We have an MD3000 storage array that did something very weird about a month ago. We have two servers attached to it via HBA adapters. The servers lost connectivity to the array and our servers disconnected. Once we rebooted our servers, everything appeared to be fine. The only errors showing were sector error on a disk. But when pulling the support information it was referring to disks that had failed 4 months prior.
Then last week the MD3000 just crashed. Totally down. Could not get onto the unit at all. Flashing Red lights... the works... It was in a constant reboot. One controller would attempt to come up, crash, and then the other controller would do the same. If we pulled the drives all out of the array we were then able to consistently ping the controllers. But could never get access to the MDSM.
Our vendor shipped us a replacement unit and we moved all the drives over 1 for 1 along with the controllers. The result was the same. Constant reboots. We were initially thinking that the backplane was the issue. Next we tried plugging the drives into another unit with new controllers. The unit did the same thing that time as well.
We then found an article that talked about putting 2 new drives into the array and powering it up. We were then able to see the drives in the array and also access MDSM. From there we put in the remaining drives 1 by 1. System stayed up the entire time. We then installed Disk 1 and finally Disk 0. The system was able to see all 10 drives we had in the system. However, it did NOT see our Disk Groups!
How can we get the controllers to read the "foreign config" off of the drives and import it to the controllers so we can access our data?
Is there some type of recovery command we can run? I see recovery data in the MDSM folder structure on my server... Is there a way to import that original config in?
HELP!!!!
Thanks,
RIch