RAID 0, and the failure there-of.

Ruckus

For some reason one of our servers' RAID 0 array partitions disappeared, and when we rebooted the server everything came back fine. Yes, I know RAID 0 has no redundancy, that is not an issue with this server. Raw read/write performance is the only concern, and RAID 0 is the best.

The only thing I can think of is that the controller lost contact with a drive and broke the array, but was able to detect the drive at reboot and the array data was intact. We need to know why this happened but there were no application or system log entries to tell us what was going on. The array software caught nothing. Anyone out there wanna toss out some suggestions as to possible causes and or know problems with this configuration?

HP ML110 G4 w/onboard SATA RAID
4x Western Digital WD800JD SATA2 harddrives, configured as one big RAID-0.

PirateJon

I've seen that happen before. If it only happened once and is fine now, I wouldn't worry much about it.

But this is a good reminder to review your recovery documentation / procedures and check your spare parts inventory in case the controller or drives are starting to die.

EggyToast

This has happened to me when my drives came online at different times, such as being asleep or when the computer boots up. Similar to your situation, when I power cycle the enclosure, they came back online (and mounted correctly). However, I did start to notice that the table was becoming corrupted and some files would "disappear," as in I could see them but as soon as I clicked on them they disappeared. I had to run a fancy disk utility in order to repair the drive, at which point I backed up the items to another drive and split the RAID0 into 2 separate drives.

If the files on the drive don't matter (such as a compilation environment, development, or whatever short-term tempo-oriented purpose you have), you shouldn't have any serious problems. It seems to be an uncommon, but present, occurrence with RAID0.