RAID Stripe, slice failure.

Question

Level 1

115 points

RAID Stripe, slice failure.

I have a RAID stripe set up on 3 disks, and I had a slice failure. Anybody know how I can remove the slice and recover the data on the other 2 disks?

Since OSX and all Unix-y OS's like to keep files contiguous, the loss should be minimal if i can only recover the data from the two remaining good drives - any suggestions?

MDD G4 dual 1GHZ, Mac OS X (10.4.9), why on earth did I set up a RAID stripe with MAXTORS???

Posted on May 31, 2007 7:51 PM

Reply

Answer 1

May 31, 2007 8:08 PM in response to psalmsmith

Hi GlenBradley;

I don't think you are going to be able to recover any worthwhile data with a slice missing. In all of the work I have done in the past with RAID 0 once a slice is lost the only way to recover the data is from a backup. You do have a backup?

No data should ever be put on a RAID array without a known good backup.

Allan

Reply

Answer 2

psalmsmith Author

Level 1

115 points

May 31, 2007 8:22 PM in response to Allan Eckert

I do have a backup, problem is, all the data I really need to recover is newer than my backup. 8-( Logically it ought to be recoverable -- because *nix file systems like to keep files contiguous, it stands to reason that the files themselves are contiguous on the two remaining good slices.

I understand and do not doubt that you speak from experience -- it's just that I understand a Windows file system being unable to recover said data - files being fragmented from one end to the other. But an *nix file system??? There is really no good reason why not.

None that I can think of anyway. It's not like Darwin fragments a file all over the place. 8-(

Reply

Answer 3

May 31, 2007 8:31 PM in response to psalmsmith

Hi GlenBradley;

The way a RAID 0 array works is that the data is divided up into blocks and the blocks are written to disks in turn. Hence if you lost a slice you have lost parts of all the files in the array. I don't know of any other way to write to RAID 0 arrays.

To the system all of the disks in the array appears as a single devices and the data is written to all of them. That is what allows for increased performance.

Allan

Reply