Merge branch 'problame/replication_refactor' into InsanePrawn-master

.circleci/config.yml

@@ -0,0 +1,79 @@
+version: 2.0
+workflows:
+  version: 2
+  build:
+    jobs:
+      - build-1.11
+      - build-1.12
+      - build-latest
+jobs:
+  # build-latest serves as the template
+  # we use YAML anchors & aliases to exchange the docker image (and hence Go version used for the build)
+  build-latest: &build-latest
+    description: Builds zrepl
+    parameters:
+      image:
+        description: "the docker image that the job should use"
+        type: string
+    docker:
+      - image: circleci/golang:latest
+    environment:
+      # required by lazy.sh
+      TERM: xterm
+    working_directory: /go/src/github.com/zrepl/zrepl
+    steps:
+      - run:
+          name: Setup environment variables
+          command: |
+            # used by pip (for docs)
+            echo 'export PATH="$HOME/.local/bin:$PATH"' >> $BASH_ENV
+
+      - restore_cache:
+          keys:
+            - source
+            - vendor
+            - protobuf
+
+      - checkout
+
+      - save_cache:
+          key: source
+          paths:
+            - ".git"
+      
+      # install deps
+      - run: wget https://github.com/protocolbuffers/protobuf/releases/download/v3.6.1/protoc-3.6.1-linux-x86_64.zip
+      - run: echo "6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip" | sha256sum -c
+      - run: sudo unzip -d /usr protoc-3.6.1-linux-x86_64.zip
+      - save_cache:
+          key: protobuf
+          paths:
+            - "/usr/include/google/protobuf"
+            
+      - run: sudo apt install python3 python3-pip libgirepository1.0-dev
+      - run: ./lazy.sh devsetup
+
+      - run: make vendordeps
+      - save_cache:
+          key: vendor
+          paths:
+            - "./vendor"
+            
+      - run: make
+      - run: make vet
+      - run: make test
+      - run: make release
+
+      - store_artifacts:
+          path: ./artifacts/release
+          when: always
+
+
+  build-1.11:
+    <<: *build-latest
+    docker:
+      - image: circleci/golang:1.11
+  build-1.12:
+    <<: *build-latest
+    docker:
+    - image: circleci/golang:1.12

.travis.yml

@@ -2,6 +2,7 @@ dist: xenial
 services:
 - docker
 
+env: # for allow_failures: https://docs.travis-ci.com/user/customizing-the-build/
 matrix:
   include:
 
@@ -15,45 +16,35 @@ matrix:
         --user "$(id -u):$(id -g)" \
         zrepl_build make vendordeps release
 
-  # all go entries vary only by go version
+  - &zrepl_build_template
-  - language: go
+    language: go
+    go_import_path: github.com/zrepl/zrepl
+    before_install:
+        - wget https://github.com/protocolbuffers/protobuf/releases/download/v3.6.1/protoc-3.6.1-linux-x86_64.zip
+        - echo "6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip" | sha256sum -c
+        - sudo unzip -d /usr protoc-3.6.1-linux-x86_64.zip
+        - ./lazy.sh godep
+        - make vendordeps
+    script:
+      - make
+      - make vet
+      - make test
+      - make artifacts/zrepl-freebsd-amd64
+      - make artifacts/zrepl-linux-amd64
+      - make artifacts/zrepl-darwin-amd64
     go:
     - "1.11"
-    go_import_path: github.com/zrepl/zrepl
-    before_install:
-        - wget https://github.com/protocolbuffers/protobuf/releases/download/v3.6.1/protoc-3.6.1-linux-x86_64.zip
-        - echo "6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip" | sha256sum -c
-        - sudo unzip -d /usr protoc-3.6.1-linux-x86_64.zip
-        - ./lazy.sh godep
-        - make vendordeps
-    script:
-      - make
-      - make vet
-      - make test
-      - make artifacts/zrepl-freebsd-amd64
-      - make artifacts/zrepl-linux-amd64
-      - make artifacts/zrepl-darwin-amd64
 
-  - language: go
+  - <<: *zrepl_build_template
+    go:
+    - "1.12"
+
+  - <<: *zrepl_build_template
     go:
     - "master"
-    go_import_path: github.com/zrepl/zrepl
-    before_install:
-        - wget https://github.com/protocolbuffers/protobuf/releases/download/v3.6.1/protoc-3.6.1-linux-x86_64.zip
-        - echo "6003de742ea3fcf703cfec1cd4a3380fd143081a2eb0e559065563496af27807  protoc-3.6.1-linux-x86_64.zip" | sha256sum -c
-        - sudo unzip -d /usr protoc-3.6.1-linux-x86_64.zip
-        - ./lazy.sh godep
-        - make vendordeps
-    script:
-      - make
-      - make vet
-      - make test
-      - make artifacts/zrepl-freebsd-amd64
-      - make artifacts/zrepl-linux-amd64
-      - make artifacts/zrepl-darwin-amd64
 
-  # all python entries vary only by python version
+  - &zrepl_docs_template 
-  - language: python
+    language: python
     python:
     - "3.4"
     install:
@@ -61,29 +52,18 @@ matrix:
     - pip install -r docs/requirements.txt
     script:
     - make docs
-  - language: python
+  - <<: *zrepl_docs_template
     python:
     - "3.5"
-    install:
+  - <<: *zrepl_docs_template
-    - sudo apt-get install libgirepository1.0-dev
-    - pip install -r docs/requirements.txt
-    script:
-    - make docs
-  - language: python
     python:
     - "3.6"
-    install:
+  - <<: *zrepl_docs_template
-    - sudo apt-get install libgirepository1.0-dev
-    - pip install -r docs/requirements.txt
-    script:
-    - make docs
-  - language: python
     python:
     - "3.7"
-    install:
-    - sudo apt-get install libgirepository1.0-dev
-    - pip install -r docs/requirements.txt
-    script:
-    - make docs
   
 
+  allow_failures:
+  - <<: *zrepl_build_template
+    go:
+    - "master"

Gopkg.lock

@@ -89,6 +89,14 @@
   revision = "aa810b61a9c79d51363740d207bb46cf8e620ed5"
   version = "v1.2.0"
 
+[[projects]]
+  digest = "1:ad92aa49f34cbc3546063c7eb2cabb55ee2278b72842eda80e2a20a8a06a8d73"
+  name = "github.com/google/uuid"
+  packages = ["."]
+  pruneopts = ""
+  revision = "0cd6bf5da1e1c83f8b45653022c74f71af0538a4"
+  version = "v1.1.1"
+
 [[projects]]
   branch = "master"
   digest = "1:cb09475f771b9167fb9333629f5d6a7161572602ea040f1094602b0dc8709878"
@@ -161,6 +169,14 @@
   revision = "3247c84500bff8d9fb6d579d800f20b3e091582c"
   version = "v1.0.0"
 
+[[projects]]
+  digest = "1:4ff67dde814694496d7aa31be44b900f9717a10c8bc9136b13f49c8ef97f439a"
+  name = "github.com/montanaflynn/stats"
+  packages = ["."]
+  pruneopts = ""
+  revision = "63fbb2597b7a13043b453a4b819945badb8f8926"
+  version = "v0.5.0"
+
 [[projects]]
   branch = "master"
   digest = "1:f60ff065b58bd53e641112b38bbda9d2684deb828393c7ffb89c69a1ee301d17"
@@ -245,6 +261,14 @@
   pruneopts = ""
   revision = "8b1c2da0d56deffdbb9e48d4414b4e674bd8083e"
 
+[[projects]]
+  digest = "1:3962f553b77bf6c03fc07cd687a22dd3b00fe11aa14d31194f5505f5bb65cdc8"
+  name = "github.com/sergi/go-diff"
+  packages = ["diffmatchpatch"]
+  pruneopts = ""
+  revision = "1744e2970ca51c86172c8190fadad617561ed6e7"
+  version = "v1.0.0"
+
 [[projects]]
   branch = "master"
   digest = "1:146327ce93be37e68bd3ff8541090d96da8cb3adc9e35d57570e9170a29f6bf6"
@@ -280,6 +304,25 @@
   revision = "93babf24513d0e8277635da8169fcc5a46ae3f6a"
   version = "v1.11.0"
 
+[[projects]]
+  digest = "1:529ed3f98838f69e13761788d0cc71b44e130058fab13bae2ce09f7a176bced4"
+  name = "github.com/yudai/gojsondiff"
+  packages = [
+    ".",
+    "formatter",
+  ]
+  pruneopts = ""
+  revision = "7b1b7adf999dab73a6eb02669c3d82dbb27a3dd6"
+  version = "1.0.0"
+
+[[projects]]
+  branch = "master"
+  digest = "1:9857bb2293f372b2181004d8b62179bbdb4ab0982ec6f762abe6cf2bfedaff85"
+  name = "github.com/yudai/golcs"
+  packages = ["."]
+  pruneopts = ""
+  revision = "ecda9a501e8220fae3b4b600c3db4b0ba22cfc68"
+
 [[projects]]
   branch = "v2"
   digest = "1:6b8a6afafde7ed31cd0c577ba40d88ce39e8f1c5eb76d7836be7d5b74f1c534a"
@@ -403,9 +446,11 @@
     "github.com/go-logfmt/logfmt",
     "github.com/golang/protobuf/proto",
     "github.com/golang/protobuf/protoc-gen-go",
+    "github.com/google/uuid",
     "github.com/jinzhu/copier",
     "github.com/kr/pretty",
     "github.com/mattn/go-isatty",
+    "github.com/montanaflynn/stats",
     "github.com/pkg/errors",
     "github.com/pkg/profile",
     "github.com/problame/go-netssh",
@@ -415,14 +460,18 @@
     "github.com/spf13/pflag",
     "github.com/stretchr/testify/assert",
     "github.com/stretchr/testify/require",
+    "github.com/yudai/gojsondiff",
+    "github.com/yudai/gojsondiff/formatter",
     "github.com/zrepl/yaml-config",
     "golang.org/x/net/context",
     "golang.org/x/sys/unix",
     "golang.org/x/tools/cmd/stringer",
     "google.golang.org/grpc",
+    "google.golang.org/grpc/codes",
     "google.golang.org/grpc/credentials",
     "google.golang.org/grpc/keepalive",
     "google.golang.org/grpc/peer",
+    "google.golang.org/grpc/status",
   ]
   solver-name = "gps-cdcl"
   solver-version = 1

Gopkg.toml

@@ -59,3 +59,7 @@ required = [
 [[constraint]]
   name = "google.golang.org/grpc"
   version = "1"
+  
+[[constraint]]
+  version = "1.1.0"
+  name = "github.com/google/uuid"

Makefile

@@ -27,7 +27,7 @@ vendordeps:
 	dep ensure -v -vendor-only
 
 generate: #not part of the build, must do that manually
-	protoc -I=replication/pdu --go_out=plugins=grpc:replication/pdu replication/pdu/pdu.proto
+	protoc -I=replication/logic/pdu --go_out=plugins=grpc:replication/logic/pdu replication/logic/pdu/pdu.proto
 	go generate -x ./...
 
 build:

client/status.go

@@ -10,8 +10,7 @@ import (
 	"github.com/zrepl/zrepl/daemon"
 	"github.com/zrepl/zrepl/daemon/job"
 	"github.com/zrepl/zrepl/daemon/pruner"
-	"github.com/zrepl/zrepl/replication"
+	"github.com/zrepl/zrepl/replication/report"
-	"github.com/zrepl/zrepl/replication/fsrep"
 	"io"
 	"math"
 	"net/http"
@@ -122,7 +121,7 @@ func wrap(s string, width int) string {
 		if idx := strings.IndexAny(s, "\n\r"); idx != -1 && idx < rem {
 			rem = idx+1
 		}
-		untilNewline := strings.TrimSpace(s[:rem])
+		untilNewline := strings.TrimRight(s[:rem], "\n\r")
 		s = s[rem:]
 		if len(untilNewline) == 0 {
 			continue
@@ -130,7 +129,7 @@ func wrap(s string, width int) string {
 		b.WriteString(untilNewline)
 		b.WriteString("\n")
 	}
-	return strings.TrimSpace(b.String())
+	return strings.TrimRight(b.String(), "\n\r")
 }
 
 func (t *tui) printfDrawIndentedAndWrappedIfMultiline(format string, a ...interface{}) {
@@ -353,74 +352,91 @@ func (t *tui) draw() {
 	termbox.Flush()
 }
 
-func (t *tui) renderReplicationReport(rep *replication.Report, history *bytesProgressHistory) {
+func (t *tui) renderReplicationReport(rep *report.Report, history *bytesProgressHistory) {
 	if rep == nil {
 		t.printf("...\n")
 		return
 	}
 
-	all := make([]*fsrep.Report, 0, len(rep.Completed)+len(rep.Pending) + 1)
+	if rep.WaitReconnectError != nil {
-	all = append(all, rep.Completed...)
+		t.printfDrawIndentedAndWrappedIfMultiline("Connectivity: %s", rep.WaitReconnectError)
-	all = append(all, rep.Pending...)
+		t.newline()
-	if rep.Active != nil {
-		all = append(all, rep.Active)
 	}
-	sort.Slice(all, func(i, j int) bool {
+	if !rep.WaitReconnectSince.IsZero() {
-		return all[i].Filesystem < all[j].Filesystem
+		delta := rep.WaitReconnectUntil.Sub(time.Now()).Round(time.Second)
+		if rep.WaitReconnectUntil.IsZero() || delta > 0 {
+			var until string
+			if rep.WaitReconnectUntil.IsZero() {
+				until = "waiting indefinitely"
+			} else {
+				until = fmt.Sprintf("hard fail in %s @ %s", delta, rep.WaitReconnectUntil)
+			}
+			t.printfDrawIndentedAndWrappedIfMultiline("Connectivity: reconnecting with exponential backoff (since %s) (%s)",
+				rep.WaitReconnectSince, until)
+		} else {
+			t.printfDrawIndentedAndWrappedIfMultiline("Connectivity: reconnects reached hard-fail timeout @ %s", rep.WaitReconnectUntil)
+		}
+		t.newline()
+	}
+
+	// TODO visualize more than the latest attempt by folding all attempts into one
+	if len(rep.Attempts) == 0 {
+		t.printf("no attempts made yet")
+		return
+	} else {
+		t.printf("Attempt #%d", len(rep.Attempts))
+		if len(rep.Attempts) > 1 {
+			t.printf(". Previous attempts failed with the follwing statuses:")
+			t.newline()
+			t.addIndent(1)
+			for i, a := range rep.Attempts[:len(rep.Attempts)-1] {
+				t.printfDrawIndentedAndWrappedIfMultiline("#%d: %s (failed at %s) (ran %s)", i + 1, a.State, a.FinishAt, a.FinishAt.Sub(a.StartAt))
+				t.newline()
+			}
+			t.addIndent(-1)
+		} else {
+			t.newline()
+		}
+	}
+
+	latest := rep.Attempts[len(rep.Attempts)-1]
+	sort.Slice(latest.Filesystems, func(i, j int) bool {
+		return latest.Filesystems[i].Info.Name < latest.Filesystems[j].Info.Name
 	})
 
-	state, err := replication.StateString(rep.Status)
+	t.printf("Status: %s", latest.State)
-	if err != nil {
-		t.printf("Status: %q (parse error: %q)\n", rep.Status, err)
-		return
-	}
-
-	t.printf("Status: %s", state)
 	t.newline()
-	if rep.Problem != "" {
+	if latest.State == report.AttemptPlanningError {
 		t.printf("Problem: ")
-		t.printfDrawIndentedAndWrappedIfMultiline("%s", rep.Problem)
+		t.printfDrawIndentedAndWrappedIfMultiline("%s", latest.PlanError)
+		t.newline()
+	} else if latest.State == report.AttemptFanOutError {
+		t.printf("Problem: one or more of the filesystems encountered errors")
 		t.newline()
-	}
-	if rep.SleepUntil.After(time.Now()) && !state.IsTerminal() {
-		t.printf("Sleeping until %s (%s left)\n", rep.SleepUntil, rep.SleepUntil.Sub(time.Now()))
 	}
 
-	if state != replication.Planning && state != replication.PlanningError {
+	if latest.State != report.AttemptPlanning && latest.State != report.AttemptPlanningError {
+		// Draw global progress bar
 		// Progress: [---------------]
-		sumUpFSRep := func(rep *fsrep.Report) (transferred, total int64) {
+		expected, replicated := latest.BytesSum()
-			for _, s := range rep.Pending {
+		rate, changeCount := history.Update(replicated)
-				transferred += s.Bytes
-				total  += s.ExpectedBytes
-			}
-			for _, s := range rep.Completed {
-				transferred += s.Bytes
-				total += s.ExpectedBytes
-			}
-			return
-		}
-		var transferred, total int64
-		for _, fs := range all {
-			fstx, fstotal := sumUpFSRep(fs)
-			transferred += fstx
-			total += fstotal
-		}
-		rate, changeCount := history.Update(transferred)
 		t.write("Progress: ")
-		t.drawBar(50, transferred, total, changeCount)
+		t.drawBar(50, replicated, expected, changeCount)
-		t.write(fmt.Sprintf(" %s / %s @ %s/s", ByteCountBinary(transferred), ByteCountBinary(total), ByteCountBinary(rate)))
+		t.write(fmt.Sprintf(" %s / %s @ %s/s", ByteCountBinary(replicated), ByteCountBinary(expected), ByteCountBinary(rate)))
 		t.newline()
-	}
 
 		var maxFSLen int
-	for _, fs := range all {
+		for _, fs := range latest.Filesystems {
-		if len(fs.Filesystem) > maxFSLen {
+			if len(fs.Info.Name) > maxFSLen {
-			maxFSLen = len(fs.Filesystem)
+				maxFSLen = len(fs.Info.Name)
 			}
 		}
-	for _, fs := range all {
+		for _, fs := range latest.Filesystems {
-		t.printFilesystemStatus(fs, fs == rep.Active, maxFSLen)
+			t.printFilesystemStatus(fs, false, maxFSLen) // FIXME bring 'active' flag back
 		}
+
+	}
+
 }
 
 func (t *tui) renderPrunerReport(r *pruner.Report) {
@@ -441,9 +457,6 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 	if r.Error != "" {
 		t.printf("Error: %s\n", r.Error)
 	}
-	if r.SleepUntil.After(time.Now()) {
-		t.printf("Sleeping until %s (%s left)\n", r.SleepUntil, r.SleepUntil.Sub(time.Now()))
-	}
 
 	type commonFS struct {
 		*pruner.FSReport
@@ -459,8 +472,7 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 
 	switch state {
 	case pruner.Plan: fallthrough
-	case pruner.PlanWait: fallthrough
+	case pruner.PlanErr:
-	case pruner.ErrPerm:
 		return
 	}
 
@@ -500,8 +512,18 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 	for _, fs := range all {
 		t.write(rightPad(fs.Filesystem, maxFSname, " "))
 		t.write(" ")
+		if !fs.SkipReason.NotSkipped() {
+			t.printf("skipped: %s\n", fs.SkipReason)
+			continue
+		}
 		if fs.LastError != "" {
-			t.printf("ERROR (%d): %s\n", fs.ErrorCount, fs.LastError) // whitespace is padding
+			if strings.ContainsAny(fs.LastError, "\r\n") {
+				t.printf("ERROR:")
+				t.printfDrawIndentedAndWrappedIfMultiline("%s\n", fs.LastError) 
+			} else {
+				t.printfDrawIndentedAndWrappedIfMultiline("ERROR: %s\n", fs.LastError) 
+			}
+			t.newline()
 			continue
 		}
 
@@ -524,25 +546,6 @@ func (t *tui) renderPrunerReport(r *pruner.Report) {
 
 }
 
-const snapshotIndent = 1
-func calculateMaxFSLength(all []*fsrep.Report) (maxFS, maxStatus int) {
-	for _, e := range all {
-		if len(e.Filesystem) > maxFS {
-			maxFS = len(e.Filesystem)
-		}
-		all2 := make([]*fsrep.StepReport, 0, len(e.Pending) + len(e.Completed))
-		all2 = append(all2, e.Pending...)
-		all2 = append(all2, e.Completed...)
-		for _, e2 := range all2 {
-			elen := len(e2.Problem) + len(e2.From) + len(e2.To) + 60 // random spacing, units, labels, etc
-			if elen > maxStatus {
-				maxStatus = elen
-			}
-		}
-	}
-	return
-}
-
 func times(str string, n int) (out string) {
 	for i := 0; i < n; i++ {
 		out += str
@@ -586,35 +589,13 @@ func (t *tui) drawBar(length int, bytes, totalBytes int64, changeCount int) {
 	t.write("]")
 }
 
-func StringStepState(s fsrep.StepState) string {
+func (t *tui) printFilesystemStatus(rep *report.FilesystemReport, active bool, maxFS int) {
-	switch s {
-	case fsrep.StepReplicationReady: return "Ready"
-	case fsrep.StepMarkReplicatedReady: return "MarkReady"
-	case fsrep.StepCompleted: return "Completed"
-	default:
-		return fmt.Sprintf("UNKNOWN %d", s)
-	}
-}
-
-func (t *tui) printFilesystemStatus(rep *fsrep.Report, active bool, maxFS int) {
-
-	bytes := int64(0)
-	totalBytes := int64(0)
-	for _, s := range rep.Pending {
-		bytes += s.Bytes
-		totalBytes += s.ExpectedBytes
-	}
-	for _, s := range rep.Completed {
-		bytes += s.Bytes
-		totalBytes += s.ExpectedBytes
-	}
-
 
+	expected, replicated := rep.BytesSum()
 	status := fmt.Sprintf("%s (step %d/%d, %s/%s)",
-		rep.Status,
+		strings.ToUpper(string(rep.State)),
-		len(rep.Completed), len(rep.Pending) + len(rep.Completed),
+		rep.CurrentStep, len(rep.Steps),
-		ByteCountBinary(bytes), ByteCountBinary(totalBytes),
+		ByteCountBinary(replicated), ByteCountBinary(expected),
-
 	)
 
 	activeIndicator := " "
@@ -623,18 +604,23 @@ func (t *tui) printFilesystemStatus(rep *fsrep.Report, active bool, maxFS int) {
 	}
 	t.printf("%s %s %s ",
 		activeIndicator,
-		rightPad(rep.Filesystem, maxFS, " "),
+		rightPad(rep.Info.Name, maxFS, " "),
 		status)
 
 	next := ""
-	if rep.Problem != "" {
+	if err := rep.Error(); err != nil {
-		next = rep.Problem
+		next = err.Err
-	} else if len(rep.Pending) > 0 {
+	} else if rep.State != report.FilesystemDone {
-		if rep.Pending[0].From != "" {
+		if nextStep := rep.NextStep(); nextStep != nil {
-			next = fmt.Sprintf("next: %s => %s", rep.Pending[0].From, rep.Pending[0].To)
+			if nextStep.IsIncremental() {
+				next = fmt.Sprintf("next: %s => %s", nextStep.Info.From, nextStep.Info.To)
 			} else {
-			next = fmt.Sprintf("next: %s (full)", rep.Pending[0].To)
+				next = fmt.Sprintf("next: %s (full)", nextStep.Info.To)
 			}
+		} else {
+			next = "" // individual FSes may still be in planning state
+		}
+
 	}
 	t.printfDrawIndentedAndWrappedIfMultiline("%s", next)
 

config/config.go

@@ -78,7 +78,38 @@ type PushJob struct {
 type PullJob struct {
 	ActiveJob `yaml:",inline"`
 	RootFS    string        `yaml:"root_fs"`
-	Interval  time.Duration `yaml:"interval,positive"`
+	Interval  PositiveDurationOrManual `yaml:"interval"`
+}
+
+type PositiveDurationOrManual struct {
+	Interval time.Duration
+	Manual   bool
+}
+
+var _ yaml.Unmarshaler = (*PositiveDurationOrManual)(nil)
+
+func (i *PositiveDurationOrManual) UnmarshalYAML(u func(interface{}, bool) error) (err error) {
+	var s string
+	if err := u(&s, true); err != nil {
+		return err
+	}
+	switch s {
+	case "manual":
+		i.Manual = true
+		i.Interval = 0
+	case "":
+		return fmt.Errorf("value must not be empty")
+	default:
+		i.Manual = false
+		i.Interval, err = time.ParseDuration(s)
+		if err != nil {
+			return err
+		}
+		if i.Interval <= 0 {
+			return fmt.Errorf("value must be a positive duration, got %q", s)
+		}
+	}
+	return nil
 }
 
 type SinkJob struct {

config/config_positiveintervalormanual_test.go

@@ -0,0 +1,41 @@
+package config
+
+import (
+	"fmt"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/zrepl/yaml-config"
+)
+
+func TestPositiveDurationOrManual(t *testing.T) {
+	cases := []struct {
+		Comment, Input string
+		Result         *PositiveDurationOrManual
+	}{
+		{"empty is error", "", nil},
+		{"negative is error", "-1s", nil},
+		{"zero seconds is error", "0s", nil},
+		{"zero is error", "0", nil},
+		{"non-manual is error", "something", nil},
+		{"positive seconds works", "1s", &PositiveDurationOrManual{Manual: false, Interval: 1 * time.Second}},
+		{"manual works", "manual", &PositiveDurationOrManual{Manual: true, Interval: 0}},
+	}
+	for _, tc := range cases {
+		t.Run(tc.Comment, func(t *testing.T) {
+			var out struct {
+				FieldName PositiveDurationOrManual `yaml:"fieldname"`
+			}
+			input := fmt.Sprintf("\nfieldname: %s\n", tc.Input)
+			err := yaml.UnmarshalStrict([]byte(input), &out)
+			if tc.Result == nil {
+				assert.Error(t, err)
+				t.Logf("%#v", out)
+			} else {
+				assert.Equal(t, *tc.Result, out.FieldName)
+			}
+		})
+	}
+
+}

daemon/job/active.go

@@ -17,10 +17,12 @@ import (
 	"github.com/zrepl/zrepl/daemon/snapper"
 	"github.com/zrepl/zrepl/endpoint"
 	"github.com/zrepl/zrepl/replication"
+	"github.com/zrepl/zrepl/replication/driver"
+	"github.com/zrepl/zrepl/replication/logic"
+	"github.com/zrepl/zrepl/replication/report"
 	"github.com/zrepl/zrepl/rpc"
 	"github.com/zrepl/zrepl/transport"
 	"github.com/zrepl/zrepl/transport/fromconfig"
-	"github.com/zrepl/zrepl/util/envconst"
 	"github.com/zrepl/zrepl/zfs"
 )
 
@@ -53,7 +55,7 @@ type activeSideTasks struct {
 	state ActiveSideState
 
 	// valid for state ActiveSideReplicating, ActiveSidePruneSender, ActiveSidePruneReceiver, ActiveSideDone
-	replication       *replication.Replication
+	replicationReport driver.ReportFunc
 	replicationCancel context.CancelFunc
 
 	// valid for state ActiveSidePruneSender, ActiveSidePruneReceiver, ActiveSideDone
@@ -79,7 +81,7 @@ func (a *ActiveSide) updateTasks(u func(*activeSideTasks)) activeSideTasks {
 type activeMode interface {
 	ConnectEndpoints(rpcLoggers rpc.Loggers, connecter transport.Connecter)
 	DisconnectEndpoints()
-	SenderReceiver() (replication.Sender, replication.Receiver)
+	SenderReceiver() (logic.Sender, logic.Receiver)
 	Type() Type
 	RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{})
 	ResetConnectBackoff()
@@ -111,7 +113,7 @@ func (m *modePush) DisconnectEndpoints() {
 	m.receiver = nil
 }
 
-func (m *modePush) SenderReceiver() (replication.Sender, replication.Receiver) {
+func (m *modePush) SenderReceiver() (logic.Sender, logic.Receiver) {
 	m.setupMtx.Lock()
 	defer m.setupMtx.Unlock()
 	return m.sender, m.receiver
@@ -151,7 +153,7 @@ type modePull struct {
 	receiver *endpoint.Receiver
 	sender   *rpc.Client
 	rootFS   *zfs.DatasetPath
-	interval time.Duration
+	interval config.PositiveDurationOrManual
 }
 
 func (m *modePull) ConnectEndpoints(loggers rpc.Loggers, connecter transport.Connecter) {
@@ -172,7 +174,7 @@ func (m *modePull) DisconnectEndpoints() {
 	m.receiver = nil
 }
 
-func (m *modePull) SenderReceiver() (replication.Sender, replication.Receiver) {
+func (m *modePull) SenderReceiver() (logic.Sender, logic.Receiver) {
 	m.setupMtx.Lock()
 	defer m.setupMtx.Unlock()
 	return m.sender, m.receiver
@@ -181,7 +183,12 @@ func (m *modePull) SenderReceiver() (replication.Sender, replication.Receiver) {
 func (*modePull) Type() Type { return TypePull }
 
 func (m *modePull) RunPeriodic(ctx context.Context, wakeUpCommon chan<- struct{}) {
-	t := time.NewTicker(m.interval)
+	if m.interval.Manual {
+		GetLogger(ctx).Info("manual pull configured, periodic pull disabled")
+		// "waiting for wakeups" is printed in common ActiveSide.do
+		return
+	}
+	t := time.NewTicker(m.interval.Interval)
 	defer t.Stop()
 	for {
 		select {
@@ -210,9 +217,6 @@ func (m *modePull) ResetConnectBackoff() {
 
 func modePullFromConfig(g *config.Global, in *config.PullJob) (m *modePull, err error) {
 	m = &modePull{}
-	if in.Interval <= 0 {
-		return nil, errors.New("interval must be positive")
-	}
 	m.interval = in.Interval
 
 	m.rootFS, err = zfs.NewDatasetPath(in.RootFS)
@@ -274,7 +278,7 @@ func (j *ActiveSide) RegisterMetrics(registerer prometheus.Registerer) {
 func (j *ActiveSide) Name() string { return j.name }
 
 type ActiveSideStatus struct {
-	Replication                    *replication.Report
+	Replication                    *report.Report
 	PruningSender, PruningReceiver *pruner.Report
 }
 
@@ -283,8 +287,8 @@ func (j *ActiveSide) Status() *Status {
 
 	s := &ActiveSideStatus{}
 	t := j.mode.Type()
-	if tasks.replication != nil {
+	if tasks.replicationReport != nil {
-		s.Replication = tasks.replication.Report()
+		s.Replication = tasks.replicationReport()
 	}
 	if tasks.prunerSender != nil {
 		s.PruningSender = tasks.prunerSender.Report()
@@ -345,78 +349,6 @@ func (j *ActiveSide) do(ctx context.Context) {
 		}
 	}()
 
-	// The code after this watchdog goroutine is sequential and transitions the state from
-	//   ActiveSideReplicating -> ActiveSidePruneSender -> ActiveSidePruneReceiver -> ActiveSideDone
-	// If any of those sequential tasks 'gets stuck' (livelock, no progress), the watchdog will eventually
-	// cancel its context.
-	// If the task is written to support context cancellation, it will return immediately (in permanent error state),
-	// and the sequential code above transitions to the next state.
-	go func() {
-
-		wdto := envconst.Duration("ZREPL_JOB_WATCHDOG_TIMEOUT", 10*time.Minute)
-		jitter := envconst.Duration("ZREPL_JOB_WATCHDOG_JITTER", 1*time.Second)
-		// shadowing!
-		log := log.WithField("watchdog_timeout", wdto.String())
-
-		log.Debug("starting watchdog")
-		defer log.Debug("watchdog stopped")
-
-		t := time.NewTicker(wdto)
-		defer t.Stop()
-
-		for {
-			select {
-			case <-ctx.Done():
-				return
-			case <-t.C: // fall
-			}
-
-			j.updateTasks(func(tasks *activeSideTasks) {
-				// Since cancelling a task will cause the sequential code to transition to the next state immediately,
-				// we cannot check for its progress right then (no fallthrough).
-				// Instead, we return (not continue because we are in a closure) and give the new state another
-				// ZREPL_JOB_WATCHDOG_TIMEOUT interval to try make some progress.
-
-				log.WithField("state", tasks.state).Debug("watchdog firing")
-
-				const WATCHDOG_ENVCONST_NOTICE = " (adjust ZREPL_JOB_WATCHDOG_TIMEOUT env variable if inappropriate)"
-
-				switch tasks.state {
-				case ActiveSideReplicating:
-					log.WithField("replication_progress", tasks.replication.Progress.String()).
-						Debug("check replication progress")
-					if tasks.replication.Progress.CheckTimeout(wdto, jitter) {
-						log.Error("replication did not make progress, cancelling" + WATCHDOG_ENVCONST_NOTICE)
-						tasks.replicationCancel()
-						return
-					}
-				case ActiveSidePruneSender:
-					log.WithField("prune_sender_progress", tasks.replication.Progress.String()).
-						Debug("check pruner_sender progress")
-					if tasks.prunerSender.Progress.CheckTimeout(wdto, jitter) {
-						log.Error("pruner_sender did not make progress, cancelling" + WATCHDOG_ENVCONST_NOTICE)
-						tasks.prunerSenderCancel()
-						return
-					}
-				case ActiveSidePruneReceiver:
-					log.WithField("prune_receiver_progress", tasks.replication.Progress.String()).
-						Debug("check pruner_receiver progress")
-					if tasks.prunerReceiver.Progress.CheckTimeout(wdto, jitter) {
-						log.Error("pruner_receiver did not make progress, cancelling" + WATCHDOG_ENVCONST_NOTICE)
-						tasks.prunerReceiverCancel()
-						return
-					}
-				case ActiveSideDone:
-					// ignore, ctx will be Done() in a few milliseconds and the watchdog will exit
-				default:
-					log.WithField("state", tasks.state).
-						Error("watchdog implementation error: unknown active side state")
-				}
-			})
-
-		}
-	}()
-
 	sender, receiver := j.mode.SenderReceiver()
 
