Powevault 220

[amscot]

Can anyone lend some advice. I've replaced two drives in our Dell Powervault 220, now the server dosen't see the array, it was set as a virtual drive. In the sever manager the drive array state shows "failed". Do I use the PERC utility to "Rebuild"? I don't want to lose the data in my array

Cheers,

Scott

 

[rclarke250]

why did you replace the drives? what was the raid level? If the array shows failed, it won't be able to rebuild from perc, you will need to restore from tape or other backup media unless you can get the array active again by putting one of the old drives back in.

 

[amscot]

One power supply had failed, the device was turned off. On restart, the virtual drive the array was made up from didn't appear on the server. Checking Dell server manager, it showed the array as "Failed" with two drives failed. It's a RAID5 of 14 disks into one virtual volume connected to a Windows 2000. No viable backup that I can find.

 

[rclarke250]

Well 2 drives fail in a raid 5, and the array is gone. with no back up, so is the data. There is no way to recover from 2 failed drives in a raid 5 array, that is why it is still imperative that there is a back up means. You can try to force one of the drives back online, and try to recover that way, but not sure if Dell has that option in the PERC configuration, and it's still iffy, but worth a look to see if the array can be recovered if no backup.

 

[jkupski]

While this is not exactly helpful to your current problem, with that many disks you might want to consider RAID6 instead, and keeping one of the disks as a hot spare.

 

[amscot]

I got it back. One of the Dell techs over at Experts Exchange came up with a solution. Did the last drive that failed force back online, combined with about 4-5 chkdsks on reboot, then replaced the first failed drive with a new one, force it to rebuild, another few reboots with chkdsk, then replaced the other one. All the data is there,as far as we can tell. I'm sure something is corrupt but the important stuff is all back.

 

[jkupski]

Glad to hear that. I suggest you now work on implementing a solid backup plan to prevent future recurrence (and seriously consider the notes above).