[BUG] Snapshotter fail to apply watch when revision is already compacted. #599
Labels
kind/bug
Bug
priority/1
Priority (lower number equals higher priority)
status/closed
Issue is closed (either delivered or triaged)
Describe the bug:
When
snapshotter
get close due to some error, so it tries to restart the snapshotter by apply the watch on etcd. But due to etcd'sauto-compaction
etcd might already compacted the revision numbersay X
and when backup-restore tries to apply the watch on revision which is<=X
(means revision number is already compacted) this will leads to error in watch connection hence watch channel get close, and backup-restore never able to restart the snapshotter.Expected behavior:
If
snapshotter
fail to apply watch when revision is already compacted then it should take a full snapshot to come out of this situation.How To Reproduce (as minimally and precisely as possible):
etcdctl compact <Revision no>
Screenshots (if applicable):
Environment (please complete the following information):
v0.22.0
Anything else we need to know?:
The text was updated successfully, but these errors were encountered: