Copyright 2015 datarecovery.com.sg
Preparation for RAID Data Recovery
This guide is not meant for you to do DIY data recovery. Rather, these are some notes on what to do when you are facing problem with your RAID server. Whatever you do will determine if you will eventually get back all your data.
1.0 Check the Status of RAID Server
The engineer normally comes to know about the disk failure of RAID server by the beeping alarm, error LED or preset mail alert. If the RAID server had fault tolerance built in such as RAID 5, the server should still be accessible when a single disk fails. The system should run in critical or degraded mode. That means one more disk failure will result in lost of RAID volume. If the critical data was little, the best thing to do is to copy out the data first without attempting any rescue such as rebuild. Rebuild frequently may trigger further disk failure (see Myth of RAID failure on why it is so).
If the data size is large, there is concern that the degraded RAID volume may be stressed beyond the failure zone during the heavy reading and copying. If the data is really important, it is better to take a clone of individual disk as additional backup. Obviously the disk could even fail during the cloning process. So how to decide the best course of action?
One may want to consider sending the disks to proferssional vendor for cloning. A well equipped service vendor would have the necessary tools and know-how to quickly extract the disk image with greater chance of success. The cost of cloning is also much cheaper than subsequent recovery should anything goes wrong without backup.
It is important to note that rebuild of RAID array can only be performed if the RAID volume is in degraded mode, ie data must still be accessible. The common mistake is a lot of engineers try to force a rebuild when the volume data is not even accessible or the system reports corrupted file system error. After much tampering, if the rebuild process indeed gets through due to faulty controller or indecent attempts, it will result in blotched rebuild. So take note that if the RAID array is not in degraded mode with all data accessible normally, rebuild will only create more damage and should never be carried out.
When the disks are taken out from the server, please ensure that you follow the correct steps as not to mix up the orders and positions of disks.
2.0 Preparation of RAID Server
Assuming the RAID volume is lost and you need to extract the disks for recovery, each disk and corresponding slot should be noted and labeled. Such labels can also be used for future reference.
In cases where there are multiple RAID volumes, it will be useful to note which disk elements are making up the corresponding volumes. In most raid controller, under the raid boot up BIOS or RAID configuration software, one may be able to locate the disk elements by selecting the disk and press "locate" function. The LED of selected disk will then "blink" to identify itself. If such function is not available, a screen capture of the RAID settings and disk elments will be useful. Take special note on RAID type, number of member disks, disk order, port ID and stripped block size if available.
All the working disks together with the faulty disks must be submitted together for recovery. It will be a plus if you are able to identify which is the first and last failed disk, and during this interval, if the RAID server is being used heavily with new data created in the system.
Accurate and complete information will reduce guessing work and help in data recovery work later.