RAID 5 Failure – What to Do?

What is the best way to recover data after RAID 5 fails?

Storage is a crucial point when it comes to setting up a server. What kind of server you should use depends on your needs based on current and future plans. If your needs are uncomplicated, the single disk should be enough, but if you are searching for advantages like data storage redundancy in combination with a high level of performance, you should think about RAID as a solution.

RAID 5 is among the other RAID levels, one of the most common used solutions. It consists of three or more hard drives with the same capacity. Three hard drives are minimum for RAID 5, and it is an advantage if all of them are the same capacity. If not, the RAID controller will establish the RAID 5 volume taking into consideration the lower capacity of hard drive. This RAID array uses a combination of striping and parity techniques.

  • Striping – the way of splitting the flow of data into blocks of a certain size and writing blocks across the RAID
  • Parity – data storage technique that combines striping and checksum methods.

The combination of those two techniques provides safer and faster data access storage.

RAID 5 advantages:

  • RAID 5 offers data redundancy, so there is a chance for recovery if one drive fails
  • single-parity data storage allows RAID 5 to offer the most usable disk space of any other redundant RAID type
  • performance and disk space features make RAID 5 ideal solution for storing videos, large data which doesn’t require frequent update
  • “Sequential write” performance and disk read performance are in the rank of other RAID levels, and sometimes even superior to other levels.

RAID 5 disadvantages:

  • very poor “random write” performance
  • if you have a large drive failure, rebuilding can take a very long time
  • when two disk fails simultaneously, you lose data
  • in order to improve RAID 5 performance, hardware RAID controllers could include dedicated “XOR Processors”, large write caches, or both

What happens if one drive fails?

If you are facing one drive fail, no need to worry, it can be fixed. In a situation like this, the system will continue to run as usually. The system administrator will be notified that drive has to be replaced. It is extremely important to replace the broken drive urgently to keep the integrity and performance of the system. Now, the question is – how to find out that one of the disks has failed? You will receive a warning message that one of the member disks has failed. RAID will be able to continue writing and reading data but the performance of read operations degrades seriously.
And what if two drives fail at the same time?

Sadly but thru, if two drives fail at the same time, you have a big problem because RAID 5 array is going down and all the array data will be lost. When this happens, the only option to get your data back is to restore them from backup. Remember that RAID is not a replacement for a backup.

It happens that people lose files stored on RAID 5. What is the main reason for that? The answer is formatting RAID 5 volume, mistakenly, of course.

Is there a chance to recover data after formatting RAID 5 volume? That depends whether the data is overwritten which happens by adding some new data. This way you occupy the space on RAID volume. So, if you haven’t added some new data, you have a chance for recovery.

How to recover data if RAID 5 fails?

The first step in process of recovering RAID 5 is to determine configuration parameters:

  • determine the number of disks in the RAID 5
  • determine the disk sequence
  • block size that was used in the array
  • at what offset the array data began
  • what the parity pattern was used in the RAID5

There are two ways to determine those parameters – manually or automatically. If your choice is manually way, you have to be aware that this process requires a lot of time and the most important, special technical skills.

To determine the array parameters manually it is necessary to determine:

  • the disk order
  • block size
  • start offset on the member disk

Unless you are a professional who sees this kind of situation every day, you will spend too much time to solve this problem, so maybe the software is a better solution.


What is the base for RAID 1?

It is based on mirroring – storage technique which allows identical copies of data to be stored on the different array member disks at the same time. This affects the fault tolerance and performance. RAID 1 consists of at least two hard disks where the smallest member disks determine the capacity of the array.

We also know this array as mirrored volume, because data stored on one hard drive will be mirrored to other hard drives in RAID 1 array. This is an advantage in the case when one drive fails. How is that possible? We already mention that the capacity of RAID 1 array is big as the smallest hard drive because the rest of the hard drives are regarded as a mirror. So, we have a situation where the data stored on RAID 1 array is actually stored on one hard drive and mirrored to the rest of the hard drives simultaneously. That gives you a chance to restore data from mirror volume in case of hard drive failure. RAID is a great choice if you need a fast and reliable way for your data protection. It is actually sad to be the most reliable way you could find for online data protection.

Two main variations

We will present you here two variations of RAID 1 – RAID 0+1 and RAID 1+0

RAID 0+1presents combination of striping and mirroring techniques which is the best way to gather RAID 0 performance and RAID 1 fault tolerance. Using a simple mirroring operation it is easy to create RAID 0+1 from the existing RAID 0. For this version, you will need a minimum of four disks because two disks are needed to create RAID 0 which is mirrored on the other array members. What array capacity you could expect here? It is simply determined by the smallest member disk multiplied by the number of the disks in the original RAID 0 version. Because of the fact that half of the array capacity is used for the redundancy, we can say that this is quite an expensive option of RAID per gigabyte storage.

RAID1+0 is a combination of RAID 1 and RAID 0 with all the characteristics of RAID 0+1. It also uses mirroring and striping techniques. In RAID 1+0, allocated disks are mirrored in pairs, and then those pairs are striped.

Even if it sounds similar for the first sight, there are some serious technical differences. The bottom line is that 1+0 is better than 0+1. The difference lies in fact that is hardly possible to create
RAID 1+0 from RAID 1 on-the-fly, because of the complete restriping.

What makes RAID 1 popular?

This solution provides writing two copies of the data at the same time on two different drives. It is called fault tolerance.

RAID 1 usually writes two copies of the data at the same time on two different drives. This is called fault tolerance. If one of the disks fails (suffers a mechanical failure or doesn’t respond), the remaining drives will function and data can be fully recovered using the other disk. 100% redundancy could improve read speed, but what happens with the write speed? If you have two-drive RAID 1, every time when you write data to the drive, a copy is written to the other drive. That means that all changes you made, have to be made twice, so the result is decreased write speed which is main disadvantage of this array.

The RAID 1 configuration could be performed by a software or hardware RAID controller.

Who needs RAID 1?

The purpose of this RAID array is suitable for accounting and other financial data, small database systems, individual users who are looking for fault tolerance, enterprise servers etc. It is extra useful because duplicated set of data is surely more secure than using parity.
Is there a way to restore data when RAID 1 fails? Of course, there is a way.
There are three reasons for losing data from RAID 1:

  • RAID 1 disk rebuild
  • RAID 1 disk failure
  • Formatting RAID 1

Now, when you are familiar with the reasons for losing data, you probably ask yourself what kinds of situations could lead to those three reasons.
Three ways that can cause data loss:

Data loss caused by human mistake

We all make mistakes for a number of reasons. Sometimes you are in a hurry, sometimes it is because of lack of concentration, stress or any other factor. The point is, that can happen to everyone and when it happen, panic is probably the first reaction. But it shouldn’t be, because if the human factor is “guilty” for data loss from RAID 1, those files can be recovered if data is not overwritten.

What are possible scenarios:

  • Mistakenly press Shift+Delete, or you delete accidentally data from Recycle bin
  • Accidentally cause disk drive formatting
  • Some other mistakes like lost files after cutting them from one drive to another

Data loss caused by RAID 1 array failure

If just one drive in the array is being corrupted, you can recover data from another hard drive due to data redundancy. On the other hand, if both drives are damaged, the best solution is to leave the data recovery to the professionals. In case of RAID 1 array failure you can expect the following messages:

  • RAID 1 drive is not readable or recognized
  • File system of RAID 1 drive turns to RAW format
  • Disk drive is not formatted do you want to format it now?

Data loss caused by hard drive physical damag

Physical damage means that disk is damaged by water, fire or it is just broken into several pieces. If this happens it is unable to recover files because the system can’t recognize the drive.