Start a Conversation

Unsolved

M

2 Posts

843

November 25th, 2019 11:00

PowerEdge 2900 drive failed, no OS, backup needed

Hi PowerPeeps,

Am running a PowerEdge 2900 on PERC 5/i with four PDs. Didn't set it up, nor can I find anyone who knows how it was originally done, but after our legacy ERP (which was running on it) crashed, I took a look and found the following:

clipboard_image_0.png

As it says above, I think the Windows 2000 server OS was installed on :02. Regardless, it stopped working and I can't boot into Windows (am guessing because the :02 PD is full). Also cannot query the database (ie. over the network) that resides on it.

Am also guessing that neither RAID 1 nor the other drives were ever configured properly - it's possible that the other three drives were never even used.

Following instructions I found on the interwebs, I tried to turn :03 into a hotspare, my thinking being to have a backup of :02 before I mess with its recovery. To do so I 'cleared' the config on Controller 0.

After clearing the config and turning :03 into a hotspare I'm left with this:

clipboard_image_1.png

Help please  What's my next step? This article doesn't explain how to safely recover the old config (or build a new one without deleting the old one), nor can I find another article that's similar enough to my issue for me to feel confident.

Also, given that I have no Windows OS, how can I backup :02 to an external drive (eg. a USB drive) before messing with the Controller config? I've created a SystemRescueCD, but neither the filesystem explorer nor GParted show the :02 drive (or any of the PowerEdge drives) or its contents. Any ideas?

Thanks in advance!

2.9K Posts

November 26th, 2019 07:00

If the data on that RAID volume is critical, I would stop and engage a data recovery service, like Drive Savers. In your second picture, you show disks 0 and 1 in a ready state, and you mentioned that the config was cleared. Just based on images, it looks like the virtual disk may have been lost. A drive that is an active array member should be in an online state, not a ready state.

Without a data recovery service, the next best option would be to do what I'll refer to as a RAID re-tag. In order to re-tag the array, you have to select the same disk set and deploy an identical array with identical configuration parameters and choose to NOT initialize. This recreates the metadata without causing the initialization to lose everything else.

If you want to boot the server into the Support Live Image that we have and run a SupportAssist collection with an included RAID log, that set of files might help provide more information as to what happened and what options look like.

2 Posts

November 27th, 2019 13:00

Hi Dylan,

Thanks for the quick reply.

I've created a bootable SLI 2.2 disk and confirmed that the image works. Still don't see any of the RAID drives in, for example, the 'Drives' app. Will try to run the SupportAssist collection.

In regard to the RAID re-tag process, do I do that through the Controller configuration utility?

2.9K Posts

December 2nd, 2019 16:00

Correct, the retag would be done within the RAID utility. This will afford you the option to NOT initialize.

No Events found!

Top