Incremental backups for MongoDB Ops Manager may be incomplete under specific load conditions
Under certain timing and workload conditions, MongoDB Ops Manager incremental (INCR) backups may not contain all changed data. This issue can occur when new data is created shortly after a snapshot is taken and is more likely in heavily loaded multi-node replica set environments. Backup jobs may complete successfully, but restores that rely on the affected incremental backup chain can result in missing or inconsistent data.
This limitation applies only to MongoDB Ops Manager incremental (INCR) backups introduced in NetBackup 11.2 and later. Full (FULL) backups and Transaction Log (TLOG) backups are not affected. There is currently no mechanism to detect this condition during backup execution. The issue may only become apparent during recovery validation.
Workaround
Avoid running incremental backups immediately after rapid data or schema changes.
Allow sufficient time for data to stabilize across replica set members before starting incremental backups.
When possible, use FULL backups or Transaction Log (TLOG) backups instead of relying solely on incremental backups until the issue is resolved.