Raid or controller?

[mbrooks]

I have a DELL PE 2550 that has recently been having issues. Apparently two the RAID arrays containers have been pronounced DEAD several times while 2 hard disks have gone to orange.

At first I though I am experiencing a drive failure. But with two drives? At the same time? Not very common.

So in the Perc BIOS I did a <restore> and scrubbed the drives and verified the disks. No clicking, appear to be in fine working condition.

All went well, no errors.

Loaded the drives back in the machine. Checked the logs. No drive errors.

Everything looked like it was back to normal. No more than 5 hours later I checked the machine. 1 of the 2 containers we pronounced dead this time. With only on drive in the orange.

So.. How can I tell if this is the drives failing or a controller issue. I am thinking the controller is slowly dying since the array re-builds and the disks verify without errors.

Any comments?

P.S. I put the drives into an identical machine to scrub. Repeat process to see if I have the same issue.

M. Brooks

http://mbrooks.info

 

[twwabw]

It could be a drive issue. I have a client w/PE2800, and we were experiences drives going offline. Simply reseating the drives would bring them back, the array would rebuild, and all would seem fine.... for a while. Then another would go out, etc. Called support, and there issues with the actual hard drive firmware on certain drives. Mine were the Maxtor Atlas 10K's. Dell has a downloadable fix that reflashes the drive firmware to fix the problem. Seems to have taken care of it for me. Might want to see if your drives are on the list.

http://support.dell.com/support/dow...1&vercnt=1&formatcnt=1&libid=33&fileid=164749

 

[mbrooks]

The drives are not on the list. What baffles me is that a week prior I check the machine and all was well.

Having 2 drives in the array fail at once is strange.

Right now I am overworking the drives (compressing a 10GB directory with over 40,000,000 files) and plan to run a disk check after to see if there are any issues afterwards.

So far so good.

M. Brooks

http://mbrooks.info

 

[mbrooks]

Update: The disk check was clean. No bad blocks.

M. Brooks

http://mbrooks.info

 

[mbrooks]

Update: I was correct. The controller is the cause of all my miseries. Also since the controller is integrated into the motherboard I guess it's time to replace the machine.

Cheers.

M. Brooks

http://mbrooks.info