You must manually run two scripts to remount preserved storage volumes and reformat any failed storage volumes. The first script remounts volumes that are properly formatted as StorageGRID storage volumes. The second script reformats any unmounted volumes, rebuilds the Cassandra database, if needed, and starts services.
About this task
To complete this procedure, you perform these high-level tasks:
- Log in to the recovered Storage Node.
- Run the sn-remount-volumes script to remount properly formatted storage volumes. When this script runs, it does the following:
- Mounts and unmounts each storage volume to replay the XFS journal.
- Performs an XFS file consistency check.
- If the file system is consistent, determines if the storage volume is a properly formatted StorageGRID storage volume.
- If the storage volume is properly formatted, remounts the storage volume. Any existing data on the volume remains intact.
- Review the script output and resolve any issues.
- Run the sn-recovery-postinstall.sh script. When this script runs, it does the following:
- Reformats any storage volumes that the sn-remount-volumes script could not mount or that were found to be improperly formatted.
Note: If a storage volume is reformatted, any data on that volume is lost. You must perform an additional procedure to restore object data from other locations in the grid, assuming that ILM rules were configured to store more than one object copy.
- Rebuilds the Cassandra database on the node, if needed.
- Starts the services on the Storage Node.
After you finish
After the sn-recovery-postinstall.sh script has started services on the node, you can restore object data to any storage volumes that were formatted by the script, as described in the next procedure.