From fffda09f673fa9927b051e3e6eb8a4219b7e6d9f Mon Sep 17 00:00:00 2001
From: Christian Schwarz <me@cschwarz.com>
Date: Sun, 21 Oct 2018 13:37:30 +0200
Subject: [PATCH] replication + pruner: progress markers during planning

---
 daemon/pruner/pruner.go | 9 ++++++++-
 replication/mainfsm.go  | 8 ++++++++
 2 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/daemon/pruner/pruner.go b/daemon/pruner/pruner.go
index 714452f..7c0572c 100644
--- a/daemon/pruner/pruner.go
+++ b/daemon/pruner/pruner.go
@@ -362,6 +362,10 @@ func onErr(u updater, e error) state {
 func statePlan(a *args, u updater) state {
 
 	ctx, target, receiver := a.ctx, a.target, a.receiver
+	var ka *watchdog.KeepAlive
+	u(func(pruner *Pruner) {
+		ka = &pruner.Progress
+	})
 
 	tfss, err := target.ListFilesystems(ctx)
 	if err != nil {
@@ -385,6 +389,8 @@ func statePlan(a *args, u updater) state {
 			l.WithError(err).Error("cannot list filesystem versions")
 			return onErr(u, err)
 		}
+		// no progress here since we could run in a live-lock (must have used target AND receiver before progress)
+
 		pfs.snaps = make([]pruning.Snapshot, 0, len(tfsvs))
 
 		rcReq := &pdu.ReplicationCursorReq{
@@ -402,6 +408,7 @@ func statePlan(a *args, u updater) state {
 			l.WithField("reqErr", rc.GetError()).Error("cannot get replication cursor")
 			return onErr(u, err)
 		}
+		ka.MadeProgress()
 
 
 		// scan from older to newer, all snapshots older than cursor are interpreted as replicated
@@ -448,7 +455,7 @@ func statePlan(a *args, u updater) state {
 
 		// Apply prune rules
 		pfs.destroyList = pruning.PruneSnapshots(pfs.snaps, a.rules)
-
+		ka.MadeProgress()
 	}
 
 	return u(func(pruner *Pruner) {
diff --git a/replication/mainfsm.go b/replication/mainfsm.go
index ff920cd..a2a8948 100644
--- a/replication/mainfsm.go
+++ b/replication/mainfsm.go
@@ -229,6 +229,7 @@ func statePlanning(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, r
 		log.WithError(err).Error("error listing sender filesystems")
 		return handlePlanningError(err)
 	}
+	// no progress here since we could run in a live-lock on connectivity issues
 
 	rfss, err := receiver.ListFilesystems(ctx)
 	if err != nil {
@@ -236,6 +237,8 @@ func statePlanning(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, r
 		return handlePlanningError(err)
 	}
 
+	ka.MadeProgress() // for both sender and receiver
+
 	q := NewReplicationQueue()
 	mainlog := log
 	for _, fs := range sfss {
@@ -249,6 +252,7 @@ func statePlanning(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, r
 			log.WithError(err).Error("cannot get remote filesystem versions")
 			return handlePlanningError(err)
 		}
+		ka.MadeProgress()
 
 		if len(sfsvs) < 1 {
 			err := errors.New("sender does not have any versions")
@@ -278,6 +282,7 @@ func statePlanning(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, r
 		} else {
 			rfsvs = []*pdu.FilesystemVersion{}
 		}
+		ka.MadeProgress()
 
 		path, conflict := IncrementalPath(rfsvs, sfsvs)
 		if conflict != nil {
@@ -291,6 +296,7 @@ func statePlanning(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, r
 				log.WithField("problem", msg).Error("cannot resolve conflict")
 			}
 		}
+		ka.MadeProgress()
 		if path == nil {
 			q.Add(fsrep.NewReplicationWithPermanentError(fs.Path, conflict))
 			continue
@@ -309,12 +315,14 @@ func statePlanning(ctx context.Context, ka *watchdog.KeepAlive, sender Sender, r
 			}
 		}
 		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.Add(qitem)
 	}