replication handle cases when entry already exists

3 files changed, 36 insertions, 11 deletions

diff --git a/weed/replication/replicator.go b/weed/replication/replicator.go
index 215be5992..834da6217 100644
--- a/weed/replication/replicator.go
+++ b/weed/replication/replicator.go
@@ -9,6 +9,7 @@ import (
 	"github.com/chrislusf/seaweedfs/weed/replication/sink/filersink"
 	"github.com/chrislusf/seaweedfs/weed/replication/source"
 	"github.com/chrislusf/seaweedfs/weed/util"
+	"github.com/chrislusf/seaweedfs/weed/glog"
 )
 
 type Replicator struct {
@@ -43,5 +44,18 @@ func (r *Replicator) Replicate(key string, message *filer_pb.EventNotification)
 	if message.OldEntry == nil && message.NewEntry != nil {
 		return r.sink.CreateEntry(key, message.NewEntry)
 	}
-	return r.sink.UpdateEntry(key, message.OldEntry, message.NewEntry, message.DeleteChunks)
+	if existingEntry, err := r.sink.LookupEntry(key); err == nil {
+		if message.OldEntry == nil && message.NewEntry == nil {
+			glog.V(0).Infof("message %+v existingEntry: %+v", message, existingEntry)
+			return r.sink.DeleteEntry(key, existingEntry, true)
+		}
+		return r.sink.UpdateEntry(key, message.OldEntry, message.NewEntry, existingEntry, message.DeleteChunks)
+	}
+
+	glog.V(0).Infof("key:%s, message %+v", key, message)
+	if message.OldEntry == nil && message.NewEntry == nil {
+		return nil
+	}
+
+	return r.sink.CreateEntry(key, message.NewEntry)
 }
diff --git a/weed/replication/sink/filersink/filer_sink.go b/weed/replication/sink/filersink/filer_sink.go
index 1cbf52864..c98c99f34 100644
--- a/weed/replication/sink/filersink/filer_sink.go
+++ b/weed/replication/sink/filersink/filer_sink.go
@@ -121,14 +121,13 @@ func (fs *FilerSink) CreateEntry(key string, entry *filer_pb.Entry) error {
 	})
 }
 
-func (fs *FilerSink) UpdateEntry(key string, oldEntry, newEntry *filer_pb.Entry, deleteIncludeChunks bool) (err error) {
+func (fs *FilerSink) LookupEntry(key string) (entry *filer_pb.Entry, err error) {
 
 	ctx := context.Background()
 
 	dir, name := filer2.FullPath(key).DirAndName()
 
 	// read existing entry
-	var entry *filer_pb.Entry
 	err = fs.withFilerClient(func(client filer_pb.SeaweedFilerClient) error {
 
 		request := &filer_pb.LookupDirectoryEntryRequest{
@@ -136,7 +135,7 @@ func (fs *FilerSink) UpdateEntry(key string, oldEntry, newEntry *filer_pb.Entry,
 			Name:      name,
 		}
 
-		glog.V(4).Infof("lookup directory entry: %v", request)
+		glog.V(4).Infof("lookup entry: %v", request)
 		resp, err := client.LookupDirectoryEntry(ctx, request)
 		if err != nil {
 			glog.V(0).Infof("lookup %s: %v", key, err)
@@ -149,10 +148,21 @@ func (fs *FilerSink) UpdateEntry(key string, oldEntry, newEntry *filer_pb.Entry,
 	})
 
 	if err != nil {
-		return fmt.Errorf("lookup when updating %s: %v", key, err)
+		return nil, fmt.Errorf("lookup %s: %v", key, err)
 	}
 
-	if filer2.ETag(newEntry.Chunks) == filer2.ETag(entry.Chunks) {
+	return entry, nil
+}
+
+func (fs *FilerSink) UpdateEntry(key string, oldEntry, newEntry, existingEntry *filer_pb.Entry, deleteIncludeChunks bool) (err error) {
+
+	ctx := context.Background()
+
+	dir, _ := filer2.FullPath(key).DirAndName()
+
+	glog.V(0).Infof("oldEntry %+v, newEntry %+v, existingEntry: %+v", oldEntry, newEntry, existingEntry)
+
+	if filer2.ETag(newEntry.Chunks) == filer2.ETag(existingEntry.Chunks) {
 		// skip if no change
 		// this usually happens when retrying the replication
 		glog.V(0).Infof("already replicated %s", key)
@@ -163,7 +173,7 @@ func (fs *FilerSink) UpdateEntry(key string, oldEntry, newEntry *filer_pb.Entry,
 		// delete the chunks that are deleted from the source
 		if deleteIncludeChunks {
 			// remove the deleted chunks. Actual data deletion happens in filer UpdateEntry FindUnusedFileChunks
-			entry.Chunks = minusChunks(entry.Chunks, deletedChunks)
+			existingEntry.Chunks = minusChunks(existingEntry.Chunks, deletedChunks)
 		}
 
 		// replicate the chunks that are new in the source
@@ -171,7 +181,7 @@ func (fs *FilerSink) UpdateEntry(key string, oldEntry, newEntry *filer_pb.Entry,
 		if err != nil {
 			return fmt.Errorf("replicte %s chunks error: %v", key, err)
 		}
-		entry.Chunks = append(entry.Chunks, replicatedChunks...)
+		existingEntry.Chunks = append(existingEntry.Chunks, replicatedChunks...)
 	}
 
 	// save updated meta data
@@ -179,11 +189,11 @@ func (fs *FilerSink) UpdateEntry(key string, oldEntry, newEntry *filer_pb.Entry,
 
 		request := &filer_pb.UpdateEntryRequest{
 			Directory: dir,
-			Entry:     entry,
+			Entry:     existingEntry,
 		}
 
 		if _, err := client.UpdateEntry(ctx, request); err != nil {
-			return fmt.Errorf("update entry %s: %v", key, err)
+			return fmt.Errorf("update existingEntry %s: %v", key, err)
 		}
 
 		return nil
diff --git a/weed/replication/sink/replication_sink.go b/weed/replication/sink/replication_sink.go
index bb4a8aa83..c33f3251b 100644
--- a/weed/replication/sink/replication_sink.go
+++ b/weed/replication/sink/replication_sink.go
@@ -8,7 +8,8 @@ import (
 type ReplicationSink interface {
 	DeleteEntry(key string, entry *filer_pb.Entry, deleteIncludeChunks bool) error
 	CreateEntry(key string, entry *filer_pb.Entry) error
-	UpdateEntry(key string, oldEntry, newEntry *filer_pb.Entry, deleteIncludeChunks bool) error
+	UpdateEntry(key string, oldEntry, newEntry, existingEntry *filer_pb.Entry, deleteIncludeChunks bool) error
+	LookupEntry(key string) (entry *filer_pb.Entry, err error)
 	GetSinkToDirectory() string
 	SetSourceFiler(s *source.FilerSource)
 }