Based on point-in-time Recovery oplog Replay Recovery data after PBM physical backup, orphan documents exist for a certain shard, and as a result, the information cannot be deleted by the CleanuPorphedcommand
Sorry but I don’t understand exactly what is the question? if you think there is a bug you can open a ticket in jira.percona.com
I’m sorry. I mean
Orphan files will be generated when data is recovered using point-in-time Recovery Oplog Replay of PBM. How can we avoid this situation
At this time you cannot restore a single shard, it’s all or nothing (although that might change in the future). See the “Important” box in this page: Running Percona Backup for MongoDB — Percona Backup for MongoDB 1.8 Documentation for more details on this.
I am the whole cluster to restore the problem