Deduplication Database Recovery post Blue Screen of Death (BSOD).

This Document is applicable if any Media Agent (MA) is affected by a BSOD error.

Things to check and steps to be taken after the BSOD error occurs on the Media Agent.

1. Check the status of all the drives on the server including but not limited to: Index cache Drive, Deduplication Database Drive, Disk libraries.
2. Check if Commvault services are online and the Media Agent is showing as ‘ready’ in the Commcell console.

Possible behaviors following a BSOD event on the Media Agent

• If the Media Agent faced BSOD errors, it's expected that the Deduplication Database (DDB) will be in inconsistent state.
  • To recover from the state, Commvault initiates DDB reconstruction operation automatically.
  • The reconstruction job can be seen on the ‘Job Controller’ for the individual DDB engine hosted on the media agent(s).
• Until the reconstruction job is completed successfully, the partition status will be showing offline in the GUI.
  • The backup operations and Aux Copy operations referring to that offline deduplication database will be in the waiting state. It may show the error: [The active DDB of current storage policy copy is not available to use.]
  • The backup operations and Aux Copy operations will be resumed automatically post the DDB reconstruction job is completed.
  • Pruning and DDB verification job can be resumed post the DDB reconstruction job is completed and the DDB is back online.

Note: Recovery operations will run without issue even if a Deduplication Database is offline.

FAQ

How long it will take for the DDB reconstruction to be completed? DDB reconstruction is a process to build the database by reading the data from the library. Hence the ETA of the reconstruction job is environmentally dependent on Mountpath, Media Agent and DDB drive performance.

Can I increase the throughput of currently running DDB reconstruction operation? No, it cannot be done on a running job. Performance enhancement can only be done based on the available resources on the environment. By default, the optimum configuration is present.

What can be done if the Mountpath / DDB drives are not accessible at the OS end? Please check with your internal teams to make sure the drives are attached and accessible at OS end.

If the DDB reconstruction job is failing for any reason, what should I do? Follow standard troubleshooting procedure for the failure, if required, a support case with Commvault can be logged.

Deduplication Database can become corrupted due to unclear shutdown The following error may be encountered: [The active DDB of current storage policy copy is not available to use.] A DDB reconstruction will launch automatically and bring the database back online.

If the DDBs are online and the backup is failing with error "Could not connect to the DDB process on the target MediaAgent", please add the following additional setting on the Dedup mediaagents and set the value to 600. Resume the backup job after that.

Name: SIDBIdleTimeout Category: MediaAgent Type: integer Value: 600

Symptoms

N/A

Causes

N/A

Resolution

N/A