I came in this morning to find one of the disks in my array (RAID 0) with an orange flashing light. I figured I'd power everything down and reseat the disk. After doing so the system came up with all green lights, but the configurator still said the disk in that spot was "missing". A dell support tech said he'd probably be able to get it online, but at the cost of the data in the whole virtual disk (about 2TB). I know I've probably got some bad practices going on right now, but would forcing the questionable disk Online, not create problems? Or could it just be a temporary communication problem solved by forcing it Online? Thanks for any feedback...I'm scared a lot of important data is gone... :(
The worst that could happen is that your data is corrupt. Best that could happen is that your data is accessible.
If the latter happens, I would highly recommend to run a new backup (unless your last backup was last night or so), destroy the raid 0, create a raid 5 and then restore the data.
Basically it comes down to, do we A) take it to a third party for data recovery, or B) take our chances with the "force Online" option... I guess what I'm asking is, can forcing it Online, make matters worse, and make a disk that might be accessible to a third party, inaccessible? Does that make sense?
If it's a RAID 0, then data writes/reads should have immediately ceased when the physical disk went offline. Forcing the disk online can result in either having the logical drive come online healthy (physical disk is healthy), or the logical disk doesn't come online (physical disk fails again). The data can still be corrupt from the event that took the disk offline, however, it should have failed whatever write may have been happening back to the host and the host operation would not complete, therefore data is not corrupt. Odds that you can recover the data if the physical disk has truly failed is beyond miniscule.
Please do yourself a favor and use parity of some sort, as the odds of this happening again are good. The more disks, the higher the odds are that one will fail.
the disk was healthy and after updating the FW on the MD1000 we were able to import the disk as a foreign configured disk. The disk was recognized as being part of the RAID and all's good....thanks for all the help....we're going to work on getting RAID 5 set up over the summer... :)
Dev Mgr
6 Operator
•
9308 Posts
•
3 Points
248
0
Posted April 25th, 2007 17:00
If the latter happens, I would highly recommend to run a new backup (unless your last backup was last night or so), destroy the raid 0, create a raid 5 and then restore the data.