daemon/control: envconst timeout for control socket server-side timeouts

refs #262

README.md

@@ -1,8 +1,9 @@
 [![GitHub license](https://img.shields.io/github/license/zrepl/zrepl.svg)](https://github.com/zrepl/zrepl/blob/master/LICENSE)
 [![Language: Go](https://img.shields.io/badge/language-Go-6ad7e5.svg)](https://golang.org/)
 [![User Docs](https://img.shields.io/badge/docs-web-blue.svg)](https://zrepl.github.io)
-[![Donate via PayPal](https://img.shields.io/badge/donate-paypal-yellow.svg)](https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96)
+[![Donate via Patreon](https://img.shields.io/endpoint.svg?url=https%3A%2F%2Fshieldsio-patreon.herokuapp.com%2Fzrepl%2Fpledges&style=flat&color=yellow)](https://www.patreon.com/zrepl)
 [![Donate via Liberapay](https://img.shields.io/liberapay/receives/zrepl.svg?logo=liberapay)](https://liberapay.com/zrepl/donate)
+[![Donate via PayPal](https://img.shields.io/badge/donate-paypal-yellow.svg)](https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=R5QSXJVYHGX96)
 [![Twitter](https://img.shields.io/twitter/url/https/github.com/zrepl/zrepl.svg?style=social)](https://twitter.com/intent/tweet?text=Wow:&url=https%3A%2F%2Fgithub.com%2Fzrepl%2Fzrepl)
 
 # zrepl
@@ -23,6 +24,7 @@ zrepl is a one-stop ZFS backup & replication solution.
    If so, think of an expressive configuration example.
 2. Think of at least one use case that generalizes from your concrete application.
 3. Open an issue on GitHub with example conf & use case attached.
+4. **Optional**: [Post a bounty](https://www.bountysource.com/teams/zrepl) on the issue, or [contact Christian Schwarz](https://cschwarz.com) for contract work.
 
 The above does not apply if you already implemented everything.
 Check out the *Coding Workflow* section below for details.
@@ -46,11 +48,8 @@ zrepl is written in [Go](https://golang.org) and uses [Go modules](https://githu
 The documentation is written in [ReStructured Text](http://docutils.sourceforge.net/rst.html) using the [Sphinx](https://www.sphinx-doc.org) framework.
 
 To get started, run `./lazy.sh devsetup` to easily install build dependencies and read `docs/installation.rst -> Compiling from Source`.
-
+`lazy.sh` uses `python3-pip` to fetch the build dependencies for the docs - you might want to use a [venv](https://docs.python.org/3/library/venv.html).
-### Overall Architecture
+If you just want to install the Go dependencies, run `./lazy.sh godep`.
-
-The application architecture is documented as part of the user docs in the *Implementation* section (`docs/content/impl`).
-Make sure to develop an understanding how zrepl is typically used by studying the user docs first.
 
 ### Project Structure
 
@@ -62,14 +61,15 @@ Make sure to develop an understanding how zrepl is typically used by studying th
 │   └── samples
 ├── daemon                  # the implementation of `zrepl daemon` subcommand
 │   ├── filters
+│   ├── hooks               # snapshot hooks
 │   ├── job                 # job implementations
 │   ├── logging             # logging outlets + formatters
 │   ├── nethelpers
 │   ├── prometheus
 │   ├── pruner              # pruner implementation
 │   ├── snapper             # snapshotter implementation
-├── docs                    # sphinx-based documentation
 ├── dist                    # supplemental material for users & package maintainers
+├── docs                    # sphinx-based documentation
 │   ├── **/*.rst            # documentation in reStructuredText
 │   ├── sphinxconf
 │   │   └── conf.py         # sphinx config (see commit 445a280 why its not in docs/)
@@ -78,6 +78,7 @@ Make sure to develop an understanding how zrepl is typically used by studying th
 │   └── public_git          # checkout of zrepl.github.io managed by above shell script
 ├── endpoint                # implementation of replication endpoints (=> package replication)
 ├── logger                  # our own logger package
+├── platformtest            # test suite for our zfs abstractions (error classification, etc)
 ├── pruning                 # pruning rules (the logic, not the actual execution)
 │   └── retentiongrid
 ├── replication
@@ -92,14 +93,13 @@ Make sure to develop an understanding how zrepl is typically used by studying th
 │   ├── transportmux        # TCP connecter and listener used to split control & data traffic
 │   └── versionhandshake    # replication protocol version handshake perfomed on newly established connections
 ├── tlsconf                 # abstraction for Go TLS server + client config
-├── transport               # transports implementation
+├── transport               # transport implementations
 │   ├── fromconfig
 │   ├── local
 │   ├── ssh
 │   ├── tcp
 │   └── tls
 ├── util
-├── vendor                  # managed by dep
 ├── version                 # abstraction for versions (filled during build by Makefile)
 └── zfs                     # zfs(8) wrappers
 ```
@@ -134,3 +134,15 @@ There will not be a big refactoring (an attempt was made, but it's destroying to
 
 However, new contributions & patches should fix naming without further notice in the commit message.
