I have a 32TB RAID (16 x2TB disks). The system was configured as RAID-6 and two Virtual Drives. One is 100GB for Operating Sysytem Area. and Another is almost 27TB for Data Strage. The OS is Centos 5.5 and it has a MegaRAID 9260 series RAID Card BIOS (MegaRAID WebBIOS).

I had 2 disks crash at once. The error message I received at the time of crash was: "... kernel: journal commit I/O error". The system is read-only. I can still log on and list files in the 100GB partition, but the operating system is not functioning. And according to the LED lights (2 steady red), there are 2 crashed hard disks. The machine has not been shut down since the crash.

I was advised to hot swap one of the failed disks, let it rebuild, and then hot swap and rebuild the 2nd failed disk (fingers crossed). However, when I hot swapped the first disk nothing really happened. A blue light flashed when the drive was seated, but after that nothing. No flashing lights to indicate that rebuilding was in process. I'm fairly certain the disks are not spinning, but the RAID is mounted on a rack and I can't be sure about the source of vibrations.

Soon after I did this, support directed me to power off the system, replace the failed disks, then power on and go into the BIOS. But, I already hot-swapped one of the failed disks? So, what should I do? Just remove it and replace it with the failed disk for shutdown and replace before power-up?

Any advice would be greatly appreciated.

Thank you.

RAID 6 can still run with two failed drives, so I'm not sure why it's in read-only mode and not running.

Either way, it doesn't really matter if you swapped out one of the failures, if it's not rebuilding it's not rebuilding. I would just power it off as-is, go into the RAID controller setup on reboot, and trigger the rebuild.

I have no experience with MegaRAID though, just LSI/3-ware and Adaptec. Every time I've lost a drive I just swap in a new one and within a couple of minutes it starts rebuilding.

It sounds like you've contacted support about this? What have they said so far? As far as the OS is concerned, the array should show up and function like nothing is wrong. It's only the card that will be angry about the failed drives. Does MegaRAID have any control utilities that run within the OS?

1 members found this post helpful.

suicidaleggroll,
Your prompt response is much appreciated. On-site support has never had the case where the replacement HD was not accepted and rebuild initiated automatically. They are checking with LIS, the manufacturer of the MegaRAID card, today for recommended course of action. I have a rep. from the firm I purchased the system from stopping by tomorrow. I'll post an update when I know more.

Thanks,

Update:
Apparently, the MegaRAID default in the case of a 2-disk failure is not to rebuild until both failures are replaced. When I removed the 2nd failure to replace it, the 1st replacement begain rebuilding. Now, awaiting results of rebuild.

The rebuild failed for both disks. The webBIOS never provided any error message, but just labeled the drives as "failed". The were both brand new disks, pre-tested for performance.

I tried rebuilding the disks in opposite order. No luck. The rebuild reaches more than 70% on the first disk and then fails.

Any recommendations on next steps? Is all hope of recovery lost?

Thanks,