 	{
@@ -426,15 +358,18 @@ func (j *ActiveSide) do(ctx context.Context) {
 		default:
 		}
 		ctx, repCancel := context.WithCancel(ctx)
-		tasks := j.updateTasks(func(tasks *activeSideTasks) {
+		var repWait driver.WaitFunc
+		j.updateTasks(func(tasks *activeSideTasks) {
 			// reset it
 			*tasks = activeSideTasks{}
 			tasks.replicationCancel = repCancel
-			tasks.replication = replication.NewReplication(j.promRepStateSecs, j.promBytesReplicated)
+			tasks.replicationReport, repWait = replication.Do(
+				ctx, logic.NewPlanner(j.promRepStateSecs, j.promBytesReplicated, sender, receiver),
+			)
 			tasks.state = ActiveSideReplicating
 		})
 		log.Info("start replication")
-		tasks.replication.Drive(ctx, sender, receiver)
+		repWait(true) // wait blocking
 		repCancel()   // always cancel to free up context resources
 	}
 

daemon/logging/build_logging.go

@@ -15,7 +15,7 @@ import (
 	"github.com/zrepl/zrepl/daemon/snapper"
 	"github.com/zrepl/zrepl/endpoint"
 	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/replication"
+	"github.com/zrepl/zrepl/replication/driver"
 	"github.com/zrepl/zrepl/rpc"
 	"github.com/zrepl/zrepl/rpc/transportmux"
 	"github.com/zrepl/zrepl/tlsconf"
@@ -79,7 +79,7 @@ const (
 )
 
 func WithSubsystemLoggers(ctx context.Context, log logger.Logger) context.Context {
-	ctx = replication.WithLogger(ctx, log.WithField(SubsysField, SubsysReplication))
+	ctx = driver.WithLogger(ctx, log.WithField(SubsysField, SubsysReplication))
 	ctx = endpoint.WithLogger(ctx, log.WithField(SubsysField, SubsyEndpoint))
 	ctx = pruner.WithLogger(ctx, log.WithField(SubsysField, SubsysPruning))
 	ctx = snapper.WithLogger(ctx, log.WithField(SubsysField, SubsysSnapshot))

daemon/pruner/pruner.go

@@ -8,10 +8,9 @@ import (
 	"github.com/zrepl/zrepl/config"
 	"github.com/zrepl/zrepl/logger"
 	"github.com/zrepl/zrepl/pruning"
-	"github.com/zrepl/zrepl/replication/pdu"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
 	"github.com/zrepl/zrepl/util/envconst"
 	"github.com/zrepl/zrepl/util/watchdog"
-	"net"
 	"sort"
 	"strings"
 	"sync"
@@ -21,6 +20,7 @@ import (
 // Try to keep it compatible with gitub.com/zrepl/zrepl/endpoint.Endpoint
 type History interface {
 	ReplicationCursor(ctx context.Context, req *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error)
+	ListFilesystems(ctx context.Context, req *pdu.ListFilesystemReq) (*pdu.ListFilesystemRes, error)
 }
 
 // Try to keep it compatible with gitub.com/zrepl/zrepl/endpoint.Endpoint
@@ -66,8 +66,7 @@ type Pruner struct {
 
 	state State
 
-	// State ErrWait|ErrPerm
+	// State PlanErr
-	sleepUntil time.Time
 	err        error
 
 	// State Exec
@@ -206,62 +205,34 @@ type State int
 
 const (
 	Plan State = 1 << iota
-	PlanWait
+	PlanErr
 	Exec
-	ExecWait
+	ExecErr
-	ErrPerm
 	Done
 )
 
-func (s State) statefunc() state {
+type updater func(func(*Pruner))
-	var statemap = map[State]state{
-		Plan:     statePlan,
-		PlanWait: statePlanWait,
-		Exec:     stateExec,
-		ExecWait: stateExecWait,
-		ErrPerm:  nil,
-		Done:     nil,
-	}
-	return statemap[s]
-}
-
-func (s State) IsTerminal() bool {
-	return s.statefunc() == nil
-}
-
-type updater func(func(*Pruner)) State
-type state func(args *args, u updater) state
 
 func (p *Pruner) Prune() {
 	p.prune(p.args)
 }
 
 func (p *Pruner) prune(args args) {
-	s := p.state.statefunc()
+	u := func(f func(*Pruner)) {
-	for s != nil {
-		pre := p.state
-		s = s(&args, func(f func(*Pruner)) State {
 			p.mtx.Lock()
 			defer p.mtx.Unlock()
 			f(p)
-			return p.state
-		})
-		post := p.state
-		GetLogger(args.ctx).
-			WithField("transition", fmt.Sprintf("%s=>%s", pre, post)).
-			Debug("state transition")
-		if err := p.Error(); err != nil {
-			GetLogger(args.ctx).
-				WithError(p.err).
-				WithField("state", post.String()).
-				Error("entering error state after error")
-		}
 		}
+	// TODO support automatic retries
+	// It is advisable to merge this code with package replication/driver before
+	// That will likely require re-modelling struct fs like replication/driver.attempt,
+	// including figuring out how to resume a plan after being interrupted by network errors
+	// The non-retrying code in this package should move straight to replication/logic.
+	doOneAttempt(&args, u)
 	}
 
 type Report struct {
 	State              string
-	SleepUntil time.Time
 	Error              string
 	Pending, Completed []FSReport
 }
@@ -269,7 +240,7 @@ type Report struct {
 type FSReport struct {
 	Filesystem                string
 	SnapshotList, DestroyList []SnapshotReport
-	ErrorCount int
+	SkipReason                FSSkipReason
 	LastError                 string
 }
 
@@ -285,14 +256,9 @@ func (p *Pruner) Report() *Report {
 
 	r := Report{State: p.state.String()}
 
-	if p.state & (PlanWait|ExecWait) != 0 {
-		r.SleepUntil = p.sleepUntil
-	}
-	if p.state & (PlanWait|ExecWait|ErrPerm) != 0 {
 		if p.err != nil {
 			r.Error = p.err.Error()
 		}
-	}
 
 	if p.execQueue != nil {
 		r.Pending, r.Completed = p.execQueue.Report()
@@ -307,20 +273,16 @@ func (p *Pruner) State() State {
 	return p.state
 }
 
-func (p *Pruner) Error() error {
-	p.mtx.Lock()
-	defer p.mtx.Unlock()
-	if p.state & (PlanWait|ExecWait|ErrPerm) != 0 {
-		return p.err
-	}
-	return nil
-}
-
 type fs struct {
 	path  string
 
 	// permanent error during planning
 	planErr        error
+	planErrContext string
+
+	// if != "", the fs was skipped for planning and the field
+	// contains the reason
+	skipReason FSSkipReason
 
 	// snapshots presented by target
 	// (type snapshot)
@@ -333,8 +295,18 @@ type fs struct {
 
 	// only during Exec state, also used by execQueue
 	execErrLast error
-	execErrCount int
+}
 
+type FSSkipReason string
+
+const (
+	NotSkipped                   = ""
+	SkipPlaceholder              = "filesystem is placeholder"
+	SkipNoCorrespondenceOnSender = "filesystem has no correspondence on sender"
+)
+
+func (r FSSkipReason) NotSkipped() bool {
+	return r == NotSkipped
 }
 
 func (f *fs) Report() FSReport {
@@ -343,7 +315,11 @@ func (f *fs) Report() FSReport {
 
 	r := FSReport{}
 	r.Filesystem = f.path
-	r.ErrorCount = f.execErrCount
+	r.SkipReason = f.skipReason
+	if !r.SkipReason.NotSkipped() {
+		return r
+	}
+
 	if f.planErr != nil {
 		r.LastError = f.planErr.Error()
 	} else  if f.execErrLast != nil {
@@ -385,39 +361,7 @@ func (s snapshot) Replicated() bool { return s.replicated }
 
 func (s snapshot) Date() time.Time { return s.date }
 
-type Error interface {
+func doOneAttempt(a *args, u updater) {
-	error
-	Temporary() bool
-}
-
-var _ Error = net.Error(nil)
-
-func shouldRetry(e error) bool {
-	if neterr, ok := e.(net.Error); ok {
-		return neterr.Temporary()
-	}
-	return false
-}
-
-func onErr(u updater, e error) state {
-	return u(func(p *Pruner) {
-		p.err = e
-		if !shouldRetry(e) {
-			p.state = ErrPerm
-			return
-		}
-		switch p.state {
-		case Plan:
-			p.state = PlanWait
-		case Exec:
-			p.state = ExecWait
-		default:
-			panic(p.state)
-		}
-	}).statefunc()
-}
-
-func statePlan(a *args, u updater) state {
 
 	ctx, target, receiver := a.ctx, a.target, a.receiver
 	var ka *watchdog.KeepAlive
@@ -425,28 +369,62 @@ func statePlan(a *args, u updater) state {
 		ka = &pruner.Progress
 	})
 
+	sfssres, err := receiver.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
+	if err != nil {
+		u(func(p *Pruner) {
+			p.state = PlanErr
+			p.err = err
+		})
+		return
+	}
+	sfss := make(map[string]*pdu.Filesystem)
+	for _, sfs := range sfssres.GetFilesystems() {
+		sfss[sfs.GetPath()] = sfs
+	}
+
 	tfssres, err := target.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
 	if err != nil {
-		return onErr(u, err)
+		u(func(p *Pruner) {
+			p.state = PlanErr
+			p.err = err
+		})
+		return
 	}
 	tfss := tfssres.GetFilesystems()
 
 	pfss := make([]*fs, len(tfss))
+tfss_loop:
 	for i, tfs := range tfss {
 
 		l := GetLogger(ctx).WithField("fs", tfs.Path)
 		l.Debug("plan filesystem")
 
-
 		pfs := &fs{
 			path: tfs.Path,
 		}
 		pfss[i] = pfs
 
+		if tfs.GetIsPlaceholder() {
+			pfs.skipReason = SkipPlaceholder
+			l.WithField("skip_reason", pfs.skipReason).Debug("skipping filesystem")
+			continue
+		} else if sfs := sfss[tfs.GetPath()]; sfs == nil {
+			pfs.skipReason = SkipNoCorrespondenceOnSender
+			l.WithField("skip_reason", pfs.skipReason).WithField("sfs", sfs.GetPath()).Debug("skipping filesystem")
+			continue
+		}
+
+		pfsPlanErrAndLog := func(err error, message string) {
+			t := fmt.Sprintf("%T", err)
+			pfs.planErr = err
+			pfs.planErrContext = message
+			l.WithField("orig_err_type", t).WithError(err).Error(fmt.Sprintf("%s: plan error, skipping filesystem", message))
+		}
+
 		tfsvsres, err := target.ListFilesystemVersions(ctx, &pdu.ListFilesystemVersionsReq{Filesystem: tfs.Path})
 		if err != nil {
-			l.WithError(err).Error("cannot list filesystem versions")
+			pfsPlanErrAndLog(err, "cannot list filesystem versions")
-			return onErr(u, err)
+			continue tfss_loop
 		}
 		tfsvs := tfsvsres.GetVersions()
 		// no progress here since we could run in a live-lock (must have used target AND receiver before progress)
@@ -461,18 +439,16 @@ func statePlan(a *args, u updater) state {
 		}
 		rc, err := receiver.ReplicationCursor(ctx, rcReq)
 		if err != nil {
-			l.WithError(err).Error("cannot get replication cursor")
+			pfsPlanErrAndLog(err, "cannot get replication cursor bookmark")
-			return onErr(u, err)
+			continue tfss_loop
 		}
 		ka.MadeProgress()
 		if rc.GetNotexist() {
-			l.Error("replication cursor does not exist, skipping")
+			err := errors.New("replication cursor bookmark does not exist (one successful replication is required before pruning works)")
-			pfs.destroyList = []pruning.Snapshot{}
+			pfsPlanErrAndLog(err, "")
-			pfs.planErr = fmt.Errorf("replication cursor bookmark does not exist (one successful replication is required before pruning works)")
+			continue tfss_loop
-			continue
 		}
 
-
 		// scan from older to newer, all snapshots older than cursor are interpreted as replicated
 		sort.Slice(tfsvs, func(i, j int) bool {
 			return tfsvs[i].CreateTXG < tfsvs[j].CreateTXG
@@ -494,11 +470,9 @@ func statePlan(a *args, u updater) state {
 			}
 			creation, err := tfsv.CreationAsTime()
 			if err != nil {
-				err := fmt.Errorf("%s%s has invalid creation date: %s", tfs, tfsv.RelName(), err)
+				err := fmt.Errorf("%s: %s", tfsv.RelName(), err)
-				l.WithError(err).
+				pfsPlanErrAndLog(err, "fs version with invalid creation date")
-					WithField("tfsv", tfsv.RelName()).
+				continue tfss_loop
-					Error("error with fileesystem version")
-				return onErr(u, err)
 			}
 			// note that we cannot use CreateTXG because target and receiver could be on different pools
 			atCursor := tfsv.Guid == rc.GetGuid()
@@ -510,9 +484,8 @@ func statePlan(a *args, u updater) state {
 			})
 		}
 		if preCursor {
-			err := fmt.Errorf("replication cursor not found in prune target filesystem versions")
+			pfsPlanErrAndLog(fmt.Errorf("replication cursor not found in prune target filesystem versions"), "")
-			l.Error(err.Error())
+			continue tfss_loop
-			return onErr(u, err)
 		}
 
 		// Apply prune rules
@@ -520,34 +493,56 @@ func statePlan(a *args, u updater) state {
 		ka.MadeProgress()
 	}
 
-	return u(func(pruner *Pruner) {
+	u(func(pruner *Pruner) {
 		pruner.Progress.MadeProgress()
 		pruner.execQueue = newExecQueue(len(pfss))
 		for _, pfs := range pfss {
 			pruner.execQueue.Put(pfs, nil, false)
 		}
 		pruner.state = Exec
-	}).statefunc()
+	})
-}
-
-func stateExec(a *args, u updater) state {
 
+	for {
 	var pfs *fs
-	state := u(func(pruner *Pruner) {
+		u(func(pruner *Pruner) {
 		pfs = pruner.execQueue.Pop()
+		})
 		if pfs == nil {
-			nextState := Done
+			break
-			if pruner.execQueue.HasCompletedFSWithErrors() {
-				nextState = ErrPerm
 		}
-			pruner.state = nextState
+		doOneAttemptExec(a, u, pfs)
-			return
+	}
+
+	var rep *Report
+	{
+		// must not hold lock for report
+		var pruner *Pruner
+		u(func(p *Pruner) {
+			pruner = p
+		})
+		rep = pruner.Report()
+	}
+	u(func(p *Pruner) {
+		if len(rep.Pending) > 0 {
+			panic("queue should not have pending items at this point")
+		}
+		hadErr := false
+		for _, fsr := range rep.Completed {
+			hadErr = hadErr || fsr.SkipReason.NotSkipped() && fsr.LastError != ""
+			}
+		if hadErr {
+			p.state = ExecErr
+		} else {
+			p.state = Done
 		}
 	})
-	if state != Exec {
+	
-		return state.statefunc()
+
 	}
 
+// attempts to exec pfs, puts it back into the queue with the result
+func doOneAttemptExec(a *args, u updater, pfs *fs) {
+
 	destroyList := make([]*pdu.FilesystemVersion, len(pfs.destroyList))
 	for i := range destroyList {
 		destroyList[i] = pfs.destroyList[i].(snapshot).fsv
@@ -566,7 +561,7 @@ func stateExec(a *args, u updater) state {
 		u(func(pruner *Pruner) {
 			pruner.execQueue.Put(pfs, err, false)
 		})
-		return onErr(u, err)
+		return
 	}
 	// check if all snapshots were destroyed
 	destroyResults := make(map[string]*pdu.DestroySnapshotRes)
@@ -607,31 +602,6 @@ func stateExec(a *args, u updater) state {
 	})
 	if err != nil {
 		GetLogger(a.ctx).WithError(err).Error("target could not destroy snapshots")
-		return onErr(u, err)
+		return
-	}
-
-	return u(func(pruner *Pruner) {
-		pruner.Progress.MadeProgress()
-	}).statefunc()
-}
-
-func stateExecWait(a *args, u updater) state {
-	return doWait(Exec, a, u)
-}
-
-func statePlanWait(a *args, u updater) state {
-	return doWait(Plan, a, u)
-}
-
-func doWait(goback State, a *args, u updater) state {
-	timer := time.NewTimer(a.retryWait)
-	defer timer.Stop()
-	select {
-	case <-timer.C:
-		return u(func(pruner *Pruner) {
-			pruner.state = goback
-		}).statefunc()
-	case <-a.ctx.Done():
-		return onErr(u, a.ctx.Err())
 	}
 }

daemon/pruner/pruner_queue.go

@@ -58,10 +58,7 @@ func (q *execQueue) Pop() *fs {
 func(q *execQueue) Put(fs *fs, err error, done bool) {
 	fs.mtx.Lock()
 	fs.execErrLast = err
-	if err != nil {
+	if done || err != nil {
-		fs.execErrCount++
-	}
-	if done || (err != nil && !shouldRetry(fs.execErrLast)) {
 		fs.mtx.Unlock()
 		q.mtx.Lock()
 		q.completed = append(q.completed, fs)
@@ -78,9 +75,6 @@ func(q *execQueue) Put(fs *fs, err error, done bool) {
 		defer q.pending[i].mtx.Unlock()
 		q.pending[j].mtx.Lock()
 		defer q.pending[j].mtx.Unlock()
-		if q.pending[i].execErrCount != q.pending[j].execErrCount {
-			return q.pending[i].execErrCount < q.pending[j].execErrCount
-		}
 		return strings.Compare(q.pending[i].path, q.pending[j].path) == -1
 	})
 	q.mtx.Unlock()

daemon/pruner/pruner_test.go

@@ -1,206 +0,0 @@
-package pruner
-
-import (
-	"context"
-	"fmt"
-	"github.com/stretchr/testify/assert"
-	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/pruning"
-	"github.com/zrepl/zrepl/replication/pdu"
-	"net"
-	"testing"
-	"time"
-)
-
-type mockFS struct {
-	path  string
-	snaps []string
-}
-
-func (m *mockFS) Filesystem() *pdu.Filesystem {
-	return &pdu.Filesystem{
-		Path: m.path,
-	}
-}
-
-func (m *mockFS) FilesystemVersions() []*pdu.FilesystemVersion {
-	versions := make([]*pdu.FilesystemVersion, len(m.snaps))
-	for i, v := range m.snaps {
-		versions[i] = &pdu.FilesystemVersion{
-			Type:     pdu.FilesystemVersion_Snapshot,
-			Name:     v,
-			Creation: pdu.FilesystemVersionCreation(time.Unix(0, 0)),
-			Guid: uint64(i),
-		}
-	}
-	return versions
-}
-
-type mockTarget struct {
-	fss                []mockFS
-	destroyed          map[string][]string
-	listVersionsErrs   map[string][]error
-	listFilesystemsErr []error
-	destroyErrs        map[string][]error
-}
-
-func (t *mockTarget) ListFilesystems(ctx context.Context, req *pdu.ListFilesystemReq) (*pdu.ListFilesystemRes, error) {
-	if len(t.listFilesystemsErr) > 0 {
-		e := t.listFilesystemsErr[0]
-		t.listFilesystemsErr = t.listFilesystemsErr[1:]
-		return nil, e
-	}
-	fss := make([]*pdu.Filesystem, len(t.fss))
-	for i := range fss {
-		fss[i] = t.fss[i].Filesystem()
-	}
-	return &pdu.ListFilesystemRes{Filesystems: fss}, nil
-}
-
-func (t *mockTarget) ListFilesystemVersions(ctx context.Context, req *pdu.ListFilesystemVersionsReq) (*pdu.ListFilesystemVersionsRes, error) {
-	fs := req.Filesystem
-	if len(t.listVersionsErrs[fs]) != 0 {
-		e := t.listVersionsErrs[fs][0]
-		t.listVersionsErrs[fs] = t.listVersionsErrs[fs][1:]
-		return nil, e
-	}
-
-	for _, mfs := range t.fss {
-		if mfs.path != fs {
-			continue
-		}
-		return &pdu.ListFilesystemVersionsRes{Versions: mfs.FilesystemVersions()}, nil
-	}
-	return nil, fmt.Errorf("filesystem %s does not exist", fs)
-}
-
-func (t *mockTarget) DestroySnapshots(ctx context.Context, req *pdu.DestroySnapshotsReq) (*pdu.DestroySnapshotsRes, error) {
-	fs, snaps := req.Filesystem, req.Snapshots
-	if len(t.destroyErrs[fs]) != 0 {
-		e := t.destroyErrs[fs][0]
-		t.destroyErrs[fs] = t.destroyErrs[fs][1:]
-		return nil, e
-	}
-	destroyed := t.destroyed[fs]
-	res := make([]*pdu.DestroySnapshotRes, len(snaps))
-	for i, s := range snaps {
-		destroyed = append(destroyed, s.Name)
-		res[i] = &pdu.DestroySnapshotRes{Error: "", Snapshot: s}
-	}
-	t.destroyed[fs] = destroyed
-	return &pdu.DestroySnapshotsRes{Results: res}, nil
-}
-
-type mockCursor struct {
-	snapname string
-	guid uint64
-}
-type mockHistory struct {
-	errs map[string][]error
-	cursors map[string]*mockCursor
-}
-
-func (r *mockHistory) ReplicationCursor(ctx context.Context, req *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error) {
-	fs := req.Filesystem
-	if len(r.errs[fs]) > 0 {
-		e := r.errs[fs][0]
-		r.errs[fs] = r.errs[fs][1:]
-		return nil, e
-	}
-	return &pdu.ReplicationCursorRes{Result: &pdu.ReplicationCursorRes_Guid{Guid: 0}}, nil
-}
-
-type stubNetErr struct {
-	msg string
-	temporary, timeout bool
-}
-
-var _ net.Error = stubNetErr{}
-
-func (e stubNetErr) Error() string {
-	return e.msg
-}
-
-func (e stubNetErr) Temporary() bool { return e.temporary }
-
-func (e stubNetErr) Timeout() bool { return e.timeout }
-
-func TestPruner_Prune(t *testing.T) {
-
-	var _ net.Error = &net.OpError{} // we use it below
-	target := &mockTarget{
-		listFilesystemsErr: []error{
-			stubNetErr{msg: "fakerror0", temporary: true},
-		},
-		listVersionsErrs: map[string][]error{
-			"zroot/foo": {
-				stubNetErr{msg: "fakeerror1", temporary: true},
-				stubNetErr{msg: "fakeerror2", temporary: true,},
-			},
-		},
-		destroyErrs: map[string][]error{
-			"zroot/baz": {
-				stubNetErr{msg: "fakeerror3", temporary: true}, // first error puts it back in the queue
-				stubNetErr{msg:"permanent error"}, // so it will be last when pruner gives up due to permanent err
-			},
-		},
-		destroyed: make(map[string][]string),
-		fss: []mockFS{
-			{
-				path: "zroot/foo",
-				snaps: []string{
-					"keep_a",
-					"keep_b",
-					"drop_c",
-					"keep_d",
-				},
-			},
-			{
-				path: "zroot/bar",
-				snaps: []string{
-					"keep_e",
-					"keep_f",
-					"drop_g",
-				},
-			},
-			{
-				path: "zroot/baz",
-				snaps: []string{
-					"keep_h",
-					"drop_i",
-				},
-			},
-		},
-	}
-	history := &mockHistory{
-		errs: map[string][]error{
-			"zroot/foo": {
-				stubNetErr{msg: "fakeerror4", temporary: true},
-			},
-		},
-	}
-
-	keepRules := []pruning.KeepRule{pruning.MustKeepRegex("^keep", false)}
-
-	p := Pruner{
-		args: args{
-			ctx: WithLogger(context.Background(), logger.NewTestLogger(t)),
-			target: target,
-			receiver: history,
-			rules: keepRules,
-			retryWait: 10*time.Millisecond,
-		},
-		state: Plan,
-	}
-	p.Prune()
-
-	exp := map[string][]string{
-		"zroot/foo": {"drop_c"},
-		"zroot/bar": {"drop_g"},
-	}
-
-	assert.Equal(t, exp, target.destroyed)
-
-	//assert.Equal(t, map[string][]error{}, target.listVersionsErrs, "retried")
-
-}

daemon/pruner/state_enumer.go

@@ -7,19 +7,17 @@ import (
 )
 
 const (
-	_StateName_0 = "PlanPlanWait"
+	_StateName_0 = "PlanPlanErr"
 	_StateName_1 = "Exec"
-	_StateName_2 = "ExecWait"
+	_StateName_2 = "ExecErr"
-	_StateName_3 = "ErrPerm"
+	_StateName_3 = "Done"
-	_StateName_4 = "Done"
 )
 
 var (
-	_StateIndex_0 = [...]uint8{0, 4, 12}
+	_StateIndex_0 = [...]uint8{0, 4, 11}
 	_StateIndex_1 = [...]uint8{0, 4}
-	_StateIndex_2 = [...]uint8{0, 8}
+	_StateIndex_2 = [...]uint8{0, 7}
-	_StateIndex_3 = [...]uint8{0, 7}
+	_StateIndex_3 = [...]uint8{0, 4}
-	_StateIndex_4 = [...]uint8{0, 4}
 )
 
 func (i State) String() string {
@@ -33,22 +31,19 @@ func (i State) String() string {
 		return _StateName_2
 	case i == 16:
 		return _StateName_3
-	case i == 32:
-		return _StateName_4
 	default:
 		return fmt.Sprintf("State(%d)", i)
 	}
 }
 
-var _StateValues = []State{1, 2, 4, 8, 16, 32}
+var _StateValues = []State{1, 2, 4, 8, 16}
 
 var _StateNameToValueMap = map[string]State{
 	_StateName_0[0:4]:  1,
-	_StateName_0[4:12]: 2,
+	_StateName_0[4:11]: 2,
 	_StateName_1[0:4]:  4,
-	_StateName_2[0:8]:  8,
+	_StateName_2[0:7]:  8,
-	_StateName_3[0:7]:  16,
+	_StateName_3[0:4]:  16,
-	_StateName_4[0:4]:  32,
 }
 
 // StateString retrieves an enum value from the enum constants string name.

docs/configuration/jobs.rst

@@ -232,7 +232,8 @@ Job Type ``pull``
       - ZFS dataset path are received to
         ``$root_fs/$client_identity``
     * - ``interval``
-      - Interval at which to pull from the source job
+      - | Interval at which to pull from the source job (e.g. ``10m``).
+        | ``manual`` disables periodic pulling, replication then only happens on :ref:`wakeup <cli-signal-wakeup>`.
     * - ``pruning``
       - |pruning-spec|
 

docs/usage.rst

@@ -13,6 +13,8 @@ CLI Overview
     The zrepl binary is self-documenting:
     run ``zrepl help`` for an overview of the available subcommands or ``zrepl SUBCOMMAND --help`` for information on available flags, etc.
 