 
+### RPC debugging
+
+Optionally, there are various RPC-related environment varibles, that if set to something != `""` will produce additional debug output on stderr:
+
+https://github.com/zrepl/zrepl/blob/master/rpc/rpc_debug.go#L11
+
+https://github.com/zrepl/zrepl/blob/master/rpc/dataconn/dataconn_debug.go#L11
+
+https://github.com/zrepl/zrepl/blob/master/rpc/dataconn/stream/stream_debug.go#L11
+
+https://github.com/zrepl/zrepl/blob/master/rpc/dataconn/heartbeatconn/heartbeatconn_debug.go#L11
+

client/signal.go

@@ -9,7 +9,7 @@ import (
 )
 
 var SignalCmd = &cli.Subcommand{
-	Use:   "signal [wakeup|reset] JOB [DATA]",
+	Use:   "signal [wakeup|reset] JOB",
 	Short: "wake up a job from wait state or abort its current invocation",
 	Run: func(subcommand *cli.Subcommand, args []string) error {
 		return runSignalCmd(subcommand.Config(), args)
@@ -17,13 +17,8 @@ var SignalCmd = &cli.Subcommand{
 }
 
 func runSignalCmd(config *config.Config, args []string) error {
-	if len(args) < 2 || len(args) > 3 {
+	if len(args) != 2 {
-		return errors.Errorf("Expected 2 arguments: [wakeup|reset|set-concurrency] JOB [DATA]")
+		return errors.Errorf("Expected 2 arguments: [wakeup|reset] JOB")
-	}
-
-	var data string
-	if len(args) == 3 {
-		data = args[2]
 	}
 
 	httpc, err := controlHttpClient(config.Global.Control.SockPath)
@@ -35,11 +30,9 @@ func runSignalCmd(config *config.Config, args []string) error {
 		struct {
 			Name string
 			Op   string
-			Data string
 		}{
 			Name: args[1],
 			Op:   args[0],
-			Data: data,
 		},
 		struct{}{},
 	)

client/testcmd.go

@@ -60,6 +60,8 @@ func runTestFilterCmd(subcommand *cli.Subcommand, args []string) error {
 		confFilter = j.Filesystems
 	case *config.PushJob:
 		confFilter = j.Filesystems
+	case *config.SnapJob:
+		confFilter = j.Filesystems
 	default:
 		return fmt.Errorf("job type %T does not have filesystems filter", j)
 	}

daemon/control.go

@@ -8,7 +8,6 @@ import (
 	"io"
 	"net"
 	"net/http"
-	"strconv"
 	"time"
 
 	"github.com/pkg/errors"
@@ -48,8 +47,6 @@ func (j *controlJob) OwnedDatasetSubtreeRoot() (p *zfs.DatasetPath, ok bool) { r
 
 func (j *controlJob) SenderConfig() *endpoint.SenderConfig { return nil }
 
-func (j *controlJob) SetConcurrency(concurrency int) error { return errors.Errorf("not supported") }
-
 var promControl struct {
 	requestBegin    *prometheus.CounterVec
 	requestFinished *prometheus.HistogramVec
@@ -129,7 +126,6 @@ func (j *controlJob) Run(ctx context.Context) {
 			type reqT struct {
 				Name string
 				Op   string
-				Data string
 			}
 			var req reqT
 			if decoder(&req) != nil {
@@ -142,14 +138,6 @@ func (j *controlJob) Run(ctx context.Context) {
 				err = j.jobs.wakeup(req.Name)
 			case "reset":
