Proliant E200 RAID controller - recover from 2 predictive failures

[Jeff J]

I have a ML350 G5 with a E200 RAID controller. 4 Disks in a RAID 5 config.

I now have 2 disks (number 1 and 4) marked as predictive failure.

I purchased a new disk, pulled disk 4 to replace it, the server immediately crashed. That seems wrong, with 3 disks, I still should have a degraded RAID 5 array, with 1 predictive failure disk.

I reinstalled the original disk 4, and installed the new disk in slot 5. The server booted up, though I had to hit F1 to accept a loss of data from the RAID controller.

Overnight, the system had to rebuild the array. In the morning I intended to mark the new disk as a hot spare to replace the 1st of my predictive failure disks. But the controller won't let me.

HP support tells me that because I have 2 disks in predictive failure, my only option is to backup the server, replace both bad disks, then reinstall all my software.

My server is currently running fine. is there a way to mark the bad disks as good, so I can replace them normally.

My guess here is that disk 1 is really failed, not just predictive, which is why the server crashed.

Any help here? I'd rather not have the hours of downtime associated with this type of rebuild, even if I use a disk image solution, it's still 400GB of data.

thanks,.

[Gary Antonio Benavides_1]

Assigning a hot spare drive to a RAID Array can only be done when the Array is OK, since you have 2 HDDs as predictive failure it won't let you.

Since you have 2 HDDs on predictive failure, by replacing one drive at the time, it might not rebuild due to second drive being on predictive failure or with read errors and it can cause the second drive to fail at any time during the rebuilding process..

If you don't have a backup, do it asap, re-install the OS, and restore from backup, it is a faster and safer way to have the server running asap, than start trying something different that can cause the server to crash at any time..

Note: The F1 data loss message you got during POST is not the data from the array you loose is the data from the memory cache module on the controller.

[Michael A. McKenney]

I would get a full backup or two. Did HP have you do firmware upgrades? It could be a firmware glitch. It could be two bad drives.

[Jeff J]

HP didn't suggest we install any new firmware.

[Michael A. McKenney]

One time HP had me install firmware on hard drive that showed failure. Get a backup first if the array is still intact and you can access it. You can call them back about upgrading firmware to see if that recovers the predictive failure.