2850 EOD76 BP Drive 4 error

[bgalvan]

Help, I'm fairly new to raid and have a problem with it. Running a dual processor 2850 with GJ871 raid board (Perc 4 edi). raid key,and the backplane daughter board. 2 300gb drives as Raid 1 and 4 300gb drives raid 5. Drive 4 seen by controllers as failed. User unlatched and reseated drive 4 with amber light still seen as failed. I replaced the drive with a 300gb replacement, rebuild did not even start, rebooted server, same results. I suspected DOA drive, replaced with another 300gb drive, same results, still seen as failed drive. Replaced the raid board, channel A and channel B cables, raid key, and backplane board and another 300gb drive, still seen as a failed drive, my thought is that the backplane daughter board could possibly be bad. Any thought out there? This is a critical server, takes a week to get permission to take it down and then only for one hour. I have run diags, every thing runs fine till it hits that drive then the basic HDD test blow off.

Thanks for your help

Ernest

 

[technome]

First I would update all firmwares on the server, ESM, Bios, raid firmware, make sure all firmware revisions on drives are the same . Clean the backplane slots with contact cleaner, check the particular slot for damage. Since you have an extra slot, if (and only if)the slot is on the same raid channel as the offending slot, move the drive to the other available slot. If you end up replacing the daughter card without success, suspect one of the original drives has an electronic component failure causing the issue.



........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[maultier]

I've had issues getting a rebuild to kick off on several Dell servers. Technome's suggestion to update the firmware and bios is a good one.

In my cases, I just had to bring up Dell Openmanage Array Manager and force the rebuild to begin. You should be able to download the iso for the install from the support.dell.com website.

 

[bgalvan]

Both are good ideas, I plan to update the firmware and drivers as soon as I get the go ahead, due to the nature of this system I can't do that till I get clearance. Also regarding open manage and forcing the rebuild to start, was your drive showing as failed at the controller?

Thanks again,

Ernest

 

[technome]

Failed to read properly, assumed you manually tried a forced rebuild, Maultier's suggestion is a very good possibility, the raid may believe the old drive is still in place. You could force the rebuild from within the raid bios, but would add a reboot and a couple more minutes of down time. All firmware updates total should be less than 25 minutes, including a a couple needed reboots, after all are downloaded.


........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[bgalvan]

Ok, thanks, if I try to force the rebuild against the failed drive am I taking the chance of clobbering the other drives or the server?

Ernest

 

[technome]

If a rebuild fails, once the "failed" drive is again removed you should be back to where you are now. There is always a chance things could go south, such as a second disk failing during the rebuild, so make sure you have a backup before proceeding.
If the original failed drive was a legit failure, chances are even if it does succeed in rebuilding, it will fail again fairly soon. If the original failure was caused by a firmware bug, you would be chucking a good drive. Personally I would not try rebuilding with the original failed drive, use one of the new drives. You might setup the "original" in a test array, creating a new array with it, destroying and creating it multiple times over a few weeks to test its worthiness.
A few months ago, a client's raid 1 kept dropping the same disk, about once a week,, so I updated the firmware and the drive has not dropped out since; as a raid 1 ( being inherently safer than raid 5), along with a image/backups, I did not sweat about losing both drives. In your case, you can not afford to take chances.
Might be worth calling Dell to find out if there are any issues with your present raid firmware revision.



........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm