Scenario 3: Recovering from a corrupt RAID

The following steps describe the recommended approach for recovering a system from a corrupted RAID. Examine the system logs (/var/log/messages or /var/log/syslog) to determine if the disks on the disk controller are failing. If the failing disks are located on a controller that is failing, installing new disks on the failing controller will not solve the problem.

This scenario assumes that the corrupted RAID is a result of multiple failed disks:

1        Determine the failed disks by executing the appropriate disk controller commands or by launching the 3Ware utility (this can also be performed in BIOS):

tw_cli info <controller> [for 3ware-based systems]

or

cat /proc/mdstat [desktops and 1U systems]

2        Insert the new disk drives. Ideally, the new disks should be the same size, type and model as the original disks.

Once the new disks have been inserted, the rebuild process should begin automatically. If it does not, use the 3Ware utility (for rack-mount units), or the rebuild_disk script (desktops and 1U systems) to add the drives and launch the rebuild process.

3        When the new device has been rebuilt successfully, create a new Unitrends Postgres database with the following command:

/usr/bp/bin/setup_postgresql.sh create

4        Perform disaster recovery from vault or archive. (See  Disaster recovery from vault and Disaster recovery from archive for instructions).

5        If applicable, apply the manual steps following Disaster Recovery (see Post-recovery considerations for details).