Partially done physical restores

When you restore a physical backup, the operation can result in the following statuses:

• Done: The restore completed successfully on all nodes.
• Error: The restore failed and could not be completed.
• Partly-Done: The restore was successful on at least one node in each replica set, but failed on some nodes.

Restore status decision flow

The decision flow is explained below:

1. Check the restore status with pbm describe-restore <restore_name> or pbm status.
2. Did the restore succeed on all nodes?
   • Yes: The status is Done. Continue with post-restore steps.
   • No: Go to the next step.
3. Did at least one node in each replica set restore successfully?
   • Yes: Is fallbackEnabled set to true?
     • Yes: Is allowPartlyDone enabled?
       • Yes: The status is PartlyDone. See Post-restore steps for partlyDone restores.
       • No: PBM triggers a fallback procedure and restores your cluster to its pre-restore state. If fallbackEnabled is disabled, the restore status is Error.
   • No: The status is Error. You’ll need to intervene manually.

Post-restore steps for partlyDone restores

If your restore finishes with the partlyDone status, you can still start the cluster and wait for the failed node to receive the data via the initial sync. Here’s what you need to do:

1. Check the restore status with pbm status or pbm describe-restore <restore_name>.
2. Start all mongod nodes. The failed nodes will perform initial sync from the healthy nodes.
3. Start pbm-agents on every node.
4. Start the balancer and all mongos nodes.
5. Monitor for the nodes to complete initial sync and report the ready status.
6. Make a fresh backup to serve as the new base for future restores.
7. Enable point-in-time recovery if you need it.

---

Last update: September 25, 2025
Created: September 25, 2025