 				err = j.jobs.reset(req.Name)
-			case "set-concurrency":
-				var concurrency int
-				concurrency, err = strconv.Atoi(req.Data) // shadow
-				if err != nil {
-					// fallthrough outer
-				} else {
-					err = j.jobs.setConcurrency(req.Name, concurrency)
-				}
 			default:
 				err = fmt.Errorf("operation %q is invalid", req.Op)
 			}
@@ -159,8 +147,8 @@ func (j *controlJob) Run(ctx context.Context) {
 	server := http.Server{
 		Handler: mux,
 		// control socket is local, 1s timeout should be more than sufficient, even on a loaded system
-		WriteTimeout: 1 * time.Second,
+		WriteTimeout: envconst.Duration("ZREPL_DAEMON_CONTROL_WRITE_TIMEOUT", 1*time.Second),
-		ReadTimeout:  1 * time.Second,
+		ReadTimeout:  envconst.Duration("ZREPL_DAEMON_CONTROL_READ_TIMEOUT", 1*time.Second),
 	}
 
 outer:

daemon/daemon.go

@@ -176,18 +176,6 @@ func (s *jobs) reset(job string) error {
 	return wu()
 }
 
-func (s *jobs) setConcurrency(jobName string, concurrency int) error {
-	s.m.RLock()
-	defer s.m.RUnlock()
-
-	job, ok := s.jobs[jobName]
-	if !ok {
-		return errors.Errorf("Job %q does not exist", job)
-	}
-
-	return job.SetConcurrency(concurrency)
-}
-
 const (
 	jobNamePrometheus = "_prometheus"
 	jobNameControl    = "_control"

daemon/job/active.go

@@ -55,12 +55,9 @@ const (
 type activeSideTasks struct {
 	state ActiveSideState
 
-	concurrency int
-
 	// valid for state ActiveSideReplicating, ActiveSidePruneSender, ActiveSidePruneReceiver, ActiveSideDone
 	replicationReport driver.ReportFunc
 	replicationCancel context.CancelFunc
-	replicationSetConcurrency driver.SetConcurrencyFunc
 
 	// valid for state ActiveSidePruneSender, ActiveSidePruneReceiver, ActiveSideDone
 	prunerSender, prunerReceiver *pruner.Pruner
@@ -281,8 +278,6 @@ func activeSide(g *config.Global, in *config.ActiveJob, configJob interface{}) (
 		return nil, errors.Wrap(err, "invalid job name")
 	}
 
-	j.tasks.concurrency = 1 // FIXME
-
 	switch v := configJob.(type) {
 	case *config.PushJob:
 		j.mode, err = modePushFromConfig(g, v, j.name) // shadow
@@ -380,22 +375,6 @@ func (j *ActiveSide) SenderConfig() *endpoint.SenderConfig {
 	return push.senderConfig
 }
 
-func (j *ActiveSide) SetConcurrency(concurrency int) (err error) {
-	j.updateTasks(func(tasks *activeSideTasks) {
-		if tasks.replicationSetConcurrency != nil {
-			err = tasks.replicationSetConcurrency(concurrency) // no shadow
-			if err == nil {
-				tasks.concurrency = concurrency
-			}
-		} else {
-			// FIXME this is not great, should always be able to set it
-			err = errors.Errorf("cannot set while not replicating")
-		}
-
-	})
-	return err
-}
-
 func (j *ActiveSide) Run(ctx context.Context) {
 	log := GetLogger(ctx)
 	ctx = logging.WithSubsystemLoggers(ctx, log)
@@ -457,14 +436,11 @@ func (j *ActiveSide) do(ctx context.Context) {
 		ctx, repCancel := context.WithCancel(ctx)
 		var repWait driver.WaitFunc
 		j.updateTasks(func(tasks *activeSideTasks) {
-			// reset it (almost)
+			// reset it
-			old := *tasks
+			*tasks = activeSideTasks{}
-			*tasks = activeSideTasks{
-				concurrency: old.concurrency,
-			}
 			tasks.replicationCancel = repCancel
-			tasks.replicationReport, repWait, tasks.replicationSetConcurrency = replication.Do(
+			tasks.replicationReport, repWait = replication.Do(
-				ctx, tasks.concurrency, logic.NewPlanner(j.promRepStateSecs, j.promBytesReplicated, sender, receiver, j.mode.PlannerPolicy()),
+				ctx, logic.NewPlanner(j.promRepStateSecs, j.promBytesReplicated, sender, receiver, j.mode.PlannerPolicy()),
 			)
 			tasks.state = ActiveSideReplicating
 		})

daemon/job/job.go

@@ -40,7 +40,6 @@ type Job interface {
 	// must return the root of that subtree as rfs and ok = true
 	OwnedDatasetSubtreeRoot() (rfs *zfs.DatasetPath, ok bool)
 	SenderConfig() *endpoint.SenderConfig
-	SetConcurrency(concurrency int) error
 }
 
 type Type string

daemon/job/passive.go

@@ -163,8 +163,6 @@ func (j *PassiveSide) SenderConfig() *endpoint.SenderConfig {
 
 func (*PassiveSide) RegisterMetrics(registerer prometheus.Registerer) {}
 
-func (*PassiveSide) SetConcurrency(concurrency int) error { return errors.Errorf("not supported") }
-
 func (j *PassiveSide) Run(ctx context.Context) {
 
 	log := GetLogger(ctx)

daemon/job/snapjob.go

@@ -117,8 +117,6 @@ outer:
 	}
 }
 
-func (*SnapJob) SetConcurrency(concurrency int) error { return errors.Errorf("not supported") }
-
 // Adaptor that implements pruner.History around a pruner.Target.
 // The ReplicationCursor method is Get-op only and always returns
 // the filesystem's most recent version's GUID.

daemon/prometheus.go

@@ -5,7 +5,6 @@ import (
 	"net"
 	"net/http"
 
-	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
 
@@ -52,8 +51,6 @@ func (j *prometheusJob) OwnedDatasetSubtreeRoot() (p *zfs.DatasetPath, ok bool)
 
 func (j *prometheusJob) SenderConfig() *endpoint.SenderConfig { return nil }
 
-func (j *prometheusJob) SetConcurrency(concurrency int) error { return errors.Errorf("not supported") }
-
 func (j *prometheusJob) RegisterMetrics(registerer prometheus.Registerer) {}
 
 func (j *prometheusJob) Run(ctx context.Context) {

docs/installation.rst

@@ -101,9 +101,9 @@ and serves as a reference for build dependencies and procedure:
 
 ::
 
-    git clone https://github.com/zrepl/zrepl.git
+    git clone https://github.com/zrepl/zrepl.git && \
-    cd zrepl
+    cd zrepl && \
-    sudo docker build -t zrepl_build -f build.Dockerfile .
+    sudo docker build -t zrepl_build -f build.Dockerfile . && \
     sudo docker run -it --rm \
         -v "${PWD}:/src" \
         --user "$(id -u):$(id -g)" \

replication/driver/replication_driver.go

@@ -89,8 +89,6 @@ type attempt struct {
 	// if both are nil, it must be assumed that Planner.Plan is active
 	planErr *timedError
 	fss     []*fs
-
-	concurrency int
 }
 
 type timedError struct {
@@ -172,12 +170,11 @@ type step struct {
 
 type ReportFunc func() *report.Report
 type WaitFunc func(block bool) (done bool)
-type SetConcurrencyFunc func(concurrency int) error
 
 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, initialConcurrency int, planner Planner) (ReportFunc, WaitFunc, SetConcurrencyFunc) {
+func Do(ctx context.Context, planner Planner) (ReportFunc, WaitFunc) {
 	log := getLog(ctx)
 	l := chainlock.New()
 	run := &run{
@@ -185,8 +182,6 @@ func Do(ctx context.Context, initialConcurrency int, planner Planner) (ReportFun
 		startedAt: time.Now(),
 	}
 
-	concurrencyChanges := make(chan concurrencyChange)
-
 	done := make(chan struct{})
 	go func() {
 		defer close(done)
@@ -203,21 +198,15 @@ func Do(ctx context.Context, initialConcurrency int, planner Planner) (ReportFun
 			run.waitReconnect.SetZero()
 			run.waitReconnectError = nil
 
-			prevConcurrency := initialConcurrency // FIXME default concurrency
-			if prev != nil {
-				prevConcurrency = prev.concurrency
-			}
-
 			// do current attempt
 			cur := &attempt{
 				l:         l,
 				startedAt: time.Now(),
 				planner:   planner,
-				concurrency: prevConcurrency,
 			}
 			run.attempts = append(run.attempts, cur)
 			run.l.DropWhile(func() {
-				cur.do(ctx, prev, concurrencyChanges)
+				cur.do(ctx, prev)
 			})
 			prev = cur
 			if ctx.Err() != nil {
@@ -288,25 +277,10 @@ func Do(ctx context.Context, initialConcurrency int, planner Planner) (ReportFun
 		defer run.l.Lock().Unlock()
 		return run.report()
 	}
-	setConcurrency := func(concurrency int) (reterr error) {
+	return report, wait
-		var wg sync.WaitGroup
-		wg.Add(1)
-		concurrencyChanges <- concurrencyChange{ concurrency, func(err error) {
-			defer wg.Done()
-			reterr = err // shadow
-		}}
-		wg.Wait()
-		return reterr
-	}
-	return report, wait, setConcurrency
 }
 
-type concurrencyChange struct {
+func (a *attempt) do(ctx context.Context, prev *attempt) {
-	value          int
-	resultCallback func(error)
-}
-
-func (a *attempt) do(ctx context.Context, prev *attempt, setConcurrency <-chan concurrencyChange) {
 	pfss, err := a.planner.Plan(ctx)
 	errTime := time.Now()
 	defer a.l.Lock().Unlock()
@@ -380,8 +354,8 @@ func (a *attempt) do(ctx context.Context, prev *attempt, setConcurrency <-chan c
 	}
 	// invariant: prevs contains an entry for each unambigious correspondence
 
-	stepQueue := newStepQueue(a.concurrency)
+	stepQueue := newStepQueue()
-	defer stepQueue.Start()()
+	defer stepQueue.Start(1)() // TODO parallel replication
 	var fssesDone sync.WaitGroup
 	for _, f := range a.fss {
 		fssesDone.Add(1)
@@ -390,27 +364,8 @@ func (a *attempt) do(ctx context.Context, prev *attempt, setConcurrency <-chan c
 			f.do(ctx, stepQueue, prevs[f])
 		}(f)
 	}
-	changeConcurrencyDone := make(chan struct{})
-	go func() {
-		for {
-			select {
-			case change := <-setConcurrency:
-				err := stepQueue.SetConcurrency(change.value)
-				go change.resultCallback(err)
-				if err == nil {
-					a.l.Lock()
-					a.concurrency = change.value
-					a.l.Unlock()
-				}
-			case <-changeConcurrencyDone:
-				return
-				// not waiting for ctx.Done here, the main job are the fsses
-			}
-		}
-	}()
 	a.l.DropWhile(func() {
 		fssesDone.Wait()
-		close(changeConcurrencyDone)
 	})
 	a.finishedAt = time.Now()
 }
@@ -506,8 +461,7 @@ func (fs *fs) do(ctx context.Context, pq *stepQueue, prev *fs) {
 		// lock must not be held while executing step in order for reporting to work
 		fs.l.DropWhile(func() {
 			targetDate := s.step.TargetDate()
-			ctx, done := pq.WaitReady(ctx, fs, targetDate)()
+			defer pq.WaitReady(fs, targetDate)()
-			defer done()
 			err = s.step.Step(ctx) // no shadow
 			errTime = time.Now()   // no shadow
 		})
@@ -528,7 +482,6 @@ func (r *run) report() *report.Report {
 		WaitReconnectSince: r.waitReconnect.begin,
 		WaitReconnectUntil: r.waitReconnect.end,
 		WaitReconnectError: r.waitReconnectError.IntoReportError(),
-		// Concurrency:        r.concurrency,
 	}
 	for i := range report.Attempts {
 		report.Attempts[i] = r.attempts[i].report()

replication/driver/replication_driver_test.go

@@ -151,7 +151,7 @@ func TestReplication(t *testing.T) {
 	ctx := context.Background()
 
 	mp := &mockPlanner{}
-	getReport, wait, _ := Do(ctx, 1, mp)
+	getReport, wait := Do(ctx, mp)
 	begin := time.Now()
 	fireAt := []time.Duration{
 		// the following values are relative to the start

replication/driver/replication_stepqueue.go

@@ -2,8 +2,6 @@ package driver
 
 import (
 	"container/heap"
-	"fmt"
-	"sync"
 	"time"
 
 	"github.com/zrepl/zrepl/util/chainlock"
@@ -13,88 +11,51 @@ type stepQueueRec struct {
 	ident      interface{}
 	targetDate time.Time
 	wakeup     chan StepCompletedFunc
-	cancelDueToConcurrencyDownsize interace{}
 }
 
 type stepQueue struct {
 	stop chan struct{}
 	reqs chan stepQueueRec
-
-	// l protects all members except the channels above
-
-	l           *chainlock.L
-	pendingCond *sync.Cond
-
-	// ident => queueItem
-	pending    *stepQueueHeap
-	active     *stepQueueHeap
-	queueItems map[interface{}]*stepQueueHeapItem // for tracking used idents in both pending and active
-
-	// stopped is used for cancellation of "wake" goroutine
-	stopped bool
-
-	concurrency int
 }
 
 type stepQueueHeapItem struct {
 	idx int
-	req *stepQueueRec
+	req stepQueueRec
-}
-type stepQueueHeap struct {
-	items   []*stepQueueHeapItem
-	reverse bool // never change after pushing first element
 }
+type stepQueueHeap []*stepQueueHeapItem
 
 func (h stepQueueHeap) Less(i, j int) bool {
-	res := h.items[i].req.targetDate.Before(h.items[j].req.targetDate)
+	return h[i].req.targetDate.Before(h[j].req.targetDate)
-	if h.reverse {
-		return !res
-	}
-	return res
 }
 
 func (h stepQueueHeap) Swap(i, j int) {
-	h.items[i], h.items[j] = h.items[j], h.items[i]
+	h[i], h[j] = h[j], h[i]
-	h.items[i].idx = i
+	h[i].idx = i
-	h.items[j].idx = j
+	h[j].idx = j
 }
 
 func (h stepQueueHeap) Len() int {
-	return len(h.items)
+	return len(h)
 }
 
 func (h *stepQueueHeap) Push(elem interface{}) {
 	hitem := elem.(*stepQueueHeapItem)
 	hitem.idx = h.Len()
-	h.items = append(h.items, hitem)
+	*h = append(*h, hitem)
 }
 