+.. _cli-signal-wakeup:
+
 .. list-table::
     :widths: 30 70
     :header-rows: 1

endpoint/endpoint.go

@@ -7,8 +7,7 @@ import (
 	"path"
 
 	"github.com/pkg/errors"
-	"github.com/zrepl/zrepl/replication"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
-	"github.com/zrepl/zrepl/replication/pdu"
 	"github.com/zrepl/zrepl/zfs"
 )
 
@@ -34,7 +33,7 @@ func (s *Sender) filterCheckFS(fs string) (*zfs.DatasetPath, error) {
 		return nil, err
 	}
 	if !pass {
-		return nil, replication.NewFilteredError(fs)
+		return nil, fmt.Errorf("endpoint does not allow access to filesystem %s", fs)
 	}
 	return dp, nil
 }
@@ -49,9 +48,10 @@ func (s *Sender) ListFilesystems(ctx context.Context, r *pdu.ListFilesystemReq)
 		rfss[i] = &pdu.Filesystem{
 			Path: fss[i].ToString(),
 			// FIXME: not supporting ResumeToken yet
+			IsPlaceholder: false, // sender FSs are never placeholders
 		}
 	}
-	res := &pdu.ListFilesystemRes{Filesystems: rfss, Empty: len(rfss) == 0}
+	res := &pdu.ListFilesystemRes{Filesystems: rfss}
 	return res, nil
 }
 
@@ -108,6 +108,21 @@ func (p *Sender) DestroySnapshots(ctx context.Context, req *pdu.DestroySnapshots
 	return doDestroySnapshots(ctx, dp, req.Snapshots)
 }
 
+func (p *Sender) Ping(ctx context.Context, req *pdu.PingReq) (*pdu.PingRes, error) {
+	res := pdu.PingRes{
+		Echo: req.GetMessage(),
+	}
+	return &res, nil
+}
+
+func (p *Sender) PingDataconn(ctx context.Context, req *pdu.PingReq) (*pdu.PingRes, error) {
+	return p.Ping(ctx, req)
+}
+
+func (p *Sender) WaitForConnectivity(ctx context.Context) error {
+	return nil
+}
+
 func (p *Sender) ReplicationCursor(ctx context.Context, req *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error) {
 	dp, err := p.filterCheckFS(req.Filesystem)
 	if err != nil {
@@ -229,7 +244,7 @@ func (s *Receiver) ListFilesystems(ctx context.Context, req *pdu.ListFilesystemR
 	if err != nil {
 		return nil, err
 	}
-	// present without prefix, and only those that are not placeholders
+	// present filesystem without the root_fs prefix
 	fss := make([]*pdu.Filesystem, 0, len(filtered))
 	for _, a := range filtered {
 		ph, err := zfs.ZFSIsPlaceholderFilesystem(a)
@@ -240,21 +255,16 @@ func (s *Receiver) ListFilesystems(ctx context.Context, req *pdu.ListFilesystemR
 				Error("inconsistent placeholder property")
 			return nil, errors.New("server error: inconsistent placeholder property") // don't leak path
 		}
-		if ph {
 		getLogger(ctx).
 			WithField("fs", a.ToString()).
-				Debug("ignoring placeholder filesystem")
+			WithField("is_placeholder", ph).
-			continue
+			Debug("filesystem")
-		}
-		getLogger(ctx).
-			WithField("fs", a.ToString()).
-			Debug("non-placeholder filesystem")
 		a.TrimPrefix(root)
-		fss = append(fss, &pdu.Filesystem{Path: a.ToString()})
+		fss = append(fss, &pdu.Filesystem{Path: a.ToString(), IsPlaceholder: ph})
 	}
 	if len(fss) == 0 {
-		getLogger(ctx).Debug("no non-placeholder filesystems")
+		getLogger(ctx).Debug("no filesystems found")
-		return &pdu.ListFilesystemRes{Empty: true}, nil
+		return &pdu.ListFilesystemRes{}, nil
 	}
 	return &pdu.ListFilesystemRes{Filesystems: fss}, nil
 }
@@ -279,6 +289,21 @@ func (s *Receiver) ListFilesystemVersions(ctx context.Context, req *pdu.ListFile
 	return &pdu.ListFilesystemVersionsRes{Versions: rfsvs}, nil
 }
 
+func (s *Receiver) Ping(ctx context.Context, req *pdu.PingReq) (*pdu.PingRes, error) {
+	res := pdu.PingRes{
+		Echo: req.GetMessage(),
+	}
+	return &res, nil
+}
+
+func (s *Receiver) PingDataconn(ctx context.Context, req *pdu.PingReq) (*pdu.PingRes, error) {
+	return s.Ping(ctx, req)
+}
+
+func (s *Receiver) WaitForConnectivity(ctx context.Context) error {
+	return nil
+}
+
 func (s *Receiver) ReplicationCursor(context.Context, *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error) {
 	return nil, fmt.Errorf("ReplicationCursor not implemented for Receiver")
 }
@@ -324,28 +349,30 @@ func (s *Receiver) Receive(ctx context.Context, req *pdu.ReceiveReq, receive zfs
 	getLogger(ctx).WithField("visitErr", visitErr).Debug("complete tree-walk")
 
 	if visitErr != nil {
-		return nil, err
+		return nil, visitErr
 	}
 
-	needForceRecv := false
+	var clearPlaceholderProperty bool
+	var recvOpts zfs.RecvOptions
 	props, err := zfs.ZFSGet(lp, []string{zfs.ZREPL_PLACEHOLDER_PROPERTY_NAME})
 	if err == nil {
 		if isPlaceholder, _ := zfs.IsPlaceholder(lp, props.Get(zfs.ZREPL_PLACEHOLDER_PROPERTY_NAME)); isPlaceholder {
-			needForceRecv = true
+			recvOpts.RollbackAndForceRecv = true
+			clearPlaceholderProperty = true
+		}
+	}
+	if clearPlaceholderProperty {
+		if err := zfs.ZFSSetNoPlaceholder(lp); err != nil {
+			return nil, fmt.Errorf("cannot clear placeholder property for forced receive: %s", err)
 		}
 	}
 
-	args := make([]string, 0, 1)
+	getLogger(ctx).WithField("opts", fmt.Sprintf("%#v", recvOpts)).Debug("start receive command")
-	if needForceRecv {
-		args = append(args, "-F")
-	}
 
-	getLogger(ctx).Debug("start receive command")
+	if err := zfs.ZFSRecv(ctx, lp.ToString(), receive, recvOpts); err != nil {
-
-	if err := zfs.ZFSRecv(ctx, lp.ToString(), receive, args...); err != nil {
 		getLogger(ctx).
 			WithError(err).
-			WithField("args", args).
+			WithField("opts", recvOpts).
 			Error("zfs receive failed")
 		return nil, err
 	}

replication/driver/errorclass_enumer.go

@@ -0,0 +1,50 @@
+// Code generated by "enumer -type=errorClass"; DO NOT EDIT.
+
+package driver
+
+import (
+	"fmt"
+)
+
+const _errorClassName = "errorClassUnknownerrorClassPermanenterrorClassTemporaryConnectivityRelated"
+
+var _errorClassIndex = [...]uint8{0, 17, 36, 74}
+
+func (i errorClass) String() string {
+	if i < 0 || i >= errorClass(len(_errorClassIndex)-1) {
+		return fmt.Sprintf("errorClass(%d)", i)
+	}
+	return _errorClassName[_errorClassIndex[i]:_errorClassIndex[i+1]]
+}
+
+var _errorClassValues = []errorClass{0, 1, 2}
+
+var _errorClassNameToValueMap = map[string]errorClass{
+	_errorClassName[0:17]:  0,
+	_errorClassName[17:36]: 1,
+	_errorClassName[36:74]: 2,
+}
+
+// errorClassString retrieves an enum value from the enum constants string name.
+// Throws an error if the param is not part of the enum.
+func errorClassString(s string) (errorClass, error) {
+	if val, ok := _errorClassNameToValueMap[s]; ok {
+		return val, nil
+	}
+	return 0, fmt.Errorf("%s does not belong to errorClass values", s)
+}
+
+// errorClassValues returns all values of the enum
+func errorClassValues() []errorClass {
+	return _errorClassValues
+}
+
+// IsAerrorClass returns "true" if the value is listed in the enum definition. "false" otherwise
+func (i errorClass) IsAerrorClass() bool {
+	for _, v := range _errorClassValues {
+		if i == v {
+			return true
+		}
+	}
+	return false
+}

replication/driver/replication_driver.go

@@ -0,0 +1,638 @@
+package driver
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"net"
+	"sort"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/zrepl/zrepl/replication/report"
+	"github.com/zrepl/zrepl/util/chainlock"
+	"github.com/zrepl/zrepl/util/envconst"
+	"google.golang.org/grpc/codes"
+	"google.golang.org/grpc/status"
+)
+
+type interval struct {
+	begin time.Time
+	end   time.Time
+}
+
+func (w *interval) SetZero() {
+	w.begin = time.Time{}
+	w.end = time.Time{}
+}
+
+// Duration of 0 means indefinite length
+func (w *interval) Set(begin time.Time, duration time.Duration) {
+	if begin.IsZero() {
+		panic("zero begin time now allowed")
+	}
+	w.begin = begin
+	w.end = begin.Add(duration)
+}
+
+// Returns the End of the interval if it has a defined length.
+// For indefinite lengths, returns the zero value.
+func (w *interval) End() time.Time {
+	return w.end
+}
+
+// Return a context with a deadline at the interval's end.
+// If the interval has indefinite length (duration 0 on Set), return ctx as is.
+// The returned context.CancelFunc can be called either way.
+func (w *interval) ContextWithDeadlineAtEnd(ctx context.Context) (context.Context, context.CancelFunc) {
+	if w.begin.IsZero() {
+		panic("must call Set before ContextWIthDeadlineAtEnd")
+	}
+	if w.end.IsZero() {
+		// indefinite length, just return context as is
+		return ctx, func() {}
+	} else {
+		return context.WithDeadline(ctx, w.end)
+	}
+}
+
+type run struct {
+	l *chainlock.L
+
+	startedAt, finishedAt time.Time
+
+	waitReconnect      interval
+	waitReconnectError *timedError
+
+	// the attempts attempted so far:
+	// All but the last in this slice must have finished with some errors.
+	// The last attempt may not be finished and may not have errors.
+	attempts []*attempt
+}
+
+type Planner interface {
+	Plan(context.Context) ([]FS, error)
+	WaitForConnectivity(context.Context) error
+}
+
+// an attempt represents a single planning & execution of fs replications
+type attempt struct {
+	planner Planner
+
+	l *chainlock.L
+
+	startedAt, finishedAt time.Time
+
+	// after Planner.Plan was called, planErr and fss are mutually exclusive with regards to nil-ness
+	// if both are nil, it must be assumed that Planner.Plan is active
+	planErr *timedError
+	fss     []*fs
+}
+
+type timedError struct {
+	Err  error
+	Time time.Time
+}
+
+func newTimedError(err error, t time.Time) *timedError {
+	if err == nil {
+		panic("error must be non-nil")
+	}
+	if t.IsZero() {
+		panic("t must be non-zero")
+	}
+	return &timedError{err, t}
+}
+
+func (e *timedError) IntoReportError() *report.TimedError {
+	if e == nil {
+		return nil
+	}
+	return report.NewTimedError(e.Err.Error(), e.Time)
+}
+
+type FS interface {
+	// Returns true if this FS and fs refer to the same filesystem returned
+	// by Planner.Plan in a previous attempt.
+	EqualToPreviousAttempt(fs FS) bool
+	// The returned steps are assumed to be dependent on exactly
+	// their direct predecessors in the returned list.
+	PlanFS(context.Context) ([]Step, error)
+	ReportInfo() *report.FilesystemInfo
+}
+
+type Step interface {
+	// Returns true iff the target snapshot is the same for this Step and other.
+	// We do not use TargetDate to avoid problems with wrong system time on
+	// snapshot creation.
+	//
+	// Implementations can assume that `other` is a step of the same filesystem,
+	// although maybe from a previous attempt.
+	// (`same` as defined by FS.EqualToPreviousAttempt)
+	//
+	// Note that TargetEquals should return true in a situation with one
+	// originally sent snapshot and a subsequent attempt's step that uses
+	// resumable send & recv.
+	TargetEquals(other Step) bool
+	TargetDate() time.Time
+	Step(context.Context) error
+	ReportInfo() *report.StepInfo
+}
+
+type fs struct {
+	fs FS
+
+	l *chainlock.L
+
+	planning struct {
+		done bool
+		err  *timedError
+	}
+
+	// valid iff planning.done && planning.err == nil
+	planned struct {
+		// valid iff planning.done && planning.err == nil
+		stepErr *timedError
+		// all steps, in the order in which they must be completed
+		steps []*step
+		// index into steps, pointing at the step that is currently executing
+		// if step >= len(steps), no more work needs to be done
+		step int
+	}
+}
+
+type step struct {
+	l    *chainlock.L
+	step Step
+}
+
+type ReportFunc func() *report.Report
+type WaitFunc func(block bool) (done bool)
+
+var maxAttempts = envconst.Int64("ZREPL_REPLICATION_MAX_ATTEMPTS", 3)
+var reconnectHardFailTimeout = envconst.Duration("ZREPL_REPLICATION_RECONNECT_HARD_FAIL_TIMEOUT", 10*time.Minute)
+
+func Do(ctx context.Context, planner Planner) (ReportFunc, WaitFunc) {
+	log := getLog(ctx)
+	l := chainlock.New()
+	run := &run{
+		l:         l,
+		startedAt: time.Now(),
+	}
+
+	done := make(chan struct{})
+	go func() {
+		defer close(done)
+
+		defer run.l.Lock().Unlock()
+		log.Debug("begin run")
+		defer log.Debug("run ended")
+		var prev *attempt
+		mainLog := log
+		for ano := 0; ano < int(maxAttempts) || maxAttempts == 0; ano++ {
+			log := mainLog.WithField("attempt_number", ano)
+			log.Debug("start attempt")
+
+			run.waitReconnect.SetZero()
+			run.waitReconnectError = nil
+
+			// do current attempt
+			cur := &attempt{
+				l:         l,
+				startedAt: time.Now(),
+				planner:   planner,
+			}
+			run.attempts = append(run.attempts, cur)
+			run.l.DropWhile(func() {
+				cur.do(ctx, prev)
+			})
+			prev = cur
+			if ctx.Err() != nil {
+				log.WithError(ctx.Err()).Info("context error")
+				return
+			}
+
+			// error classification, bail out if done / permanent error
+			rep := cur.report()
+			log.WithField("attempt_state", rep.State).Debug("attempt state")
+			errRep := cur.errorReport()
+
+			if rep.State == report.AttemptDone {
+				log.Debug("attempt completed successfully")
+				break
+			}
+
+			mostRecentErr, mostRecentErrClass := errRep.MostRecent()
+			log.WithField("most_recent_err", mostRecentErr).WithField("most_recent_err_class", mostRecentErrClass).Debug("most recent error used for re-connect decision")
+			if mostRecentErr == nil {
+				// inconsistent reporting, let's bail out
+				log.Warn("attempt does not report done but error report does not report errors, aborting run")
+				break
+			}
+			log.WithError(mostRecentErr.Err).Error("most recent error in this attempt")
+			shouldReconnect := mostRecentErrClass == errorClassTemporaryConnectivityRelated
+			log.WithField("reconnect_decision", shouldReconnect).Debug("reconnect decision made")
+			if shouldReconnect {
+				run.waitReconnect.Set(time.Now(), reconnectHardFailTimeout)
+				log.WithField("deadline", run.waitReconnect.End()).Error("temporary connectivity-related error identified, start waiting for reconnect")
+				var connectErr error
+				var connectErrTime time.Time
+				run.l.DropWhile(func() {
+					ctx, cancel := run.waitReconnect.ContextWithDeadlineAtEnd(ctx)
+					defer cancel()
+					connectErr = planner.WaitForConnectivity(ctx)
+					connectErrTime = time.Now()
+				})
+				if connectErr == nil {
+					log.Error("reconnect successful") // same level as 'begin with reconnect' message above
+					continue
+				} else {
+					run.waitReconnectError = newTimedError(connectErr, connectErrTime)
+					log.WithError(connectErr).Error("reconnecting failed, aborting run")
+					break
+				}
+			} else {
+				log.Error("most recent error cannot be solved by reconnecting, aborting run")
+				return
+			}
+
+		}
+
+	}()
+
+	wait := func(block bool) bool {
+		if block {
+			<-done
+		}
+		select {
+		case <-done:
+			return true
+		default:
+			return false
+		}
+	}
+	report := func() *report.Report {
+		defer run.l.Lock().Unlock()
+		return run.report()
+	}
+	return report, wait
+}
+
+func (a *attempt) do(ctx context.Context, prev *attempt) {
+	pfss, err := a.planner.Plan(ctx)
+	errTime := time.Now()
+	defer a.l.Lock().Unlock()
+	if err != nil {
+		a.planErr = newTimedError(err, errTime)
+		a.fss = nil
+		a.finishedAt = time.Now()
+		return
+	}
+
+	for _, pfs := range pfss {
+		fs := &fs{
+			fs: pfs,
+			l:  a.l,
+		}
+		a.fss = append(a.fss, fs)
+	}
+
+	prevs := make(map[*fs]*fs)
+	{
+		prevFSs := make(map[*fs][]*fs, len(pfss))
+		if prev != nil {
+			debug("previous attempt has %d fss", len(a.fss))
+			for _, fs := range a.fss {
+				for _, prevFS := range prev.fss {
+					if fs.fs.EqualToPreviousAttempt(prevFS.fs) {
+						l := prevFSs[fs]
+						l = append(l, prevFS)
+						prevFSs[fs] = l
+					}
+				}
+			}
+		}
+		type inconsistency struct {
+			cur   *fs
+			prevs []*fs
+		}
+		var inconsistencies []inconsistency
+		for cur, fss := range prevFSs {
+			if len(fss) > 1 {
+				inconsistencies = append(inconsistencies, inconsistency{cur, fss})
+			}
+		}
+		sort.SliceStable(inconsistencies, func(i, j int) bool {
+			return inconsistencies[i].cur.fs.ReportInfo().Name < inconsistencies[j].cur.fs.ReportInfo().Name
+		})
+		if len(inconsistencies) > 0 {
+			var msg strings.Builder
+			msg.WriteString("cannot determine filesystem correspondences between different attempts:\n")
+			var inconsistencyLines []string
+			for _, i := range inconsistencies {
+				var prevNames []string
+				for _, prev := range i.prevs {
+					prevNames = append(prevNames, prev.fs.ReportInfo().Name)
+				}
+				l := fmt.Sprintf("  %s => %v", i.cur.fs.ReportInfo().Name, prevNames)
+				inconsistencyLines = append(inconsistencyLines, l)
+			}
+			fmt.Fprintf(&msg, strings.Join(inconsistencyLines, "\n"))
+			now := time.Now()
+			a.planErr = newTimedError(errors.New(msg.String()), now)
+			a.fss = nil
+			a.finishedAt = now
+			return
+		}
+		for cur, fss := range prevFSs {
+			if len(fss) > 0 {
+				prevs[cur] = fss[0]
+			}
+		}
+	}
+	// invariant: prevs contains an entry for each unambigious correspondence
+
+	stepQueue := newStepQueue()
+	defer stepQueue.Start(1)() // TODO parallel replication
+	var fssesDone sync.WaitGroup
+	for _, f := range a.fss {
+		fssesDone.Add(1)
+		go func(f *fs) {
+			defer fssesDone.Done()
+			f.do(ctx, stepQueue, prevs[f])
+		}(f)
+	}
+	a.l.DropWhile(func() {
+		fssesDone.Wait()
+	})
+	a.finishedAt = time.Now()
+}
+
+func (fs *fs) do(ctx context.Context, pq *stepQueue, prev *fs) {
+	psteps, err := fs.fs.PlanFS(ctx)
+	errTime := time.Now()
+	defer fs.l.Lock().Unlock()
+	debug := debugPrefix("fs=%s", fs.fs.ReportInfo().Name)
+	fs.planning.done = true
+	if err != nil {
+		fs.planning.err = newTimedError(err, errTime)
+		return
+	}
+	for _, pstep := range psteps {
+		step := &step{
+			l:    fs.l,
+			step: pstep,
+		}
+		fs.planned.steps = append(fs.planned.steps, step)
+	}
+	debug("iniital len(fs.planned.steps) = %d", len(fs.planned.steps))
+
+	// for not-first attempts, only allow fs.planned.steps
+	// up to including the originally planned target snapshot
+	if prev != nil && prev.planning.done && prev.planning.err == nil {
+		prevUncompleted := prev.planned.steps[prev.planned.step:]
+		if len(prevUncompleted) == 0 {
+			debug("prevUncompleted is empty")
+			return
+		}
+		if len(fs.planned.steps) == 0 {
+			debug("fs.planned.steps is empty")
+			return
+		}
+		prevFailed := prevUncompleted[0]
+		curFirst := fs.planned.steps[0]
+		// we assume that PlanFS retries prevFailed (using curFirst)
+		if !prevFailed.step.TargetEquals(curFirst.step) {
+			debug("Targets don't match")
+			// Two options:
+			// A: planning algorithm is broken
+			// B: manual user intervention inbetween
+			// Neither way will we make progress, so let's error out
+			stepFmt := func(step *step) string {
+				r := step.report()
+				s := r.Info
+				if r.IsIncremental() {
+					return fmt.Sprintf("%s=>%s", s.From, s.To)
+				} else {
+					return fmt.Sprintf("full=>%s", s.To)
+				}
+			}
+			msg := fmt.Sprintf("last attempt's uncompleted step %s does not correspond to this attempt's first planned step %s",
+				stepFmt(prevFailed), stepFmt(curFirst))
+			fs.planned.stepErr = newTimedError(errors.New(msg), time.Now())
+			return
+		}
+		// only allow until step targets diverge
+		min := len(prevUncompleted)
+		if min > len(fs.planned.steps) {
+			min = len(fs.planned.steps)
+		}
+		diverge := 0
+		for ; diverge < min; diverge++ {
+			debug("diverge compare iteration %d", diverge)
+			if !fs.planned.steps[diverge].step.TargetEquals(prevUncompleted[diverge].step) {
+				break
+			}
+		}
+		debug("diverge is %d", diverge)
+		fs.planned.steps = fs.planned.steps[0:diverge]
+	}
+	debug("post-prev-merge len(fs.planned.steps) = %d", len(fs.planned.steps))
+
+	for i, s := range fs.planned.steps {
+		var (
+			err     error
+			errTime time.Time
+		)
+		// lock must not be held while executing step in order for reporting to work
+		fs.l.DropWhile(func() {
+			targetDate := s.step.TargetDate()
+			defer pq.WaitReady(fs, targetDate)()
+			err = s.step.Step(ctx) // no shadow
+			errTime = time.Now()   // no shadow
+		})
+		if err != nil {
+			fs.planned.stepErr = newTimedError(err, errTime)
+			break
+		}
+		fs.planned.step = i + 1 // fs.planned.step must be == len(fs.planned.steps) if all went OK
+	}
+}
+
+// caller must hold lock l
+func (r *run) report() *report.Report {
+	report := &report.Report{
+		Attempts:           make([]*report.AttemptReport, len(r.attempts)),
+		StartAt:            r.startedAt,
+		FinishAt:           r.finishedAt,
+		WaitReconnectSince: r.waitReconnect.begin,
+		WaitReconnectUntil: r.waitReconnect.end,
+		WaitReconnectError: r.waitReconnectError.IntoReportError(),
+	}
+	for i := range report.Attempts {
+		report.Attempts[i] = r.attempts[i].report()
+	}
+	return report
+}
+
+// caller must hold lock l
+func (a *attempt) report() *report.AttemptReport {
+
+	r := &report.AttemptReport{
+		// State is set below
+		Filesystems: make([]*report.FilesystemReport, len(a.fss)),
+		StartAt:     a.startedAt,
+		FinishAt:    a.finishedAt,
+		PlanError:   a.planErr.IntoReportError(),
+	}
+
+	for i := range r.Filesystems {
+		r.Filesystems[i] = a.fss[i].report()
+	}
+
+	state := report.AttemptPlanning
+	if a.planErr != nil {
+		state = report.AttemptPlanningError
+	} else if a.fss != nil {
+		if a.finishedAt.IsZero() {
+			state = report.AttemptFanOutFSs
+		} else {
+			fsWithError := false
+			for _, s := range r.Filesystems {
+				fsWithError = fsWithError || s.Error() != nil
+			}
+			state = report.AttemptDone
+			if fsWithError {
+				state = report.AttemptFanOutError
+			}
+		}
+	}
+	r.State = state
+
+	return r
+}
+
+// caller must hold lock l
+func (f *fs) report() *report.FilesystemReport {
+	state := report.FilesystemPlanningErrored
+	if f.planning.err == nil {
+		if f.planning.done {
+			if f.planned.stepErr != nil {
+				state = report.FilesystemSteppingErrored
+			} else if f.planned.step < len(f.planned.steps) {
+				state = report.FilesystemStepping
+			} else {
+				state = report.FilesystemDone
+			}
+		} else {
+			state = report.FilesystemPlanning
+		}
+	}
+	r := &report.FilesystemReport{
+		Info:        f.fs.ReportInfo(),
+		State:       state,
+		PlanError:   f.planning.err.IntoReportError(),
+		StepError:   f.planned.stepErr.IntoReportError(),
+		Steps:       make([]*report.StepReport, len(f.planned.steps)),
+		CurrentStep: f.planned.step,
+	}
+	for i := range r.Steps {
+		r.Steps[i] = f.planned.steps[i].report()
+	}
+	return r
+}
+
+// caller must hold lock l
+func (s *step) report() *report.StepReport {
+	r := &report.StepReport{
+		Info: s.step.ReportInfo(),
+	}
+	return r
+}
+
+type stepErrorReport struct {
+	err  *timedError
+	step int
+}
+
+//go:generate enumer -type=errorClass
+type errorClass int
+
+const (
+	errorClassUnknown errorClass = iota
+	errorClassPermanent
+	errorClassTemporaryConnectivityRelated
+)
+
+type errorReport struct {
+	flattened []*timedError
+	// sorted DESCending by err time
+	byClass map[errorClass][]*timedError
+}
+
+// caller must hold lock l
+func (a *attempt) errorReport() *errorReport {
+	r := &errorReport{}
+	if a.planErr != nil {
+		r.flattened = append(r.flattened, a.planErr)
+	}
+	for _, fs := range a.fss {
+		if fs.planning.done && fs.planning.err != nil {
+			r.flattened = append(r.flattened, fs.planning.err)
+		} else if fs.planning.done && fs.planned.stepErr != nil {
+			r.flattened = append(r.flattened, fs.planned.stepErr)
+		}
+	}
+
+	// build byClass
+	{
+		r.byClass = make(map[errorClass][]*timedError)
+		putClass := func(err *timedError, class errorClass) {
+			errs := r.byClass[class]
+			errs = append(errs, err)
+			r.byClass[class] = errs
+		}
+		for _, err := range r.flattened {
+			if neterr, ok := err.Err.(net.Error); ok && neterr.Temporary() {
+				putClass(err, errorClassTemporaryConnectivityRelated)
+				continue
+			}
+			if st, ok := status.FromError(err.Err); ok && st.Code() == codes.Unavailable {
+				// technically, codes.Unavailable could be returned by the gRPC endpoint, indicating overload, etc.
+				// for now, let's assume it only happens for connectivity issues, as specified in
+				// https://grpc.io/grpc/core/md_doc_statuscodes.html
+				putClass(err, errorClassTemporaryConnectivityRelated)
+				continue
+			}
+			putClass(err, errorClassPermanent)
+		}
+		for _, errs := range r.byClass {
+			sort.Slice(errs, func(i, j int) bool {
+				return errs[i].Time.After(errs[j].Time) // sort descendingly
+			})
+		}
+	}
+
+	return r
+}
+
+func (r *errorReport) AnyError() *timedError {
+	for _, err := range r.flattened {
+		if err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func (r *errorReport) MostRecent() (err *timedError, errClass errorClass) {
+	for class, errs := range r.byClass {
+		// errs are sorted descendingly during construction
+		if len(errs) > 0 && (err == nil || errs[0].Time.After(err.Time)) {
+			err = errs[0]
+			errClass = class
+		}
+	}
+	return
+}

replication/driver/replication_driver_debug.go

@@ -0,0 +1,29 @@
+package driver
+
+import (
+	"fmt"
+	"os"
+)
+
+var debugEnabled bool = false
+
+func init() {
+	if os.Getenv("ZREPL_REPLICATION_DRIVER_DEBUG") != "" {
+		debugEnabled = true
+	}
+}
+
+func debug(format string, args ...interface{}) {
+	if debugEnabled {
+		fmt.Fprintf(os.Stderr, "repl: driver: %s\n", fmt.Sprintf(format, args...))
+	}
+}
+
+type debugFunc func(format string, args ...interface{})
+
+func debugPrefix(prefixFormat string, prefixFormatArgs ...interface{}) debugFunc {
+	prefix := fmt.Sprintf(prefixFormat, prefixFormatArgs...)
+	return func(format string, args ...interface{}) {
+		debug("%s: %s", prefix, fmt.Sprintf(format, args))
+	}
+}

replication/driver/replication_driver_logging.go

@@ -0,0 +1,25 @@
+package driver
+
+import (
+	"context"
+
+	"github.com/zrepl/zrepl/logger"
+)
+
+type Logger = logger.Logger
+
+type contexKey int
+
+const contexKeyLogger contexKey = iota + 1
+
+func getLog(ctx context.Context) Logger {
+	l, ok := ctx.Value(contexKeyLogger).(Logger)
+	if !ok {
+		l = logger.NewNullLogger()
+	}
+	return l
+}
+
+func WithLogger(ctx context.Context, log Logger) context.Context {
+	return context.WithValue(ctx, contexKeyLogger, log)
+}

replication/driver/replication_driver_test.go

@@ -0,0 +1,215 @@
+package driver
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"sort"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+	"github.com/zrepl/zrepl/replication/report"
+
+	"github.com/stretchr/testify/assert"
+
+	jsondiff "github.com/yudai/gojsondiff"
+	jsondiffformatter "github.com/yudai/gojsondiff/formatter"
+)
+
+type mockPlanner struct {
+	stepCounter uint32
+	fss         []FS // *mockFS
+}
+
+func (p *mockPlanner) Plan(ctx context.Context) ([]FS, error) {
+	time.Sleep(1 * time.Second)
+	p.fss = []FS{
+		&mockFS{
+			&p.stepCounter,
+			"zroot/one",
+			nil,
+		},
+		&mockFS{
+			&p.stepCounter,
+			"zroot/two",
+			nil,
+		},
+	}
+	return p.fss, nil
+}
+
+func (p *mockPlanner) WaitForConnectivity(context.Context) error {
+	return nil
+}
+
+type mockFS struct {
+	globalStepCounter *uint32
+	name              string
+	steps             []Step
+}
+
+func (f *mockFS) EqualToPreviousAttempt(other FS) bool {
+	return f.name == other.(*mockFS).name
+}
+
+func (f *mockFS) PlanFS(ctx context.Context) ([]Step, error) {
+	if f.steps != nil {
+		panic("PlanFS used twice")
+	}
+	switch f.name {
+	case "zroot/one":
+		f.steps = []Step{
+			&mockStep{
+				fs:         f,
+				ident:      "a",
+				duration:   1 * time.Second,
+				targetDate: time.Unix(2, 0),
+			},
+			&mockStep{
+				fs:         f,
+				ident:      "b",
+				duration:   1 * time.Second,
+				targetDate: time.Unix(10, 0),
+			},
+			&mockStep{
+				fs:         f,
+				ident:      "c",
+				duration:   1 * time.Second,
+				targetDate: time.Unix(20, 0),
+			},
+		}
+	case "zroot/two":
+		f.steps = []Step{
+			&mockStep{
+				fs:         f,
+				ident:      "u",
+				duration:   500 * time.Millisecond,
+				targetDate: time.Unix(15, 0),
+			},
+			&mockStep{
+				fs:         f,
+				duration:   500 * time.Millisecond,
+				ident:      "v",
+				targetDate: time.Unix(30, 0),
+			},
+		}
+	default:
+		panic("unimplemented")
+	}
+
+	return f.steps, nil
+}
+
+func (f *mockFS) ReportInfo() *report.FilesystemInfo {
+	return &report.FilesystemInfo{Name: f.name}
+}
+
+type mockStep struct {
+	fs         *mockFS
+	ident      string
+	duration   time.Duration
+	targetDate time.Time
+
+	// filled by method Step
+	globalCtr uint32
+}
+
+func (f *mockStep) String() string {
+	return fmt.Sprintf("%s{%s} targetDate=%s globalCtr=%v", f.fs.name, f.ident, f.targetDate, f.globalCtr)
+}
+
+func (f *mockStep) Step(ctx context.Context) error {
+	f.globalCtr = atomic.AddUint32(f.fs.globalStepCounter, 1)
+	time.Sleep(f.duration)
+	return nil
+}
+
+func (f *mockStep) TargetEquals(s Step) bool {
+	return f.ident == s.(*mockStep).ident
+}
+
+func (f *mockStep) TargetDate() time.Time {
+	return f.targetDate
+}
+
+func (f *mockStep) ReportInfo() *report.StepInfo {
+	return &report.StepInfo{From: f.ident, To: f.ident, BytesExpected: 100, BytesReplicated: 25}
+}
+
+// TODO: add meaningful validation (i.e. actual checks)
+// Since the stepqueue is not deterministic due to scheduler jitter,
+// we cannot test for any definitive sequence of steps here.
+// Such checks would further only be sensible for a non-concurrent step-queue,
+// but we're going to have concurrent replication in the future.
+//
+// For the time being, let's just exercise the code a bit.
+func TestReplication(t *testing.T) {
+
+	ctx := context.Background()
+
+	mp := &mockPlanner{}
+	getReport, wait := Do(ctx, mp)
+	begin := time.Now()
+	fireAt := []time.Duration{
+		// the following values are relative to the start
+		500 * time.Millisecond,  // planning
+		1500 * time.Millisecond, // nothing is done, a is running
+		2500 * time.Millisecond, // a done,  b running
+		3250 * time.Millisecond, // a,b done, u running
+		3750 * time.Millisecond, // a,b,u done, c running
+		4750 * time.Millisecond, // a,b,u,c done, v running
+		5250 * time.Millisecond, // a,b,u,c,v done
+	}
+	reports := make([]*report.Report, len(fireAt))
+	for i := range fireAt {
+		sleepUntil := begin.Add(fireAt[i])
+		time.Sleep(sleepUntil.Sub(time.Now()))
+		reports[i] = getReport()
+		// uncomment for viewing non-diffed results
+		// t.Logf("report @ %6.4f:\n%s", fireAt[i].Seconds(), pretty.Sprint(reports[i]))
+	}
+	waitBegin := time.Now()
+	wait(true)
+	waitDuration := time.Now().Sub(waitBegin)
+	assert.True(t, waitDuration < 10*time.Millisecond, "%v", waitDuration) // and that's gratious
+
+	prev, err := json.Marshal(reports[0])
+	require.NoError(t, err)
+	for _, r := range reports[1:] {
+		this, err := json.Marshal(r)
+		require.NoError(t, err)
+		differ := jsondiff.New()
+		diff, err := differ.Compare(prev, this)
+		require.NoError(t, err)
+		df := jsondiffformatter.NewDeltaFormatter()
+		_, err = df.Format(diff)
+		require.NoError(t, err)
+		// uncomment the following line to get json diffs between each captured step
+		// t.Logf("%s", res)
+		prev, err = json.Marshal(r)
+		require.NoError(t, err)
+	}
+
+	steps := make([]*mockStep, 0)
+	for _, fs := range mp.fss {
+		for _, step := range fs.(*mockFS).steps {
+			steps = append(steps, step.(*mockStep))
+		}
+	}
+
+	// sort steps in pq order (although, remember, pq is not deterministic)
+	sort.Slice(steps, func(i, j int) bool {
+		return steps[i].targetDate.Before(steps[j].targetDate)
+	})
+
+	// manual inspection of the globalCtr value should show that, despite
+	// scheduler-dependent behavior of pq, steps should generally be taken
+	// from oldest to newest target date (globally, not per FS).
+	t.Logf("steps sorted by target date:")
+	for _, step := range steps {
+		t.Logf("\t%s", step)
+	}
+
+}

replication/driver/replication_stepqueue.go

@@ -0,0 +1,163 @@
+package driver
+
+import (
+	"container/heap"
+	"time"
+
+	"github.com/zrepl/zrepl/util/chainlock"
+)
+
+type stepQueueRec struct {
+	ident      interface{}
+	targetDate time.Time
+	wakeup     chan StepCompletedFunc
+}
+
+type stepQueue struct {
+	stop chan struct{}
+	reqs chan stepQueueRec
+}
+
+type stepQueueHeapItem struct {
+	idx int
+	req stepQueueRec
+}
+type stepQueueHeap []*stepQueueHeapItem
+
+func (h stepQueueHeap) Less(i, j int) bool {
+	return h[i].req.targetDate.Before(h[j].req.targetDate)
+}
+
+func (h stepQueueHeap) Swap(i, j int) {
+	h[i], h[j] = h[j], h[i]
+	h[i].idx = i
+	h[j].idx = j
+}
+
+func (h stepQueueHeap) Len() int {
+	return len(h)
+}
+
+func (h *stepQueueHeap) Push(elem interface{}) {
+	hitem := elem.(*stepQueueHeapItem)
+	hitem.idx = h.Len()
+	*h = append(*h, hitem)
+}
+
+func (h *stepQueueHeap) Pop() interface{} {
+	elem := (*h)[h.Len()-1]
+	elem.idx = -1
+	*h = (*h)[:h.Len()-1]
+	return elem
+}
+
+// returned stepQueue must be closed with method Close
+func newStepQueue() *stepQueue {
+	q := &stepQueue{
+		stop: make(chan struct{}),
+		reqs: make(chan stepQueueRec),
+	}
+	return q
+}
+
+// the returned done function must be called to free resources
+// allocated by the call to Start
+//
+// No WaitReady calls must be active at the time done is called
+// The behavior of calling WaitReady after done was called is undefined
+func (q *stepQueue) Start(concurrency int) (done func()) {
+	if concurrency < 1 {
+		panic("concurrency must be >= 1")
+	}
+	// l protects pending and queueItems
+	l := chainlock.New()
+	pendingCond := l.NewCond()
+	// priority queue
+	pending := &stepQueueHeap{}
+	// ident => queueItem
+	queueItems := make(map[interface{}]*stepQueueHeapItem)
+	// stopped is used for cancellation of "wake" goroutine
+	stopped := false
+	active := 0
+	go func() { // "stopper" goroutine
+		<-q.stop
+		defer l.Lock().Unlock()
+		stopped = true
+		pendingCond.Broadcast()
+	}()
+	go func() { // "reqs" goroutine
+		for {
+			select {
+			case <-q.stop:
+				select {
+				case <-q.reqs:
+					panic("WaitReady call active while calling Close")
+				default:
+					return
+				}
+			case req := <-q.reqs:
+				func() {
+					defer l.Lock().Unlock()
+					if _, ok := queueItems[req.ident]; ok {
+						panic("WaitReady must not be called twice for the same ident")
+					}
+					qitem := &stepQueueHeapItem{
+						req: req,
+					}
+					queueItems[req.ident] = qitem
+					heap.Push(pending, qitem)
+					pendingCond.Broadcast()
+				}()
+			}
+		}
+	}()
+	go func() { // "wake" goroutine
+		defer l.Lock().Unlock()
+		for {
+
+			for !stopped && (active >= concurrency || pending.Len() == 0) {
+				pendingCond.Wait()
+			}
+			if stopped {
+				return
+			}
+			if pending.Len() <= 0 {
+				return
+			}
+			active++
+			next := heap.Pop(pending).(*stepQueueHeapItem).req
+			delete(queueItems, next.ident)
+
+			next.wakeup <- func() {
+				defer l.Lock().Unlock()
+				active--
+				pendingCond.Broadcast()
+			}
+		}
+	}()
+
+	done = func() {
+		close(q.stop)
+	}
+	return done
+}
+
+type StepCompletedFunc func()
+
+func (q *stepQueue) sendAndWaitForWakeup(ident interface{}, targetDate time.Time) StepCompletedFunc {
+	req := stepQueueRec{
+		ident,
+		targetDate,
+		make(chan StepCompletedFunc),
+	}
+	q.reqs <- req
+	return <-req.wakeup
+}
+
+// Wait for the ident with targetDate to be selected to run.
+func (q *stepQueue) WaitReady(ident interface{}, targetDate time.Time) StepCompletedFunc {
+	if targetDate.IsZero() {
+		panic("targetDate of zero is reserved for marking Done")
+	}
+	return q.sendAndWaitForWakeup(ident, targetDate)
+}

replication/driver/replication_stepqueue_test.go

@@ -0,0 +1,166 @@
+package driver
+
+import (
+	"fmt"
+	"math"
+	"sort"
+	"sync"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/montanaflynn/stats"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestPqNotconcurrent(t *testing.T) {
+
+	var ctr uint32
+	q := newStepQueue()
+	var wg sync.WaitGroup
+	wg.Add(3)
+	go func() {
+		defer wg.Done()
+		defer q.WaitReady("1", time.Unix(1, 0))()
+		ret := atomic.AddUint32(&ctr, 1)
+		assert.Equal(t, uint32(1), ret)
+	}()
+	go func() {
+		defer wg.Done()
+		defer q.WaitReady("2", time.Unix(2, 0))()
+		ret := atomic.AddUint32(&ctr, 1)
+		assert.Equal(t, uint32(2), ret)
+	}()
+	go func() {
+		defer wg.Done()
+		defer q.WaitReady("3", time.Unix(3, 0))()
+		ret := atomic.AddUint32(&ctr, 1)
+		assert.Equal(t, uint32(3), ret)
+	}()
+
+	time.Sleep(1 * time.Second)
+	defer q.Start(1)()
+	wg.Wait()
+
+}
+
+type record struct {
+	fs        int
+	step      int
+	globalCtr uint32
+	wakeAt    time.Duration // relative to begin
+}
+
+func (r record) String() string {
+	return fmt.Sprintf("fs %08d step %08d globalCtr %08d wakeAt %2.8f", r.fs, r.step, r.globalCtr, r.wakeAt.Seconds())
+}
+
+// This tests uses stepPq concurrently, simulating the following scenario:
+// Given a number of filesystems F, each filesystem has N steps to take.
+// The number of concurrent steps is limited to C.
+// The target date for each step is the step number N.
+// Hence, there are always F filesystems runnable (calling WaitReady)
+// The priority queue prioritizes steps with lower target data (= lower step number).
+// Hence, all steps with lower numbers should be woken up before steps with higher numbers.
+// However, scheduling is not 100% deterministic (runtime, OS scheduler, etc).
+// Hence, perform some statistics on the wakeup times and assert that the mean wakeup
+// times for each step are close together.
+func TestPqConcurrent(t *testing.T) {
+
+	q := newStepQueue()
+	var wg sync.WaitGroup
+	filesystems := 100
+	stepsPerFS := 20
+	sleepTimePerStep := 50 * time.Millisecond
+	wg.Add(filesystems)
+	var globalCtr uint32
+
+	begin := time.Now()
+	records := make(chan []record, filesystems)
+	for fs := 0; fs < filesystems; fs++ {
+		go func(fs int) {
+			defer wg.Done()
+			recs := make([]record, 0)
+			for step := 0; step < stepsPerFS; step++ {
+				pos := atomic.AddUint32(&globalCtr, 1)
+				t := time.Unix(int64(step), 0)
+				done := q.WaitReady(fs, t)
+				wakeAt := time.Now().Sub(begin)
+				time.Sleep(sleepTimePerStep)
+				done()
+				recs = append(recs, record{fs, step, pos, wakeAt})
+			}
+			records <- recs
+		}(fs)
+	}
+	concurrency := 5
+	defer q.Start(concurrency)()
+	wg.Wait()
+	close(records)
+	t.Logf("loop done")
+
+	flattenedRecs := make([]record, 0)
+	for recs := range records {
+		flattenedRecs = append(flattenedRecs, recs...)
+	}
+
+	sort.Slice(flattenedRecs, func(i, j int) bool {
+		return flattenedRecs[i].globalCtr < flattenedRecs[j].globalCtr
+	})
+
+	wakeTimesByStep := map[int][]float64{}
+	for _, rec := range flattenedRecs {
+		wakeTimes, ok := wakeTimesByStep[rec.step]
+		if !ok {
+			wakeTimes = []float64{}
+		}
+		wakeTimes = append(wakeTimes, rec.wakeAt.Seconds())
+		wakeTimesByStep[rec.step] = wakeTimes
+	}
+
+	meansByStepId := make([]float64, stepsPerFS)
+	interQuartileRangesByStepIdx := make([]float64, stepsPerFS)
+	for step := 0; step < stepsPerFS; step++ {
+		t.Logf("step %d", step)
+		mean, _ := stats.Mean(wakeTimesByStep[step])
+		meansByStepId[step] = mean
+		t.Logf("\tmean: %v", mean)
+		median, _ := stats.Median(wakeTimesByStep[step])
+		t.Logf("\tmedian: %v", median)
+		midhinge, _ := stats.Midhinge(wakeTimesByStep[step])
+		t.Logf("\tmidhinge: %v", midhinge)
+		min, _ := stats.Min(wakeTimesByStep[step])
+		t.Logf("\tmin: %v", min)
+		max, _ := stats.Max(wakeTimesByStep[step])
+		t.Logf("\tmax: %v", max)
+		quartiles, _ := stats.Quartile(wakeTimesByStep[step])
+		t.Logf("\t%#v", quartiles)
+		interQuartileRange, _ := stats.InterQuartileRange(wakeTimesByStep[step])
+		t.Logf("\tinter-quartile range: %v", interQuartileRange)
+		interQuartileRangesByStepIdx[step] = interQuartileRange
+	}
+
+	iqrMean, _ := stats.Mean(interQuartileRangesByStepIdx)
+	t.Logf("inter-quartile-range mean: %v", iqrMean)
+	iqrDev, _ := stats.StandardDeviation(interQuartileRangesByStepIdx)
+	t.Logf("inter-quartile-range deviation: %v", iqrDev)
+
+	// each step should have the same "distribution" (=~ "spread")
+	assert.True(t, iqrDev < 0.01)
+
+	minTimeForAllStepsWithIdxI := sleepTimePerStep.Seconds() * float64(filesystems) / float64(concurrency)
+	t.Logf("minTimeForAllStepsWithIdxI = %11.8f", minTimeForAllStepsWithIdxI)
+	for i, mean := range meansByStepId {
+		// we can't just do (i + 0.5) * minTimeforAllStepsWithIdxI
+		// because this doesn't account for drift
+		idealMean := 0.5 * minTimeForAllStepsWithIdxI
+		if i > 0 {
+			previousMean := meansByStepId[i-1]
+			idealMean = previousMean + minTimeForAllStepsWithIdxI
+		}
+		deltaFromIdeal := idealMean - mean
+		t.Logf("step %02d delta from ideal mean wake time: %11.8f - %11.8f = %11.8f", i, idealMean, mean, deltaFromIdeal)
+		assert.True(t, math.Abs(deltaFromIdeal) < 0.05)
+	}
+
+}

replication/fsrep/fsfsm.go

@@ -1,557 +0,0 @@
-// Package fsrep implements replication of a single file system with existing versions
-// from a sender to a receiver.
-package fsrep
-
-import (
-	"context"
-	"errors"
-	"fmt"
-	"net"
-	"sync"
-	"time"
-
-	"github.com/prometheus/client_golang/prometheus"
-	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/replication/pdu"
-	"github.com/zrepl/zrepl/util/bytecounter"
-	"github.com/zrepl/zrepl/util/watchdog"
-	"github.com/zrepl/zrepl/zfs"
-)
-
-type contextKey int
-
-const (
-	contextKeyLogger contextKey = iota
-)
-
-type Logger = logger.Logger
-
-func WithLogger(ctx context.Context, log Logger) context.Context {
-	return context.WithValue(ctx, contextKeyLogger, log)
-}
-
-func getLogger(ctx context.Context) Logger {
-	l, ok := ctx.Value(contextKeyLogger).(Logger)
-	if !ok {
-		l = logger.NewNullLogger()
-	}
-	return l
-}
-
-// A Sender is usually part of a github.com/zrepl/zrepl/replication.Endpoint.
-type Sender interface {
-	// If a non-nil io.ReadCloser is returned, it is guaranteed to be closed before
-	// any next call to the parent github.com/zrepl/zrepl/replication.Endpoint.
-	// If the send request is for dry run the io.ReadCloser will be nil
-	Send(ctx context.Context, r *pdu.SendReq) (*pdu.SendRes, zfs.StreamCopier, error)
-	ReplicationCursor(ctx context.Context, req *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error)
-}
-
-// A Sender is usually part of a github.com/zrepl/zrepl/replication.Endpoint.
-type Receiver interface {
-	// Receive sends r and sendStream (the latter containing a ZFS send stream)
-	// to the parent github.com/zrepl/zrepl/replication.Endpoint.
-	Receive(ctx context.Context, req *pdu.ReceiveReq, receive zfs.StreamCopier) (*pdu.ReceiveRes, error)
-}
-
-type StepReport struct {
-	From, To string
-	Status   StepState
-	Problem  string
-	Bytes    int64
-	ExpectedBytes int64 // 0 means no size estimate possible
-}
-
-type Report struct {
-	Filesystem         string
-	Status             string
-	Problem            string
-	Completed, Pending []*StepReport
-}
-
-//go:generate enumer -type=State
-type State uint
-
-const (
-	Ready State = 1 << iota
-	Completed
-)
-
-type Error interface {
-	error
-	Temporary() bool
-	ContextErr() bool
-	LocalToFS() bool
-}
-
-type Replication struct {
-	promBytesReplicated prometheus.Counter
-
-	fs                 string
-
-	// lock protects all fields below it in this struct, but not the data behind pointers
-	lock               sync.Mutex
-	state              State
-	err                Error
-	completed, pending []*ReplicationStep
-}
-
-func (f *Replication) State() State {
-	f.lock.Lock()
-	defer f.lock.Unlock()
-	return f.state
-}
-
-func (f *Replication) FS() string { return f.fs }
-
-// returns zero value time.Time{} if no more pending steps
-func (f *Replication) NextStepDate() time.Time {
-	if len(f.pending) == 0 {
-		return time.Time{}
-	}
-	return f.pending[0].to.SnapshotTime()
-}
-
-func (f *Replication) Err() Error {
-	f.lock.Lock()
-	defer f.lock.Unlock()
-	return f.err
-}
-
-func (f *Replication) CanRetry() bool {
-	f.lock.Lock()
-	defer f.lock.Unlock()
-	if f.state == Completed {
-		return false
-	}
-	if f.state != Ready {
-		panic(fmt.Sprintf("implementation error: %v", f.state))
-	}
-	if f.err == nil {
-		return true
-	}
-	return f.err.Temporary()
-}
-
-func (f *Replication) UpdateSizeEsitmate(ctx context.Context, sender Sender) error {
-	f.lock.Lock()
-	defer f.lock.Unlock()
-	for _, e := range f.pending {
-		if err := e.updateSizeEstimate(ctx, sender); err != nil {
-			return err
-		}
-	}
-	return nil
-}
-
-type ReplicationBuilder struct {
-	r *Replication
-}
-
-func BuildReplication(fs string, promBytesReplicated prometheus.Counter) *ReplicationBuilder {
-	return &ReplicationBuilder{&Replication{fs: fs, promBytesReplicated: promBytesReplicated}}
-}
-
-func (b *ReplicationBuilder) AddStep(from, to FilesystemVersion) *ReplicationBuilder {
-	step := &ReplicationStep{
-		state:  StepReplicationReady,
-		parent: b.r,
-		from:   from,
-		to:     to,
-	}
-	b.r.pending = append(b.r.pending, step)
-	return b
-}
-
-func (b *ReplicationBuilder) Done() (r *Replication) {
-	if len(b.r.pending) > 0 {
-		b.r.state = Ready
-	} else {
-		b.r.state = Completed
-	}
-	r = b.r
-	b.r = nil
-	return r
-}
-
-type ReplicationConflictError struct {
-	Err error
-}
-
-func (e *ReplicationConflictError) Timeout() bool { return false }
-
-func (e *ReplicationConflictError) Temporary() bool { return false }
-
-func (e *ReplicationConflictError) Error() string { return fmt.Sprintf("permanent error: %s", e.Err.Error()) }
-
-func (e *ReplicationConflictError) LocalToFS() bool { return true }
-
-func (e *ReplicationConflictError) ContextErr() bool { return false }
-
-func NewReplicationConflictError(fs string, err error) *Replication {
-	return &Replication{
-		state: Completed,
-		fs:    fs,
-		err:   &ReplicationConflictError{Err: err},
-	}
-}
-
-//go:generate enumer -type=StepState
-type StepState uint
-
-const (
-	StepReplicationReady StepState = 1 << iota
-	StepMarkReplicatedReady
-	StepCompleted
-)
-
-func (s StepState) IsTerminal() bool { return s == StepCompleted }
-
-type FilesystemVersion interface {
-	SnapshotTime() time.Time
-	GetName() string // name without @ or #
-	RelName() string // name with @ or #
-}
-
-type ReplicationStep struct {
-	// only protects state, err
-	// from, to and parent are assumed to be immutable
-	lock sync.Mutex
-
-	state    StepState
-	from, to FilesystemVersion
-	parent   *Replication
-
-	// both retry and permanent error
-	err error
-
-	byteCounter  bytecounter.StreamCopier
-	expectedSize int64 // 0 means no size estimate present / possible
-}
-
-func (f *Replication) Retry(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver) Error {
-
-	var u updater = func(fu func(*Replication)) State {
-		f.lock.Lock()
-		defer f.lock.Unlock()
-		if fu != nil {
-			fu(f)
-		}
-		return f.state
-	}
-
-
-	var current *ReplicationStep
-	pre := u(nil)
-	getLogger(ctx).WithField("fsrep_state", pre).Debug("begin fsrep.Retry")
-	defer func() {
-		post := u(nil)
-		getLogger(ctx).WithField("fsrep_transition", post).Debug("end fsrep.Retry")
-	}()
-
-	st := u(func(f *Replication) {
-		if len(f.pending) == 0 {
-			f.state = Completed
-			return
-		}
-		current = f.pending[0]
-	})
-	if st == Completed {
-		return nil
-	}
-	if st != Ready {
-		panic(fmt.Sprintf("implementation error: %v", st))
-	}
-
-	stepCtx := WithLogger(ctx, getLogger(ctx).WithField("step", current))
-	getLogger(stepCtx).Debug("take step")
-	err := current.Retry(stepCtx, ka, sender, receiver)
-	if err != nil {
-		getLogger(stepCtx).WithError(err).Error("step could not be completed")
-	}
-
-	u(func(fsr *Replication) {
-		if err != nil {
-			f.err = &StepError{stepStr: current.String(), err: err}
-			return
-		}
-		if err == nil && current.state != StepCompleted {
-			panic(fmt.Sprintf("implementation error: %v", current.state))
-		}
-		f.err = nil
-		f.completed = append(f.completed, current)
-		f.pending = f.pending[1:]
-		if len(f.pending) > 0 {
-			f.state = Ready
-		} else {
-			f.state = Completed
-		}
-	})
-	var retErr Error = nil
-	u(func(fsr *Replication) {
-		retErr = fsr.err
-	})
-	return retErr
-}
-
-type updater func(func(fsr *Replication)) State
-
-type state func(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver, u updater) state
-
-type StepError struct {
-	stepStr string
-	err     error
-}
-
-var _ Error = &StepError{}
-
-func (e StepError) Error() string {
-	if e.LocalToFS() {
-		return fmt.Sprintf("step %s failed: %s", e.stepStr, e.err)
-	}
-	return e.err.Error()
-}
-
-func (e StepError) Timeout() bool {
-	if neterr, ok := e.err.(net.Error); ok {
-		return neterr.Timeout()
-	}
-	return false
-}
-
-func (e StepError) Temporary() bool {
-	if neterr, ok := e.err.(net.Error); ok {
-		return neterr.Temporary()
-	}
-	return false
-}
-
-func (e StepError) LocalToFS() bool {
-	if _, ok := e.err.(net.Error); ok {
-		return false
-	}
-	return true // conservative approximation: we'd like to check for specific errors returned over RPC here...
-}
-
-func (e StepError) ContextErr() bool {
-	switch e.err {
-	case context.Canceled:
-		return true
-	case context.DeadlineExceeded:
-		return true
-	}
-	return false
-}
-
-func (fsr *Replication) Report() *Report {
-	fsr.lock.Lock()
-	defer fsr.lock.Unlock()
-
-	rep := Report{
-		Filesystem: fsr.fs,
-		Status:     fsr.state.String(),
-	}
-
-	if fsr.err != nil && fsr.err.LocalToFS() {
-		rep.Problem = fsr.err.Error()
-	}
-
-	rep.Completed = make([]*StepReport, len(fsr.completed))
-	for i := range fsr.completed {
-		rep.Completed[i] = fsr.completed[i].Report()
-	}
-	rep.Pending = make([]*StepReport, len(fsr.pending))
-	for i := range fsr.pending {
-		rep.Pending[i] = fsr.pending[i].Report()
-	}
-
-	return &rep
-}
-
-func (s *ReplicationStep) Retry(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver) error {
-	switch s.state {
-	case StepReplicationReady:
-		return s.doReplication(ctx, ka, sender, receiver)
-	case StepMarkReplicatedReady:
-		return s.doMarkReplicated(ctx, ka, sender)
-	case StepCompleted:
-		return nil
-	}
-	panic(fmt.Sprintf("implementation error: %v", s.state))
-}
-
-func (s *ReplicationStep) Error() error {
-	if s.state & (StepReplicationReady|StepMarkReplicatedReady) != 0 {
-		return s.err
-	}
-	return nil
-}
-
-func (s *ReplicationStep) doReplication(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver) error {
-
-	if s.state != StepReplicationReady {
-		panic(fmt.Sprintf("implementation error: %v", s.state))
-	}
-
-	fs := s.parent.fs
-
-	log := getLogger(ctx)
-	sr := s.buildSendRequest(false)
-
-	log.Debug("initiate send request")
-	sres, sstreamCopier, err := sender.Send(ctx, sr)
-	if err != nil {
-		log.WithError(err).Error("send request failed")
-		return err
-	}
-	if sstreamCopier == nil {
-		err := errors.New("send request did not return a stream, broken endpoint implementation")
-		return err
-	}
-	defer sstreamCopier.Close()
-
-	// Install a byte counter to track progress + for status report
-	s.byteCounter = bytecounter.NewStreamCopier(sstreamCopier)
-	byteCounterStopProgress := make(chan struct{})
-	defer close(byteCounterStopProgress)
-	go func() {
-		var lastCount int64
-		t := time.NewTicker(1 * time.Second)
-		defer t.Stop()
-		for {
-			select {
-			case <-byteCounterStopProgress:
-				return
-			case <-t.C:
-				newCount := s.byteCounter.Count()
-				if lastCount != newCount {
-					ka.MadeProgress()
-				} else {
-					lastCount = newCount
-				}
-			}
-		}
-	}()
-	defer func() {
-		s.parent.promBytesReplicated.Add(float64(s.byteCounter.Count()))
-	}()
-
-	rr := &pdu.ReceiveReq{
-		Filesystem:       fs,
-		ClearResumeToken: !sres.UsedResumeToken,
-	}
-	log.Debug("initiate receive request")
-	_, err = receiver.Receive(ctx, rr, s.byteCounter)
-	if err != nil {
-		log.
-			WithError(err).
-			WithField("errType", fmt.Sprintf("%T", err)).
-			Error("receive request failed (might also be error on sender)")
-		// This failure could be due to
-		// 	- an unexpected exit of ZFS on the sending side
-		//  - an unexpected exit of ZFS on the receiving side
-		//  - a connectivity issue
-		return err
-	}
-	log.Debug("receive finished")
-	ka.MadeProgress()
-
-	s.state = StepMarkReplicatedReady
-	return s.doMarkReplicated(ctx, ka, sender)
-
-}
-
-func (s *ReplicationStep) doMarkReplicated(ctx context.Context, ka *watchdog.KeepAlive, sender Sender) error {
-
-	if s.state != StepMarkReplicatedReady {
-		panic(fmt.Sprintf("implementation error: %v", s.state))
-	}
-
-	log := getLogger(ctx)
-
-	log.Debug("advance replication cursor")
-	req := &pdu.ReplicationCursorReq{
-		Filesystem: s.parent.fs,
-		Op: &pdu.ReplicationCursorReq_Set{
-			Set: &pdu.ReplicationCursorReq_SetOp{
-				Snapshot:   s.to.GetName(),
-			},
-		},
-	}
-	_, err := sender.ReplicationCursor(ctx, req)
-	if err != nil {
-		log.WithError(err).Error("error advancing replication cursor")
-		return err
-	}
-	ka.MadeProgress()
-
-	s.state = StepCompleted
-	return err
-}
-
-func (s *ReplicationStep) updateSizeEstimate(ctx context.Context, sender Sender) error {
-
-	log := getLogger(ctx)
-
-	sr := s.buildSendRequest(true)
-
-	log.Debug("initiate dry run send request")
-	sres, _, err := sender.Send(ctx, sr)
-	if err != nil {
-		log.WithError(err).Error("dry run send request failed")
-		return err
-	}
-	s.expectedSize = sres.ExpectedSize
-	return nil
-}
-
-func (s *ReplicationStep) buildSendRequest(dryRun bool) (sr *pdu.SendReq) {
-	fs := s.parent.fs
-	if s.from == nil {
-		sr = &pdu.SendReq{
-			Filesystem: fs,
-			To:         s.to.RelName(),
-			DryRun:     dryRun,
-		}
-	} else {
-		sr = &pdu.SendReq{
-			Filesystem: fs,
-			From:       s.from.RelName(),
-			To:         s.to.RelName(),
-			DryRun:     dryRun,
-		}
-	}
-	return sr
-}
-
-func (s *ReplicationStep) String() string {
-	if s.from == nil { // FIXME: ZFS semantics are that to is nil on non-incremental send
-		return fmt.Sprintf("%s%s (full)", s.parent.fs, s.to.RelName())
-	} else {
-		return fmt.Sprintf("%s(%s => %s)", s.parent.fs, s.from.RelName(), s.to.RelName())
-	}
-}
-
-func (s *ReplicationStep) Report() *StepReport {
-	var from string // FIXME follow same convention as ZFS: to should be nil on full send
-	if s.from != nil {
-		from = s.from.RelName()
-	}
-	bytes := int64(0)
-	if s.byteCounter != nil {
-		bytes = s.byteCounter.Count()
-	}
-	problem := ""
-	if s.err != nil {
-		problem = s.err.Error()
-	}
-	rep := StepReport{
-		From:   from,
-		To:     s.to.RelName(),
-		Status: s.state,
-		Problem: problem,
-		Bytes:  bytes,
-		ExpectedBytes: s.expectedSize,
-	}
-	return &rep
-}

replication/fsrep/state_enumer.go

@@ -1,50 +0,0 @@
-// Code generated by "enumer -type=State"; DO NOT EDIT.
-
-package fsrep
-
-import (
-	"fmt"
-)
-
-const _StateName = "ReadyCompleted"
-
-var _StateIndex = [...]uint8{0, 5, 14}
-
-func (i State) String() string {
-	i -= 1
-	if i >= State(len(_StateIndex)-1) {
-		return fmt.Sprintf("State(%d)", i+1)
-	}
-	return _StateName[_StateIndex[i]:_StateIndex[i+1]]
-}
-
-var _StateValues = []State{1, 2}
-
-var _StateNameToValueMap = map[string]State{
-	_StateName[0:5]:  1,
-	_StateName[5:14]: 2,
-}
-
-// StateString retrieves an enum value from the enum constants string name.
-// Throws an error if the param is not part of the enum.
-func StateString(s string) (State, error) {
-	if val, ok := _StateNameToValueMap[s]; ok {
-		return val, nil
-	}
-	return 0, fmt.Errorf("%s does not belong to State values", s)
-}
-
-// StateValues returns all values of the enum
-func StateValues() []State {
-	return _StateValues
-}
-
-// IsAState returns "true" if the value is listed in the enum definition. "false" otherwise
-func (i State) IsAState() bool {
-	for _, v := range _StateValues {
-		if i == v {
-			return true
-		}
-	}
-	return false
-}

replication/fsrep/stepstate_enumer.go

@@ -1,61 +0,0 @@
-// Code generated by "enumer -type=StepState"; DO NOT EDIT.
-
-package fsrep
-
-import (
-	"fmt"
-)
-
-const (
-	_StepStateName_0 = "StepReplicationReadyStepMarkReplicatedReady"
-	_StepStateName_1 = "StepCompleted"
-)
-
-var (
-	_StepStateIndex_0 = [...]uint8{0, 20, 43}
-	_StepStateIndex_1 = [...]uint8{0, 13}
-)
-
-func (i StepState) String() string {
-	switch {
-	case 1 <= i && i <= 2:
-		i -= 1
-		return _StepStateName_0[_StepStateIndex_0[i]:_StepStateIndex_0[i+1]]
-	case i == 4:
-		return _StepStateName_1
-	default:
-		return fmt.Sprintf("StepState(%d)", i)
-	}
-}
-
-var _StepStateValues = []StepState{1, 2, 4}
-
-var _StepStateNameToValueMap = map[string]StepState{
-	_StepStateName_0[0:20]:  1,
-	_StepStateName_0[20:43]: 2,
-	_StepStateName_1[0:13]:  4,
-}
-
-// StepStateString retrieves an enum value from the enum constants string name.
-// Throws an error if the param is not part of the enum.
-func StepStateString(s string) (StepState, error) {
-	if val, ok := _StepStateNameToValueMap[s]; ok {
-		return val, nil
-	}
-	return 0, fmt.Errorf("%s does not belong to StepState values", s)
-}
-
-// StepStateValues returns all values of the enum
-func StepStateValues() []StepState {
-	return _StepStateValues
-}
-
-// IsAStepState returns "true" if the value is listed in the enum definition. "false" otherwise
-func (i StepState) IsAStepState() bool {
-	for _, v := range _StepStateValues {
-		if i == v {
-			return true
-		}
-	}
-	return false
-}

replication/logic/diff/diff.go

@@ -1,9 +1,11 @@
-package mainfsm
+package diff
 
 import (
+	"fmt"
 	"sort"
+	"strings"
 
-	. "github.com/zrepl/zrepl/replication/pdu"
+	. "github.com/zrepl/zrepl/replication/logic/pdu"
 )
 
 type ConflictNoCommonAncestor struct {
@@ -11,7 +13,19 @@ type ConflictNoCommonAncestor struct {
 }
 
 func (c *ConflictNoCommonAncestor) Error() string {
-	return "no common snapshot or suitable bookmark between sender and receiver"
+	var buf strings.Builder
+	buf.WriteString("no common snapshot or suitable bookmark between sender and receiver")
+	if len(c.SortedReceiverVersions) > 0 || len(c.SortedSenderVersions) > 0 {
+		buf.WriteString(":\n  sorted sender versions:\n")
+		for _, v := range c.SortedSenderVersions {
+			fmt.Fprintf(&buf, "    %s\n", v.RelName())
+		}
+		buf.WriteString("  sorted receiver versions:\n")
+		for _, v := range c.SortedReceiverVersions {
+			fmt.Fprintf(&buf, "    %s\n", v.RelName())
+		}
+	}
+	return buf.String()
 }
 
 type ConflictDiverged struct {
@@ -21,7 +35,18 @@ type ConflictDiverged struct {
 }
 
 func (c *ConflictDiverged) Error() string {
-	return "the receiver's latest snapshot is not present on sender"
+	var buf strings.Builder
+	buf.WriteString("the receiver's latest snapshot is not present on sender:\n")
+	fmt.Fprintf(&buf, "  last common: %s\n", c.CommonAncestor.RelName())
+	fmt.Fprintf(&buf, "  sender-only:\n")
+	for _, v := range c.SenderOnly {
+		fmt.Fprintf(&buf, "    %s\n", v.RelName())
+	}
+	fmt.Fprintf(&buf, "  receiver-only:\n")
+	for _, v := range c.ReceiverOnly {
+		fmt.Fprintf(&buf, "    %s\n", v.RelName())
+	}
+	return buf.String()
 }
 
 func SortVersionListByCreateTXGThenBookmarkLTSnapshot(fsvslice []*FilesystemVersion) []*FilesystemVersion {

replication/logic/diff/diff_test.go

@@ -0,0 +1,130 @@
+package diff
+
+import (
+	"strconv"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	. "github.com/zrepl/zrepl/replication/logic/pdu"
+)
+
+func fsvlist(fsv ...string) (r []*FilesystemVersion) {
+
+	r = make([]*FilesystemVersion, len(fsv))
+	for i, f := range fsv {
+
+		// parse the id from fsvlist. it is used to derivce Guid,CreateTXG and Creation attrs
+		split := strings.Split(f, ",")
+		if len(split) != 2 {
+			panic("invalid fsv spec")
+		}
+		id, err := strconv.Atoi(split[1])
+		if err != nil {
+			panic(err)
+		}
+		creation := func(id int) string {
+			return FilesystemVersionCreation(time.Unix(0, 0).Add(time.Duration(id) * time.Second))
+		}
+		if strings.HasPrefix(f, "#") {
+			r[i] = &FilesystemVersion{
+				Name:      strings.TrimPrefix(f, "#"),
+				Type:      FilesystemVersion_Bookmark,
+				Guid:      uint64(id),
+				CreateTXG: uint64(id),
+				Creation:  creation(id),
+			}
+		} else if strings.HasPrefix(f, "@") {
+			r[i] = &FilesystemVersion{
+				Name:      strings.TrimPrefix(f, "@"),
+				Type:      FilesystemVersion_Snapshot,
+				Guid:      uint64(id),
+				CreateTXG: uint64(id),
+				Creation:  creation(id),
+			}
+		} else {
+			panic("invalid character")
+		}
+	}
+	return
+}
+
+func doTest(receiver, sender []*FilesystemVersion, validate func(incpath []*FilesystemVersion, conflict error)) {
+	p, err := IncrementalPath(receiver, sender)
+	validate(p, err)
+}
+
+func TestIncrementalPath_SnapshotsOnly(t *testing.T) {
+
+	l := fsvlist
+
+	// basic functionality
+	doTest(l("@a,1", "@b,2"), l("@a,1", "@b,2", "@c,3", "@d,4"), func(path []*FilesystemVersion, conflict error) {
+		assert.Equal(t, l("@b,2", "@c,3", "@d,4"), path)
+	})
+
+	// no common ancestor
+	doTest(l(), l("@a,1"), func(path []*FilesystemVersion, conflict error) {
+		assert.Nil(t, path)
+		ca, ok := conflict.(*ConflictNoCommonAncestor)
+		require.True(t, ok)
+		assert.Equal(t, l("@a,1"), ca.SortedSenderVersions)
+	})
+	doTest(l("@a,1", "@b,2"), l("@c,3", "@d,4"), func(path []*FilesystemVersion, conflict error) {
+		assert.Nil(t, path)
+		ca, ok := conflict.(*ConflictNoCommonAncestor)
+		require.True(t, ok)
+		assert.Equal(t, l("@a,1", "@b,2"), ca.SortedReceiverVersions)
+		assert.Equal(t, l("@c,3", "@d,4"), ca.SortedSenderVersions)
+	})
+
+	// divergence is detected
+	doTest(l("@a,1", "@b1,2"), l("@a,1", "@b2,3"), func(path []*FilesystemVersion, conflict error) {
+		assert.Nil(t, path)
+		cd, ok := conflict.(*ConflictDiverged)
+		require.True(t, ok)
+		assert.Equal(t, l("@a,1")[0], cd.CommonAncestor)
+		assert.Equal(t, l("@b1,2"), cd.ReceiverOnly)
+		assert.Equal(t, l("@b2,3"), cd.SenderOnly)
+	})
+
+	// gaps before most recent common ancestor do not matter
+	doTest(l("@a,1", "@b,2", "@c,3"), l("@a,1", "@c,3", "@d,4"), func(path []*FilesystemVersion, conflict error) {
+		assert.Equal(t, l("@c,3", "@d,4"), path)
+	})
+
+	// sender with earlier but also current version as sender is not a conflict
+	doTest(l("@c,3"), l("@a,1", "@b,2", "@c,3") , func(path []*FilesystemVersion, conflict error) {
+		t.Logf("path: %#v", path)
+		t.Logf("conflict: %#v", conflict)
+		assert.Empty(t, path)
+		assert.Nil(t, conflict)
+	})
+
+}
+
+func TestIncrementalPath_BookmarkSupport(t *testing.T) {
+	l := fsvlist
+
+	// bookmarks are used
+	doTest(l("@a,1"), l("#a,1", "@b,2"), func(path []*FilesystemVersion, conflict error) {
+		assert.Equal(t, l("#a,1", "@b,2"), path)
+	})
+
+	// boomarks are stripped from IncrementalPath (cannot send incrementally)
+	doTest(l("@a,1"), l("#a,1", "#b,2", "@c,3"), func(path []*FilesystemVersion, conflict error) {
+		assert.Equal(t, l("#a,1", "@c,3"), path)
+	})
+
+	// test that snapshots are preferred over bookmarks in IncrementalPath
+	doTest(l("@a,1"), l("#a,1", "@a,1", "@b,2"), func(path []*FilesystemVersion, conflict error) {
+		assert.Equal(t, l("@a,1", "@b,2"), path)
+	})
+	doTest(l("@a,1"), l("@a,1", "#a,1", "@b,2"), func(path []*FilesystemVersion, conflict error) {
+		assert.Equal(t, l("@a,1", "@b,2"), path)
+	})
+
+}

replication/logic/pdu/pdu.pb.go

@@ -43,7 +43,7 @@ func (x FilesystemVersion_VersionType) String() string {
 	return proto.EnumName(FilesystemVersion_VersionType_name, int32(x))
 }
 func (FilesystemVersion_VersionType) EnumDescriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{5, 0}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{5, 0}
 }
 
 type ListFilesystemReq struct {
@@ -56,7 +56,7 @@ func (m *ListFilesystemReq) Reset()         { *m = ListFilesystemReq{} }
 func (m *ListFilesystemReq) String() string { return proto.CompactTextString(m) }
 func (*ListFilesystemReq) ProtoMessage()    {}
 func (*ListFilesystemReq) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{0}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{0}
 }
 func (m *ListFilesystemReq) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ListFilesystemReq.Unmarshal(m, b)
@@ -78,7 +78,6 @@ var xxx_messageInfo_ListFilesystemReq proto.InternalMessageInfo
 
 type ListFilesystemRes struct {
 	Filesystems          []*Filesystem `protobuf:"bytes,1,rep,name=Filesystems,proto3" json:"Filesystems,omitempty"`
-	Empty                bool          `protobuf:"varint,2,opt,name=Empty,proto3" json:"Empty,omitempty"`
 	XXX_NoUnkeyedLiteral struct{}      `json:"-"`
 	XXX_unrecognized     []byte        `json:"-"`
 	XXX_sizecache        int32         `json:"-"`
@@ -88,7 +87,7 @@ func (m *ListFilesystemRes) Reset()         { *m = ListFilesystemRes{} }
 func (m *ListFilesystemRes) String() string { return proto.CompactTextString(m) }
 func (*ListFilesystemRes) ProtoMessage()    {}
 func (*ListFilesystemRes) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{1}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{1}
 }
 func (m *ListFilesystemRes) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ListFilesystemRes.Unmarshal(m, b)
@@ -115,16 +114,10 @@ func (m *ListFilesystemRes) GetFilesystems() []*Filesystem {
 	return nil
 }
 
-func (m *ListFilesystemRes) GetEmpty() bool {
-	if m != nil {
-		return m.Empty
-	}
-	return false
-}
-
 type Filesystem struct {
 	Path                 string   `protobuf:"bytes,1,opt,name=Path,proto3" json:"Path,omitempty"`
 	ResumeToken          string   `protobuf:"bytes,2,opt,name=ResumeToken,proto3" json:"ResumeToken,omitempty"`
+	IsPlaceholder        bool     `protobuf:"varint,3,opt,name=IsPlaceholder,proto3" json:"IsPlaceholder,omitempty"`
 	XXX_NoUnkeyedLiteral struct{} `json:"-"`
 	XXX_unrecognized     []byte   `json:"-"`
 	XXX_sizecache        int32    `json:"-"`
@@ -134,7 +127,7 @@ func (m *Filesystem) Reset()         { *m = Filesystem{} }
 func (m *Filesystem) String() string { return proto.CompactTextString(m) }
 func (*Filesystem) ProtoMessage()    {}
 func (*Filesystem) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{2}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{2}
 }
 func (m *Filesystem) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_Filesystem.Unmarshal(m, b)
@@ -168,6 +161,13 @@ func (m *Filesystem) GetResumeToken() string {
 	return ""
 }
 
+func (m *Filesystem) GetIsPlaceholder() bool {
+	if m != nil {
+		return m.IsPlaceholder
+	}
+	return false
+}
+
 type ListFilesystemVersionsReq struct {
 	Filesystem           string   `protobuf:"bytes,1,opt,name=Filesystem,proto3" json:"Filesystem,omitempty"`
 	XXX_NoUnkeyedLiteral struct{} `json:"-"`
@@ -179,7 +179,7 @@ func (m *ListFilesystemVersionsReq) Reset()         { *m = ListFilesystemVersion
 func (m *ListFilesystemVersionsReq) String() string { return proto.CompactTextString(m) }
 func (*ListFilesystemVersionsReq) ProtoMessage()    {}
 func (*ListFilesystemVersionsReq) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{3}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{3}
 }
 func (m *ListFilesystemVersionsReq) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ListFilesystemVersionsReq.Unmarshal(m, b)
@@ -217,7 +217,7 @@ func (m *ListFilesystemVersionsRes) Reset()         { *m = ListFilesystemVersion
 func (m *ListFilesystemVersionsRes) String() string { return proto.CompactTextString(m) }
 func (*ListFilesystemVersionsRes) ProtoMessage()    {}
 func (*ListFilesystemVersionsRes) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{4}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{4}
 }
 func (m *ListFilesystemVersionsRes) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ListFilesystemVersionsRes.Unmarshal(m, b)
@@ -259,7 +259,7 @@ func (m *FilesystemVersion) Reset()         { *m = FilesystemVersion{} }
 func (m *FilesystemVersion) String() string { return proto.CompactTextString(m) }
 func (*FilesystemVersion) ProtoMessage()    {}
 func (*FilesystemVersion) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{5}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{5}
 }
 func (m *FilesystemVersion) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_FilesystemVersion.Unmarshal(m, b)
@@ -339,7 +339,7 @@ func (m *SendReq) Reset()         { *m = SendReq{} }
 func (m *SendReq) String() string { return proto.CompactTextString(m) }
 func (*SendReq) ProtoMessage()    {}
 func (*SendReq) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{6}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{6}
 }
 func (m *SendReq) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_SendReq.Unmarshal(m, b)
@@ -420,7 +420,7 @@ func (m *Property) Reset()         { *m = Property{} }
 func (m *Property) String() string { return proto.CompactTextString(m) }
 func (*Property) ProtoMessage()    {}
 func (*Property) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{7}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{7}
 }
 func (m *Property) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_Property.Unmarshal(m, b)
@@ -470,7 +470,7 @@ func (m *SendRes) Reset()         { *m = SendRes{} }
 func (m *SendRes) String() string { return proto.CompactTextString(m) }
 func (*SendRes) ProtoMessage()    {}
 func (*SendRes) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{8}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{8}
 }
 func (m *SendRes) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_SendRes.Unmarshal(m, b)
@@ -524,7 +524,7 @@ func (m *ReceiveReq) Reset()         { *m = ReceiveReq{} }
 func (m *ReceiveReq) String() string { return proto.CompactTextString(m) }
 func (*ReceiveReq) ProtoMessage()    {}
 func (*ReceiveReq) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{9}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{9}
 }
 func (m *ReceiveReq) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ReceiveReq.Unmarshal(m, b)
@@ -568,7 +568,7 @@ func (m *ReceiveRes) Reset()         { *m = ReceiveRes{} }
 func (m *ReceiveRes) String() string { return proto.CompactTextString(m) }
 func (*ReceiveRes) ProtoMessage()    {}
 func (*ReceiveRes) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{10}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{10}
 }
 func (m *ReceiveRes) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ReceiveRes.Unmarshal(m, b)
@@ -601,7 +601,7 @@ func (m *DestroySnapshotsReq) Reset()         { *m = DestroySnapshotsReq{} }
 func (m *DestroySnapshotsReq) String() string { return proto.CompactTextString(m) }
 func (*DestroySnapshotsReq) ProtoMessage()    {}
 func (*DestroySnapshotsReq) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{11}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{11}
 }
 func (m *DestroySnapshotsReq) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_DestroySnapshotsReq.Unmarshal(m, b)
@@ -647,7 +647,7 @@ func (m *DestroySnapshotRes) Reset()         { *m = DestroySnapshotRes{} }
 func (m *DestroySnapshotRes) String() string { return proto.CompactTextString(m) }
 func (*DestroySnapshotRes) ProtoMessage()    {}
 func (*DestroySnapshotRes) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{12}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{12}
 }
 func (m *DestroySnapshotRes) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_DestroySnapshotRes.Unmarshal(m, b)
@@ -692,7 +692,7 @@ func (m *DestroySnapshotsRes) Reset()         { *m = DestroySnapshotsRes{} }
 func (m *DestroySnapshotsRes) String() string { return proto.CompactTextString(m) }
 func (*DestroySnapshotsRes) ProtoMessage()    {}
 func (*DestroySnapshotsRes) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{13}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{13}
 }
 func (m *DestroySnapshotsRes) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_DestroySnapshotsRes.Unmarshal(m, b)
@@ -734,7 +734,7 @@ func (m *ReplicationCursorReq) Reset()         { *m = ReplicationCursorReq{} }
 func (m *ReplicationCursorReq) String() string { return proto.CompactTextString(m) }
 func (*ReplicationCursorReq) ProtoMessage()    {}
 func (*ReplicationCursorReq) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{14}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{14}
 }
 func (m *ReplicationCursorReq) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ReplicationCursorReq.Unmarshal(m, b)
@@ -882,7 +882,7 @@ func (m *ReplicationCursorReq_GetOp) Reset()         { *m = ReplicationCursorReq
 func (m *ReplicationCursorReq_GetOp) String() string { return proto.CompactTextString(m) }
 func (*ReplicationCursorReq_GetOp) ProtoMessage()    {}
 func (*ReplicationCursorReq_GetOp) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{14, 0}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{14, 0}
 }
 func (m *ReplicationCursorReq_GetOp) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ReplicationCursorReq_GetOp.Unmarshal(m, b)
@@ -913,7 +913,7 @@ func (m *ReplicationCursorReq_SetOp) Reset()         { *m = ReplicationCursorReq
 func (m *ReplicationCursorReq_SetOp) String() string { return proto.CompactTextString(m) }
 func (*ReplicationCursorReq_SetOp) ProtoMessage()    {}
 func (*ReplicationCursorReq_SetOp) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{14, 1}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{14, 1}
 }
 func (m *ReplicationCursorReq_SetOp) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ReplicationCursorReq_SetOp.Unmarshal(m, b)
@@ -954,7 +954,7 @@ func (m *ReplicationCursorRes) Reset()         { *m = ReplicationCursorRes{} }
 func (m *ReplicationCursorRes) String() string { return proto.CompactTextString(m) }
 func (*ReplicationCursorRes) ProtoMessage()    {}
 func (*ReplicationCursorRes) Descriptor() ([]byte, []int) {
-	return fileDescriptor_pdu_89315d819a6e0938, []int{15}
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{15}
 }
 func (m *ReplicationCursorRes) XXX_Unmarshal(b []byte) error {
 	return xxx_messageInfo_ReplicationCursorRes.Unmarshal(m, b)
@@ -1079,6 +1079,83 @@ func _ReplicationCursorRes_OneofSizer(msg proto.Message) (n int) {
 	return n
 }
 
+type PingReq struct {
+	Message              string   `protobuf:"bytes,1,opt,name=Message,proto3" json:"Message,omitempty"`
+	XXX_NoUnkeyedLiteral struct{} `json:"-"`
+	XXX_unrecognized     []byte   `json:"-"`
+	XXX_sizecache        int32    `json:"-"`
+}
+
+func (m *PingReq) Reset()         { *m = PingReq{} }
+func (m *PingReq) String() string { return proto.CompactTextString(m) }
+func (*PingReq) ProtoMessage()    {}
+func (*PingReq) Descriptor() ([]byte, []int) {
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{16}
+}
+func (m *PingReq) XXX_Unmarshal(b []byte) error {
+	return xxx_messageInfo_PingReq.Unmarshal(m, b)
+}
+func (m *PingReq) XXX_Marshal(b []byte, deterministic bool) ([]byte, error) {
+	return xxx_messageInfo_PingReq.Marshal(b, m, deterministic)
+}
+func (dst *PingReq) XXX_Merge(src proto.Message) {
+	xxx_messageInfo_PingReq.Merge(dst, src)
+}
+func (m *PingReq) XXX_Size() int {
+	return xxx_messageInfo_PingReq.Size(m)
+}
+func (m *PingReq) XXX_DiscardUnknown() {
+	xxx_messageInfo_PingReq.DiscardUnknown(m)
+}
+
+var xxx_messageInfo_PingReq proto.InternalMessageInfo
+
+func (m *PingReq) GetMessage() string {
+	if m != nil {
+		return m.Message
+	}
+	return ""
+}
+
+type PingRes struct {
+	// Echo must be PingReq.Message
+	Echo                 string   `protobuf:"bytes,1,opt,name=Echo,proto3" json:"Echo,omitempty"`
+	XXX_NoUnkeyedLiteral struct{} `json:"-"`
+	XXX_unrecognized     []byte   `json:"-"`
+	XXX_sizecache        int32    `json:"-"`
+}
+
+func (m *PingRes) Reset()         { *m = PingRes{} }
+func (m *PingRes) String() string { return proto.CompactTextString(m) }
+func (*PingRes) ProtoMessage()    {}
+func (*PingRes) Descriptor() ([]byte, []int) {
+	return fileDescriptor_pdu_83b7e2a28d820622, []int{17}
+}
+func (m *PingRes) XXX_Unmarshal(b []byte) error {
+	return xxx_messageInfo_PingRes.Unmarshal(m, b)
+}
+func (m *PingRes) XXX_Marshal(b []byte, deterministic bool) ([]byte, error) {
+	return xxx_messageInfo_PingRes.Marshal(b, m, deterministic)
+}
+func (dst *PingRes) XXX_Merge(src proto.Message) {
+	xxx_messageInfo_PingRes.Merge(dst, src)
+}
+func (m *PingRes) XXX_Size() int {
+	return xxx_messageInfo_PingRes.Size(m)
+}
+func (m *PingRes) XXX_DiscardUnknown() {
+	xxx_messageInfo_PingRes.DiscardUnknown(m)
+}
+
+var xxx_messageInfo_PingRes proto.InternalMessageInfo
+
+func (m *PingRes) GetEcho() string {
+	if m != nil {
+		return m.Echo
+	}
+	return ""
+}
+
 func init() {
 	proto.RegisterType((*ListFilesystemReq)(nil), "ListFilesystemReq")
 	proto.RegisterType((*ListFilesystemRes)(nil), "ListFilesystemRes")
@@ -1098,6 +1175,8 @@ func init() {
 	proto.RegisterType((*ReplicationCursorReq_GetOp)(nil), "ReplicationCursorReq.GetOp")
 	proto.RegisterType((*ReplicationCursorReq_SetOp)(nil), "ReplicationCursorReq.SetOp")
 	proto.RegisterType((*ReplicationCursorRes)(nil), "ReplicationCursorRes")
+	proto.RegisterType((*PingReq)(nil), "PingReq")
+	proto.RegisterType((*PingRes)(nil), "PingRes")
 	proto.RegisterEnum("FilesystemVersion_VersionType", FilesystemVersion_VersionType_name, FilesystemVersion_VersionType_value)
 }
 
@@ -1113,6 +1192,7 @@ const _ = grpc.SupportPackageIsVersion4
 //
 // For semantics around ctx use and closing/ending streaming RPCs, please refer to https://godoc.org/google.golang.org/grpc#ClientConn.NewStream.
 type ReplicationClient interface {
+	Ping(ctx context.Context, in *PingReq, opts ...grpc.CallOption) (*PingRes, error)
 	ListFilesystems(ctx context.Context, in *ListFilesystemReq, opts ...grpc.CallOption) (*ListFilesystemRes, error)
 	ListFilesystemVersions(ctx context.Context, in *ListFilesystemVersionsReq, opts ...grpc.CallOption) (*ListFilesystemVersionsRes, error)
 	DestroySnapshots(ctx context.Context, in *DestroySnapshotsReq, opts ...grpc.CallOption) (*DestroySnapshotsRes, error)
@@ -1127,6 +1207,15 @@ func NewReplicationClient(cc *grpc.ClientConn) ReplicationClient {
 	return &replicationClient{cc}
 }
 
+func (c *replicationClient) Ping(ctx context.Context, in *PingReq, opts ...grpc.CallOption) (*PingRes, error) {
+	out := new(PingRes)
+	err := c.cc.Invoke(ctx, "/Replication/Ping", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
 func (c *replicationClient) ListFilesystems(ctx context.Context, in *ListFilesystemReq, opts ...grpc.CallOption) (*ListFilesystemRes, error) {
 	out := new(ListFilesystemRes)
 	err := c.cc.Invoke(ctx, "/Replication/ListFilesystems", in, out, opts...)
@@ -1165,6 +1254,7 @@ func (c *replicationClient) ReplicationCursor(ctx context.Context, in *Replicati
 
 // ReplicationServer is the server API for Replication service.
 type ReplicationServer interface {
+	Ping(context.Context, *PingReq) (*PingRes, error)
 	ListFilesystems(context.Context, *ListFilesystemReq) (*ListFilesystemRes, error)
 	ListFilesystemVersions(context.Context, *ListFilesystemVersionsReq) (*ListFilesystemVersionsRes, error)
 	DestroySnapshots(context.Context, *DestroySnapshotsReq) (*DestroySnapshotsRes, error)
@@ -1175,6 +1265,24 @@ func RegisterReplicationServer(s *grpc.Server, srv ReplicationServer) {
 	s.RegisterService(&_Replication_serviceDesc, srv)
 }
 
+func _Replication_Ping_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(PingReq)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(ReplicationServer).Ping(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/Replication/Ping",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(ReplicationServer).Ping(ctx, req.(*PingReq))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
 func _Replication_ListFilesystems_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
 	in := new(ListFilesystemReq)
 	if err := dec(in); err != nil {
@@ -1251,6 +1359,10 @@ var _Replication_serviceDesc = grpc.ServiceDesc{
 	ServiceName: "Replication",
 	HandlerType: (*ReplicationServer)(nil),
 	Methods: []grpc.MethodDesc{
+		{
+			MethodName: "Ping",
+			Handler:    _Replication_Ping_Handler,
+		},
 		{
 			MethodName: "ListFilesystems",
 			Handler:    _Replication_ListFilesystems_Handler,
@@ -1272,54 +1384,58 @@ var _Replication_serviceDesc = grpc.ServiceDesc{
 	Metadata: "pdu.proto",
 }
 
-func init() { proto.RegisterFile("pdu.proto", fileDescriptor_pdu_89315d819a6e0938) }
+func init() { proto.RegisterFile("pdu.proto", fileDescriptor_pdu_83b7e2a28d820622) }
 
-var fileDescriptor_pdu_89315d819a6e0938 = []byte{
+var fileDescriptor_pdu_83b7e2a28d820622 = []byte{
-	// 735 bytes of a gzipped FileDescriptorProto
+	// 785 bytes of a gzipped FileDescriptorProto
-	0x1f, 0x8b, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0xff, 0x8c, 0x55, 0xdd, 0x6e, 0xda, 0x4a,
+	0x1f, 0x8b, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0xff, 0x8c, 0x55, 0xd1, 0x8e, 0xe3, 0x34,
-	0x10, 0xc6, 0x60, 0xc0, 0x0c, 0x51, 0x42, 0x36, 0x9c, 0xc8, 0xc7, 0xe7, 0x28, 0x42, 0xdb, 0x1b,
+	0x14, 0x9d, 0xb4, 0x69, 0x9b, 0xde, 0x0e, 0xbb, 0x1d, 0x4f, 0x59, 0x85, 0x00, 0xab, 0xca, 0xcb,
-	0x52, 0xa9, 0x6e, 0x45, 0x7b, 0x53, 0x55, 0xaa, 0x54, 0x42, 0x7e, 0xa4, 0x56, 0x69, 0xb4, 0xd0,
+	0x43, 0x17, 0x89, 0x80, 0x0a, 0x2f, 0x08, 0x09, 0x89, 0x4e, 0x67, 0x67, 0x10, 0xb0, 0x54, 0x6e,
-	0x28, 0xca, 0x1d, 0x0d, 0xa3, 0xc4, 0x0a, 0xb0, 0xce, 0xee, 0xba, 0x0a, 0xbd, 0xec, 0x7b, 0xf4,
+	0x59, 0xad, 0xf6, 0x2d, 0x34, 0x57, 0x6d, 0x34, 0x6d, 0x9d, 0xb5, 0x13, 0xb4, 0xe5, 0x91, 0xbf,
-	0x41, 0xfa, 0x0e, 0xbd, 0xec, 0x03, 0x55, 0xbb, 0x60, 0xe3, 0x60, 0x23, 0x71, 0xe5, 0xfd, 0xbe,
+	0x9a, 0x7f, 0xe0, 0x91, 0x0f, 0x42, 0x76, 0xe3, 0x34, 0x6d, 0x52, 0xa9, 0x4f, 0xf1, 0x39, 0xf7,
-	0x9d, 0x9d, 0x9d, 0xf9, 0x76, 0x66, 0x0c, 0xb5, 0x70, 0x14, 0xf9, 0xa1, 0xe0, 0x8a, 0xd3, 0x3d,
+	0xda, 0x3e, 0xf7, 0xd8, 0xd7, 0x81, 0x76, 0x1c, 0xa6, 0x7e, 0x2c, 0x78, 0xc2, 0xe9, 0x35, 0x5c,
-	0xd8, 0xfd, 0x14, 0x48, 0x75, 0x12, 0x8c, 0x51, 0xce, 0xa4, 0xc2, 0x09, 0xc3, 0x07, 0x7a, 0x95,
+	0xfd, 0x1a, 0xc9, 0xe4, 0x55, 0xb4, 0x42, 0xb9, 0x95, 0x09, 0xae, 0x19, 0xbe, 0xa7, 0xa3, 0x32,
-	0x25, 0x25, 0x79, 0x01, 0xf5, 0x25, 0x21, 0x5d, 0xab, 0x55, 0x6a, 0xd7, 0x3b, 0x75, 0x3f, 0x65,
+	0x29, 0xc9, 0x57, 0xd0, 0xd9, 0x13, 0xd2, 0xb5, 0xfa, 0xf5, 0x41, 0x67, 0xd8, 0xf1, 0x0b, 0x49,
-	0x94, 0xde, 0x27, 0x4d, 0x28, 0x1f, 0x4f, 0x42, 0x35, 0x73, 0x8b, 0x2d, 0xab, 0xed, 0xb0, 0x39,
+	0xc5, 0x38, 0x5d, 0x02, 0xec, 0x21, 0x21, 0x60, 0x4f, 0x82, 0x64, 0xe9, 0x5a, 0x7d, 0x6b, 0xd0,
-	0xa0, 0x5d, 0x80, 0xa5, 0x11, 0x21, 0x60, 0x5f, 0x0c, 0xd5, 0x9d, 0x6b, 0xb5, 0xac, 0x76, 0x8d,
+	0x66, 0x7a, 0x4c, 0xfa, 0xd0, 0x61, 0x28, 0xd3, 0x35, 0xce, 0xf8, 0x03, 0x6e, 0xdc, 0x9a, 0x0e,
-	0x99, 0x35, 0x69, 0x41, 0x9d, 0xa1, 0x8c, 0x26, 0x38, 0xe0, 0xf7, 0x38, 0x35, 0xa7, 0x6b, 0x2c,
+	0x15, 0x29, 0xf2, 0x05, 0x7c, 0xf4, 0xb3, 0x9c, 0xac, 0x82, 0x39, 0x2e, 0xf9, 0x2a, 0x44, 0xe1,
-	0x4d, 0xd1, 0x77, 0xf0, 0xef, 0xd3, 0xe8, 0x2e, 0x51, 0xc8, 0x80, 0x4f, 0x25, 0xc3, 0x07, 0x72,
+	0xd6, 0xfb, 0xd6, 0xc0, 0x61, 0x87, 0x24, 0xfd, 0x01, 0x3e, 0x39, 0x54, 0xfb, 0x06, 0x85, 0x8c,
-	0x90, 0xbe, 0x60, 0xe1, 0x38, 0xc5, 0xd0, 0x8f, 0xeb, 0x0f, 0x4b, 0xe2, 0x83, 0x13, 0xc3, 0x45,
+	0xf8, 0x46, 0x32, 0x7c, 0x4f, 0x9e, 0x17, 0x65, 0x64, 0xdb, 0x17, 0x18, 0xfa, 0xcb, 0xe9, 0xc9,
-	0x7e, 0xc4, 0xcf, 0x58, 0xb2, 0xc4, 0x86, 0xfe, 0xb1, 0x60, 0x37, 0xb3, 0x4f, 0x3a, 0x60, 0x0f,
+	0x92, 0xf8, 0xe0, 0x18, 0x98, 0xd5, 0x4b, 0xfc, 0x52, 0x26, 0xcb, 0x73, 0xe8, 0x7f, 0x16, 0x5c,
-	0x66, 0x21, 0x9a, 0xcb, 0xb7, 0x3b, 0x07, 0x59, 0x0f, 0xfe, 0xe2, 0xab, 0xad, 0x98, 0xb1, 0xd5,
+	0x95, 0xe2, 0x64, 0x08, 0xf6, 0x6c, 0x1b, 0xa3, 0xde, 0xfc, 0xc9, 0xf0, 0x79, 0x79, 0x05, 0x3f,
-	0x4a, 0x9c, 0x0f, 0x27, 0xb8, 0x48, 0xd7, 0xac, 0x35, 0x77, 0x1a, 0x05, 0x23, 0xb7, 0xd4, 0xb2,
+	0xfb, 0xaa, 0x2c, 0xa6, 0x73, 0x95, 0x5f, 0xaf, 0x83, 0x35, 0x66, 0xa6, 0xe8, 0xb1, 0xe2, 0xee,
-	0xda, 0x36, 0x33, 0x6b, 0xf2, 0x3f, 0xd4, 0x8e, 0x04, 0x0e, 0x15, 0x0e, 0xae, 0x4e, 0x5d, 0xdb,
+	0xd2, 0x28, 0xd4, 0x26, 0xd8, 0x4c, 0x8f, 0xc9, 0x67, 0xd0, 0xbe, 0x11, 0x18, 0x24, 0x38, 0x7b,
-	0x6c, 0x2c, 0x09, 0xe2, 0x81, 0x63, 0x40, 0xc0, 0xa7, 0x6e, 0xd9, 0x78, 0x4a, 0x30, 0x3d, 0x84,
+	0x7b, 0xe7, 0xda, 0x3a, 0xb0, 0x27, 0x88, 0x07, 0x8e, 0x06, 0x11, 0xdf, 0xb8, 0x0d, 0xbd, 0x52,
-	0x7a, 0xea, 0x5a, 0xb2, 0x05, 0x4e, 0x7f, 0x3a, 0x0c, 0xe5, 0x1d, 0x57, 0x8d, 0x82, 0x46, 0x5d,
+	0x8e, 0xe9, 0x4b, 0xe8, 0x14, 0xb6, 0x25, 0x97, 0xe0, 0x4c, 0x37, 0x41, 0x2c, 0x97, 0x3c, 0xe9,
-	0xce, 0xef, 0x27, 0x43, 0x71, 0xdf, 0xb0, 0xe8, 0x2f, 0x0b, 0xaa, 0x7d, 0x9c, 0x8e, 0x36, 0xd0,
+	0x5e, 0x28, 0x34, 0xe2, 0xfc, 0x61, 0x1d, 0x88, 0x87, 0xae, 0x45, 0x1f, 0x2d, 0x68, 0x4d, 0x71,
-	0x53, 0x07, 0x79, 0x22, 0xf8, 0x24, 0x0e, 0x5c, 0xaf, 0xc9, 0x36, 0x14, 0x07, 0xdc, 0x84, 0x5d,
+	0x13, 0x9e, 0xe1, 0xa7, 0x12, 0xf9, 0x4a, 0xf0, 0xb5, 0x11, 0xae, 0xc6, 0xe4, 0x09, 0xd4, 0x66,
-	0x63, 0xc5, 0x01, 0x5f, 0x7d, 0x52, 0x3b, 0xf3, 0xa4, 0x26, 0x70, 0x3e, 0x09, 0x05, 0x4a, 0x69,
+	0x5c, 0xcb, 0x6e, 0xb3, 0xda, 0x8c, 0x1f, 0x1f, 0xbc, 0x5d, 0x3e, 0x78, 0x25, 0x9c, 0xaf, 0x63,
-	0x02, 0x77, 0x58, 0x82, 0x75, 0x21, 0xf5, 0x70, 0x14, 0x85, 0x6e, 0x65, 0x5e, 0x48, 0x06, 0x90,
+	0x81, 0x52, 0x6a, 0xe1, 0x0e, 0xcb, 0x31, 0xe9, 0x41, 0x63, 0x8c, 0x61, 0x1a, 0xbb, 0x4d, 0x1d,
-	0x7d, 0xa8, 0xf4, 0xc4, 0x8c, 0x45, 0x53, 0xb7, 0x6a, 0xe8, 0x05, 0xa2, 0x6f, 0xc0, 0xb9, 0x10,
+	0xd8, 0x01, 0xf2, 0x0c, 0x9a, 0x63, 0xb1, 0x65, 0xe9, 0xc6, 0x6d, 0x69, 0x3a, 0x43, 0xf4, 0x3b,
-	0x3c, 0x44, 0xa1, 0x66, 0x89, 0xa8, 0x56, 0x4a, 0xd4, 0x26, 0x94, 0x2f, 0x87, 0xe3, 0x28, 0x56,
+	0x70, 0x26, 0x82, 0xc7, 0x28, 0x92, 0x6d, 0x6e, 0xaa, 0x55, 0x30, 0xb5, 0x07, 0x8d, 0x37, 0xc1,
-	0x7a, 0x0e, 0xe8, 0x8f, 0x24, 0x63, 0x49, 0xda, 0xb0, 0xf3, 0x45, 0xe2, 0x68, 0xb5, 0x08, 0x1d,
+	0x2a, 0x35, 0x4e, 0xef, 0x00, 0xfd, 0x27, 0xaf, 0x58, 0x92, 0x01, 0x3c, 0xfd, 0x43, 0x62, 0x78,
-	0xb6, 0x4a, 0x13, 0x0a, 0x5b, 0xc7, 0x8f, 0x21, 0xde, 0x28, 0x1c, 0xf5, 0x83, 0xef, 0x68, 0x32,
+	0x7c, 0x55, 0x1d, 0x76, 0x4c, 0x13, 0x0a, 0x97, 0xb7, 0x1f, 0x62, 0x9c, 0x27, 0x18, 0x4e, 0xa3,
-	0x2e, 0xb1, 0x27, 0x1c, 0x39, 0x04, 0x58, 0xc4, 0x13, 0xa0, 0x74, 0x6d, 0x53, 0x54, 0x35, 0x3f,
+	0xbf, 0x51, 0x57, 0x5c, 0x67, 0x07, 0x1c, 0x79, 0x09, 0x90, 0xe9, 0x89, 0x50, 0xba, 0xb6, 0xbe,
-	0x0e, 0x91, 0xa5, 0x36, 0xe9, 0x15, 0x00, 0xc3, 0x1b, 0x0c, 0xbe, 0xe1, 0x26, 0xc2, 0x3f, 0x87,
+	0x54, 0x6d, 0xdf, 0x48, 0x64, 0x85, 0x20, 0x7d, 0x0b, 0xc0, 0x70, 0x8e, 0xd1, 0x5f, 0x78, 0x8e,
-	0xc6, 0xd1, 0x18, 0x87, 0x22, 0x1b, 0x67, 0x86, 0xa7, 0x5b, 0x29, 0xcf, 0x92, 0xde, 0xc2, 0x5e,
+	0xf1, 0x5f, 0x42, 0xf7, 0x66, 0x85, 0x81, 0x28, 0xeb, 0x2c, 0xf1, 0xf4, 0xb2, 0xb0, 0xb2, 0xa4,
-	0x0f, 0xa5, 0x12, 0x7c, 0x16, 0x57, 0xc0, 0x26, 0x9d, 0x43, 0x5e, 0x41, 0x2d, 0xb1, 0x77, 0x8b,
+	0x0b, 0xb8, 0x1e, 0xa3, 0x4c, 0x04, 0xdf, 0x9a, 0x1b, 0x70, 0x4e, 0xe7, 0x90, 0x6f, 0xa0, 0x9d,
-	0x6b, 0xbb, 0x63, 0x69, 0x44, 0xaf, 0x81, 0xac, 0x5c, 0xb4, 0x68, 0xb2, 0x18, 0x9a, 0x5b, 0xd6,
+	0xe7, 0xbb, 0xb5, 0x93, 0xdd, 0xb1, 0x4f, 0xa2, 0xef, 0x80, 0x1c, 0x6d, 0x94, 0x35, 0x99, 0x81,
-	0x34, 0x59, 0x6c, 0x63, 0x06, 0x89, 0x10, 0x5c, 0xc4, 0x2f, 0x66, 0x00, 0xed, 0xe5, 0x25, 0xa1,
+	0x7a, 0x97, 0x13, 0x4d, 0x66, 0x72, 0xd4, 0x89, 0xdd, 0x0a, 0xc1, 0x85, 0x39, 0x31, 0x0d, 0xe8,
-	0x87, 0x54, 0x55, 0x27, 0x3e, 0x56, 0x71, 0x03, 0xef, 0xf9, 0xd9, 0x10, 0x58, 0x6c, 0x43, 0x7f,
+	0xb8, 0xaa, 0x08, 0xf5, 0x68, 0xb5, 0x54, 0xe1, 0xab, 0xc4, 0x34, 0xf0, 0xb5, 0x5f, 0x96, 0xc0,
-	0x5b, 0xd0, 0x64, 0x18, 0x8e, 0x83, 0x1b, 0xd3, 0x24, 0x47, 0x91, 0x90, 0x5c, 0x6c, 0x22, 0xc6,
+	0x4c, 0x0e, 0xfd, 0xd7, 0x82, 0x1e, 0xc3, 0x78, 0x15, 0xcd, 0x75, 0x93, 0xdc, 0xa4, 0x42, 0x72,
-	0x4b, 0x28, 0xdd, 0xa2, 0x32, 0x21, 0xd5, 0x3b, 0xff, 0xf9, 0x79, 0x3e, 0xfc, 0x53, 0x54, 0x9f,
+	0x71, 0x8e, 0x19, 0x5f, 0x43, 0x7d, 0x81, 0x89, 0x96, 0xd4, 0x19, 0x7e, 0xea, 0x57, 0xad, 0xe1,
-	0xc3, 0xb3, 0x02, 0xd3, 0x96, 0xfa, 0x80, 0x44, 0x65, 0x4a, 0x64, 0xed, 0x81, 0x7e, 0x7c, 0x40,
+	0xdf, 0x61, 0xf2, 0x7b, 0x7c, 0x7f, 0xc1, 0x54, 0xa6, 0x9a, 0x20, 0x31, 0xd1, 0x57, 0xe4, 0xe4,
-	0xa2, 0xf2, 0xaa, 0x50, 0x36, 0x0e, 0xbc, 0x67, 0x50, 0x36, 0x1b, 0xba, 0x49, 0x12, 0xe1, 0xe6,
+	0x84, 0xa9, 0x99, 0x20, 0x31, 0xf1, 0x5a, 0xd0, 0xd0, 0x0b, 0x78, 0x2f, 0xa0, 0xa1, 0x03, 0xaa,
-	0x5a, 0x24, 0xb8, 0x6b, 0x43, 0x91, 0x87, 0x74, 0x90, 0x9b, 0x8d, 0x6e, 0xa1, 0xf9, 0x24, 0xd1,
+	0x49, 0x72, 0xe3, 0x76, 0x5e, 0xe4, 0x78, 0x64, 0x43, 0x8d, 0xc7, 0x74, 0x56, 0x59, 0x8d, 0x6a,
-	0x79, 0xd8, 0x67, 0x85, 0x64, 0x96, 0x38, 0xe7, 0x5c, 0xe1, 0x63, 0x20, 0xe7, 0xfe, 0x9c, 0xb3,
+	0xa1, 0xdd, 0x4b, 0xa2, 0xea, 0xb0, 0xef, 0x2f, 0xf2, 0xb7, 0xc4, 0x79, 0xcd, 0x13, 0xfc, 0x10,
-	0x02, 0x4b, 0x98, 0xae, 0x03, 0x95, 0xb9, 0x4a, 0x9d, 0x9f, 0x45, 0xdd, 0xbf, 0x89, 0x5b, 0xf2,
+	0xc9, 0xdd, 0x7a, 0xce, 0xfd, 0x05, 0xcb, 0x99, 0x91, 0x03, 0xcd, 0x9d, 0x4b, 0xf4, 0x05, 0xb4,
-	0x16, 0x76, 0x9e, 0x8e, 0x50, 0x49, 0x88, 0x9f, 0xf9, 0x89, 0x78, 0x59, 0x4e, 0x92, 0x0b, 0xd8,
+	0x26, 0xd1, 0x66, 0xa1, 0x6c, 0x71, 0xa1, 0xf5, 0x1b, 0x4a, 0x19, 0x2c, 0x4c, 0x53, 0x19, 0x48,
-	0xcf, 0x9f, 0xbe, 0xc4, 0xf3, 0xd7, 0xce, 0x74, 0x6f, 0xfd, 0x9e, 0x24, 0xef, 0xa1, 0xb1, 0x5a,
+	0x3f, 0x37, 0x49, 0x52, 0xb5, 0xdd, 0xed, 0x7c, 0xc9, 0x4d, 0xdb, 0xa9, 0xf1, 0xf0, 0xb1, 0xa6,
-	0x07, 0xa4, 0xe9, 0xe7, 0xd4, 0xb7, 0x97, 0xc7, 0x4a, 0xf2, 0x01, 0x76, 0x33, 0x92, 0x91, 0x7f,
+	0xde, 0x80, 0x5c, 0x1a, 0xf1, 0xc0, 0x56, 0xe9, 0xc4, 0xf1, 0xb3, 0xa5, 0x3d, 0x33, 0x92, 0xe4,
-	0x72, 0xdf, 0xc7, 0xcb, 0xa5, 0x65, 0xb7, 0x7c, 0x5d, 0x0a, 0x47, 0xd1, 0xd7, 0x8a, 0xf9, 0xa1,
+	0x7b, 0x78, 0x7a, 0xf8, 0x44, 0x4b, 0x42, 0xfc, 0xd2, 0x4f, 0xcb, 0x2b, 0x73, 0x92, 0x4c, 0xe0,
-	0xbe, 0xfe, 0x1b, 0x00, 0x00, 0xff, 0xff, 0xa3, 0xba, 0x8e, 0x63, 0x5d, 0x07, 0x00, 0x00,
+	0x59, 0xf5, 0xeb, 0x4e, 0x3c, 0xff, 0xe4, 0x3f, 0xc3, 0x3b, 0x1d, 0x93, 0xe4, 0x47, 0xe8, 0x1e,
+	0xdf, 0x33, 0xd2, 0xf3, 0x2b, 0xfa, 0xc7, 0xab, 0x62, 0x25, 0xf9, 0x09, 0xae, 0x4a, 0x47, 0x42,
+	0x3e, 0xae, 0x3c, 0x7f, 0xaf, 0x92, 0x96, 0xa3, 0xc6, 0xbb, 0x7a, 0x1c, 0xa6, 0x7f, 0x36, 0xf5,
+	0x0f, 0xfc, 0xdb, 0xff, 0x03, 0x00, 0x00, 0xff, 0xff, 0x37, 0x0e, 0xf2, 0xe4, 0xcd, 0x07, 0x00,
+	0x00,
 }

replication/logic/pdu/pdu.proto

@@ -2,6 +2,7 @@ syntax = "proto3";
 option go_package = "pdu";
 
 service Replication {
+    rpc Ping (PingReq) returns (PingRes);
     rpc ListFilesystems (ListFilesystemReq) returns (ListFilesystemRes);
     rpc ListFilesystemVersions (ListFilesystemVersionsReq) returns (ListFilesystemVersionsRes);
     rpc DestroySnapshots (DestroySnapshotsReq) returns (DestroySnapshotsRes);
@@ -13,12 +14,12 @@ message ListFilesystemReq {}
 
 message ListFilesystemRes {
     repeated Filesystem Filesystems = 1;
-    bool Empty = 2;
 }
 
 message Filesystem {
     string Path = 1;
     string ResumeToken = 2;
+    bool IsPlaceholder = 3;
 }
 
 message ListFilesystemVersionsReq {
@@ -120,3 +121,12 @@ message ReplicationCursorRes {
         bool Notexist = 2;
     }
 }
+
+message PingReq {
+    string Message = 1;
+}
+
+message PingRes {
+    // Echo must be PingReq.Message
+    string Echo = 1;
+}

replication/logic/replication_logic.go

@@ -0,0 +1,495 @@
+package logic
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"sync"
+	"time"
+
+	"github.com/prometheus/client_golang/prometheus"
+
+	"github.com/zrepl/zrepl/replication/driver"
+	. "github.com/zrepl/zrepl/replication/logic/diff"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
+	"github.com/zrepl/zrepl/replication/report"
+	"github.com/zrepl/zrepl/util/bytecounter"
+	"github.com/zrepl/zrepl/zfs"
+)
+
+// Endpoint represents one side of the replication.
+//
+// An endpoint is either in Sender or Receiver mode, represented by the correspondingly
+// named interfaces defined in this package.
+type Endpoint interface {
+	// Does not include placeholder filesystems
+	ListFilesystems(ctx context.Context, req *pdu.ListFilesystemReq) (*pdu.ListFilesystemRes, error)
+	ListFilesystemVersions(ctx context.Context, req *pdu.ListFilesystemVersionsReq) (*pdu.ListFilesystemVersionsRes, error)
+	DestroySnapshots(ctx context.Context, req *pdu.DestroySnapshotsReq) (*pdu.DestroySnapshotsRes, error)
+	WaitForConnectivity(ctx context.Context) (error)
+}
+
+type Sender interface {
+	Endpoint
+	// If a non-nil io.ReadCloser is returned, it is guaranteed to be closed before
+	// any next call to the parent github.com/zrepl/zrepl/replication.Endpoint.
+	// If the send request is for dry run the io.ReadCloser will be nil
+	Send(ctx context.Context, r *pdu.SendReq) (*pdu.SendRes, zfs.StreamCopier, error)
+	ReplicationCursor(ctx context.Context, req *pdu.ReplicationCursorReq) (*pdu.ReplicationCursorRes, error)
+}
+
+type Receiver interface {
+	Endpoint
+	// Receive sends r and sendStream (the latter containing a ZFS send stream)
+	// to the parent github.com/zrepl/zrepl/replication.Endpoint.
+	Receive(ctx context.Context, req *pdu.ReceiveReq, receive zfs.StreamCopier) (*pdu.ReceiveRes, error)
+}
+
+type Planner struct {
+	sender   Sender
+	receiver Receiver
+
+	promSecsPerState    *prometheus.HistogramVec // labels: state
+	promBytesReplicated *prometheus.CounterVec   // labels: filesystem
+}
+
+func (p *Planner) Plan(ctx context.Context) ([]driver.FS, error) {
+	fss, err := p.doPlanning(ctx)
+	if err != nil {
+		return nil, err
+	}
+	dfss := make([]driver.FS, len(fss))
+	for i := range dfss {
+		dfss[i] = fss[i]
+	}
+	return dfss, nil
+}
+
+func (p *Planner) WaitForConnectivity(ctx context.Context) error {
+	var wg sync.WaitGroup
+	doPing := func(endpoint Endpoint, errOut *error) {
+		defer wg.Done()
+		err := endpoint.WaitForConnectivity(ctx)
+		if err != nil {
+			*errOut = err
+		} else {
+			*errOut = nil
+		}
+	}
+	wg.Add(2)
+	var senderErr, receiverErr error
+	go doPing(p.sender, &senderErr)
+	go doPing(p.receiver, &receiverErr)
+	wg.Wait()
+	if senderErr == nil && receiverErr == nil {
+		return nil
+	} else if senderErr != nil && receiverErr != nil {
+		if senderErr.Error() == receiverErr.Error() {
+			return fmt.Errorf("sender and receiver are not reachable: %s", senderErr.Error())
+		} else {
+			return fmt.Errorf("sender and receiver are not reachable:\n  sender: %s\n  receiver: %s", senderErr, receiverErr)
+		}
+	} else {
+		var side string
+		var err *error
+		if senderErr != nil {
+			side = "sender"
+			err = &senderErr
+		} else {
+			side = "receiver"
+			err = &receiverErr
+		}
+		return fmt.Errorf("%s is not reachable: %s", side, *err)
+	}
+}
+
+type Filesystem struct {
+	sender   Sender
+	receiver Receiver
+
+	Path                string             // compat
+	receiverFS          *pdu.Filesystem
+	promBytesReplicated prometheus.Counter // compat
+}
+
+func (f *Filesystem) EqualToPreviousAttempt(other driver.FS) bool {
+	g, ok := other.(*Filesystem)
+	if !ok {
+		return false
+	}
+	// TODO: use GUIDs (issued by zrepl, not those from ZFS)
+	return f.Path == g.Path
+}
+
+func (f *Filesystem) PlanFS(ctx context.Context) ([]driver.Step, error) {
+	steps, err := f.doPlanning(ctx)
+	if err != nil {
+		return nil, err
+	}
+	dsteps := make([]driver.Step, len(steps))
+	for i := range dsteps {
+		dsteps[i] = steps[i]
+	}
+	return dsteps, nil
+}
+func (f *Filesystem) ReportInfo() *report.FilesystemInfo {
+	return &report.FilesystemInfo{Name: f.Path} // FIXME compat name
+}
+
+type Step struct {
+	sender   Sender
+	receiver Receiver
+
+	parent   *Filesystem
+	from, to *pdu.FilesystemVersion // compat
+
+	byteCounter  bytecounter.StreamCopier
+	expectedSize int64 // 0 means no size estimate present / possible
+}
+
+func (s *Step) TargetEquals(other driver.Step) bool {
+	t, ok := other.(*Step)
+	if !ok {
+		return false
+	}
+	if !s.parent.EqualToPreviousAttempt(t.parent) {
+		panic("Step interface promise broken: parent filesystems must be same")
+	}
+	return s.from.GetGuid() == t.from.GetGuid() &&
+		s.to.GetGuid() == t.to.GetGuid()
+}
+
+func (s *Step) TargetDate() time.Time {
+	return s.to.SnapshotTime() // FIXME compat name
+}
+
+func (s *Step) Step(ctx context.Context) error {
+	return s.doReplication(ctx)
+}
+
+func (s *Step) ReportInfo() *report.StepInfo {
+	var byteCounter int64
+	if s.byteCounter != nil {
+		byteCounter = s.byteCounter.Count()
+	}
+	// FIXME stick to zfs convention of from and to
+	from := ""
+	if s.from != nil {
+		from = s.from.RelName()
+	}
+	return &report.StepInfo{
+		From:            from,
+		To:              s.to.RelName(),
+		BytesExpected:   s.expectedSize,
+		BytesReplicated: byteCounter,
+	}
+}
+
+func NewPlanner(secsPerState *prometheus.HistogramVec, bytesReplicated *prometheus.CounterVec, sender Sender, receiver Receiver) *Planner {
+	return &Planner{
+		sender:              sender,
+		receiver:            receiver,
+		promSecsPerState:    secsPerState,
+		promBytesReplicated: bytesReplicated,
+	}
+}
+func resolveConflict(conflict error) (path []*pdu.FilesystemVersion, msg string) {
+	if noCommonAncestor, ok := conflict.(*ConflictNoCommonAncestor); ok {
+		if len(noCommonAncestor.SortedReceiverVersions) == 0 {
+			// TODO this is hard-coded replication policy: most recent snapshot as source
+			var mostRecentSnap *pdu.FilesystemVersion
+			for n := len(noCommonAncestor.SortedSenderVersions) - 1; n >= 0; n-- {
+				if noCommonAncestor.SortedSenderVersions[n].Type == pdu.FilesystemVersion_Snapshot {
+					mostRecentSnap = noCommonAncestor.SortedSenderVersions[n]
+					break
+				}
+			}
+			if mostRecentSnap == nil {
+				return nil, "no snapshots available on sender side"
+			}
+			return []*pdu.FilesystemVersion{mostRecentSnap}, fmt.Sprintf("start replication at most recent snapshot %s", mostRecentSnap.RelName())
+		}
+	}
+	return nil, "no automated way to handle conflict type"
+}
+
+func (p *Planner) doPlanning(ctx context.Context) ([]*Filesystem, error) {
+
+	log := getLogger(ctx)
+
+	log.Info("start planning")
+
+	slfssres, err := p.sender.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
+	if err != nil {
+		log.WithError(err).WithField("errType", fmt.Sprintf("%T", err)).Error("error listing sender filesystems")
+		return nil, err
+	}
+	sfss := slfssres.GetFilesystems()
+	// no progress here since we could run in a live-lock on connectivity issues
+
+	rlfssres, err := p.receiver.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
+	if err != nil {
+		log.WithError(err).WithField("errType", fmt.Sprintf("%T", err)).Error("error listing receiver filesystems")
+		return nil, err
+	}
+	rfss := rlfssres.GetFilesystems()
+
+	q := make([]*Filesystem, 0, len(sfss))
+	for _, fs := range sfss {
+
+		var receiverFS *pdu.Filesystem
+		for _, rfs := range rfss {
+			if rfs.Path == fs.Path {
+				receiverFS = rfs
+			}
+		}
+
+		ctr := p.promBytesReplicated.WithLabelValues(fs.Path)
+
+		q = append(q, &Filesystem{
+			sender:              p.sender,
+			receiver:            p.receiver,
+			Path:                fs.Path,
+			receiverFS:          receiverFS,
+			promBytesReplicated: ctr,
+		})
+	}
+
+	return q, nil
+}
+
+func (fs *Filesystem) doPlanning(ctx context.Context) ([]*Step, error) {
+
+	log := getLogger(ctx).WithField("filesystem", fs.Path)
+
+	log.Debug("assessing filesystem")
+
+	sfsvsres, err := fs.sender.ListFilesystemVersions(ctx, &pdu.ListFilesystemVersionsReq{Filesystem: fs.Path})
+	if err != nil {
+		log.WithError(err).Error("cannot get remote filesystem versions")
+		return nil, err
+	}
+	sfsvs := sfsvsres.GetVersions()
+
+	if len(sfsvs) < 1 {
+		err := errors.New("sender does not have any versions")
+		log.Error(err.Error())
+		return nil, err
+	}
+
+	var rfsvs []*pdu.FilesystemVersion
+	if fs.receiverFS != nil && !fs.receiverFS.GetIsPlaceholder() {
+		rfsvsres, err := fs.receiver.ListFilesystemVersions(ctx, &pdu.ListFilesystemVersionsReq{Filesystem: fs.Path})
+		if err != nil {
+			log.WithError(err).Error("receiver error")
+			return nil, err
+		}
+		rfsvs = rfsvsres.GetVersions()
+	} else {
+		rfsvs = []*pdu.FilesystemVersion{}
+	}
+
+	path, conflict := IncrementalPath(rfsvs, sfsvs)
+	if conflict != nil {
+		var msg string
+		path, msg = resolveConflict(conflict) // no shadowing allowed!
+		if path != nil {
+			log.WithField("conflict", conflict).Info("conflict")
+			log.WithField("resolution", msg).Info("automatically resolved")
+		} else {
+			log.WithField("conflict", conflict).Error("conflict")
+			log.WithField("problem", msg).Error("cannot resolve conflict")
+		}
+	}
+	if len(path) == 0 {
+		return nil, conflict
+	}
+
+	steps := make([]*Step, 0, len(path))
+	// FIXME unify struct declarations => initializer?
+	if len(path) == 1 {
+		steps = append(steps, &Step{
+			parent:   fs,
+			sender:   fs.sender,
+			receiver: fs.receiver,
+			from:     nil,
+			to:       path[0],
+		})
+	} else {
+		for i := 0; i < len(path)-1; i++ {
+			steps = append(steps, &Step{
+				parent:   fs,
+				sender:   fs.sender,
+				receiver: fs.receiver,
+				from:     path[i],
+				to:       path[i+1],
+			})
+		}
+	}
+
+	log.Debug("compute send size estimate")
+	errs := make(chan error, len(steps))
+	var wg sync.WaitGroup
+	fanOutCtx, fanOutCancel := context.WithCancel(ctx)
+	defer fanOutCancel()
+	for _, step := range steps {
+		wg.Add(1)
+		go func(step *Step) {
+			defer wg.Done()
+			err := step.updateSizeEstimate(fanOutCtx)
+			if err != nil {
+				log.WithError(err).WithField("step", step).Error("error computing size estimate")
+				fanOutCancel()
+			}
+			errs <- err
+		}(step)
+	}
+	wg.Wait()
+	close(errs)
+	var significantErr error = nil
+	for err := range errs {
+		if err != nil {
+			if significantErr == nil || significantErr == context.Canceled {
+				significantErr = err
+			}
+		}
+	}
+	if significantErr != nil {
+		return nil, significantErr
+	}
+
+	log.Debug("filesystem planning finished")
+	return steps, nil
+}
+
+// type FilesystemsReplicationFailedError struct {
+// 	FilesystemsWithError []*fsrep.Replication
+// }
+
+// func (e FilesystemsReplicationFailedError) Error() string {
+// 	allSame := true
+// 	lastErr := e.FilesystemsWithError[0].Err().Error()
+// 	for _, fs := range e.FilesystemsWithError {
+// 		fsErr := fs.Err().Error()
+// 		allSame = allSame && lastErr == fsErr
+// 	}
+
+// 	fsstr := "multiple filesystems"
+// 	if len(e.FilesystemsWithError) == 1 {
+// 		fsstr = fmt.Sprintf("filesystem %s", e.FilesystemsWithError[0].FS())
+// 	}
+// 	errorStr := lastErr
+// 	if !allSame {
+// 		errorStr = "multiple different errors"
+// 	}
+// 	return fmt.Sprintf("%s could not be replicated: %s", fsstr, errorStr)
+// }
+
+func (s *Step) updateSizeEstimate(ctx context.Context) error {
+
+	log := getLogger(ctx)
+
+	sr := s.buildSendRequest(true)
+
+	log.Debug("initiate dry run send request")
+	sres, _, err := s.sender.Send(ctx, sr)
+	if err != nil {
+		log.WithError(err).Error("dry run send request failed")
+		return err
+	}
+	s.expectedSize = sres.ExpectedSize
+	return nil
+}
+
+func (s *Step) buildSendRequest(dryRun bool) (sr *pdu.SendReq) {
+	fs := s.parent.Path
+	if s.from == nil {
+		sr = &pdu.SendReq{
+			Filesystem: fs,
+			To:         s.to.RelName(),
+			DryRun:     dryRun,
+		}
+	} else {
+		sr = &pdu.SendReq{
+			Filesystem: fs,
+			From:       s.from.RelName(),
+			To:         s.to.RelName(),
+			DryRun:     dryRun,
+		}
+	}
+	return sr
+}
+
+func (s *Step) doReplication(ctx context.Context) error {
+
+	fs := s.parent.Path
+
+	log := getLogger(ctx)
+	sr := s.buildSendRequest(false)
+
+	log.Debug("initiate send request")
+	sres, sstreamCopier, err := s.sender.Send(ctx, sr)
+	if err != nil {
+		log.WithError(err).Error("send request failed")
+		return err
+	}
+	if sstreamCopier == nil {
+		err := errors.New("send request did not return a stream, broken endpoint implementation")
+		return err
+	}
+	defer sstreamCopier.Close()
+
+	// Install a byte counter to track progress + for status report
+	s.byteCounter = bytecounter.NewStreamCopier(sstreamCopier)
+	defer func() {
+		s.parent.promBytesReplicated.Add(float64(s.byteCounter.Count()))
+	}()
+
+	rr := &pdu.ReceiveReq{
+		Filesystem:       fs,
+		ClearResumeToken: !sres.UsedResumeToken,
+	}
+	log.Debug("initiate receive request")
+	_, err = s.receiver.Receive(ctx, rr, s.byteCounter)
+	if err != nil {
+		log.
+			WithError(err).
+			WithField("errType", fmt.Sprintf("%T", err)).
+			Error("receive request failed (might also be error on sender)")
+		// This failure could be due to
+		// 	- an unexpected exit of ZFS on the sending side
+		//  - an unexpected exit of ZFS on the receiving side
+		//  - a connectivity issue
+		return err
+	}
+	log.Debug("receive finished")
+
+	log.Debug("advance replication cursor")
+	req := &pdu.ReplicationCursorReq{
+		Filesystem: fs,
+		Op: &pdu.ReplicationCursorReq_Set{
+			Set: &pdu.ReplicationCursorReq_SetOp{
+				Snapshot: s.to.GetName(),
+			},
+		},
+	}
+	_, err = s.sender.ReplicationCursor(ctx, req)
+	if err != nil {
+		log.WithError(err).Error("error advancing replication cursor")
+		// If this fails and replication planning restarts, the diff algorithm will find
+		// that cursor out of place. This is not a problem because then, it would just use another FS
+		// However, we FIXME have no means to just update the cursor in a
+		// second replication attempt right after this one where we don't have new snaps yet
+		return err
+	}
+
+	return err
+}
+
+func (s *Step) String() string {
+	if s.from == nil { // FIXME: ZFS semantics are that to is nil on non-incremental send
+		return fmt.Sprintf("%s%s (full)", s.parent.Path, s.to.RelName())
+	} else {
+		return fmt.Sprintf("%s(%s => %s)", s.parent.Path, s.from.RelName(), s.to.RelName())
+	}
+}

replication/logic/replication_logic_context.go

@@ -1,9 +1,9 @@
-package replication
+package logic
 
 import (
 	"context"
+
 	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/replication/fsrep"
 )
 
 type contextKey int
@@ -16,7 +16,6 @@ type Logger = logger.Logger
 
 func WithLogger(ctx context.Context, l Logger) context.Context {
 	ctx = context.WithValue(ctx, contextKeyLog, l)
-	ctx = fsrep.WithLogger(ctx, l)
 	return ctx
 }
 

replication/mainfsm.go

@@ -1,560 +0,0 @@
-// Package replication implements replication of filesystems with existing
-// versions (snapshots) from a sender to a receiver.
-package replication
-
-import (
-	"context"
-	"errors"
-	"fmt"
-	"github.com/prometheus/client_golang/prometheus"
-	"github.com/zrepl/zrepl/daemon/job/wakeup"
-	"github.com/zrepl/zrepl/util/envconst"
-	"github.com/zrepl/zrepl/util/watchdog"
-	"math/bits"
-	"net"
-	"sort"
-	"sync"
-	"time"
-
-	"github.com/zrepl/zrepl/replication/fsrep"
-	. "github.com/zrepl/zrepl/replication/internal/diff"
-	"github.com/zrepl/zrepl/replication/pdu"
-)
-
-//go:generate enumer -type=State
-type State uint
-
-const (
-	Planning State = 1 << iota
-	PlanningError
-	Working
-	WorkingWait
-	Completed
-	PermanentError
-)
-
-func (s State) rsf() state {
-	idx := bits.TrailingZeros(uint(s))
-	if idx == bits.UintSize {
-		panic(s) // invalid value
-	}
-	m := []state{
-		statePlanning,
-		statePlanningError,
-		stateWorking,
-		stateWorkingWait,
-		nil,
-		nil,
-	}
-	return m[idx]
-}
-
-func (s State) IsTerminal() bool {
-	return s.rsf() == nil
-}
-
-// Replication implements the replication of multiple file systems from a Sender to a Receiver.
-//
-// It is a state machine that is driven by the Drive method
-// and provides asynchronous reporting via the Report method (i.e. from another goroutine).
-type Replication struct {
-	// not protected by lock
-	promSecsPerState *prometheus.HistogramVec // labels: state
-	promBytesReplicated *prometheus.CounterVec // labels: filesystem
-
-	Progress watchdog.KeepAlive
-
-	// lock protects all fields of this struct (but not the fields behind pointers!)
-	lock sync.Mutex
-
-	state State
-
-	// Working, WorkingWait, Completed, ContextDone
-	queue     []*fsrep.Replication
-	completed []*fsrep.Replication
-	active    *fsrep.Replication // == queue[0] or nil, unlike in Report
-
-	// for PlanningError, WorkingWait and ContextError and Completed
-	err error
-
-	// PlanningError, WorkingWait
-	sleepUntil time.Time
-}
-
-type Report struct {
-	Status    string
-	Problem   string
-	SleepUntil time.Time
-	Completed []*fsrep.Report
-	Pending   []*fsrep.Report
-	Active    *fsrep.Report // not contained in Pending, unlike in struct Replication
-}
-
-func NewReplication(secsPerState *prometheus.HistogramVec, bytesReplicated *prometheus.CounterVec) *Replication {
-	r := Replication{
-		promSecsPerState: secsPerState,
-		promBytesReplicated: bytesReplicated,
-		state:            Planning,
-	}
-	return &r
-}
-
-// Endpoint represents one side of the replication.
-//
-// An endpoint is either in Sender or Receiver mode, represented by the correspondingly
-// named interfaces defined in this package.
-type Endpoint interface {
-	// Does not include placeholder filesystems
-	ListFilesystems(ctx context.Context, req *pdu.ListFilesystemReq) (*pdu.ListFilesystemRes, error)
-	ListFilesystemVersions(ctx context.Context, req *pdu.ListFilesystemVersionsReq) (*pdu.ListFilesystemVersionsRes, error)
-	DestroySnapshots(ctx context.Context, req *pdu.DestroySnapshotsReq) (*pdu.DestroySnapshotsRes, error)
-}
-
-type Sender interface {
-	Endpoint
-	fsrep.Sender
-}
-
-type Receiver interface {
-	Endpoint
-	fsrep.Receiver
-}
-
-type FilteredError struct{ fs string }
-
-func NewFilteredError(fs string) *FilteredError {
-	return &FilteredError{fs}
-}
-
-func (f FilteredError) Error() string { return "endpoint does not allow access to filesystem " + f.fs }
-
-type updater func(func(*Replication)) (newState State)
-type state func(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver, u updater) state
-
-// Drive starts the state machine and returns only after replication has finished (with or without errors).
-// The Logger in ctx is used for both debug and error logging, but is not guaranteed to be stable
-// or end-user friendly.
-// User-facing replication progress reports and can be obtained using the Report method,
-// whose output will not change after Drive returns.
-//
-// FIXME: Drive may be only called once per instance of Replication
-func (r *Replication) Drive(ctx context.Context, sender Sender, receiver Receiver) {
-
-	var u updater = func(f func(*Replication)) State {
-		r.lock.Lock()
-		defer r.lock.Unlock()
-		if f != nil {
-			f(r)
-		}
-		return r.state
-	}
-
-	var s state = statePlanning
-	var pre, post State
-	for s != nil {
-		preTime := time.Now()
-		pre = u(nil)
-		s = s(ctx, &r.Progress, sender, receiver, u)
-		delta := time.Now().Sub(preTime)
-		r.promSecsPerState.WithLabelValues(pre.String()).Observe(delta.Seconds())
-		post = u(nil)
-		getLogger(ctx).
-			WithField("transition", fmt.Sprintf("%s => %s", pre, post)).
-			WithField("duration", delta).
-			Debug("main state transition")
-		if post == Working && pre != post {
-			getLogger(ctx).Info("start working")
-		}
-	}
-
-	getLogger(ctx).
-		WithField("final_state", post).
-		Debug("main final state")
-}
-
-func resolveConflict(conflict error) (path []*pdu.FilesystemVersion, msg string) {
-	if noCommonAncestor, ok := conflict.(*ConflictNoCommonAncestor); ok {
-		if len(noCommonAncestor.SortedReceiverVersions) == 0 {
-			// TODO this is hard-coded replication policy: most recent snapshot as source
-			var mostRecentSnap *pdu.FilesystemVersion
-			for n := len(noCommonAncestor.SortedSenderVersions) - 1; n >= 0; n-- {
-				if noCommonAncestor.SortedSenderVersions[n].Type == pdu.FilesystemVersion_Snapshot {
-					mostRecentSnap = noCommonAncestor.SortedSenderVersions[n]
-					break
-				}
-			}
-			if mostRecentSnap == nil {
-				return nil, "no snapshots available on sender side"
-			}
-			return []*pdu.FilesystemVersion{mostRecentSnap}, fmt.Sprintf("start replication at most recent snapshot %s", mostRecentSnap.RelName())
-		}
-	}
-	return nil, "no automated way to handle conflict type"
-}
-
-var RetryInterval = envconst.Duration("ZREPL_REPLICATION_RETRY_INTERVAL", 10 * time.Second)
-
-type Error interface {
-	error
-	Temporary() bool
-}
-
-var _ Error = fsrep.Error(nil)
-var _ Error = net.Error(nil)
-
-func isPermanent(err error) bool {
-	if e, ok := err.(Error); ok {
-		return !e.Temporary()
-	}
-	return true
-}
-
-func statePlanning(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver, u updater) state {
-
-	log := getLogger(ctx)
-
-	log.Info("start planning")
-
-	handlePlanningError := func(err error) state {
-		return u(func(r *Replication) {
-			ge := GlobalError{Err: err, Temporary: !isPermanent(err)}
-			log.WithError(ge).Error("encountered global error while planning replication")
-			r.err = ge
-			if !ge.Temporary {
-				r.state = PermanentError
-			} else {
-				r.sleepUntil = time.Now().Add(RetryInterval)
-				r.state = PlanningError
-			}
-		}).rsf()
-	}
-
-	slfssres, err := sender.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
-	if err != nil {
-		log.WithError(err).WithField("errType", fmt.Sprintf("%T", err)).Error("error listing sender filesystems")
-		return handlePlanningError(err)
-	}
-	sfss := slfssres.GetFilesystems()
-	// no progress here since we could run in a live-lock on connectivity issues
-
-	rlfssres, err := receiver.ListFilesystems(ctx, &pdu.ListFilesystemReq{})
-	if err != nil {
-		log.WithError(err).WithField("errType", fmt.Sprintf("%T", err)).Error("error listing receiver filesystems")
-		return handlePlanningError(err)
-	}
-	rfss := rlfssres.GetFilesystems()
-	ka.MadeProgress() // for both sender and receiver
-
-	q := make([]*fsrep.Replication, 0, len(sfss))
-	mainlog := log
-	for _, fs := range sfss {
-
-		log := mainlog.WithField("filesystem", fs.Path)
-
-		log.Debug("assessing filesystem")
-
-		sfsvsres, err := sender.ListFilesystemVersions(ctx, &pdu.ListFilesystemVersionsReq{Filesystem: fs.Path})
-		if err != nil {
-			log.WithError(err).Error("cannot get remote filesystem versions")
-			return handlePlanningError(err)
-		}
-		sfsvs := sfsvsres.GetVersions()
-		ka.MadeProgress()
-
-		if len(sfsvs) < 1 {
-			err := errors.New("sender does not have any versions")
-			log.Error(err.Error())
-			q = append(q, fsrep.NewReplicationConflictError(fs.Path, err))
-			continue
-		}
-
-		receiverFSExists := false
-		for _, rfs := range rfss {
-			if rfs.Path == fs.Path {
-				receiverFSExists = true
-			}
-		}
-
-		var rfsvs []*pdu.FilesystemVersion
-		if receiverFSExists {
-			rfsvsres, err := receiver.ListFilesystemVersions(ctx, &pdu.ListFilesystemVersionsReq{Filesystem: fs.Path})
-			if err != nil {
-				if _, ok := err.(*FilteredError); ok {
-					log.Info("receiver ignores filesystem")
-					continue
-				}
-				log.WithError(err).Error("receiver error")
-				return handlePlanningError(err)
-			}
-			rfsvs = rfsvsres.GetVersions()
-		} else {
-			rfsvs = []*pdu.FilesystemVersion{}
-		}
-		ka.MadeProgress()
-
-		path, conflict := IncrementalPath(rfsvs, sfsvs)
-		if conflict != nil {
-			var msg string
-			path, msg = resolveConflict(conflict) // no shadowing allowed!
-			if path != nil {
-				log.WithField("conflict", conflict).Info("conflict")
-				log.WithField("resolution", msg).Info("automatically resolved")
-			} else {
-				log.WithField("conflict", conflict).Error("conflict")
-				log.WithField("problem", msg).Error("cannot resolve conflict")
-			}
-		}
-		ka.MadeProgress()
-		if path == nil {
-			q = append(q, fsrep.NewReplicationConflictError(fs.Path, conflict))
-			continue
-		}
-
-		var promBytesReplicated *prometheus.CounterVec
-		u(func(replication *Replication) { // FIXME args struct like in pruner (also use for sender and receiver)
-			promBytesReplicated = replication.promBytesReplicated
-		})
-		fsrfsm := fsrep.BuildReplication(fs.Path, promBytesReplicated.WithLabelValues(fs.Path))
-		if len(path) == 1 {
-			fsrfsm.AddStep(nil, path[0])
-		} else {
-			for i := 0; i < len(path)-1; i++ {
-				fsrfsm.AddStep(path[i], path[i+1])
-			}
-		}
-		qitem := fsrfsm.Done()
-		ka.MadeProgress()
-
-		log.Debug("compute send size estimate")
-		if err = qitem.UpdateSizeEsitmate(ctx, sender); err != nil {
-			log.WithError(err).Error("error computing size estimate")
-			return handlePlanningError(err)
-		}
-		ka.MadeProgress()
-
-		q = append(q, qitem)
-	}
-
-	ka.MadeProgress()
-
-	return u(func(r *Replication) {
-		r.completed = nil
-		r.queue = q
-		r.err = nil
-		r.state = Working
-	}).rsf()
-}
-
-func statePlanningError(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver, u updater) state {
-	var sleepUntil time.Time
-	u(func(r *Replication) {
-		sleepUntil = r.sleepUntil
-	})
-	t := time.NewTimer(sleepUntil.Sub(time.Now()))
-	getLogger(ctx).WithField("until", sleepUntil).Info("retry wait after planning error")
-	defer t.Stop()
-	select {
-	case <-ctx.Done():
-		return u(func(r *Replication) {
-			r.state = PermanentError
-			r.err = ctx.Err()
-		}).rsf()
-	case <-t.C:
-	case <-wakeup.Wait(ctx):
-	}
-	return u(func(r *Replication) {
-		r.state = Planning
-	}).rsf()
-}
-
-type GlobalError struct {
-	Err       error
-	Temporary bool
-}
-
-func (e GlobalError) Error() string {
-	errClass := "temporary"
-	if !e.Temporary {
-		errClass = "permanent"
-	}
-	return fmt.Sprintf("%s global error: %s", errClass, e.Err)
-}
-
-type FilesystemsReplicationFailedError struct {
-	FilesystemsWithError []*fsrep.Replication
-}
-
-func (e FilesystemsReplicationFailedError) Error() string {
-	allSame := true
-	lastErr := e.FilesystemsWithError[0].Err().Error()
-	for _, fs := range e.FilesystemsWithError {
-		fsErr := fs.Err().Error()
-		allSame = allSame && lastErr == fsErr
-	}
-
-	fsstr := "multiple filesystems"
-	if len(e.FilesystemsWithError) == 1 {
-		fsstr = fmt.Sprintf("filesystem %s", e.FilesystemsWithError[0].FS())
-	}
-	errorStr := lastErr
-	if !allSame {
-		errorStr = "multiple different errors"
-	}
-	return fmt.Sprintf("%s could not be replicated: %s", fsstr, errorStr)
-}
-
-func stateWorking(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver, u updater) state {
-
-	var active *fsrep.Replication
-	rsfNext := u(func(r *Replication) {
-
-		r.err = nil
-
-		newq := make([]*fsrep.Replication, 0, len(r.queue))
-		for i := range r.queue {
-			if r.queue[i].CanRetry() {
-				newq = append(newq, r.queue[i])
-			} else {
-				r.completed = append(r.completed, r.queue[i])
-			}
-		}
-		sort.SliceStable(newq, func(i, j int) bool {
-			return newq[i].NextStepDate().Before(newq[j].NextStepDate())
-		})
-		r.queue = newq
-
-		if len(r.queue) == 0 {
-			r.state = Completed
-			fsWithErr := FilesystemsReplicationFailedError{ // prepare it
-				FilesystemsWithError: make([]*fsrep.Replication, 0, len(r.completed)),
-			}
-			for _, fs := range r.completed {
-				if fs.CanRetry() {
-					panic(fmt.Sprintf("implementation error: completed contains retryable FS %s %#v",
-						fs.FS(), fs.Err()))
-				}
-				if fs.Err() != nil {
-					fsWithErr.FilesystemsWithError = append(fsWithErr.FilesystemsWithError, fs)
-				}
-			}
-			if len(fsWithErr.FilesystemsWithError) > 0 {
-				r.err = fsWithErr
-				r.state = PermanentError
-			}
-			return
-		}
-
-		active =  r.queue[0] // do not dequeue: if it's done, it will be sorted the next time we check for more work
-		r.active = active
-	}).rsf()
-
-	if active == nil {
-		return rsfNext
-	}
-
-	activeCtx := fsrep.WithLogger(ctx, getLogger(ctx).WithField("fs", active.FS()))
-	err := active.Retry(activeCtx, ka, sender, receiver)
-	u(func(r *Replication) {
-		r.active = nil
-	}).rsf()
-
-	if err != nil {
-		if err.ContextErr() && ctx.Err() != nil {
-			getLogger(ctx).WithError(err).
-				Info("filesystem replication was cancelled")
-			u(func(r*Replication) {
-				r.err = GlobalError{Err: err, Temporary: false}
-				r.state = PermanentError
-			})
-		} else if err.LocalToFS() {
-			getLogger(ctx).WithError(err).
-				Error("filesystem replication encountered a filesystem-specific error")
-			// we stay in this state and let the queuing logic above de-prioritize this failing FS
-		} else if err.Temporary() {
-			getLogger(ctx).WithError(err).
-				Error("filesystem encountered a non-filesystem-specific temporary error, enter retry-wait")
-			u(func(r *Replication) {
-				r.err = GlobalError{Err: err, Temporary: true}
-				r.sleepUntil = time.Now().Add(RetryInterval)
-				r.state = WorkingWait
-			}).rsf()
-		} else {
-			getLogger(ctx).WithError(err).
-				Error("encountered a permanent non-filesystem-specific error")
-			u(func(r *Replication) {
-				r.err = GlobalError{Err: err, Temporary: false}
-				r.state = PermanentError
-			}).rsf()
-		}
-	}
-
-	return u(nil).rsf()
-}
-
-func stateWorkingWait(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, receiver Receiver, u updater) state {
-	var sleepUntil time.Time
-	u(func(r *Replication) {
-		sleepUntil = r.sleepUntil
-	})
-	t := time.NewTimer(RetryInterval)
-	getLogger(ctx).WithField("until", sleepUntil).Info("retry wait after error")
-	defer t.Stop()
-	select {
-	case <-ctx.Done():
-		return u(func(r *Replication) {
-			r.state = PermanentError
-			r.err = ctx.Err()
-		}).rsf()
-
-	case <-t.C:
-	case <-wakeup.Wait(ctx):
-	}
-	return u(func(r *Replication) {
-		r.state = Working
-	}).rsf()
-}
-
-// Report provides a summary of the progress of the Replication,
-// i.e., a condensed dump of the internal state machine.
-// Report is safe to be called asynchronously while Drive is running.
-func (r *Replication) Report() *Report {
-	r.lock.Lock()
-	defer r.lock.Unlock()
-
-	rep := Report{
-		Status: r.state.String(),
-		SleepUntil: r.sleepUntil,
-	}
-
-	if r.err != nil {
-		rep.Problem = r.err.Error()
-	}
-
-	if r.state&(Planning|PlanningError) != 0 {
-		return &rep
-	}
-
-	rep.Pending = make([]*fsrep.Report, 0, len(r.queue))
-	rep.Completed = make([]*fsrep.Report, 0, len(r.completed)) // room for active (potentially)
-
-	// since r.active == r.queue[0], do not contain it in pending output
-	pending := r.queue
-	if r.active != nil {
-		rep.Active = r.active.Report()
-		pending = r.queue[1:]
-	}
-	for _, fsr := range pending {
-		rep.Pending= append(rep.Pending, fsr.Report())
-	}
-	for _, fsr := range r.completed {
-		rep.Completed = append(rep.Completed, fsr.Report())
-	}
-
-	return &rep
-}
-
-func (r *Replication) State() State {
-	r.lock.Lock()
-	defer r.lock.Unlock()
-	return r.state
-}

replication/replication.go

@@ -0,0 +1,13 @@
+// Package replication implements replication of filesystems with existing
+// versions (snapshots) from a sender to a receiver.
+package replication
+
+import (
+	"context"
+
+	"github.com/zrepl/zrepl/replication/driver"
+)
+
+func Do(ctx context.Context, planner driver.Planner) (driver.ReportFunc, driver.WaitFunc) {
+	return driver.Do(ctx, planner)
+}

replication/report/replication_report.go

@@ -0,0 +1,152 @@
+package report
+
+import (
+	"encoding/json"
+	"time"
+)
+
+type Report struct {
+	StartAt, FinishAt                      time.Time
+	WaitReconnectSince, WaitReconnectUntil time.Time
+	WaitReconnectError                     *TimedError
+	Attempts                               []*AttemptReport
+}
+
+var _, _ = json.Marshal(&Report{})
+
+type TimedError struct {
+	Err  string
+	Time time.Time
+}
+
+func NewTimedError(err string, t time.Time) *TimedError {
+	if err == "" {
+		panic("error must be empty")
+	}
+	if t.IsZero() {
+		panic("t must be non-zero")
+	}
+	return &TimedError{err, t}
+}
+
+func (s *TimedError) Error() string {
+	return s.Err
+}
+
+var _, _ = json.Marshal(&TimedError{})
+
+type AttemptReport struct {
+	State             AttemptState
+	StartAt, FinishAt time.Time
+	PlanError         *TimedError
+	Filesystems       []*FilesystemReport
+}
+
+type AttemptState string
+
+const (
+	AttemptPlanning      AttemptState = "planning"
+	AttemptPlanningError AttemptState = "planning-error"
+	AttemptFanOutFSs     AttemptState = "fan-out-filesystems"
+	AttemptFanOutError   AttemptState = "filesystem-error"
+	AttemptDone          AttemptState = "done"
+)
+
+type FilesystemState string
+
+const (
+	FilesystemPlanning        FilesystemState = "planning"
+	FilesystemPlanningErrored FilesystemState = "planning-error"
+	FilesystemStepping        FilesystemState = "stepping"
+	FilesystemSteppingErrored FilesystemState = "step-error"
+	FilesystemDone            FilesystemState = "done"
+)
+
+type FilesystemReport struct {
+	Info *FilesystemInfo
+
+	State FilesystemState
+
+	// Valid in State = FilesystemPlanningErrored
+	PlanError *TimedError
+	// Valid in State = FilesystemSteppingErrored
+	StepError *TimedError
+
+	// Valid in State = FilesystemStepping
+	CurrentStep int
+	Steps       []*StepReport
+}
+
+type FilesystemInfo struct {
+	Name string
+}
+
+type StepReport struct {
+	Info *StepInfo
+}
+
+type StepInfo struct {
+	From, To        string
+	BytesExpected   int64
+	BytesReplicated int64
+}
+
+func (a *AttemptReport) BytesSum() (expected, replicated int64) {
+	for _, fs := range a.Filesystems {
+		e, r := fs.BytesSum()
+		expected += e
+		replicated += r
+	}
+	return expected, replicated
+}
+
+func (f *FilesystemReport) BytesSum() (expected, replicated int64) {
+	for _, step := range f.Steps {
+		expected += step.Info.BytesExpected
+		replicated += step.Info.BytesReplicated
+	}
+	return
+}
+
+func (f *AttemptReport) FilesystemsByState() map[FilesystemState][]*FilesystemReport {
+	r := make(map[FilesystemState][]*FilesystemReport, 4)
+	for _, fs := range f.Filesystems {
+		l := r[fs.State]
+		l = append(l, fs)
+		r[fs.State] = l
+	}
+	return r
+}
+
+func (f *FilesystemReport) Error() *TimedError {
+	switch f.State {
+	case FilesystemPlanningErrored:
+		return f.PlanError
+	case FilesystemSteppingErrored:
+		return f.StepError
+	}
+	return nil
+}
+
+// may return nil
+func (f *FilesystemReport) NextStep() *StepReport {
+	switch f.State {
+	case FilesystemDone:
+		return nil
+	case FilesystemPlanningErrored:
+		return nil
+	case FilesystemSteppingErrored:
+		return nil
+	case FilesystemPlanning:
+		return nil
+	case FilesystemStepping:
+		// invariant is that this is always correct
+		// TODO what about 0-length Steps but short intermediary state?
+		return f.Steps[f.CurrentStep]
+	}
+	panic("unreachable")
+}
+
+func (f *StepReport) IsIncremental() bool {
+	return f.Info.From != "" // FIXME change to ZFS semantics (To != "")
+}

replication/state_enumer.go

@@ -1,76 +0,0 @@
-// Code generated by "enumer -type=State"; DO NOT EDIT.
-
-package replication
-
-import (
-	"fmt"
-)
-
-const (
-	_StateName_0 = "PlanningPlanningError"
-	_StateName_1 = "Working"
-	_StateName_2 = "WorkingWait"
-	_StateName_3 = "Completed"
-	_StateName_4 = "PermanentError"
-)
-
-var (
-	_StateIndex_0 = [...]uint8{0, 8, 21}
-	_StateIndex_1 = [...]uint8{0, 7}
-	_StateIndex_2 = [...]uint8{0, 11}
-	_StateIndex_3 = [...]uint8{0, 9}
-	_StateIndex_4 = [...]uint8{0, 14}
-)
-
-func (i State) String() string {
-	switch {
-	case 1 <= i && i <= 2:
-		i -= 1
-		return _StateName_0[_StateIndex_0[i]:_StateIndex_0[i+1]]
-	case i == 4:
-		return _StateName_1
-	case i == 8:
-		return _StateName_2
-	case i == 16:
-		return _StateName_3
-	case i == 32:
-		return _StateName_4
-	default:
-		return fmt.Sprintf("State(%d)", i)
-	}
-}
-
-var _StateValues = []State{1, 2, 4, 8, 16, 32}
-
-var _StateNameToValueMap = map[string]State{
-	_StateName_0[0:8]:  1,
-	_StateName_0[8:21]: 2,
-	_StateName_1[0:7]:  4,
-	_StateName_2[0:11]: 8,
-	_StateName_3[0:9]:  16,
-	_StateName_4[0:14]: 32,
-}
-
-// StateString retrieves an enum value from the enum constants string name.
-// Throws an error if the param is not part of the enum.
-func StateString(s string) (State, error) {
-	if val, ok := _StateNameToValueMap[s]; ok {
-		return val, nil
-	}
-	return 0, fmt.Errorf("%s does not belong to State values", s)
-}
-
-// StateValues returns all values of the enum
-func StateValues() []State {
-	return _StateValues
-}
-
-// IsAState returns "true" if the value is listed in the enum definition. "false" otherwise
-func (i State) IsAState() bool {
-	for _, v := range _StateValues {
-		if i == v {
-			return true
-		}
-	}
-	return false
-}

rpc/dataconn/dataconn_client.go

@@ -7,7 +7,7 @@ import (
 	"strings"
 
 	"github.com/golang/protobuf/proto"
-	"github.com/zrepl/zrepl/replication/pdu"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
 	"github.com/zrepl/zrepl/rpc/dataconn/stream"
 	"github.com/zrepl/zrepl/transport"
 	"github.com/zrepl/zrepl/zfs"
@@ -213,3 +213,23 @@ func (c *Client) ReqRecv(ctx context.Context, req *pdu.ReceiveReq, streamCopier
 
 	return res.res, cause
 }
+
+
+func (c *Client) ReqPing(ctx context.Context, req *pdu.PingReq) (*pdu.PingRes, error) {
+	conn, err := c.getWire(ctx)
+	if err != nil {
+		return nil, err
+	}
+			defer c.putWire(conn)
+
+	if err := c.send(ctx, conn, EndpointPing, req, nil); err != nil {
+		return nil, err
+	}
+
+	var res pdu.PingRes
+	if err := c.recv(ctx, conn, &res); err != nil {
+		return nil, err
+	}
+
+	return &res, nil
+}

rpc/dataconn/dataconn_server.go

@@ -7,7 +7,7 @@ import (
 
 	"github.com/golang/protobuf/proto"
 	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/replication/pdu"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
 	"github.com/zrepl/zrepl/rpc/dataconn/stream"
 	"github.com/zrepl/zrepl/transport"
 	"github.com/zrepl/zrepl/zfs"
@@ -25,6 +25,8 @@ type Handler interface {
 	// It is guaranteed that Server calls Receive with a stream that holds the IdleConnTimeout
 	// configured in ServerConfig.Shared.IdleConnTimeout.
 	Receive(ctx context.Context, r *pdu.ReceiveReq, receive zfs.StreamCopier) (*pdu.ReceiveRes, error)
+	// PingDataconn handles a PingReq
+	PingDataconn(ctx context.Context, r *pdu.PingReq) (*pdu.PingRes, error)
 }
 
 type Logger = logger.Logger
@@ -125,6 +127,13 @@ func (s *Server) serveConn(nc *transport.AuthConn) {
 			return
 		}
 		res, handlerErr = s.h.Receive(ctx, &req, &streamCopier{streamConn: c, closeStreamOnClose: false}) // SHADOWING
+	case EndpointPing:
+		var req pdu.PingReq
+		if err := proto.Unmarshal(reqStructured, &req); err != nil {
+			s.log.WithError(err).Error("cannot unmarshal ping request")
+			return
+		}
+		res, handlerErr = s.h.PingDataconn(ctx, &req) // SHADOWING
 	default:
 		s.log.WithField("endpoint", endpoint).Error("unknown endpoint")
 		handlerErr = fmt.Errorf("requested endpoint does not exist")
@@ -137,6 +146,10 @@ func (s *Server) serveConn(nc *transport.AuthConn) {
 	// if marshaling fails. We consider failed marshaling a handler error
 	var protobuf *bytes.Buffer
 	if handlerErr == nil {
+		if res == nil {
+			handlerErr = fmt.Errorf("implementation error: handler for endpoint %q returns nil error and nil result", endpoint)
+			s.log.WithError(err).Error("handle implementation error")
+		} else {
 			protobufBytes, err := proto.Marshal(res)
 			if err != nil {
 				s.log.WithError(err).Error("cannot marshal handler protobuf")
@@ -144,6 +157,7 @@ func (s *Server) serveConn(nc *transport.AuthConn) {
 			}
 			protobuf = bytes.NewBuffer(protobufBytes) // SHADOWING
 		}
+	}
 
 	var resHeaderBuf bytes.Buffer
 	if handlerErr == nil {

rpc/dataconn/dataconn_shared.go

@@ -10,6 +10,7 @@ import (
 )
 
 const (
+	EndpointPing string = "/v1/ping"
 	EndpointSend string = "/v1/send"
 	EndpointRecv string = "/v1/recv"
 )

rpc/dataconn/microbenchmark/microbenchmark.go

@@ -24,7 +24,7 @@ import (
 	"github.com/pkg/profile"
 
 	"github.com/zrepl/zrepl/logger"
-	"github.com/zrepl/zrepl/replication/pdu"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
 	"github.com/zrepl/zrepl/rpc/dataconn"
 	"github.com/zrepl/zrepl/rpc/dataconn/timeoutconn"
 	"github.com/zrepl/zrepl/transport"
@@ -77,6 +77,12 @@ func (devNullHandler) Receive(ctx context.Context, r *pdu.ReceiveReq, stream zfs
 	return &res, err
 }
 
+func (devNullHandler) PingDataconn(ctx context.Context, r *pdu.PingReq) (*pdu.PingRes, error) {
+	return &pdu.PingRes{
+		Echo: r.GetMessage(),
+	}, nil
+}
+
 type tcpConnecter struct {
 	addr string
 }

rpc/grpcclientidentity/authlistener_grpc_adaptor.go

@@ -105,11 +105,12 @@ func NewInterceptors(logger Logger, clientIdentityKey interface{}) (unary grpc.U
 		if !ok {
 			panic("peer.FromContext expected to return a peer in grpc.UnaryServerInterceptor")
 		}
-		logger.WithField("peer", fmt.Sprintf("%v", p)).Debug("peer")
+		logger.WithField("peer_addr", fmt.Sprintf("%s", p.Addr)).Debug("peer addr")
 		a, ok := p.AuthInfo.(*authConnAuthType)
 		if !ok {
 			panic(fmt.Sprintf("NewInterceptors must be used in combination with grpc.NewTransportCredentials, but got auth type %T", p.AuthInfo))
 		}
+		logger.WithField("peer_client_identity", a.clientIdentity).Debug("peer client identity")
 		ctx = context.WithValue(ctx, clientIdentityKey, a.clientIdentity)
 		return handler(ctx, req)
 	}

rpc/grpcclientidentity/grpchelper/authlistener_grpc_adaptor_wrapper.go

@@ -36,7 +36,7 @@ func ClientConn(cn transport.Connecter, log Logger) *grpc.ClientConn {
 	})
 	dialerOption := grpc.WithDialer(grpcclientidentity.NewDialer(log, cn))
 	cred := grpc.WithTransportCredentials(grpcclientidentity.NewTransportCredentials(log))
-	ctx, cancel := context.WithTimeout(context.TODO(), 5*time.Second)
+	ctx, cancel := context.WithTimeout(context.TODO(), 5*time.Second) // FIXME constant
 	defer cancel()
 	cc, err := grpc.DialContext(ctx, "doesn't matter done by dialer", dialerOption, cred, ka)
 	if err != nil {

rpc/rpc_client.go

@@ -2,13 +2,18 @@ package rpc
 
 import (
 	"context"
+	"errors"
+	"fmt"
 	"net"
+	"sync"
+	"sync/atomic"
 	"time"
 
 	"google.golang.org/grpc"
 
-	"github.com/zrepl/zrepl/replication"
+	"github.com/google/uuid"
-	"github.com/zrepl/zrepl/replication/pdu"
+	"github.com/zrepl/zrepl/replication/logic"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
 	"github.com/zrepl/zrepl/rpc/dataconn"
 	"github.com/zrepl/zrepl/rpc/grpcclientidentity/grpchelper"
 	"github.com/zrepl/zrepl/rpc/versionhandshake"
@@ -24,11 +29,12 @@ type Client struct {
 	controlClient pdu.ReplicationClient // this the grpc client instance, see constructor
 	controlConn   *grpc.ClientConn
 	loggers       Loggers
+	closed        chan struct{}
 }
 
-var _ replication.Endpoint = &Client{}
+var _ logic.Endpoint = &Client{}
-var _ replication.Sender = &Client{}
+var _ logic.Sender = &Client{}
-var _ replication.Receiver = &Client{}
+var _ logic.Receiver = &Client{}
 
 type DialContextFunc = func(ctx context.Context, network string, addr string) (net.Conn, error)
 
@@ -41,14 +47,21 @@ func NewClient(cn transport.Connecter, loggers Loggers) *Client {
 
 	c := &Client{
 		loggers: loggers,
+		closed:  make(chan struct{}),
 	}
 	grpcConn := grpchelper.ClientConn(muxedConnecter.control, loggers.Control)
 
 	go func() {
-		for {
+		ctx, cancel := context.WithCancel(context.Background())
+		go func() {
+			<-c.closed
+			cancel()
+		}()
+		defer cancel()
+		for ctx.Err() == nil {
 			state := grpcConn.GetState()
 			loggers.General.WithField("grpc_state", state.String()).Debug("grpc state change")
-			grpcConn.WaitForStateChange(context.TODO(), state)
+			grpcConn.WaitForStateChange(ctx, state)
 		}
 	}()
 	c.controlClient = pdu.NewReplicationClient(grpcConn)
@@ -59,8 +72,9 @@ func NewClient(cn transport.Connecter, loggers Loggers) *Client {
 }
 
 func (c *Client) Close() {
+	close(c.closed)
 	if err := c.controlConn.Close(); err != nil {
-		c.loggers.General.WithError(err).Error("cannot cloe control connection")
+		c.loggers.General.WithError(err).Error("cannot close control connection")
 	}
 	// TODO c.dataClient should have Close()
 }
@@ -101,6 +115,72 @@ func (c *Client) ReplicationCursor(ctx context.Context, in *pdu.ReplicationCurso
 	return c.controlClient.ReplicationCursor(ctx, in)
 }
 
+func (c *Client) WaitForConnectivity(ctx context.Context) error {
+	ctx, cancel := context.WithCancel(ctx)
+	defer cancel()
+	msg := uuid.New().String()
+	req := pdu.PingReq{Message: msg}
+	var ctrlOk, dataOk int32
+	loggers := GetLoggersOrPanic(ctx)
+	var wg sync.WaitGroup
+	wg.Add(2)
+	checkRes := func(res *pdu.PingRes, err error, logger Logger, okVar *int32) {
+		if err == nil && res.GetEcho() != req.GetMessage() {
+			err = errors.New("pilot message not echoed correctly")
+		}
+		if err == context.Canceled {
+			err = nil
+		}
+		if err != nil {
+			logger.WithError(err).Error("ping failed")
+			atomic.StoreInt32(okVar, 0)
+			cancel()
+		} else {
+			atomic.StoreInt32(okVar, 1)
+		}
+	}
+	go func() {
+		defer wg.Done()
+		ctrl, ctrlErr := c.controlClient.Ping(ctx, &req, grpc.FailFast(false))
+		checkRes(ctrl, ctrlErr, loggers.Control, &ctrlOk)
+	}()
+	go func() {
+		defer wg.Done()
+		for ctx.Err() == nil {
+			data, dataErr := c.dataClient.ReqPing(ctx, &req)
+			// dataClient uses transport.Connecter, which doesn't expose FailFast(false)
+			// => we need to mask dial timeouts
+			if err, ok := dataErr.(interface{ Temporary() bool }); ok && err.Temporary() {
+				// Rate-limit pings here in case Temporary() is a mis-classification
+				// or returns immediately (this is a tight loop in that case)
+				// TODO keep this in lockstep with controlClient
+				// 		=> don't use FailFast for control, but check that both control and data worked
+				time.Sleep(envconst.Duration("ZREPL_RPC_DATACONN_PING_SLEEP", 1*time.Second))
+				continue
+			}
+			// it's not a dial timeout, 
+			checkRes(data, dataErr, loggers.Data, &dataOk)
+			return
+		}
+	}()
+	wg.Wait()
+	var what string
+	if ctrlOk == 1 && dataOk == 1 {
+		return nil
+	}
+	if ctrlOk == 0 {
+		what += "control"
+	}
+	if dataOk == 0 {
+		if len(what) > 0 {
+			what += " and data"
+		} else {
+			what += "data"
+		}
+	}
+	return fmt.Errorf("%s rpc failed to respond to ping rpcs", what)
+}
+
 func (c *Client) ResetConnectBackoff() {
 	c.controlConn.ResetConnectBackoff()
 }

rpc/rpc_server.go

@@ -7,7 +7,7 @@ import (
 	"google.golang.org/grpc"
 
 	"github.com/zrepl/zrepl/endpoint"
-	"github.com/zrepl/zrepl/replication/pdu"
+	"github.com/zrepl/zrepl/replication/logic/pdu"
 	"github.com/zrepl/zrepl/rpc/dataconn"
 	"github.com/zrepl/zrepl/rpc/grpcclientidentity"
 	"github.com/zrepl/zrepl/rpc/netadaptor"

rpc/transportmux/transportmux.go

@@ -7,10 +7,10 @@ package transportmux
 
 import (
 	"context"
+	"fmt"
 	"io"
 	"net"
 	"time"
-	"fmt"
 
 	"github.com/zrepl/zrepl/logger"
 	"github.com/zrepl/zrepl/transport"
@@ -111,7 +111,7 @@ func Demux(ctx context.Context, rawListener transport.AuthenticatedListener, lab
 				if ctx.Err() != nil {
 					return
 				}
-				getLog(ctx).WithError(err).Error("accept error")
+				getLog(ctx).WithError(err).WithField("errType", fmt.Sprintf("%T", err)).Error("accept error")
 				continue
 			}
 			closeConn := func() {

rpc/versionhandshake/versionhandshake.go

@@ -26,14 +26,22 @@ type HandshakeError struct {
 	msg string
 	// If not nil, the underlying IO error that caused the handshake to fail.
 	IOError error
+	isAcceptError bool
 }
 
 var _ net.Error = &HandshakeError{}
 
 func (e HandshakeError) Error() string { return e.msg }
 
-// Always true to enable usage in a net.Listener.
+// Like with net.OpErr (Go issue 6163), a client failing to handshake
-func (e HandshakeError) Temporary() bool { return true }
+// should be a temporary Accept error toward the Listener .
+func (e HandshakeError) Temporary() bool {
+	if  e.isAcceptError {
+		return true
+	}
+	te, ok := e.IOError.(interface{ Temporary() bool });
+	return ok && te.Temporary()
+}
 
 // If the underlying IOError was net.Error.Timeout(), Timeout() returns that value.
 // Otherwise false.
@@ -142,14 +150,14 @@ func (m *HandshakeMessage) DecodeReader(r io.Reader, maxLen int) error {
 	return nil
 }
 
-func DoHandshakeCurrentVersion(conn net.Conn, deadline time.Time) error {
+func DoHandshakeCurrentVersion(conn net.Conn, deadline time.Time) *HandshakeError {
 	// current protocol version is hardcoded here
 	return DoHandshakeVersion(conn, deadline, 1)
 }
 
 const HandshakeMessageMaxLen = 16 * 4096
 
-func DoHandshakeVersion(conn net.Conn, deadline time.Time, version int) error {
+func DoHandshakeVersion(conn net.Conn, deadline time.Time, version int) *HandshakeError {
 	ours := HandshakeMessage{
 		ProtocolVersion: version,
 		Extensions: nil,

rpc/versionhandshake/versionhandshake_transport_wrappers.go

@@ -55,6 +55,7 @@ func (l HandshakeListener) Accept(ctx context.Context) (*transport.AuthConn, err
 		dl = time.Now().Add(l.timeout) // shadowing
 	}
 	if err := DoHandshakeCurrentVersion(conn, dl); err != nil {
+		err.isAcceptError = true
 		conn.Close()
 		return nil, err
 	}

util/chainlock/chainlock.go

@@ -0,0 +1,42 @@
+// package chainlock implements a mutex whose Lock and Unlock
+// methods return the lock itself, to enable chaining.
+//
+// Intended Usage
+//
+//   defer s.lock().unlock()
+//   // drop lock while waiting for wait group
+//   func() {
+//	     defer a.l.Unlock().Lock()
+//	     fssesDone.Wait()
+//	 }()
+//
+package chainlock
+
+import "sync"
+
+type L struct {
+	mtx sync.Mutex
+}
+
+func New() *L {
+	return &L{}
+}
+
+func (l *L) Lock() *L {
+	l.mtx.Lock()
+	return l
+}
+
+func (l *L) Unlock() *L {
+	l.mtx.Unlock()
+	return l
+}
+
+func (l *L) NewCond() *sync.Cond {
+	return sync.NewCond(&l.mtx)
+}
+
+func (l *L) DropWhile(f func()) {
+	defer l.Unlock().Lock()
+	f()
+}

util/envconst/envconst.go

@@ -40,3 +40,19 @@ func Int64(varname string, def int64) int64 {
 	cache.Store(varname, d)
 	return d
 }
+
+func Bool(varname string, def bool) bool {
+	if v, ok := cache.Load(varname); ok {
+		return v.(bool)
+	}
+	e := os.Getenv(varname)
+	if e == "" {
+		return def
+	}
+	d, err := strconv.ParseBool(e)
+	if err != nil {
+		panic(err)
+	}
+	cache.Store(varname, d)
+	return d
+}

zfs/conflict_string.go

@@ -1,16 +0,0 @@
-// Code generated by "stringer -type=Conflict"; DO NOT EDIT.
-
-package zfs
-
-import "strconv"
-
-const _Conflict_name = "ConflictIncrementalConflictAllRightConflictNoCommonAncestorConflictDiverged"
-
-var _Conflict_index = [...]uint8{0, 19, 35, 59, 75}
-
-func (i Conflict) String() string {
-	if i < 0 || i >= Conflict(len(_Conflict_index)-1) {
-		return "Conflict(" + strconv.FormatInt(int64(i), 10) + ")"
-	}
-	return _Conflict_name[_Conflict_index[i]:_Conflict_index[i+1]]
-}

zfs/diff.go

@@ -1,284 +0,0 @@
-package zfs
-
-import (
-	"bytes"
-	"crypto/sha512"
-	"encoding/hex"
-	"fmt"
-	"io"
-	"os/exec"
-	"sort"
-)
-
-type fsbyCreateTXG []FilesystemVersion
-
-func (l fsbyCreateTXG) Len() int      { return len(l) }
-func (l fsbyCreateTXG) Swap(i, j int) { l[i], l[j] = l[j], l[i] }
-func (l fsbyCreateTXG) Less(i, j int) bool {
-	return l[i].CreateTXG < l[j].CreateTXG
-}
-
-//go:generate stringer -type=Conflict
-type Conflict int
-
-const (
-	ConflictIncremental Conflict = iota // no conflict, incremental repl possible
-	ConflictAllRight                    // no conflict, initial repl possible
-	ConflictNoCommonAncestor
-	ConflictDiverged
-)
-
-/* The receiver (left) wants to know if the sender (right) has more recent versions
-
-	Left :         | C |
-	Right: | A | B | C | D | E |
-	=>   :         | C | D | E |
-
-	Left:         | C |
-	Right:			  | D | E |
-	=>   :  <empty list>, no common ancestor
-
-	Left :         | C | D | E |
-	Right: | A | B | C |
-	=>   :  <empty list>, the left has newer versions
-
-	Left : | A | B | C |       | F |
-	Right:         | C | D | E |
-	=>   :         | C |	   | F | => diverged => <empty list>
-
-IMPORTANT: since ZFS currently does not export dataset UUIDs, the best heuristic to
-		   identify a filesystem version is the tuple (name,creation)
-*/
-type FilesystemDiff struct {
-
-	// Which kind of conflict / "way forward" is possible.
-	// Check this first to determine the semantics of this struct's remaining members
-	Conflict Conflict
-
-	// Conflict = Incremental | AllRight
-	// 		The incremental steps required to get left up to right's most recent version
-	// 		0th element is the common ancestor, ordered by birthtime, oldest first
-	// 		If len() < 2, left and right are at same most recent version
-	// Conflict = otherwise
-	// 		nil; there is no incremental path for left to get to right's most recent version
-	IncrementalPath []FilesystemVersion
-
-	// Conflict = Incremental | AllRight: nil
-	// Conflict = NoCommonAncestor: left as passed as input
-	// Conflict = Diverged: contains path from left most recent common ancestor (mrca) to most
-	//						recent version on left
-	MRCAPathLeft []FilesystemVersion
-	// Conflict = Incremental | AllRight: nil
-	// Conflict = NoCommonAncestor: right as passed as input
-	// Conflict = Diverged: contains path from right most recent common ancestor (mrca)
-	// 						to most recent version on right
-	MRCAPathRight []FilesystemVersion
-}
-
-func (f FilesystemDiff) String() (str string) {
-	var b bytes.Buffer
-
-	fmt.Fprintf(&b, "%s, ", f.Conflict)
-
-	switch f.Conflict {
-	case ConflictIncremental:
-		fmt.Fprintf(&b, "incremental path length %v, common ancestor at %s", len(f.IncrementalPath)-1, f.IncrementalPath[0])
-	case ConflictAllRight:
-		fmt.Fprintf(&b, "%v versions, most recent is %s", len(f.MRCAPathRight)-1, f.MRCAPathRight[len(f.MRCAPathRight)-1])
-	case ConflictDiverged:
-		fmt.Fprintf(&b, "diverged at %s", f.MRCAPathRight[0]) // right always has at least one snap...?
-	case ConflictNoCommonAncestor:
-		fmt.Fprintf(&b, "no diff to show")
-	default:
-		fmt.Fprintf(&b, "unknown conflict type, likely a bug")
-	}
-
-	return b.String()
-}
-
-// we must assume left and right are ordered ascendingly by ZFS_PROP_CREATETXG and that
-// names are unique (bas ZFS_PROP_GUID replacement)
-func MakeFilesystemDiff(left, right []FilesystemVersion) (diff FilesystemDiff) {
-
-	if right == nil {
-		panic("right must not be nil")
-	}
-	if left == nil {
-		diff = FilesystemDiff{
-			IncrementalPath: nil,
-			Conflict:        ConflictAllRight,
-			MRCAPathLeft:    left,
-			MRCAPathRight:   right,
-		}
-		return
-	}
-
-	// Assert both left and right are sorted by createtxg
-	{
-		var leftSorted, rightSorted fsbyCreateTXG
-		leftSorted = left
-		rightSorted = right
-		if !sort.IsSorted(leftSorted) {
-			panic("cannot make filesystem diff: unsorted left")
-		}
-		if !sort.IsSorted(rightSorted) {
-			panic("cannot make filesystem diff: unsorted right")
-		}
-	}
-
-	// Find most recent common ancestor by name, preferring snapshots over bookmarks
-	mrcaLeft := len(left) - 1
-	var mrcaRight int
-outer:
-	for ; mrcaLeft >= 0; mrcaLeft-- {
-		for i := len(right) - 1; i >= 0; i-- {
-			if left[mrcaLeft].Guid == right[i].Guid {
-				mrcaRight = i
-				if i-1 >= 0 && right[i-1].Guid == right[i].Guid && right[i-1].Type == Snapshot {
-					// prefer snapshots over bookmarks
-					mrcaRight = i - 1
-				}
-				break outer
-			}
-		}
-	}
-
-	// no common ancestor?
-	if mrcaLeft == -1 {
-		diff = FilesystemDiff{
-			IncrementalPath: nil,
-			Conflict:        ConflictNoCommonAncestor,
-			MRCAPathLeft:    left,
-			MRCAPathRight:   right,
-		}
-		return
-	}
-
-	// diverged?
-	if mrcaLeft != len(left)-1 {
-		diff = FilesystemDiff{
-			IncrementalPath: nil,
-			Conflict:        ConflictDiverged,
-			MRCAPathLeft:    left[mrcaLeft:],
-			MRCAPathRight:   right[mrcaRight:],
-		}
-		return
-	}
-
-	if mrcaLeft != len(left)-1 {
-		panic("invariant violated: mrca on left must be the last item in the left list")
-	}
-
-	// incPath must not contain bookmarks except initial one,
-	//   and only if that initial bookmark's snapshot is gone
-	incPath := make([]FilesystemVersion, 0, len(right))
-	incPath = append(incPath, right[mrcaRight])
-	// right[mrcaRight] may be a bookmark if there's no equally named snapshot
-	for i := mrcaRight + 1; i < len(right); i++ {
-		if right[i].Type != Bookmark {
-			incPath = append(incPath, right[i])
-		}
-	}
-
-	diff = FilesystemDiff{
-		IncrementalPath: incPath,
-	}
-	return
-}
-
-const ZREPL_PLACEHOLDER_PROPERTY_NAME string = "zrepl:placeholder"
-
-type FilesystemState struct {
-	Placeholder bool
-	// TODO extend with resume token when that feature is finally added
-}
-
-// A somewhat efficient way to determine if a filesystem exists on this host.
-// Particularly useful if exists is called more than once (will only fork exec once and cache the result)
-func ZFSListFilesystemState() (localState map[string]FilesystemState, err error) {
-
-	var actual [][]string
-	if actual, err = ZFSList([]string{"name", ZREPL_PLACEHOLDER_PROPERTY_NAME}, "-t", "filesystem,volume"); err != nil {
-		return
-	}
-
-	localState = make(map[string]FilesystemState, len(actual))
-	for _, e := range actual {
-		dp, err := NewDatasetPath(e[0])
-		if err != nil {
-			return nil, fmt.Errorf("ZFS does not return parseable dataset path: %s", e[0])
-		}
-		placeholder, _ := IsPlaceholder(dp, e[1])
-		localState[e[0]] = FilesystemState{
-			placeholder,
-		}
-	}
-	return
-
-}
-
-// Computes the value for the ZREPL_PLACEHOLDER_PROPERTY_NAME ZFS user property
-// to mark the given DatasetPath p as a placeholder
-//
-// We cannot simply use booleans here since user properties are always
-// inherited.
-//
-// We hash the DatasetPath and use it to check for a given path if it is the
-// one originally marked as placeholder.
-//
-// However, this prohibits moving datasets around via `zfs rename`. The
-// placeholder attribute must be re-computed for the dataset path after the
-// move.
-//
-// TODO better solution available?
-func PlaceholderPropertyValue(p *DatasetPath) string {
-	ps := []byte(p.ToString())
-	sum := sha512.Sum512_256(ps)
-	return hex.EncodeToString(sum[:])
-}
-
-func IsPlaceholder(p *DatasetPath, placeholderPropertyValue string) (isPlaceholder bool, err error) {
-	expected := PlaceholderPropertyValue(p)
-	isPlaceholder = expected == placeholderPropertyValue
-	if !isPlaceholder {
-		err = fmt.Errorf("expected %s, has %s", expected, placeholderPropertyValue)
-	}
-	return
-}
-
-// for nonexistent FS, isPlaceholder == false && err == nil
-func ZFSIsPlaceholderFilesystem(p *DatasetPath) (isPlaceholder bool, err error) {
-	props, err := zfsGet(p.ToString(), []string{ZREPL_PLACEHOLDER_PROPERTY_NAME}, sourceAny)
-	if err == io.ErrUnexpectedEOF {
-		// interpret this as an early exit of the zfs binary due to the fs not existing
-		return false, nil
-	} else if err != nil {
-		return false, err
-	}
-	isPlaceholder, _ = IsPlaceholder(p, props.Get(ZREPL_PLACEHOLDER_PROPERTY_NAME))
-	return
-}
-
-func ZFSCreatePlaceholderFilesystem(p *DatasetPath) (err error) {
-	v := PlaceholderPropertyValue(p)
-	cmd := exec.Command(ZFS_BINARY, "create",
-		"-o", fmt.Sprintf("%s=%s", ZREPL_PLACEHOLDER_PROPERTY_NAME, v),
-		"-o", "mountpoint=none",
-		p.ToString())
-
-	stderr := bytes.NewBuffer(make([]byte, 0, 1024))
-	cmd.Stderr = stderr
-
-	if err = cmd.Start(); err != nil {
-		return err
-	}
-
-	if err = cmd.Wait(); err != nil {
-		err = &ZFSError{
-			Stderr:  stderr.Bytes(),
-			WaitErr: err,
-		}
-	}
-
-	return
-}

zfs/diff_test.go

@@ -1,112 +0,0 @@
-package zfs
-
-import (
-	"github.com/stretchr/testify/assert"
-	"strconv"
-	"strings"
-	"testing"
-	"time"
-)
-
-func fsvlist(fsv ...string) (r []FilesystemVersion) {
-
-	r = make([]FilesystemVersion, len(fsv))
-	for i, f := range fsv {
-
-		// parse the id from fsvlist. it is used to derivce Guid,CreateTXG and Creation attrs
-		split := strings.Split(f, ",")
-		if len(split) != 2 {
-			panic("invalid fsv spec")
-		}
-		id, err := strconv.Atoi(split[1])
-		if err != nil {
-			panic(err)
-		}
-
-		if strings.HasPrefix(f, "#") {
-			r[i] = FilesystemVersion{
-				Name:      strings.TrimPrefix(f, "#"),
-				Type:      Bookmark,
-				Guid:      uint64(id),
-				CreateTXG: uint64(id),
-				Creation:  time.Unix(0, 0).Add(time.Duration(id) * time.Second),
-			}
-		} else if strings.HasPrefix(f, "@") {
-			r[i] = FilesystemVersion{
-				Name:      strings.TrimPrefix(f, "@"),
-				Type:      Snapshot,
-				Guid:      uint64(id),
-				CreateTXG: uint64(id),
-				Creation:  time.Unix(0, 0).Add(time.Duration(id) * time.Second),
-			}
-		} else {
-			panic("invalid character")
-		}
-	}
-	return
-}
-
-func doTest(left, right []FilesystemVersion, validate func(d FilesystemDiff)) {
-	var d FilesystemDiff
-	d = MakeFilesystemDiff(left, right)
-	validate(d)
-}
-
-func TestMakeFilesystemDiff_IncrementalSnapshots(t *testing.T) {
-
-	l := fsvlist
-
-	// basic functionality
-	doTest(l("@a,1", "@b,2"), l("@a,1", "@b,2", "@c,3", "@d,4"), func(d FilesystemDiff) {
-		assert.Equal(t, l("@b,2", "@c,3", "@d,4"), d.IncrementalPath)
-	})
-
-	// no common ancestor
-	doTest(l(), l("@a,1"), func(d FilesystemDiff) {
-		assert.Nil(t, d.IncrementalPath)
-		assert.EqualValues(t, d.Conflict, ConflictNoCommonAncestor)
-		assert.Equal(t, l("@a,1"), d.MRCAPathRight)
-	})
-	doTest(l("@a,1", "@b,2"), l("@c,3", "@d,4"), func(d FilesystemDiff) {
-		assert.Nil(t, d.IncrementalPath)
-		assert.EqualValues(t, d.Conflict, ConflictNoCommonAncestor)
-		assert.Equal(t, l("@c,3", "@d,4"), d.MRCAPathRight)
-	})
-
-	// divergence is detected
-	doTest(l("@a,1", "@b1,2"), l("@a,1", "@b2,3"), func(d FilesystemDiff) {
-		assert.Nil(t, d.IncrementalPath)
-		assert.EqualValues(t, d.Conflict, ConflictDiverged)
-		assert.Equal(t, l("@a,1", "@b1,2"), d.MRCAPathLeft)
-		assert.Equal(t, l("@a,1", "@b2,3"), d.MRCAPathRight)
-	})
-
-	// gaps before most recent common ancestor do not matter
-	doTest(l("@a,1", "@b,2", "@c,3"), l("@a,1", "@c,3", "@d,4"), func(d FilesystemDiff) {
-		assert.Equal(t, l("@c,3", "@d,4"), d.IncrementalPath)
-	})
-
-}
-
-func TestMakeFilesystemDiff_BookmarksSupport(t *testing.T) {
-	l := fsvlist
-
-	// bookmarks are used
-	doTest(l("@a,1"), l("#a,1", "@b,2"), func(d FilesystemDiff) {
-		assert.Equal(t, l("#a,1", "@b,2"), d.IncrementalPath)
-	})
-
-	// boomarks are stripped from IncrementalPath (cannot send incrementally)
-	doTest(l("@a,1"), l("#a,1", "#b,2", "@c,3"), func(d FilesystemDiff) {
-		assert.Equal(t, l("#a,1", "@c,3"), d.IncrementalPath)
-	})
-
-	// test that snapshots are preferred over bookmarks in IncrementalPath
-	doTest(l("@a,1"), l("#a,1", "@a,1", "@b,2"), func(d FilesystemDiff) {
-		assert.Equal(t, l("@a,1", "@b,2"), d.IncrementalPath)
-	})
-	doTest(l("@a,1"), l("@a,1", "#a,1", "@b,2"), func(d FilesystemDiff) {
-		assert.Equal(t, l("@a,1", "@b,2"), d.IncrementalPath)
-	})
-
-}

zfs/placeholder.go

@@ -0,0 +1,113 @@
+package zfs
+
+import (
+	"bytes"
+	"crypto/sha512"
+	"encoding/hex"
+	"fmt"
+	"io"
+	"os/exec"
+)
+
+const ZREPL_PLACEHOLDER_PROPERTY_NAME string = "zrepl:placeholder"
+
+type FilesystemState struct {
+	Placeholder bool
+	// TODO extend with resume token when that feature is finally added
+}
+
+// A somewhat efficient way to determine if a filesystem exists on this host.
+// Particularly useful if exists is called more than once (will only fork exec once and cache the result)
+func ZFSListFilesystemState() (localState map[string]FilesystemState, err error) {
+
+	var actual [][]string
+	if actual, err = ZFSList([]string{"name", ZREPL_PLACEHOLDER_PROPERTY_NAME}, "-t", "filesystem,volume"); err != nil {
+		return
+	}
+
+	localState = make(map[string]FilesystemState, len(actual))
+	for _, e := range actual {
+		dp, err := NewDatasetPath(e[0])
+		if err != nil {
+			return nil, fmt.Errorf("ZFS does not return parseable dataset path: %s", e[0])
+		}
+		placeholder, _ := IsPlaceholder(dp, e[1])
+		localState[e[0]] = FilesystemState{
+			placeholder,
+		}
+	}
+	return
+
+}
+
+// Computes the value for the ZREPL_PLACEHOLDER_PROPERTY_NAME ZFS user property
+// to mark the given DatasetPath p as a placeholder
+//
+// We cannot simply use booleans here since user properties are always
+// inherited.
+//
+// We hash the DatasetPath and use it to check for a given path if it is the
+// one originally marked as placeholder.
+//
+// However, this prohibits moving datasets around via `zfs rename`. The
+// placeholder attribute must be re-computed for the dataset path after the
+// move.
+//
+// TODO better solution available?
+func PlaceholderPropertyValue(p *DatasetPath) string {
+	ps := []byte(p.ToString())
+	sum := sha512.Sum512_256(ps)
+	return hex.EncodeToString(sum[:])
+}
+
+func IsPlaceholder(p *DatasetPath, placeholderPropertyValue string) (isPlaceholder bool, err error) {
+	expected := PlaceholderPropertyValue(p)
+	isPlaceholder = expected == placeholderPropertyValue
+	if !isPlaceholder {
+		err = fmt.Errorf("expected %s, has %s", expected, placeholderPropertyValue)
+	}
+	return
+}
+
+// for nonexistent FS, isPlaceholder == false && err == nil
+func ZFSIsPlaceholderFilesystem(p *DatasetPath) (isPlaceholder bool, err error) {
+	props, err := zfsGet(p.ToString(), []string{ZREPL_PLACEHOLDER_PROPERTY_NAME}, sourceAny)
+	if err == io.ErrUnexpectedEOF {
+		// interpret this as an early exit of the zfs binary due to the fs not existing
+		return false, nil
+	} else if err != nil {
+		return false, err
+	}
+	isPlaceholder, _ = IsPlaceholder(p, props.Get(ZREPL_PLACEHOLDER_PROPERTY_NAME))
+	return
+}
+
+func ZFSCreatePlaceholderFilesystem(p *DatasetPath) (err error) {
+	v := PlaceholderPropertyValue(p)
+	cmd := exec.Command(ZFS_BINARY, "create",
+		"-o", fmt.Sprintf("%s=%s", ZREPL_PLACEHOLDER_PROPERTY_NAME, v),
+		"-o", "mountpoint=none",
+		p.ToString())
+
+	stderr := bytes.NewBuffer(make([]byte, 0, 1024))
+	cmd.Stderr = stderr
+
+	if err = cmd.Start(); err != nil {
+		return err
+	}
+
+	if err = cmd.Wait(); err != nil {
+		err = &ZFSError{
+			Stderr:  stderr.Bytes(),
+			WaitErr: err,
+		}
+	}
+
+	return
+}
+
+func ZFSSetNoPlaceholder(p *DatasetPath) error {
+	props := NewZFSProperties()
+	props.Set(ZREPL_PLACEHOLDER_PROPERTY_NAME, "off")
+	return zfsSet(p.ToString(), props)
+}

zfs/zfs.go

@@ -9,6 +9,7 @@ import (
 	"io"
 	"os"
 	"os/exec"
+	"sort"
 	"strings"
 	"sync"
 	"time"
@@ -691,17 +692,62 @@ type StreamCopier interface {
 	Close() error
 }
 
+type RecvOptions struct {
+	// Rollback to the oldest snapshot, destroy it, then perform `recv -F`.
+	// Note that this doesn't change property values, i.e. an existing local property value will be kept.
+	RollbackAndForceRecv bool
+}
 
-func ZFSRecv(ctx context.Context, fs string, streamCopier StreamCopier, additionalArgs ...string) (err error) {
+func ZFSRecv(ctx context.Context, fs string, streamCopier StreamCopier, opts RecvOptions) (err error) {
 
 	if err := validateZFSFilesystem(fs); err != nil {
 		return err
 	}
+	fsdp, err := NewDatasetPath(fs)
+	if err != nil {
+		return err
+	}
+
+	if opts.RollbackAndForceRecv {
+		// destroy all snapshots before `recv -F` because `recv -F`
+		// does not perform a rollback unless `send -R` was used (which we assume hasn't been the case)
+		var snaps []FilesystemVersion
+		{
+			vs, err := ZFSListFilesystemVersions(fsdp, nil)
+			if err != nil {
+				err = fmt.Errorf("cannot list versions to rollback is required: %s", err)
+			}
+			for _, v := range vs {
+				if v.Type == Snapshot {
+					snaps = append(snaps, v)
+				}
+			}
+			sort.Slice(snaps, func(i, j int) bool {
+				return snaps[i].CreateTXG < snaps[j].CreateTXG
+			})
+		}
+		// bookmarks are rolled back automatically
+		if len(snaps) > 0 {
+			// use rollback to efficiently destroy all but the earliest snapshot
+			// then destroy that earliest snapshot
+			// afterwards, `recv -F` will work
+			rollbackTarget := snaps[0]
+			rollbackTargetAbs := rollbackTarget.ToAbsPath(fsdp)
+			debug("recv: rollback to %q", rollbackTargetAbs)
+			if err := ZFSRollback(fsdp, rollbackTarget, "-r"); err != nil {
+				return fmt.Errorf("cannot rollback %s to %s for forced receive: %s", fsdp.ToString(), rollbackTarget, err)
+			}
+			debug("recv: destroy %q", rollbackTargetAbs)
+			if err := ZFSDestroy(rollbackTargetAbs); err != nil {
+				return fmt.Errorf("cannot destroy %s for forced receive: %s", rollbackTargetAbs, err)
+			}
+		}
+	}
 
 	args := make([]string, 0)
 	args = append(args, "recv")
-	if len(args) > 0 {
+	if opts.RollbackAndForceRecv {
-		args = append(args, additionalArgs...)
+		args = append(args, "-F")
 	}
 	args = append(args, fs)
 
@@ -1038,3 +1084,33 @@ func ZFSBookmark(fs *DatasetPath, snapshot, bookmark string) (err error) {
 	return
 
 }
+
+func ZFSRollback(fs *DatasetPath, snapshot FilesystemVersion, rollbackArgs ...string) (err error) {
+
+	snapabs := snapshot.ToAbsPath(fs)
+	if snapshot.Type != Snapshot {
+		return fmt.Errorf("can only rollback to snapshots, got %s", snapabs)
+	}
+
+	args := []string{"rollback"}
+	args = append(args, rollbackArgs...)
+	args = append(args, snapabs)
+
+	cmd := exec.Command(ZFS_BINARY, args...)
+
+	stderr := bytes.NewBuffer(make([]byte, 0, 1024))
+	cmd.Stderr = stderr
+
+	if err = cmd.Start(); err != nil {
+		return err
+	}
+
+	if err = cmd.Wait(); err != nil {
+		err = &ZFSError{
+			Stderr:  stderr.Bytes(),
+			WaitErr: err,
+		}
+	}
+
+	return err
+}