Since the introduction of the concept of RAID, it has redefined how storage systems manage and store data. The basic concept of this technology is actually quite simple: A RAID array is a configuration of many different physical hard disks that together create the used RAID architecture (RAID level 0, 1, 5 etc.). The RAID architecture then distributes data over all disks because the RAID is one single disk to the operating system. Even though the different RAID-levels have certain protection mechanisms against data loss due to physical failure of one or more hard disks, the technology is not bullet proof. Therefore, it’s wise to prepare yourself against data loss when using a RAID array and at least consider the following tips.
A backup is a must-have and not a nice-to-have.
Whether you have a system failure or a data loss event, always have a current backup of your data ready. Many vendors of small, medium or high-end storage systems based on RAID, sell the systems as being some kind of “physical” backup. Don’t fool yourself! RAID – regardless of the level type – is not a backup. If you have a serious RAID failure, the data can disappear. A backup, if up-to-date and stored correctly, is the best tool to get your data back.
Choose the right RAID configuration.
Since different RAID configurations have different redundancies to prevent data loss, it’s a good plan to create a conceptual design of your storage system which truly covers your data and data loss provisions. Single drives will fail at some point in their lifetime. When this happens, assuming it’s RAID 1 or greater, the faulty drive is replaceable and the data storage map can be rebuilt with zero data loss. If you have lots of data, it makes no sense to use RAID 0, where no drive can fail. Chances are good that more than just one drive will fail at a certain point. If a drive failure exceeds the redundancy capacity of the RAID array and hard disks fail, get a specialist help to rebuild your RAID array and recover your data.
Don’t ever try to rebuild a system when too many disks have failed and the RAID level cannot support itself.
If you have a failure on two drives on a RAID 5 based system, it makes no sense to replace one of the failed drives and run a rebuild. You will most likely cause permanent data loss. The rebuild operation, as the term suggests, rebuilds the RAID array in case of a drive problem. Many disk drives offer the “hot plug” that allows you to remove and replace the hard drive without the need to shut down the system and consequently to terminate the service. The operation to rebuild usually takes a long time, but could solve the problem by rebuilding the RAID after replacing the failed disk with a new one. However, if something fails in the procedure, the rebuild operation will result in additional damages. The execution of a rebuild is not without risks and should be done only if you can rely on an updated backup.
Always prepare for a system or hardware failure.
If you have purchased a RAID-based system by one manufacturer, it’s most likely that all the hard disks used are from the same batch and production date. Therefore, it’s not uncommon that they reach their end-of-life at almost the same time. Due to this fact, it’s absolutely necessary to check and monitor the status of the hard drives at selected time intervals. If one hard drive in a RAID 5 fails, don’t hesitate to replace it immediately. Waiting here too long can cause severe data loss because the chances increase of another drive dropping out. Keeping track of the usage of the hard disks inside the RAID storage array is a must-do and should be in a business continuity and data recovery plan. Additionally, these plans should also cover what to do when a RAID storage array failure and data loss occurs. In severe cases the built-in recovery functions may not work and can cause even more damage. In those cases the contact information about a reliable and professional data recovery service provider should be handy.
If you want to DIY rebuild of a failed hard drive, make an image of the content of the hard disks inside the whole storage array before you start the rebuild – and label them.
With an image of all drives available, you extend the chances of a data recovery later if the rebuild fails. Imagine, for example, a scenario where a rebuild process stops at only 5% and then a data loss occurs. In this instance, the content originally stored on several disks is overwritten and the data is destroyed forever, without ever having the chance to recover them again. With the RAID images, professional data recovery experts can at least try to rebuild the multiple pieces out of both the hard disks from the system in addition to the images to reconstruct the original RAID system and data structure. Additionally, the imaged disks should be labeled in the correct order of their original state in the storage array, so the stripe order of data can be reconstructed easier.
If you don’t know why the storage array or hardware failed – and a data loss occurred, don’t attempt a DIY rebuild or data recovery.
If you don’t have enough knowledge on how to get the system running again, or what procedures have to be taken and can how to access the lost data, it’s best to contact professional help than risk destroying valuable data forever!
The more complex the storage system, the more complicated it becomes to recover data once a data loss happens. It’s not just the RAID arrays themselves, but a combination of several complex technologies included in the systems, such as virtualization and deduplication. These are technologies which, in combination, can cause failures that cannot be solved by a single operating software function itself, but only after a deep analysis from experts, who then can decide which layer of complexity is to be recovered first, second, third and so on, until the real data can be restored. If you’re not sure that you have the needed knowledge to recover the lost data from your RAID system – or other technology-based storage system, then contact a data recovery specialist like Kroll Ontrack. In most cases, it pays off in the end!