Switch to full style
Data recovery and disk repair questions and discussions related to old-fashioned SATA, SAS, SCSI, IDE, MFM hard drives - any type of storage device that has moving parts
Post a reply

When does a RAID 5 beak?

June 2nd, 2016, 18:57

Hi,

I have a general question about RAID 5. Lets presume, I have 3 HDD in my system running RAID 5. Very often I had the situation, that one HDD had bad blocks and was thrown out of the array. Now the RAID 5 was running on just 2 HDDs. But how tolerant is a raid (in general) in this state? What’s going to happen, when I try to restore the RAID and during the restore process, I got Bad Blocks on one of the two HDDs? Does the RAID Controller / Software going to throw the last “more or less” working disk out of the array or is the RAID “whatever” tolerant in this moment, so the user can at least get the data, that’s not corrupt?

Re: When does a RAID 5 break?

June 2nd, 2016, 20:08

D_R wrote:Lets presume, I have 3 HDD in my system running RAID 5. Very often I had the situation, that one HDD had bad blocks and was thrown out of the array. Now the RAID 5 was running on just 2 HDDs. But how tolerant is a raid (in general) in this state?
It is not tolerant at all. Independently on the number of drives, RAID5 can tolerate only one lost drive.
Moreover, in this state RAID array becomes (N-1) times less reliable than a single hard drive, where N is the initial number of drives.

D_R wrote:What’s going to happen, when I try to restore the RAID and during the restore process, I got Bad Blocks on one of the two HDDs?
Briefly, that would be an issue. Cases when one more drive fails during RAID rebuild aren't uncommon at all.
Besides the performance drop, working in degraded mode puts extra load on remaining drives, as the missing data needs to be reconstructed from parity.

So that's why using RAID 5 for large arrays isn't the best idea. At some point chances of the above-mentioned scenario are becoming unacceptably high.

Re: When does a RAID 5 beak?

June 3rd, 2016, 4:57

Hmm...

Let me rewrite my question a little bit. Let's take a normal working RAID 5, software raid on linux. Now lets say, that I want to access a file stored on this raid. But on one of the HDDs I have a bad block, so the file (or a chunk of the file) can't be read from it. AFAIK now the RAID goes into the N-1 state and reads the file from another HDD.

Lets go on. I still have a degraded array. But what's going to happen, when I'm going to read a file, that has a bad block on one of the two HDDs left? Is the array going to the N-2 state aka commit suicide or does the array now ignore the error giving an error message, that the file can't be read just like on a conventional single HDD?

Re: When does a RAID 5 beak?

June 3rd, 2016, 11:29

The RAID5 will continue to operate in degraded mode when the state is N-1.

In state N-2 your RAID5 will crash and you lose access to your data. However having worked on many N-2 cases I can say recovery is almost always possible with minimal, if any, data loss.

Re: When does a RAID 5 beak?

June 3rd, 2016, 12:27

S.Haran wrote:The RAID5 will continue to operate in degraded mode when the state is N-1.

In state N-2 your RAID5 will crash and you lose access to your data. However having worked on many N-2 cases I can say recovery is almost always possible with minimal, if any, data loss.


Well,
Recently i did a NAS 4TB [ 1TB X 4 ] Drives From a IOMEGA Product That Was RAID 5 ,One of The Seagate HDD Was In Degraded State And Still NAS Was Bricked .This Was RAID 5 Too

Re: When does a RAID 5 beak?

June 4th, 2016, 11:30

Certainly failure is possible, just rare. Perhaps someone tried to "fix" the problem before you and instead bricked the NAS. I've worked on many Iomega's mostly the IX4-200D and all were recoverable. Did you attempt recovery using Linux? I perform remote data recovery using a custom Linux boot CD I developed. I'm happy to take a quick look if you still have the NAS drives.
Post a reply