []byte, err error) {
+	begin := bytes.IndexFunc(data, isNoSpace)
+	if begin == -1 {
+		return len(data), nil, nil
+	}
+	if data[begin] == '"' {
+		end := begin + 1
+		for end < len(data) {
+			endCandidate := bytes.Index(data[end:], []byte(`"`))
+			if endCandidate == -1 {
+				return 0, nil, nil
+			}
+			end += endCandidate
+			if data[end-1] != '\\' {
+				// unescaped quote, end of this string
+				// remove backslash-escapes
+				withBackslash := data[begin+1 : end]
+				withoutBaskslash := bytes.Replace(withBackslash, []byte("\\\""), []byte("\""), -1)
+				return end + 1, withoutBaskslash, nil
+			} else {
+				// continue to next quote
+				end += 1
+			}
+		}
+	} else {
+		endOffset := bytes.IndexFunc(data[begin:], unicode.IsSpace)
+		var end int
+		if endOffset == -1 {
+			if !atEOF {
+				return 0, nil, nil
+			} else {
+				end = len(data)
+			}
+		} else {
+			end = begin + endOffset
+		}
+		return end, data[begin:end], nil
+	}
+	return 0, nil, fmt.Errorf("unexpected")
+}
+
+func parseSequence(rootds, stmtsStr string) (stmts []Stmt, err error) {
+	scan := bufio.NewScanner(strings.NewReader(stmtsStr))
+nextLine:
+	for scan.Scan() {
+		if len(bytes.TrimSpace(scan.Bytes())) == 0 {
+			continue
+		}
+		comps := bufio.NewScanner(bytes.NewReader(scan.Bytes()))
+		comps.Split(splitQuotedWords)
+
+		expectMoreTokens := func() error {
+			if !comps.Scan() {
+				return &LineError{scan.Text(), "unexpected EOL"}
+			}
+			return nil
+		}
+
+		// Op
+		if err := expectMoreTokens(); err != nil {
+			return nil, err
+		}
+		var op Op
+		switch comps.Text() {
+
+		case string(RunCmd):
+			script := strings.TrimPrefix(strings.TrimSpace(scan.Text()), string(RunCmd))
+			stmts = append(stmts, &RunOp{RootDS: rootds, Script: script})
+			continue nextLine
+
+		case string(DestroyRoot):
+			if comps.Scan() {
+				return nil, &LineError{scan.Text(), fmt.Sprintf("unexpected tokens at EOL")}
+			}
+			stmts = append(stmts, &DestroyRootOp{rootds})
+			continue nextLine
+
+		case string(CreateRoot):
+			if comps.Scan() {
+				return nil, &LineError{scan.Text(), fmt.Sprintf("unexpected tokens at EOL")}
+			}
+			stmts = append(stmts, &FSOp{Op: Add, Path: rootds})
+			continue nextLine
+
+		case string(Add):
+			op = Add
+		case string(Del):
+			op = Del
+		case string(AssertExists):
+			op = AssertExists
+		case string(AssertNotExists):
+			op = AssertNotExists
+		default:
+			return nil, &LineError{scan.Text(), fmt.Sprintf("invalid op %q", comps.Text())}
+		}
+
+		// FS / SNAP
+		if err := expectMoreTokens(); err != nil {
+			return nil, err
+		}
+		if strings.ContainsAny(comps.Text(), "@") {
+			stmts = append(stmts, &SnapOp{Op: op, Path: fmt.Sprintf("%s/%s", rootds, comps.Text())})
+		} else {
+			stmts = append(stmts, &FSOp{Op: op, Path: fmt.Sprintf("%s/%s", rootds, comps.Text())})
+		}
+
+		if comps.Scan() {
+			return nil, &LineError{scan.Text(), fmt.Sprintf("unexpected tokens at EOL")}
+		}
+	}
+	return stmts, nil
+}

platformtest/platformtest_parser_test.go

@@ -0,0 +1,33 @@
+package platformtest
+
+import (
+	"bufio"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestSplitQuotedWords(t *testing.T) {
+	s := bufio.NewScanner(strings.NewReader(`
+	foo "bar baz" blah "foo \"with single escape" "blah baz" "\"foo" "foo\""
+	`))
+	s.Split(splitQuotedWords)
+	var words []string
+	for s.Scan() {
+		words = append(words, s.Text())
+	}
+	assert.Equal(
+		t,
+		[]string{
+			"foo",
+			"bar baz",
+			"blah",
+			"foo \"with single escape",
+			"blah baz",
+			"\"foo",
+			"foo\"",
+		},
+		words)
+
+}

platformtest/platformtest_zpool.go

@@ -0,0 +1,92 @@
+package platformtest
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+
+	"github.com/pkg/errors"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+type Zpool struct {
+	args ZpoolCreateArgs
+}
+
+type ZpoolCreateArgs struct {
+	PoolName   string
+	ImagePath  string
+	ImageSize  int64
+	Mountpoint string
+}
+
+func (a ZpoolCreateArgs) Validate() error {
+	if !filepath.IsAbs(a.ImagePath) {
+		return errors.Errorf("ImagePath must be absolute, got %q", a.ImagePath)
+	}
+	const minImageSize = 1024
+	if a.ImageSize < minImageSize {
+		return errors.Errorf("ImageSize must be > %v, got %v", minImageSize, a.ImageSize)
+	}
+	if a.Mountpoint != "none" {
+		return errors.Errorf("Mountpoint must be \"none\"")
+	}
+	if a.PoolName == "" {
+		return errors.Errorf("PoolName must not be emtpy")
+	}
+	return nil
+}
+
+func CreateOrReplaceZpool(ctx context.Context, e Execer, args ZpoolCreateArgs) (*Zpool, error) {
+	if err := args.Validate(); err != nil {
+		return nil, errors.Wrap(err, "zpool create args validation error")
+	}
+
+	// export pool if it already exists (idempotence)
+	if _, err := zfs.ZFSGetRawAnySource(args.PoolName, []string{"name"}); err != nil {
+		if _, ok := err.(*zfs.DatasetDoesNotExist); ok {
+			// we'll create it shortly
+		} else {
+			return nil, errors.Wrapf(err, "cannot determine whether test pool %q exists", args.PoolName)
+		}
+	} else {
+		// exists, export it, OpenFile will destroy it
+		if err := e.RunExpectSuccessNoOutput(ctx, "zpool", "export", args.PoolName); err != nil {
+			return nil, errors.Wrapf(err, "cannot destroy test pool %q", args.PoolName)
+		}
+	}
+
+	// idempotently (re)create the pool image
+	image, err := os.OpenFile(args.ImagePath, os.O_CREATE|os.O_RDWR, 0600)
+	if err != nil {
+		return nil, errors.Wrap(err, "create image file")
+	}
+	defer image.Close()
+	if err := image.Truncate(args.ImageSize); err != nil {
+		return nil, errors.Wrap(err, "create image: truncate")
+	}
+	image.Close()
+
+	// create the pool
+	err = e.RunExpectSuccessNoOutput(ctx, "zpool", "create", "-O", "mountpoint=none", args.PoolName, args.ImagePath)
+	if err != nil {
+		return nil, errors.Wrap(err, "zpool create")
+	}
+
+	return &Zpool{args}, nil
+}
+
+func (p *Zpool) Name() string { return p.args.PoolName }
+
+func (p *Zpool) Destroy(ctx context.Context, e Execer) error {
+
+	if err := e.RunExpectSuccessNoOutput(ctx, "zpool", "export", p.args.PoolName); err != nil {
+		return errors.Wrapf(err, "export pool %q", p.args.PoolName)
+	}
+
+	if err := os.Remove(p.args.ImagePath); err != nil {
+		return errors.Wrapf(err, "remove pool image")
+	}
+
+	return nil
+}

platformtest/tests/batchDestroy.go

@@ -0,0 +1,54 @@
+package tests
+
+import (
+	"fmt"
+	"strings"
+
+	"github.com/zrepl/zrepl/platformtest"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+func BatchDestroy(ctx *platformtest.Context) {
+
+	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
+		DESTROYROOT
+		CREATEROOT
+		+  "foo bar"
+		+  "foo bar@1"
+		+  "foo bar@2"
+		+  "foo bar@3"
+		R  zfs hold zrepl_platformtest "${ROOTDS}/foo bar@2"
+	`)
+
+	reqs := []*zfs.DestroySnapOp{
+		&zfs.DestroySnapOp{
+			ErrOut:     new(error),
+			Filesystem: fmt.Sprintf("%s/foo bar", ctx.RootDataset),
+			Name:       "3",
+		},
+		&zfs.DestroySnapOp{
+			ErrOut:     new(error),
+			Filesystem: fmt.Sprintf("%s/foo bar", ctx.RootDataset),
+			Name:       "2",
+		},
+	}
+	zfs.ZFSDestroyFilesystemVersions(reqs)
+	if *reqs[0].ErrOut != nil {
+		panic("expecting no error")
+	}
+	err := (*reqs[1].ErrOut).Error()
+	if !strings.Contains(err, fmt.Sprintf("%s/foo bar@2", ctx.RootDataset)) {
+		panic(fmt.Sprintf("expecting error about being unable to destroy @2: %T\n%s", err, err))
+	}
+
+	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
+	!N "foo bar@3"
+	!E "foo bar@1"
+	!E "foo bar@2"
+	R zfs release zrepl_platformtest "${ROOTDS}/foo bar@2"
+	-  "foo bar@2"
+	-  "foo bar@1"
+	-  "foo bar"
+	`)
+
+}

