It is safe to say that out of all the RAID 5 data recovery jobs we receive, approximately eighty percent of them will need to be recovered with one drive missing. All RAID 5 jobs we receive have drives out but there are many times that the drives that were active just before the RAID went down may be impossible to recover for several reasons.
- Rebooting Over and Over
It is a normal circumstance when a RAID 5 degrades to the point of crashing that access to the data becomes a primary concern. Most of the RAIDs we see are on very large networks and access to data residing on the server could cost several thousand dollars every day the RAID is down. In desperation the server admin will reboot the server over and over again in hopes that it will come up and all will be right with the world again. If it does, then they can pull a backup off and then fix the RAID before things get totally out of hand. It is this rebooting process that usually causes the offending drive to grind and grind upon itself and eventually damages the platters beyond repair.
- Ignoring or procrastinating when a drive degrades the array
RAID 5, by design, is very fault tolerant. The method in which it stores data, the ability to lose an entire drive and continue working, the firmware constantly checking for inconsistent data, makes RAID 5 an excellent choice when the data being stored on the server is critical. The difficulty occurs when the server admin either does not check the array on a daily basis looking for a degraded drive, or perhaps is not even onsite and does not set the RAID up to email log files or notification that the RAID is degraded and has lost a drive. It may also be that the company does not have an onsite technician and use the ‘keep your fingers crossed’ method of support. In any event, this will only lead to disaster. It has been my experience of the several decades I have been in this business that if there are two drives down in an array they did not go down together. One of them went before the other which means one is stale and must be virtually built.
- Configuration change
This particular occurrence is when a RAID goes down, the admin uses the remaining drives to build an array but builds it short one drive. In other words, the RAID runs one drive down and that is how they do business. We have received RAIDs that have had as many as three RAID 5s and all of them were functioning one drive short. This particular anomaly does not happen that often, but often enough that it deserves a mention.
With these examples there are two ways to address this through software. First would be to take the remaining drives and using software set up a virtual drive to complete the set. Sort of like the way firmware works on a RAID board when the drive drops out. There are many software companies that use this method but I have found that these same pieces of software have tremendous configuration issues, especially if the RAID is the least bit odd. In addition, depending upon the software, on the fly virtualization can over tax already volatile code and can hide other problems that are not apparent during the virtualization.
The second method is to just build the drive with software and then evaluate the RAID. This way there are no mistakes in virtual logic, you can look at the image independent or the virtualizing software, and the destripe and evaluation process becomes much smoother if the entire array is physically available.
RAID Sleuth offers the facility to build a RAID 5 image from the remaining drives in the set so that it can be used in the destripe and/or rebuild process. It also allows for the build of a drive that can then be placed back into the array and possibly give access to the array since the drive that was down is now part of the array.
The function is simple to use and offers the technician an array of possibilities that will allow them multiple avenues of RAID 5 data recovery.
Software that calculates a possible configuration on the fly is all well and good, but they are not tools that can be used every day to optimize recovery chances. RAID Sleuth puts the tools in the hand of the technician so that they can use their expertise to bring the RAID online, or simply recover the data.