 func (h *stepQueueHeap) Pop() interface{} {
-	elem := h.items[h.Len()-1]
+	elem := (*h)[h.Len()-1]
 	elem.idx = -1
-	h.items = h.items[:h.Len()-1]
+	*h = (*h)[:h.Len()-1]
 	return elem
 }
 
 // returned stepQueue must be closed with method Close
-func newStepQueue(concurrency int) *stepQueue {
+func newStepQueue() *stepQueue {
-	l := chainlock.New()
 	q := &stepQueue{
 		stop: make(chan struct{}),
 		reqs: make(chan stepQueueRec),
-		l:           l,
-		pendingCond: l.NewCond(),
-		// priority queue
-		pending: &stepQueueHeap{reverse: false},
-		active:  &stepQueueHeap{reverse: true},
-		// ident => queueItem
-		queueItems: make(map[interface{}]*stepQueueHeapItem),
-		// stopped is used for cancellation of "wake" goroutine
-		stopped: false,
-	}
-	err := q.setConcurrencyLocked(concurrency)
-	if err != nil {
-		panic(err)
 	}
 	return q
 }
@@ -104,12 +65,25 @@ func newStepQueue(concurrency int) *stepQueue {
 //
 // 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() (done func()) {
+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 q.l.Lock().Unlock()
+		defer l.Lock().Unlock()
-		q.stopped = true
+		stopped = true
-		q.pendingCond.Broadcast()
+		pendingCond.Broadcast()
 	}()
 	go func() { // "reqs" goroutine
 		for {
@@ -123,52 +97,41 @@ func (q *stepQueue) Start() (done func()) {
 				}
 			case req := <-q.reqs:
 				func() {
-					defer q.l.Lock().Unlock()
+					defer l.Lock().Unlock()
-					if _, ok := q.queueItems[req.ident]; ok {
+					if _, ok := queueItems[req.ident]; ok {
 						panic("WaitReady must not be called twice for the same ident")
 					}
 					qitem := &stepQueueHeapItem{
 						req: req,
 					}
-					q.queueItems[req.ident] = qitem
+					queueItems[req.ident] = qitem
-					heap.Push(q.pending, qitem)
+					heap.Push(pending, qitem)
-					q.pendingCond.Broadcast()
+					pendingCond.Broadcast()
 				}()
 			}
 		}
 	}()
 	go func() { // "wake" goroutine
-		defer q.l.Lock().Unlock()
+		defer l.Lock().Unlock()
 		for {
 
-			for !q.stopped && (q.active.Len() >= q.concurrency || q.pending.Len() == 0) {
+			for !stopped && (active >= concurrency || pending.Len() == 0) {
-				q.pendingCond.Wait()
+				pendingCond.Wait()
 			}
-			if q.stopped {
+			if stopped {
 				return
 			}
-			if q.pending.Len() <= 0 {
+			if pending.Len() <= 0 {
 				return
 			}
+			active++
+			next := heap.Pop(pending).(*stepQueueHeapItem).req
+			delete(queueItems, next.ident)
 
-			// pop from tracked items
+			next.wakeup <- func() {
-			next := heap.Pop(q.pending).(*stepQueueHeapItem)
+				defer l.Lock().Unlock()
-
+				active--
-			next.req.cancelDueToConcurrencyDownsize = 
+				pendingCond.Broadcast()
-
-			heap.Push(q.active, next)
-
-			next.req.wakeup <- func() {
-				defer q.l.Lock().Unlock()
-
-				//
-				qitem := &stepQueueHeapItem{
-					req: req,
-				}
-
-				// delete(q.queueItems, next.req.ident) // def
-
-				q.pendingCond.Broadcast()
 			}
 		}
 	}()
@@ -198,24 +161,3 @@ func (q *stepQueue) WaitReady(ident interface{}, targetDate time.Time) StepCompl
 	}
 	return q.sendAndWaitForWakeup(ident, targetDate)
 }
-
-// caller must hold lock
-func (q *stepQueue) setConcurrencyLocked(newConcurrency int) error {
-	if !(newConcurrency >= 1) {
-		return fmt.Errorf("concurrency must be >= 1 but requested %v", newConcurrency)
-	}
