Recovering an IBM server single disk failure

RAID 5 distributes data across all disks with redundancy. If one disk fails, the data can be recovered from the others.

When a single disk fails, the array becomes ‘degraded’. The data can be rebuilt using parity across the remaining disks. With a ‘hot-spare’ disk, the controller will rebuild the data automatically.

Check the motherboard by powering off, removing all drives and powering on. If it’s still blinking or not responding, it may be a motherboard fault. If you hear a bleep and the power LED is static then the motherboard is healthy.

What if you don’t have a ‘hot-spare’?

In this case, the server cannot recover the data automatically and a rebuild is necessary. The procedure for doing this is as follows:

1. Run iprconfig
2. Select work with disk unit recovery
3. Select rebuild disk unit data
4. Highlight the disks you need to rebuild
5. Press enter to rebuild or ‘q’ to cancel

What if multiple disks have failed?

In this case the array fails and the data becomes inaccessible. The array can be rebuilt using new disks and the data can be restored from backup. If there’s no backup the only option is either a hardware repair of the disks, or data recovery from the remaining disks. If you’re attempting to recover the data, create a RAID structured image and a backup on a different volume.


Further reading

The best RAID level for performance

What is RAID redundancy?

Redundancy and RAID 10