Can anyone tell me how to restore a replica set using filesystem snapshot (copy of dbpath) without downtime ?
I have simulated by following below process.
- Took entire dbpath snapshot
- Dropped a database from Primary
- Tried restoring it in rolling manner
It’s obvious that even if we will replace secondaries data files with the snapshot, it will again re-issue drop command from Primary’s oplog.
I have found out below process but it will need completed downtime.
- Stop all replica set nodes
- Replaca data files with recent snapshot
- Start all nodes
Is there any other workaround to get back the original data without downtime using only the snapshot ?
Also, is it possible to recover specific DB from snapshots ?