-	q.concurrency = newConcurrency
-	q.pendingCond.Broadcast() // wake up waiters who could make progress
-
-	for q.active.Len() > q.concurrency {
-		item := heap.Pop(q.active).(*stepQueueHeapItem)
-		item.req.cancelDueToConcurrencyDownsize()
-		heap.Push(q.pending, item)
-	}
-	return nil
-}
-
-func (q *stepQueue) SetConcurrency(new int) error {
-	defer q.l.Lock().Unlock()
-	return q.setConcurrencyLocked(new)
-}

replication/driver/replication_stepqueue_test.go

@@ -17,7 +17,7 @@ import (
 // (relies on scheduler responsivity of < 500ms)
 func TestPqNotconcurrent(t *testing.T) {
 	var ctr uint32
-	q := newStepQueue(1)
+	q := newStepQueue()
 	var wg sync.WaitGroup
 	wg.Add(4)
 	go func() {
@@ -29,7 +29,7 @@ func TestPqNotconcurrent(t *testing.T) {
 	}()
 
 	// give goroutine "1" 500ms to enter queue, get the active slot and enter time.Sleep
-	defer q.Start()()
+	defer q.Start(1)()
 	time.Sleep(500 * time.Millisecond)
 
 	// while "1" is still running, queue in "2", "3" and "4"
@@ -78,8 +78,7 @@ func (r record) String() string {
 // times for each step are close together.
 func TestPqConcurrent(t *testing.T) {
 
-	concurrency := 5
+	q := newStepQueue()
-	q := newStepQueue(concurrency)
 	var wg sync.WaitGroup
 	filesystems := 100
 	stepsPerFS := 20
@@ -105,7 +104,8 @@ func TestPqConcurrent(t *testing.T) {
 			records <- recs
 		}(fs)
 	}
-	defer q.Start()()
+	concurrency := 5
+	defer q.Start(concurrency)()
 	wg.Wait()
 	close(records)
 	t.Logf("loop done")

replication/logic/replication_logic.go

@@ -262,7 +262,6 @@ func (p *Planner) doPlanning(ctx context.Context) ([]*Filesystem, error) {
 		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 {

replication/replication.go

@@ -8,6 +8,6 @@ import (
 	"github.com/zrepl/zrepl/replication/driver"
 )
 
-func Do(ctx context.Context, initialConcurrency int, planner driver.Planner) (driver.ReportFunc, driver.WaitFunc, driver.SetConcurrencyFunc) {
+func Do(ctx context.Context, planner driver.Planner) (driver.ReportFunc, driver.WaitFunc) {
-	return driver.Do(ctx, initialConcurrency, planner)
+	return driver.Do(ctx, planner)
 }

replication/report/replication_report.go

@@ -10,7 +10,6 @@ type Report struct {
 	WaitReconnectSince, WaitReconnectUntil time.Time
 	WaitReconnectError                     *TimedError
 	Attempts                               []*AttemptReport
-	Concurrency                            int
 }
 
 var _, _ = json.Marshal(&Report{})

rpc/dataconn/timeoutconn/timeoutconn.go

@@ -49,7 +49,7 @@ type Wire interface {
 	// No data that could otherwise be Read is lost as a consequence of this call.
 	// The use case for this API is abortive connection shutdown.
 	// To provide any value over draining Wire using io.Read, an implementation
-	// will likely use out-of-bounds messaging mechanisms.
+	// will likely use out-of-band messaging mechanisms.
 	// TODO WaitForPeerClose() (supported bool, err error)
 }
 

transport/tls/serve_tls.go

@@ -32,7 +32,7 @@ func TLSListenerFactoryFromConfig(c *config.Global, in *config.TLSServe) (transp
 
 	serverCert, err := tls.LoadX509KeyPair(in.Cert, in.Key)
 	if err != nil {
-		return nil, errors.Wrap(err, "cannot parse cer/key pair")
+		return nil, errors.Wrap(err, "cannot parse cert/key pair")
 	}
 
 	clientCNs := make(map[string]struct{}, len(in.ClientCNs))