platformtest/tests/getNonexistent.go

@@ -0,0 +1,54 @@
+package tests
+
+import (
+	"fmt"
+
+	"github.com/zrepl/zrepl/platformtest"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+func GetNonexistent(ctx *platformtest.Context) {
+
+	platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
+		DESTROYROOT
+		CREATEROOT
+		+  "foo bar"
+		+  "foo bar@1"
+	`)
+	defer platformtest.Run(ctx, platformtest.PanicErr, ctx.RootDataset, `
+		DESTROYROOT
+	`)
+
+	// test raw
+	_, err := zfs.ZFSGetRawAnySource(fmt.Sprintf("%s/foo bar", ctx.RootDataset), []string{"name"})
+	if err != nil {
+		panic(err)
+	}
+
+	// test nonexistent filesystem
+	nonexistent := fmt.Sprintf("%s/nonexistent filesystem", ctx.RootDataset)
+	props, err := zfs.ZFSGetRawAnySource(nonexistent, []string{"name"})
+	if err == nil {
+		panic(props)
+	}
+	dsne, ok := err.(*zfs.DatasetDoesNotExist)
+	if !ok {
+		panic(err)
+	} else if dsne.Path != nonexistent {
+		panic(err)
+	}
+
+	// test nonexistent snapshot
+	nonexistent = fmt.Sprintf("%s/foo bar@non existent", ctx.RootDataset)
+	props, err = zfs.ZFSGetRawAnySource(nonexistent, []string{"name"})
+	if err == nil {
+		panic(props)
+	}
+	dsne, ok = err.(*zfs.DatasetDoesNotExist)
+	if !ok {
+		panic(err)
+	} else if dsne.Path != nonexistent {
+		panic(err)
+	}
+
+}

platformtest/tests/tests.go

@@ -0,0 +1,20 @@
+package tests
+
+import (
+	"reflect"
+	"runtime"
+
+	"github.com/zrepl/zrepl/platformtest"
+)
+
+type Case func(*platformtest.Context)
+
+func (c Case) String() string {
+	return runtime.FuncForPC(reflect.ValueOf(c).Pointer()).Name()
+}
+
+var Cases = []Case{
+	BatchDestroy,
+	UndestroyableSnapshotParsing,
+	GetNonexistent,
+}

platformtest/tests/undestroyableSnapshotParsing.go

@@ -0,0 +1,40 @@
+package tests
+
+import (
+	"fmt"
+
+	"github.com/stretchr/testify/require"
+	"github.com/zrepl/zrepl/platformtest"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+func UndestroyableSnapshotParsing(t *platformtest.Context) {
+	platformtest.Run(t, platformtest.PanicErr, t.RootDataset, `
+		DESTROYROOT
+		CREATEROOT
+		+  "foo bar"
+		+  "foo bar@1 2 3"
+		+  "foo bar@4 5 6"
+		+  "foo bar@7 8 9"
+		R  zfs hold zrepl_platformtest "${ROOTDS}/foo bar@4 5 6"
+	`)
+	defer platformtest.Run(t, platformtest.PanicErr, t.RootDataset, `
+	R zfs release zrepl_platformtest "${ROOTDS}/foo bar@4 5 6"
+	DESTROYROOT
+	`)
+
+	err := zfs.ZFSDestroy(fmt.Sprintf("%s/foo bar@1 2 3,4 5 6,7 8 9", t.RootDataset))
+	if err == nil {
+		panic("expecting destroy error due to hold")
+	}
+	if dse, ok := err.(*zfs.DestroySnapshotsError); !ok {
+		panic(fmt.Sprintf("expecting *zfs.DestroySnapshotsError, got %T\n%v\n%s", err, err, err))
+	} else {
+		if dse.Filesystem != fmt.Sprintf("%s/foo bar", t.RootDataset) {
+			panic(dse.Filesystem)
+		}
+		require.Equal(t, []string{"4 5 6"}, dse.Undestroyable)
+		require.Equal(t, []string{"dataset is busy"}, dse.Reason)
+	}
+
+}

replication/driver/errorclass_enumer.go

@@ -1,5 +1,6 @@
 // Code generated by "enumer -type=errorClass"; DO NOT EDIT.
 
+//
 package driver
 
 import (

replication/logic/replication_logic.go

@@ -14,6 +14,7 @@ import (
 	"github.com/zrepl/zrepl/replication/logic/pdu"
 	"github.com/zrepl/zrepl/replication/report"
 	"github.com/zrepl/zrepl/util/bytecounter"
+	"github.com/zrepl/zrepl/util/chainlock"
 	"github.com/zrepl/zrepl/util/envconst"
 	"github.com/zrepl/zrepl/util/semaphore"
 	"github.com/zrepl/zrepl/zfs"
@@ -147,8 +148,12 @@ type Step struct {
 	parent   *Filesystem
 	from, to *pdu.FilesystemVersion // compat
 
-	byteCounter  bytecounter.StreamCopier
 	expectedSize int64 // 0 means no size estimate present / possible
+
+	// byteCounter is nil initially, and set later in Step.doReplication
+	// => concurrent read of that pointer from Step.ReportInfo must be protected
+	byteCounter    bytecounter.StreamCopier
+	byteCounterMtx chainlock.L
 }
 
 func (s *Step) TargetEquals(other driver.Step) bool {
@@ -172,10 +177,15 @@ func (s *Step) Step(ctx context.Context) error {
 }
 
 func (s *Step) ReportInfo() *report.StepInfo {
+
+	// get current byteCounter value
 	var byteCounter int64
+	s.byteCounterMtx.Lock()
 	if s.byteCounter != nil {
 		byteCounter = s.byteCounter.Count()
 	}
+	s.byteCounterMtx.Unlock()
+
 	// FIXME stick to zfs convention of from and to
 	from := ""
 	if s.from != nil {
@@ -457,8 +467,12 @@ func (s *Step) doReplication(ctx context.Context) error {
 	defer sstreamCopier.Close()
 
 	// Install a byte counter to track progress + for status report
-	s.byteCounter = bytecounter.NewStreamCopier(sstreamCopier)
+	byteCountingStream := bytecounter.NewStreamCopier(sstreamCopier)
+	s.byteCounterMtx.Lock()
+	s.byteCounter = byteCountingStream
+	s.byteCounterMtx.Unlock()
 	defer func() {
+		defer s.byteCounterMtx.Lock().Unlock()
 		s.parent.promBytesReplicated.Add(float64(s.byteCounter.Count()))
 	}()
 
@@ -467,7 +481,7 @@ func (s *Step) doReplication(ctx context.Context) error {
 		ClearResumeToken: !sres.UsedResumeToken,
 	}
 	log.Debug("initiate receive request")
-	_, err = s.receiver.Receive(ctx, rr, s.byteCounter)
+	_, err = s.receiver.Receive(ctx, rr, byteCountingStream)
 	if err != nil {
 		log.
 			WithError(err).

rpc/dataconn/base2bufpool/nofitbehavior_enumer.go

@@ -1,5 +1,6 @@
 // Code generated by "enumer -type NoFitBehavior"; DO NOT EDIT.
 
+//
 package base2bufpool
 
 import (

rpc/dataconn/dataconn_client.go

@@ -141,7 +141,7 @@ func (c *Client) ReqSend(ctx context.Context, req *pdu.SendReq) (*pdu.SendRes, z
 
 func (c *Client) ReqRecv(ctx context.Context, req *pdu.ReceiveReq, streamCopier zfs.StreamCopier) (*pdu.ReceiveRes, error) {
 
-	defer c.log.Info("ReqRecv returns")
+	defer c.log.Debug("ReqRecv returns")
 	conn, err := c.getWire(ctx)
 	if err != nil {
 		return nil, err

rpc/dataconn/frameconn/frameconn.go

@@ -8,6 +8,7 @@ import (
 	"io/ioutil"
 	"net"
 	"sync"
+	"syscall"
 	"time"
 
 	"github.com/prometheus/client_golang/prometheus"
@@ -222,6 +223,10 @@ func (c *Conn) writeFrame(payload []byte, frameType uint32) error {
 	return nil
 }
 
+func (c *Conn) ResetWriteTimeout() error {
+	return c.nc.RenewWriteDeadline()
+}
+
 func (c *Conn) Shutdown(deadline time.Time) error {
 	// TCP connection teardown is a bit wonky if we are in a situation
 	// where there is still data in flight (DIF) to our side:
@@ -291,11 +296,22 @@ func (c *Conn) Shutdown(deadline time.Time) error {
 
 	closeWire := func(step string) error {
 		// TODO SetLinger(0) or similiar (we want RST frames here, not FINS)
-		if closeErr := c.nc.Close(); closeErr != nil {
-			prom.ShutdownCloseErrors.WithLabelValues("close").Inc()
-			return closeErr
+		closeErr := c.nc.Close()
+		if closeErr == nil {
+			return nil
 		}
-		return nil
+
+		// TODO go1.13: https://github.com/zrepl/zrepl/issues/190
+		//              https://github.com/golang/go/issues/8319
+		// (use errors.Is(closeErr, syscall.ECONNRESET))
+		if pe, ok := closeErr.(*net.OpError); ok && pe.Err == syscall.ECONNRESET {
+			// connection reset by peer on FreeBSD, see https://github.com/zrepl/zrepl/issues/190
+			// We know from kernel code reading that the FD behind c.nc is closed, so let's not consider this an error
+			return nil
+		}
+
+		prom.ShutdownCloseErrors.WithLabelValues("close").Inc()
+		return closeErr
 	}
 
 	hardclose := func(err error, step string) error {
@@ -305,7 +321,7 @@ func (c *Conn) Shutdown(deadline time.Time) error {
 
 	c.shutdown.Begin()
 	// new calls to c.ReadFrame and c.WriteFrame will now return ErrShutdown
-	// Aquiring writeMtx and readMtx ensures that the last calls exit successfully
+	// Aquiring writeMtx and readMtx afterwards ensures that already-running calls exit successfully
 
 	// disable renewing timeouts now, enforce the requested deadline instead
 	// we need to do this before aquiring locks to enforce the timeout on slow

rpc/dataconn/heartbeatconn/heartbeatconn.go

@@ -123,7 +123,17 @@ func (c *Conn) readFrameFiltered() (frameconn.Frame, error) {
 			return frameconn.Frame{}, fmt.Errorf("unknown frame type %x", f.Header.Type)
 		}
 		// drop heartbeat frame
-		debug("received heartbeat")
+		debug("received heartbeat, resetting write timeout")
+		// the peer's heartbeat proves to us that the peer is still live
+		// => trust the peer at this point (DoS risks are ignored ATM)
+		// => we assume that the connection is symmetric duplex, i.e., if receiving works for us,
+		//    sending works for us, too.
+		// So, let's grant the peer another write timeout.
+		err = c.fc.ResetWriteTimeout()
+		debug("renew frameconn write timeout returned errT=%T err=%s", err, err)
+		if err != nil {
+			return frameconn.Frame{}, err
+		}
 		continue
 	}
 }

rpc/dataconn/heartbeatconn/integration_test_variablereceiverate/heartbeatconn_integration_variablereceiverate.go

@@ -0,0 +1,144 @@
+// This integration test exercises the behavior of heartbeatconn
+// where the server is slow at handling the data received over the connection.
+// Note that the server is still sending heartbeats to the client, it's just the
+// data handling (usually I/O in case of zrepl endpoint.Receiver) that is slow.
+//
+// In commit 082335df5d85e1b0b9faa35ff182c71886142d3e and earlier, heartbeatconn would fail
+// this benchmark with a writev I/O timeout (here the ss(8) output at the time of failure)
+//
+//  ESTAB        33369    0                                        127.0.0.1:12345                                                127.0.0.1:57282                    users:(("heartbeatconn_i",pid=25953,fd=5))
+//	 cubic wscale:7,7 rto:203 rtt:2.992/5.849 ato:162 mss:32768 pmtu:65535 rcvmss:32741 advmss:65483 cwnd:10 bytes_sent:48 bytes_acked:48 bytes_received:195401 segs_out:44 segs_in:57 data_segs_out:6 data_segs_in:34 send 876.1Mbps lastsnd:125 lastrcv:9390 lastack:125 pacing_rate 1752.0Mbps delivery_rate 6393.8Mbps delivered:7 app_limited busy:42ms rcv_rtt:1 rcv_space:65483 rcv_ssthresh:65483 minrtt:0.029
+//	 --
+//	 ESTAB        0        3956805                                  127.0.0.1:57282                                                127.0.0.1:12345                    users:(("heartbeatconn_i",pid=26100,fd=3))
+//		  cubic wscale:7,7 rto:211 backoff:5 rtt:10.38/16.937 ato:40 mss:32768 pmtu:65535 rcvmss:536 advmss:65483 cwnd:10 bytes_sent:195401 bytes_acked:195402 bytes_received:48 segs_out:57 segs_in:45 data_segs_out:34 data_segs_in:6 send 252.5Mbps lastsnd:9390 lastrcv:125 lastack:125 pacing_rate 505.1Mbps delivery_rate 1971.0Mbps delivered:35 busy:30127ms rwnd_limited:30086ms(99.9%) rcv_space:65495 rcv_ssthresh:65495 notsent:3956805 minrtt:0.007
+//	 panic: writev tcp 127.0.0.1:57282->127.0.0.1:12345: i/o timeout
+//
+// The assumed reason for those writev timeouts is the following:
+// - Sporadic server stalls (sever data handling, usually I/O) cause TCP exponential backoff on the client for client->server
+// - Go runtime unblocks after the deadline expires, resultin gin writev I/O timeout
+// - That is, even though the client observed heartbeats from the server
+// -> TCP doesn't assume symmetric connection behavior, but our implementation does.
+//
+// The fix contained in the commit this message was committed with resets the deadline whenever
+// a heartbeat is received from the server.
+//
+//
+// How to run this integration test:
+//
+//
+//  Terminal 1:
+//     $ ZREPL_RPC_DATACONN_HEARTBEATCONN_DEBUG=1 go run heartbeatconn_integration_variablereceiverate.go  -mode server -addr 127.0.0.1:12345
+//     rpc/dataconn/heartbeatconn: send heartbeat
+//     rpc/dataconn/heartbeatconn: send heartbeat
+//     ...
+//
+//  Terminal 2:
+//     $ ZREPL_RPC_DATACONN_HEARTBEATCONN_DEBUG=1 go run heartbeatconn_integration_variablereceiverate.go  -mode  client -addr 127.0.0.1:12345
+//     rpc/dataconn/heartbeatconn: received heartbeat, resetting write timeout
+//     rpc/dataconn/heartbeatconn: renew frameconn write timeout returned errT=<nil> err=%!s(<nil>)
+//     rpc/dataconn/heartbeatconn: send heartbeat
+//     rpc/dataconn/heartbeatconn: received heartbeat, resetting write timeout
+//     rpc/dataconn/heartbeatconn: renew frameconn write timeout returned errT=<nil> err=%!s(<nil>)
+//     rpc/dataconn/heartbeatconn: received heartbeat, resetting write timeout
+//     ...
+//
+//  You should observe
+package main
+
+import (
+	"bytes"
+	"flag"
+	"fmt"
+	"io"
+	"log"
+	"math/rand"
+	"net"
+	"os"
+	"os/exec"
+	"path"
+	"time"
+
+	"github.com/zrepl/zrepl/util/devnoop"
+
+	"github.com/zrepl/zrepl/rpc/dataconn/heartbeatconn"
+)
+
+func orDie(err error) {
+	if err != nil {
+		grepfield := path.Base(os.Args[0])[:10]
+		fmt.Fprintf(os.Stderr, "grepping for %s\n", grepfield)
+		sh := fmt.Sprintf("ss -ntpi | grep -A1 %s", grepfield)
+		cmd := exec.Command("bash", "-c", sh)
+		o, _ := cmd.CombinedOutput()
+		buf := bytes.NewBuffer(o)
+		_, _ = io.Copy(os.Stderr, buf)
+		panic(err)
+	}
+}
+
+var mode string
+var addr string
+
+func main() {
+
+	flag.StringVar(&mode, "mode", "", "server|client")
+	flag.StringVar(&addr, "addr", "INVALID", "")
+	flag.Parse()
+
+	modemap := map[string]func(){
+		"server": server,
+		"client": client,
+	}
+	modemap[mode]()
+
+}
+
+func server() {
+	ln, err := net.Listen("tcp", addr)
+	orDie(err)
+	l := ln.(*net.TCPListener)
+	for {
+		c, err := l.AcceptTCP()
+		if err != nil {
+			log.Printf("accept err: %s", err)
+			continue
+		}
+		hc := heartbeatconn.Wrap(c, 5*time.Second, 10*time.Second)
+
+		for {
+			f, err := hc.ReadFrame()
+			orDie(err)
+			// _, err = buf.Write(f.Buffer.Bytes())
+			// orDie(err)
+			sleep := time.Duration(rand.NormFloat64()*500) * time.Millisecond
+			time.Sleep(sleep)
+			f.Buffer.Free()
+		}
+
+	}
+
+}
+
+func client() {
+	c, err := net.Dial("tcp", addr)
+	orDie(err)
+	hc := heartbeatconn.Wrap(c.(*net.TCPConn), 5*time.Second, 10*time.Second)
+
+	// follow API requirements to always ReadFrame
+	go func() {
+		for {
+			f, err := hc.ReadFrame()
+			orDie(err)
+			f.Buffer.Free()
+		}
+	}()
+
+	dn := devnoop.Get()
+	var buf [1 << 10]byte
+	for {
+		n, err := dn.Read(buf[:])
+		orDie(err)
+		err = hc.WriteFrame(buf[:n], 23)
+		orDie(err)
+	}
+}

rpc/dataconn/stream/stream.go

@@ -7,6 +7,7 @@ import (
 	"io"
 	"net"
 	"strings"
+	"sync/atomic"
 	"unicode/utf8"
 
 	"github.com/zrepl/zrepl/logger"
@@ -81,8 +82,10 @@ func doWriteStream(ctx context.Context, c *heartbeatconn.Conn, stream io.Reader,
 	}
 
 	reads := make(chan read, 5)
+	var stopReading uint32
 	go func() {
-		for {
+		defer close(reads)
+		for atomic.LoadUint32(&stopReading) == 0 {
 			buffer := bufpool.Get(1 << FramePayloadShift)
 			bufferBytes := buffer.Bytes()
 			n, err := io.ReadFull(stream, bufferBytes)
@@ -97,12 +100,21 @@ func doWriteStream(ctx context.Context, c *heartbeatconn.Conn, stream io.Reader,
 			}
 			if err != nil {
 				reads <- read{err: err} // RULE1
-				close(reads)
 				return
 			}
 		}
 	}()
 
+	defer func() {
+		// stop reading
+		atomic.StoreUint32(&stopReading, 1)
+		// drain in-flight reads
+		for read := range reads {
+			debug("doWriteStream: drain read channel")
+			read.buf.Free()
+		}
+	}()
+
 	for read := range reads {
 		if read.err == nil {
 			// RULE 1: read.buf is valid
@@ -201,10 +213,17 @@ type readFrameResult struct {
 	err error
 }
 
-func readFrames(reads chan<- readFrameResult, c *heartbeatconn.Conn) {
-	for {
+// readFrames reads from c into reads
+// if a read from c encounters an error, noMoreReads is closed before sending the result into reads
+func readFrames(reads chan<- readFrameResult, noMoreReads chan<- struct{}, c *heartbeatconn.Conn) {
+	// noMoreReads is already closed, don't re-close it
+	defer close(reads)
+	for { // only exits after a read error, make sure noMoreReads is closed
 		var r readFrameResult
 		r.f, r.err = c.ReadFrame()
+		if r.err != nil && noMoreReads != nil {
+			close(noMoreReads)
+		}
 		reads <- r
 		if r.err != nil {
 			return

rpc/dataconn/stream/stream_conn.go

@@ -5,9 +5,13 @@ import (
 	"context"
 	"fmt"
 	"io"
+	"net"
 	"sync"
+	"sync/atomic"
 	"time"
 
+	"github.com/pkg/errors"
+
 	"github.com/zrepl/zrepl/rpc/dataconn/heartbeatconn"
 	"github.com/zrepl/zrepl/rpc/dataconn/timeoutconn"
 	"github.com/zrepl/zrepl/zfs"
@@ -21,6 +25,8 @@ type Conn struct {
 	// filled by per-conn readFrames goroutine
 	frameReads chan readFrameResult
 
+	closeState closeState
+
 	// readMtx serializes read stream operations because we inherently only
 	// support a single stream at a time over hc.
 	readMtx   sync.Mutex
@@ -63,15 +69,29 @@ func isConnCleanAfterWrite(err error) bool {
 	return err == nil
 }
 
-var ErrReadFramesStopped = fmt.Errorf("stream: reading frames stopped")
-
 func (c *Conn) readFrames() {
-	defer close(c.waitReadFramesDone)
-	defer close(c.frameReads)
-	readFrames(c.frameReads, c.hc)
+	readFrames(c.frameReads, c.waitReadFramesDone, c.hc)
 }
 
-func (c *Conn) ReadStreamedMessage(ctx context.Context, maxSize uint32, frameType uint32) ([]byte, *ReadStreamError) {
+func (c *Conn) ReadStreamedMessage(ctx context.Context, maxSize uint32, frameType uint32) (_ []byte, err *ReadStreamError) {
+
+	// if we are closed while reading, return that as an error
+	if closeGuard, cse := c.closeState.RWEntry(); cse != nil {
+		return nil, &ReadStreamError{
+			Kind: ReadStreamErrorKindConn,
+			Err:  cse,
+		}
+	} else {
+		defer func(err **ReadStreamError) {
+			if closed := closeGuard.RWExit(); closed != nil {
+				*err = &ReadStreamError{
+					Kind: ReadStreamErrorKindConn,
+					Err:  closed,
+				}
+			}
+		}(&err)
+	}
+
 	c.readMtx.Lock()
 	defer c.readMtx.Unlock()
 	if !c.readClean {
@@ -92,7 +112,7 @@ func (c *Conn) ReadStreamedMessage(ctx context.Context, maxSize uint32, frameTyp
 			panic(err)
 		}
 	}()
-	err := readStream(c.frameReads, c.hc, w, frameType)
+	err = readStream(c.frameReads, c.hc, w, frameType)
 	c.readClean = isConnCleanAfterRead(err)
 	_ = w.CloseWithError(readMessageSentinel) // always returns nil
 	wg.Wait()
@@ -104,23 +124,47 @@ func (c *Conn) ReadStreamedMessage(ctx context.Context, maxSize uint32, frameTyp
 }
 
 // WriteStreamTo reads a stream from Conn and writes it to w.
-func (c *Conn) ReadStreamInto(w io.Writer, frameType uint32) zfs.StreamCopierError {
+func (c *Conn) ReadStreamInto(w io.Writer, frameType uint32) (err zfs.StreamCopierError) {
+
+	// if we are closed while writing, return that as an error
+	if closeGuard, cse := c.closeState.RWEntry(); cse != nil {
+		return cse
+	} else {
+		defer func(err *zfs.StreamCopierError) {
+			if closed := closeGuard.RWExit(); closed != nil {
+				*err = closed
+			}
+		}(&err)
+	}
+
 	c.readMtx.Lock()
 	defer c.readMtx.Unlock()
 	if !c.readClean {
 		return writeStreamToErrorUnknownState{}
 	}
-	var err *ReadStreamError = readStream(c.frameReads, c.hc, w, frameType)
-	c.readClean = isConnCleanAfterRead(err)
+	var rse *ReadStreamError = readStream(c.frameReads, c.hc, w, frameType)
+	c.readClean = isConnCleanAfterRead(rse)
 
 	// https://golang.org/doc/faq#nil_error
-	if err == nil {
+	if rse == nil {
 		return nil
 	}
-	return err
+	return rse
 }
 
-func (c *Conn) WriteStreamedMessage(ctx context.Context, buf io.Reader, frameType uint32) error {
+func (c *Conn) WriteStreamedMessage(ctx context.Context, buf io.Reader, frameType uint32) (err error) {
+
+	// if we are closed while writing, return that as an error
+	if closeGuard, cse := c.closeState.RWEntry(); cse != nil {
+		return cse
+	} else {
+		defer func(err *error) {
+			if closed := closeGuard.RWExit(); closed != nil {
+				*err = closed
+			}
+		}(&err)
+	}
+
 	c.writeMtx.Lock()
 	defer c.writeMtx.Unlock()
 	if !c.writeClean {
@@ -134,7 +178,19 @@ func (c *Conn) WriteStreamedMessage(ctx context.Context, buf io.Reader, frameTyp
 	return errConn
 }
 
-func (c *Conn) SendStream(ctx context.Context, src zfs.StreamCopier, frameType uint32) error {
+func (c *Conn) SendStream(ctx context.Context, src zfs.StreamCopier, frameType uint32) (err error) {
+
+	// if we are closed while reading, return that as an error
+	if closeGuard, cse := c.closeState.RWEntry(); cse != nil {
+		return cse
+	} else {
+		defer func(err *error) {
+			if closed := closeGuard.RWExit(); closed != nil {
+				*err = closed
+			}
+		}(&err)
+	}
+
 	c.writeMtx.Lock()
 	defer c.writeMtx.Unlock()
 	if !c.writeClean {
@@ -186,8 +242,75 @@ func (c *Conn) SendStream(ctx context.Context, src zfs.StreamCopier, frameType u
 	}
 }
 
-func (c *Conn) Close() error {
-	err := c.hc.Shutdown()
-	<-c.waitReadFramesDone
-	return err
+type closeState struct {
+	closeCount uint32
+}
+
+type closeStateErrConnectionClosed struct{}
+
+var _ zfs.StreamCopierError = (*closeStateErrConnectionClosed)(nil)
+var _ error = (*closeStateErrConnectionClosed)(nil)
+var _ net.Error = (*closeStateErrConnectionClosed)(nil)
+
+func (e *closeStateErrConnectionClosed) Error() string {
+	return "connection closed"
+}
+func (e *closeStateErrConnectionClosed) IsReadError() bool  { return true }
+func (e *closeStateErrConnectionClosed) IsWriteError() bool { return true }
+func (e *closeStateErrConnectionClosed) Timeout() bool      { return false }
+func (e *closeStateErrConnectionClosed) Temporary() bool    { return false }
+
+func (s *closeState) CloseEntry() error {
+	firstCloser := atomic.AddUint32(&s.closeCount, 1) == 1
+	if !firstCloser {
+		return errors.New("duplicate close")
+	}
+	return nil
+}
+
+type closeStateEntry struct {
+	s          *closeState
+	entryCount uint32
+}
+
+func (s *closeState) RWEntry() (e *closeStateEntry, err zfs.StreamCopierError) {
+	entry := &closeStateEntry{s, atomic.LoadUint32(&s.closeCount)}
+	if entry.entryCount > 0 {
+		return nil, &closeStateErrConnectionClosed{}
+	}
+	return entry, nil
+}
+
+func (e *closeStateEntry) RWExit() zfs.StreamCopierError {
+	if atomic.LoadUint32(&e.entryCount) == e.entryCount {
+		// no calls to Close() while running rw operation
+		return nil
+	}
+	return &closeStateErrConnectionClosed{}
+}
+
+func (c *Conn) Close() error {
+	if err := c.closeState.CloseEntry(); err != nil {
+		return errors.Wrap(err, "stream conn close")
+	}
+
+	// Shutdown c.hc, which will cause c.readFrames to close c.waitReadFramesDone
+	err := c.hc.Shutdown()
+	// However, c.readFrames may be blocking on a filled c.frameReads
+	// and since the connection is closed, nobody is going to read from it
+	for read := range c.frameReads {
+		debug("Conn.Close() draining queued read")
+		read.f.Buffer.Free()
+	}
+	// if we can't close, don't expect c.readFrames to terminate
+	// this might leak c.readFrames, but we can't do something useful at this point
+	if err != nil {
+		return err
+	}
+	// shutdown didn't report errors, so c.readFrames will exit due to read error
+	// This behavior is the contract we have with c.hc.
+	// If that contract is broken, this read will block indefinitely and
+	// cause an easily diagnosable goroutine leak (of this goroutine)
+	<-c.waitReadFramesDone
+	return nil
 }

rpc/dataconn/stream/stream_test.go

@@ -57,7 +57,7 @@ func TestStreamer(t *testing.T) {
 		wg.Add(1)
 		go func() {
 			defer wg.Done()
-			readFrames(ch, b)
+			readFrames(ch, nil, b)
 		}()
 		err := readStream(ch, b, &buf, stype)
 		log.WithField("errType", fmt.Sprintf("%T %v", err, err)).Debug("ReadStream returned")
@@ -113,7 +113,7 @@ func TestMultiFrameStreamErrTraileror(t *testing.T) {
 		wg.Add(1)
 		go func() {
 			defer wg.Done()
-			readFrames(ch, b)
+			readFrames(ch, nil, b)
 		}()
 		err := readStream(ch, b, &buf, stype)
 		t.Logf("%s", err)

rpc/dataconn/timeoutconn/timeoutconn.go

@@ -80,7 +80,7 @@ func (c *Conn) renewReadDeadline() error {
 	return c.SetReadDeadline(time.Now().Add(c.idleTimeout))
 }
 
-func (c *Conn) renewWriteDeadline() error {
+func (c *Conn) RenewWriteDeadline() error {
 	if atomic.LoadInt32(&c.renewDeadlinesDisabled) != 0 {
 		return nil
 	}
@@ -107,7 +107,7 @@ restart:
 func (c Conn) Write(p []byte) (n int, err error) {
 	n = 0
 restart:
-	if err := c.renewWriteDeadline(); err != nil {
+	if err := c.RenewWriteDeadline(); err != nil {
 		return n, err
 	}
 	var nCurWrite int
@@ -127,7 +127,7 @@ restart:
 func (c Conn) WritevFull(bufs net.Buffers) (n int64, err error) {
 	n = 0
 restart:
-	if err := c.renewWriteDeadline(); err != nil {
+	if err := c.RenewWriteDeadline(); err != nil {
 		return n, err
 	}
 	var nCurWrite int64
@@ -162,10 +162,14 @@ func buildIovecs(buffers net.Buffers) (totalLen int64, vecs []syscall.Iovec) {
 		if len(buffers[i]) == 0 {
 			continue
 		}
-		vecs = append(vecs, syscall.Iovec{
+
+		v := syscall.Iovec{
 			Base: &buffers[i][0],
-			Len:  uint64(len(buffers[i])),
-		})
+		}
+		// syscall.Iovec.Len has platform-dependent size, thus use SetLen
+		v.SetLen(len(buffers[i]))
+
+		vecs = append(vecs, v)
 	}
 	return totalLen, vecs
 }
@@ -244,6 +248,14 @@ func (c Conn) readv(rawConn syscall.RawConn, iovecs []syscall.Iovec) (n int64, e
 func (c Conn) doOneReadv(rawConn syscall.RawConn, iovecs *[]syscall.Iovec) (n int64, err error) {
 	rawReadErr := rawConn.Read(func(fd uintptr) (done bool) {
 		// iovecs, n and err must not be shadowed!
+
+		// NOTE: unsafe.Pointer safety rules
+		// 		https://tip.golang.org/pkg/unsafe/#Pointer
+		//
+		//		(4) Conversion of a Pointer to a uintptr when calling syscall.Syscall.
+		// 		...
+		//		uintptr() conversions must appear within the syscall.Syscall argument list.
+		//      (even though we are not the escape analysis Likely not )
 		thisReadN, _, errno := syscall.Syscall(
 			syscall.SYS_READV,
 			fd,
@@ -260,16 +272,32 @@ func (c Conn) doOneReadv(rawConn syscall.RawConn, iovecs *[]syscall.Iovec) (n in
 		if int(thisReadN) < 0 {
 			panic("unexpected return value")
 		}
-		n += int64(thisReadN)
+		n += int64(thisReadN) // TODO check overflow
+
 		// shift iovecs forward
-		for left := int64(thisReadN); left > 0; {
-			curVecNewLength := int64((*iovecs)[0].Len) - left // TODO assert conversion
-			if curVecNewLength <= 0 {
-				left -= int64((*iovecs)[0].Len)
+		for left := int(thisReadN); left > 0; {
+			// conversion to uint does not change value, see TestIovecLenFieldIsMachineUint, and left > 0
+			thisIovecConsumedCompletely := uint((*iovecs)[0].Len) <= uint(left)
+			if thisIovecConsumedCompletely {
+				// Update left, cannot go below 0 due to
+				// a) definition of thisIovecConsumedCompletely
+				// b) left > 0 due to loop invariant
+				// Convertion .Len to int64 is thus also safe now, because it is < left < INT_MAX
+				left -= int((*iovecs)[0].Len)
 				*iovecs = (*iovecs)[1:]
 			} else {
+				// trim this iovec to remaining length
+
+				// NOTE: unsafe.Pointer safety rules
+				// 		https://tip.golang.org/pkg/unsafe/#Pointer
+				// 		(3) Conversion of a Pointer to a uintptr and back, with arithmetic.
+				// 		...
+				//		Note that both conversions must appear in the same expression,
+				//		with only the intervening arithmetic between them:
 				(*iovecs)[0].Base = (*byte)(unsafe.Pointer(uintptr(unsafe.Pointer((*iovecs)[0].Base)) + uintptr(left)))
-				(*iovecs)[0].Len = uint64(curVecNewLength)
+				curVecNewLength := uint((*iovecs)[0].Len) - uint(left) // casts to uint do not change value
+				(*iovecs)[0].SetLen(int(curVecNewLength))              // int and uint have the same size, no change of value
+
 				break // inner
 			}
 		}

rpc/dataconn/timeoutconn/timeoutconn_test.go

@@ -5,8 +5,10 @@ import (
 	"io"
 	"net"
 	"sync"
+	"syscall"
 	"testing"
 	"time"
+	"unsafe"
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -176,3 +178,14 @@ func TestNoPartialWritesDueToDeadline(t *testing.T) {
 	require.True(t, ok)
 	assert.True(t, netErr.Timeout())
 }
+
+func TestIovecLenFieldIsMachineUint(t *testing.T) {
+	iov := syscall.Iovec{}
+	_ = iov // make linter happy (unsafe.Sizeof not recognized as usage)
+	size_t := unsafe.Sizeof(iov.Len)
+	if size_t != unsafe.Sizeof(uint(23)) {
+		t.Fatalf("expecting (struct iov)->Len to be sizeof(uint)")
+	}
+	// ssize_t is defined to be the signed version of size_t,
+	// so we know sizeof(ssize_t) == sizeof(int)
+}

tools.go

@@ -0,0 +1,11 @@
+// +build tools
+
+package main
+
+import (
+	_ "golang.org/x/tools/cmd/stringer"
+	_ "github.com/golang/protobuf/protoc-gen-go"
+	_ "github.com/alvaroloes/enumer"
+	_ "golang.org/x/tools/cmd/goimports"
+	_ "github.com/golangci/golangci-lint/cmd/golangci-lint"
+)

transport/local/connect_local.go

@@ -3,6 +3,7 @@ package local
 import (
 	"context"
 	"fmt"
+	"time"
 
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/transport"
@@ -11,6 +12,7 @@ import (
 type LocalConnecter struct {
 	listenerName   string
 	clientIdentity string
+	dialTimeout    time.Duration
 }
 
 func LocalConnecterFromConfig(in *config.LocalConnect) (*LocalConnecter, error) {
@@ -20,10 +22,27 @@ func LocalConnecterFromConfig(in *config.LocalConnect) (*LocalConnecter, error)
 	if in.ListenerName == "" {
 		return nil, fmt.Errorf("ListenerName must not be empty")
 	}
-	return &LocalConnecter{listenerName: in.ListenerName, clientIdentity: in.ClientIdentity}, nil
+	if in.DialTimeout < 0 {
+		return nil, fmt.Errorf("DialTimeout must be zero or positive")
+	}
+	cn := &LocalConnecter{
+		listenerName:   in.ListenerName,
+		clientIdentity: in.ClientIdentity,
+		dialTimeout:    in.DialTimeout,
+	}
+	return cn, nil
 }
 
 func (c *LocalConnecter) Connect(dialCtx context.Context) (transport.Wire, error) {
 	l := GetLocalListener(c.listenerName)
-	return l.Connect(dialCtx, c.clientIdentity)
+	if c.dialTimeout > 0 {
+		ctx, cancel := context.WithTimeout(dialCtx, c.dialTimeout)
+		defer cancel()
+		dialCtx = ctx // shadow
+	}
+	w, err := l.Connect(dialCtx, c.clientIdentity)
+	if err == context.DeadlineExceeded {
+		return nil, fmt.Errorf("local listener %q not reachable", c.listenerName)
+	}
+	return w, err
 }

transport/local/serve_local.go

@@ -61,7 +61,9 @@ func (l *LocalListener) Connect(dialCtx context.Context, clientIdentity string)
 	// place request
 	req := connectRequest{
 		clientIdentity: clientIdentity,
-		callback:       make(chan connectResult),
+		// ensure non-blocking send in Accept, we don't necessarily read from callback before
+		// Accept writes to it
+		callback: make(chan connectResult, 1),
 	}
 	select {
 	case l.connects <- req:
@@ -93,19 +95,35 @@ func (l *LocalListener) Addr() net.Addr { return localAddr{"<listening>"} }
 
 func (l *LocalListener) Accept(ctx context.Context) (*transport.AuthConn, error) {
 	respondToRequest := func(req connectRequest, res connectResult) (err error) {
+
 		transport.GetLogger(ctx).
 			WithField("res.conn", res.conn).WithField("res.err", res.err).
 			Debug("responding to client request")
+
+		// contract bewteen Connect and Accept is that Connect sends a req.callback
+		// into which we can send one result non-blockingly.
+		// We want to panic if that contract is violated (impl error)
+		//
+		// However, Connect also supports timeouts through context cancellation:
+		// Connect closes the channel into which we might send, which will panic.
+		//
+		// ==> distinguish those cases in defer
+		const clientCallbackBlocked = "client-provided callback did block on send"
 		defer func() {
 			errv := recover()
-			transport.GetLogger(ctx).WithField("recover_err", errv).
-				Debug("panic on send to client callback, likely a legitimate client-side timeout")
+			if errv == clientCallbackBlocked {
+				// this would be a violation of contract betwee Connect and Accept, see above
+				panic(clientCallbackBlocked)
+			} else {
+				transport.GetLogger(ctx).WithField("recover_err", errv).
+					Debug("panic on send to client callback, likely a legitimate client-side timeout")
+			}
 		}()
 		select {
 		case req.callback <- res:
 			err = nil
 		default:
-			err = fmt.Errorf("client-provided callback did block on send")
+			panic(clientCallbackBlocked)
 		}
 		close(req.callback)
 		return err

util/circlog/circlog.go

@@ -0,0 +1,167 @@
+package circlog
+
+import (
+	"fmt"
+	"math/bits"
+	"sync"
+)
+
+const CIRCULARLOG_INIT_SIZE int = 32 << 10
+
+type CircularLog struct {
+	buf         []byte
+	size        int
+	max         int
+	written     int
+	writeCursor int
+	/*
+		Mutex prevents:
+		concurrent writes:
+			buf, size, written, writeCursor in Write([]byte)
+			buf, writeCursor in Reset()
+		data races vs concurrent Write([]byte) calls:
+			size in Size()
+			size, writeCursor in Len()
+			buf, size, written, writeCursor in Bytes()
+	*/
+	mtx sync.Mutex
+}
+
+func NewCircularLog(max int) (*CircularLog, error) {
+	if max <= 0 {
+		return nil, fmt.Errorf("max must be positive")
+	}
+
+	return &CircularLog{
+		size: CIRCULARLOG_INIT_SIZE,
+		buf:  make([]byte, CIRCULARLOG_INIT_SIZE),
+		max:  max,
+	}, nil
+}
+
+func nextPow2Int(n int) int {
+	if n < 1 {
+		panic("can only round up positive integers")
+	}
+	r := uint(n)
+	// Credit: https://graphics.stanford.edu/~seander/bithacks.html#RoundUpPowerOf2
+	r--
+	// Assume at least a 32 bit integer
+	r |= r >> 1
+	r |= r >> 2
+	r |= r >> 4
+	r |= r >> 8
+	r |= r >> 16
+	if bits.UintSize == 64 {
+		r |= r >> 32
+	}
+	r++
+	// Can't exceed max positive int value
+	if r > ^uint(0)>>1 {
+		panic("rounded to larger than int()")
+	}
+	return int(r)
+}
+
+func (cl *CircularLog) Write(data []byte) (int, error) {
+	cl.mtx.Lock()
+	defer cl.mtx.Unlock()
+	n := len(data)
+
+	// Keep growing the buffer by doubling until
+	// hitting cl.max
+	bufAvail := cl.size - cl.writeCursor
+	// If cl.writeCursor wrapped on the last write,
+	// then the buffer is full, not empty.
+	if cl.writeCursor == 0 && cl.written > 0 {
+		bufAvail = 0
+	}
+	if n > bufAvail && cl.size < cl.max {
+		// Add to size, not writeCursor, so as
+		// to not resize multiple times if this
+		// Write() immediately fills up the buffer
+		newSize := nextPow2Int(cl.size + n)
+		if newSize > cl.max {
+			newSize = cl.max
+		}
+		newBuf := make([]byte, newSize)
+		// Reset write cursor to old size if wrapped
+		if cl.writeCursor == 0 && cl.written > 0 {
+			cl.writeCursor = cl.size
+		}
+		copy(newBuf, cl.buf[:cl.writeCursor])
+		cl.buf = newBuf
+		cl.size = newSize
+	}
+
+	// If data to be written is larger than the max size,
+	// discard all but the last cl.size bytes
+	if n > cl.size {
+		data = data[n-cl.size:]
+		// Overwrite the beginning of data
+		// with a string indicating truncation
+		copy(data, []byte("(...)"))
+	}
+	// First copy data to the end of buf. If that wasn't
+	// all of data, then copy the rest to the beginning
+	// of buf.
+	copied := copy(cl.buf[cl.writeCursor:], data)
+	if copied < n {
+		copied += copy(cl.buf, data[copied:])
+	}
+
+	cl.writeCursor = ((cl.writeCursor + copied) % cl.size)
+	cl.written += copied
+	return copied, nil
+}
+
+func (cl *CircularLog) Size() int {
+	cl.mtx.Lock()
+	defer cl.mtx.Unlock()
+	return cl.size
+}
+
+func (cl *CircularLog) Len() int {
+	cl.mtx.Lock()
+	defer cl.mtx.Unlock()
+	if cl.written >= cl.size {
+		return cl.size
+	} else {
+		return cl.writeCursor
+	}
+}
+
+func (cl *CircularLog) TotalWritten() int {
+	cl.mtx.Lock()
+	defer cl.mtx.Unlock()
+
+	return cl.written
+}
+
+func (cl *CircularLog) Reset() {
+	cl.mtx.Lock()
+	defer cl.mtx.Unlock()
+	cl.writeCursor = 0
+	cl.buf = make([]byte, CIRCULARLOG_INIT_SIZE)
+	cl.size = CIRCULARLOG_INIT_SIZE
+}
+
+func (cl *CircularLog) Bytes() []byte {
+	cl.mtx.Lock()
+	defer cl.mtx.Unlock()
+	switch {
+	case cl.written >= cl.size && cl.writeCursor == 0:
+		return cl.buf
+	case cl.written > cl.size:
+		ret := make([]byte, cl.size)
+		copy(ret, cl.buf[cl.writeCursor:])
+		copy(ret[cl.size-cl.writeCursor:], cl.buf[:cl.writeCursor])
+		return ret
+	default:
+		return cl.buf[:cl.writeCursor]
+	}
+}
+
+func (cl *CircularLog) String() string {
+	return string(cl.Bytes())
+}

util/circlog/circlog_test.go

@@ -0,0 +1,144 @@
+package circlog_test
+
+import (
+	"bytes"
+	"io"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+
+	"github.com/zrepl/zrepl/util/circlog"
+)
+
+func TestCircularLog(t *testing.T) {
+	var maxCircularLogSize int = 1 << 20
+
+	writeFixedSize := func(w io.Writer, size int) (int, error) {
+		pattern := []byte{0xDE, 0xAD, 0xBE, 0xEF}
+		writeBytes := bytes.Repeat(pattern, size/4)
+		if len(writeBytes) < size {
+			writeBytes = append(writeBytes, pattern[:size-len(writeBytes)]...)
+		}
+		n, err := w.Write(writeBytes)
+		if err != nil {
+			return n, err
+		}
+		return n, nil
+	}
+
+	t.Run("negative-size-error", func(t *testing.T) {
+		_, err := circlog.NewCircularLog(-1)
+		require.EqualError(t, err, "max must be positive")
+	})
+
+	t.Run("no-resize", func(t *testing.T) {
+		log, err := circlog.NewCircularLog(maxCircularLogSize)
+		require.NoError(t, err)
+
+		initSize := log.Size()
+		writeSize := circlog.CIRCULARLOG_INIT_SIZE - 1
+		_, err = writeFixedSize(log, writeSize)
+		require.NoError(t, err)
+		finalSize := log.Size()
+		require.Equal(t, initSize, finalSize)
+		require.Equal(t, writeSize, log.Len())
+	})
+	t.Run("one-resize", func(t *testing.T) {
+		log, err := circlog.NewCircularLog(maxCircularLogSize)
+		require.NoError(t, err)
+
+		initSize := log.Size()
+		writeSize := circlog.CIRCULARLOG_INIT_SIZE + 1
+		_, err = writeFixedSize(log, writeSize)
+		require.NoError(t, err)
+		finalSize := log.Size()
+		require.Greater(t, finalSize, initSize)
+		require.LessOrEqual(t, finalSize, maxCircularLogSize)
+		require.Equal(t, writeSize, log.Len())
+	})
+	t.Run("reset", func(t *testing.T) {
+		log, err := circlog.NewCircularLog(maxCircularLogSize)
+		require.NoError(t, err)
+
+		initSize := log.Size()
+		_, err = writeFixedSize(log, maxCircularLogSize)
+		require.NoError(t, err)
+		log.Reset()
+		finalSize := log.Size()
+		require.Equal(t, initSize, finalSize)
+	})
+	t.Run("wrap-exactly-maximum", func(t *testing.T) {
+		log, err := circlog.NewCircularLog(maxCircularLogSize)
+		require.NoError(t, err)
+
+		initSize := log.Size()
+		writeSize := maxCircularLogSize / 4
+		for written := 0; written < maxCircularLogSize; {
+			n, err := writeFixedSize(log, writeSize)
+			written += n
+			require.NoError(t, err)
+		}
+		finalSize := log.Size()
+		require.Greater(t, finalSize, initSize)
+		require.Equal(t, finalSize, maxCircularLogSize)
+		require.Equal(t, finalSize, log.Len())
+	})
+	t.Run("wrap-partial", func(t *testing.T) {
+		log, err := circlog.NewCircularLog(maxCircularLogSize)
+		require.NoError(t, err)
+
+		initSize := log.Size()
+
+		startSentinel := []byte{0x00, 0x00, 0x00, 0x00}
+		_, err = log.Write(startSentinel)
+		require.NoError(t, err)
+
+		nWritesToFill := 4
+		writeSize := maxCircularLogSize / nWritesToFill
+		for i := 0; i < (nWritesToFill + 1); i++ {
+			_, err := writeFixedSize(log, writeSize)
+			require.NoError(t, err)
+		}
+
+		endSentinel := []byte{0xFF, 0xFF, 0xFF, 0xFF}
+		_, err = log.Write(endSentinel)
+		require.NoError(t, err)
+
+		finalSize := log.Size()
+		require.Greater(t, finalSize, initSize)
+		require.Greater(t, log.TotalWritten(), finalSize)
+		require.Equal(t, finalSize, maxCircularLogSize)
+		require.Equal(t, finalSize, log.Len())
+
+		logBytes := log.Bytes()
+
+		require.Equal(t, endSentinel, logBytes[len(logBytes)-len(endSentinel):])
+		require.NotEqual(t, startSentinel, logBytes[:len(startSentinel)])
+	})
+	t.Run("overflow-write", func(t *testing.T) {
+		log, err := circlog.NewCircularLog(maxCircularLogSize)
+		require.NoError(t, err)
+
+		initSize := log.Size()
+		writeSize := maxCircularLogSize + 1
+		n, err := writeFixedSize(log, writeSize)
+		require.NoError(t, err)
+		finalSize := log.Size()
+		require.Less(t, n, writeSize)
+		require.Greater(t, finalSize, initSize)
+		require.Equal(t, finalSize, maxCircularLogSize)
+		logBytes := log.Bytes()
+		require.Equal(t, []byte("(...)"), logBytes[:5])
+	})
+	t.Run("stringify", func(t *testing.T) {
+		log, err := circlog.NewCircularLog(maxCircularLogSize)
+		require.NoError(t, err)
+
+		writtenString := "A harmful truth is better than a useful lie."
+		n, err := log.Write([]byte(writtenString))
+		require.NoError(t, err)
+		require.Equal(t, len(writtenString), n)
+		loggedString := log.String()
+		require.Equal(t, writtenString, loggedString)
+	})
+}

util/circlog/nextpow2int_test.go

@@ -0,0 +1,16 @@
+package circlog
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func TestNextPow2Int(t *testing.T) {
+	require.Equal(t, 512, nextPow2Int(512), "a power of 2 should round to itself")
+	require.Equal(t, 1024, nextPow2Int(513), "should round up to the next power of 2")
+	require.PanicsWithValue(t, "can only round up positive integers", func() { nextPow2Int(0) }, "unimplemented: zero is not positive; corner case")
+	require.PanicsWithValue(t, "can only round up positive integers", func() { nextPow2Int(-1) }, "unimplemented: cannot round up negative numbers")
+	maxInt := int((^uint(0)) >> 1)
+	require.PanicsWithValue(t, "rounded to larger than int()", func() { nextPow2Int(maxInt - 1) }, "cannot round to a number bigger than the int type")
+}

util/envconst/envconst.go

@@ -25,6 +25,23 @@ func Duration(varname string, def time.Duration) time.Duration {
 	return d
 }
 
+func Int(varname string, def int) int {
+	if v, ok := cache.Load(varname); ok {
+		return v.(int)
+	}
+	e := os.Getenv(varname)
+	if e == "" {
+		return def
+	}
+	d64, err := strconv.ParseInt(e, 10, strconv.IntSize)
+	if err != nil {
+		panic(err)
+	}
+	d := int(d64)
+	cache.Store(varname, d)
+	return d
+}
+
 func Int64(varname string, def int64) int64 {
 	if v, ok := cache.Load(varname); ok {
 		return v.(int64)
@@ -56,3 +73,15 @@ func Bool(varname string, def bool) bool {
 	cache.Store(varname, d)
 	return d
 }
+
+func String(varname string, def string) string {
+	if v, ok := cache.Load(varname); ok {
+		return v.(string)
+	}
+	e := os.Getenv(varname)
+	if e == "" {
+		return def
+	}
+	cache.Store(varname, e)
+	return e
+}

zfs/placeholder.go

@@ -78,6 +78,9 @@ func ZFSGetFilesystemPlaceholderState(p *DatasetPath) (state *FilesystemPlacehol
 }
 
 func ZFSCreatePlaceholderFilesystem(p *DatasetPath) (err error) {
+	if p.Length() == 1 {
+		return fmt.Errorf("cannot create %q: pools cannot be created with zfs create", p.ToString())
+	}
 	cmd := exec.Command(ZFS_BINARY, "create",
 		"-o", fmt.Sprintf("%s=%s", PlaceholderPropertyName, placeholderPropertyOn),
 		"-o", "mountpoint=none",

zfs/versions.go

@@ -154,15 +154,3 @@ func ZFSListFilesystemVersions(fs *DatasetPath, filter FilesystemVersionFilter)
 	}
 	return
 }
-
-func ZFSDestroyFilesystemVersion(filesystem *DatasetPath, version *FilesystemVersion) (err error) {
-
-	datasetPath := version.ToAbsPath(filesystem)
-
-	// Sanity check...
-	if !strings.ContainsAny(datasetPath, "@#") {
-		return fmt.Errorf("sanity check failed: no @ or # character found in %q", datasetPath)
-	}
-
-	return ZFSDestroy(datasetPath)
-}

zfs/versions_destroy.go

@@ -0,0 +1,234 @@
+package zfs
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"os/exec"
+	"sort"
+	"strings"
+	"sync"
+	"syscall"
+
+	"github.com/zrepl/zrepl/util/envconst"
+)
+
+func ZFSDestroyFilesystemVersion(filesystem *DatasetPath, version *FilesystemVersion) (err error) {
+
+	datasetPath := version.ToAbsPath(filesystem)
+
+	// Sanity check...
+	if !strings.ContainsAny(datasetPath, "@#") {
+		return fmt.Errorf("sanity check failed: no @ or # character found in %q", datasetPath)
+	}
+
+	return ZFSDestroy(datasetPath)
+}
+
+var destroyerSingleton = destroyerImpl{}
+
+type DestroySnapOp struct {
+	Filesystem string
+	Name       string
+	ErrOut     *error
+}
+
+func (o *DestroySnapOp) String() string {
+	return fmt.Sprintf("destroy operation %s@%s", o.Filesystem, o.Name)
+}
+
+func ZFSDestroyFilesystemVersions(reqs []*DestroySnapOp) {
+	doDestroy(context.TODO(), reqs, destroyerSingleton)
+}
+
+func setDestroySnapOpErr(b []*DestroySnapOp, err error) {
+	for _, r := range b {
+		*r.ErrOut = err
+	}
+}
+
+type destroyer interface {
+	Destroy(args []string) error
+	DestroySnapshotsCommaSyntaxSupported() (bool, error)
+}
+
+func doDestroy(ctx context.Context, reqs []*DestroySnapOp, e destroyer) {
+
+	var validated []*DestroySnapOp
+	for _, req := range reqs {
+		// Filesystem and Snapshot should not be empty
+		// ZFS will generally fail because those are invalid destroy arguments,
+		// but we'd rather apply defensive programming here (doing destroy after all)
+		if req.Filesystem == "" {
+			*req.ErrOut = fmt.Errorf("Filesystem must not be an empty string")
+		} else if req.Name == "" {
+			*req.ErrOut = fmt.Errorf("Name must not be an empty string")
+		} else {
+			validated = append(validated, req)
+		}
+	}
+	reqs = validated
+
+	commaSupported, err := e.DestroySnapshotsCommaSyntaxSupported()
+	if err != nil {
+		debug("destroy: comma syntax support detection failed: %s", err)
+		setDestroySnapOpErr(reqs, err)
+		return
+	}
+
+	if !commaSupported {
+		doDestroySeq(ctx, reqs, e)
+	} else {
+		doDestroyBatched(ctx, reqs, e)
+	}
+}
+
+func doDestroySeq(ctx context.Context, reqs []*DestroySnapOp, e destroyer) {
+	for _, r := range reqs {
+		*r.ErrOut = e.Destroy([]string{fmt.Sprintf("%s@%s", r.Filesystem, r.Name)})
+	}
+}
+
+func doDestroyBatched(ctx context.Context, reqs []*DestroySnapOp, d destroyer) {
+	perFS := buildBatches(reqs)
+	for _, fsbatch := range perFS {
+		doDestroyBatchedRec(ctx, fsbatch, d)
+	}
+}
+
+func buildBatches(reqs []*DestroySnapOp) [][]*DestroySnapOp {
+	if len(reqs) == 0 {
+		return nil
+	}
+	sorted := make([]*DestroySnapOp, len(reqs))
+	copy(sorted, reqs)
+	sort.SliceStable(sorted, func(i, j int) bool {
+		// by filesystem, then snap name
+		fscmp := strings.Compare(sorted[i].Filesystem, sorted[j].Filesystem)
+		if fscmp != 0 {
+			return fscmp == -1
+		}
+		return strings.Compare(sorted[i].Name, sorted[j].Name) == -1
+	})
+
+	// group by fs
+	var perFS [][]*DestroySnapOp
+	consumed := 0
+	for consumed < len(sorted) {
+		batchConsumedUntil := consumed
+		for ; batchConsumedUntil < len(sorted) && sorted[batchConsumedUntil].Filesystem == sorted[consumed].Filesystem; batchConsumedUntil++ {
+		}
+		perFS = append(perFS, sorted[consumed:batchConsumedUntil])
+		consumed = batchConsumedUntil
+	}
+	return perFS
+}
+
+// batch must be on same Filesystem, panics otherwise
+func tryBatch(ctx context.Context, batch []*DestroySnapOp, d destroyer) error {
+	batchFS := batch[0].Filesystem
+	batchNames := make([]string, len(batch))
+	for i := range batchNames {
+		batchNames[i] = batch[i].Name
+		if batchFS != batch[i].Filesystem {
+			panic("inconsistent batch")
+		}
+	}
+	batchArg := fmt.Sprintf("%s@%s", batchFS, strings.Join(batchNames, ","))
+	return d.Destroy([]string{batchArg})
+}
+
+// fsbatch must be on same filesystem
+func doDestroyBatchedRec(ctx context.Context, fsbatch []*DestroySnapOp, d destroyer) {
+	if len(fsbatch) <= 1 {
+		doDestroySeq(ctx, fsbatch, d)
+		return
+	}
+
+	err := tryBatch(ctx, fsbatch, d)
+	if err == nil {
+		setDestroySnapOpErr(fsbatch, nil)
+		return
+	}
+
+	if pe, ok := err.(*os.PathError); ok && pe.Err == syscall.E2BIG {
+		// see TestExcessiveArgumentsResultInE2BIG
+		// try halving batch size, assuming snapshots names are roughly the same length
+		debug("batch destroy: E2BIG encountered: %s", err)
+		doDestroyBatchedRec(ctx, fsbatch[0:len(fsbatch)/2], d)
+		doDestroyBatchedRec(ctx, fsbatch[len(fsbatch)/2:], d)
+		return
+	}
+
+	singleRun := fsbatch // the destroys that will be tried sequentially after "smart" error handling below
+
+	if err, ok := err.(*DestroySnapshotsError); ok {
+		// eliminate undestroyable datasets from batch and try it once again
+		strippedBatch, remaining := make([]*DestroySnapOp, 0, len(fsbatch)), make([]*DestroySnapOp, 0, len(fsbatch))
+
+		for _, b := range fsbatch {
+			isUndestroyable := false
+			for _, undestroyable := range err.Undestroyable {
+				if undestroyable == b.Name {
+					isUndestroyable = true
+					break
+				}
+			}
+			if isUndestroyable {
+				remaining = append(remaining, b)
+			} else {
+				strippedBatch = append(strippedBatch, b)
+			}
+		}
+
+		err := tryBatch(ctx, strippedBatch, d)
+		if err != nil {
+			// run entire batch sequentially if the stripped one fails
+			// (it shouldn't because we stripped erronous datasets)
+			singleRun = fsbatch // shadow
+		} else {
+			setDestroySnapOpErr(strippedBatch, nil) // these ones worked
+			singleRun = remaining                   // shadow
+		}
+		// fallthrough
+	}
+
+	doDestroySeq(ctx, singleRun, d)
+
+}
+
+type destroyerImpl struct{}
+
+func (d destroyerImpl) Destroy(args []string) error {
+	if len(args) != 1 {
+		// we have no use case for this at the moment, so let's crash (safer than destroying something unexpectedly)
+		panic(fmt.Sprintf("unexpected number of arguments: %v", args))
+	}
+	// we know that we are only using this for snapshots, so also sanity check for an @ in args[0]
+	if !strings.ContainsAny(args[0], "@") {
+		panic(fmt.Sprintf("sanity check: expecting '@' in call to Destroy, got %q", args[0]))
+	}
+	return ZFSDestroy(args[0])
+}
+
+var batchDestroyFeatureCheck struct {
+	once   sync.Once
+	enable bool
+	err    error
+}
+
+func (d destroyerImpl) DestroySnapshotsCommaSyntaxSupported() (bool, error) {
+	batchDestroyFeatureCheck.once.Do(func() {
+		// "feature discovery"
+		cmd := exec.Command("zfs", "destroy")
+		output, err := cmd.CombinedOutput()
+		if _, ok := err.(*exec.ExitError); !ok {
+			debug("destroy feature check failed: %T %s", err, err)
+			batchDestroyFeatureCheck.err = err
+		}
+		def := strings.Contains(string(output), "<filesystem|volume>@<snap>[%<snap>][,...]")
+		batchDestroyFeatureCheck.enable = envconst.Bool("ZREPL_EXPERIMENTAL_ZFS_COMMA_SYNTAX_SUPPORTED", def)
+		debug("destroy feature check complete %#v", &batchDestroyFeatureCheck)
+	})
+	return batchDestroyFeatureCheck.enable, batchDestroyFeatureCheck.err
+}