From cf52691959f6707337466d8c7d6c774c9ce4c6e6 Mon Sep 17 00:00:00 2001
From: Anis Elleuch <vadmeste@users.noreply.github.com>
Date: Fri, 2 Sep 2022 00:53:36 +0100
Subject: [PATCH] Save resync status in the backend using a last update
 timestamp (#15638)

Currently, there is a short time window where the code is allowed
to save the status of a replication resync. Currently, the window is
`now.Sub(st.EndTime) <= resyncTimeInterval`. Also, any failure to
write in the backend disks is not retried.

Refactor the code a little bit to rely on the last timestamp of a
successful write of the resync status of any given bucket in the
backend disks.
---
 cmd/bucket-replication.go | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

diff --git a/cmd/bucket-replication.go b/cmd/bucket-replication.go
index ce37f4627..1b08144f4 100644
--- a/cmd/bucket-replication.go
+++ b/cmd/bucket-replication.go
@@ -1987,25 +1987,32 @@ func (p *ReplicationPool) updateResyncStatus(ctx context.Context, objectAPI Obje
 	resyncTimer := time.NewTimer(resyncTimeInterval)
 	defer resyncTimer.Stop()
 
+	// For each bucket name, store the last timestamp of the
+	// successful save of replication status in the backend disks.
+	lastResyncStatusSave := make(map[string]time.Time)
+
 	for {
 		select {
 		case <-resyncTimer.C:
-			now := UTCNow()
 			p.resyncState.RLock()
 			for bucket, brs := range p.resyncState.statusMap {
 				var updt bool
+				// Save the replication status if one resync to any bucket target is still not finished
 				for _, st := range brs.TargetsMap {
-					// if resync in progress or just ended, needs to save to disk
-					if st.EndTime.Equal(timeSentinel) || now.Sub(st.EndTime) <= resyncTimeInterval {
+					if st.EndTime.Equal(timeSentinel) {
 						updt = true
 						break
 					}
 				}
+				// Save the replication status if a new stats update is found and not saved in the backend yet
+				if brs.LastUpdate.After(lastResyncStatusSave[bucket]) {
+					updt = true
+				}
 				if updt {
-					brs.LastUpdate = now
 					if err := saveResyncStatus(ctx, bucket, brs, objectAPI); err != nil {
 						logger.LogIf(ctx, fmt.Errorf("Could not save resync metadata to drive for %s - %w", bucket, err))
-						continue
+					} else {
+						lastResyncStatusSave[bucket] = brs.LastUpdate
 					}
 				}
 			}
@@ -2031,6 +2038,7 @@ func resyncBucket(ctx context.Context, bucket, arn string, heal bool, objectAPI
 		st.EndTime = UTCNow()
 		st.ResyncStatus = resyncStatus
 		m.TargetsMap[arn] = st
+		m.LastUpdate = UTCNow()
 		globalReplicationPool.resyncState.statusMap[bucket] = m
 		globalReplicationPool.resyncState.Unlock()
 	}()
@@ -2140,6 +2148,7 @@ func resyncBucket(ctx context.Context, bucket, arn string, heal bool, objectAPI
 			st.ReplicatedSize += roi.Size
 		}
 		m.TargetsMap[arn] = st
+		m.LastUpdate = UTCNow()
 		globalReplicationPool.resyncState.statusMap[bucket] = m
 		globalReplicationPool.resyncState.Unlock()
 	}