I'm sure I probably know the answer to this question, and am just looking for some reassurance!

We have a Dell PE2950 with a RAID-1 (2 x 300GB) for the OS and a RAID-5 (4 x 300GB) housing some SQL data. Disk 0 in the RAID-1 configuration has failed, with the Virtual Disk showing as 'Degraded' in OMSA and the physical disk showing as 'Failed'.

If it were one of the disks in the data-housing RAID-5, I would quite happily wander into the server room, take the drive out and replace it with a new one, and wait for it to rebuild. However, I'm getting a bit jumpy with the idea of pulling one of the OS disks (and for some reason, the fact that it's 'Disk 0' adds to that, even though I know it makes no difference if it's Disk 0 or Disk 3..).

Is there anything I should be aware of, or is it simply a case of pulling the disk out and replacing it while the server's running? I presume the OS won't be writing to that disk since it's failed?

Any advice would be appreciated!


since the partitions are mirrored in this setup, you can happily take the disk offline, replace it (or hotswap if you have capable SCSI/SAS drives) and let the array rebuild itself.

Do not be scared by the disk number, it only locates its physical position on the controller.

And a "get well soon" to your array!

  • Thanks for the response, I can't take it offline through OMSA as it's showing as failed, so presumably the only steps to take are 1) remove drive and 2) replace drive! @ewwhite - I have a spare drive sat on my desk at the minute – Chris May 24 '13 at 12:15
  • please let us know how it turned out ;) – Ali V May 24 '13 at 12:17
  • 2
    Does exactly what it says on the tin - swapped the drives over with the server running, and looking in OMSA the new drive is currently 'rebuilding'. Thanks for the replies everyone, enjoy your weekends! – Chris May 24 '13 at 13:39

You can replace the disk without incident in a hardware RAID scenario like this. The drives are meant to be hot-swappable and the process simplified to this level.

Do you have a spare or warranty replacement drive handy?



I just replaced a disk in a similar setup on a PE2960. Mine was a RAID5 and not 1 though but should make no difference. IF the hot swap doesn't work, here are some other things you may need to try:

Pull the disk out and put the new one in. Check in OMSA that it actually shows up and in the Ready state. When I popped mine in it did not actually start to rebuild the array at all. I had to go to Storage -> "Raid Controller" -> Virtual Disk and click on the available tasks next to the Virtual Disk I had and then clicked Reconfigure. I had to add the disk back to the VD. It took a while to rebuild.

I also got a Foreign Disk message on the drive I inserted. To fix this you will want to select the controller and click on "information/configuration" tab in the right pane window of OSM. Then go to Controller Tasks -> Foreign Configuration Operations and click Execute then Clear on the next page.

Also, if you plan to add additional disks to the array I would recommend waiting to add them until the array is completely rebuilt. There will be a lot of disk I/O at this time and I did experience a considerable amount of server degradation. Since it was a production and necessary server I could not take it offline and had to do the rebuild during office hours.

  • 1
    That's not normal. Many systems run without Dell agents, so having to interact with them for a hardware-level hotswap, shouldn't be (and normally isn't) a requirement. – ewwhite May 24 '13 at 12:42
  • 1
    You'll only get a "Foreign Disk" message if the disk was part of another RAID set and was not wiped. If you insert an empty RAW disk, there's nothing you need to do other that sit back and watch it automatically rebuild. – MDMarra May 24 '13 at 12:45
  • 1
    Nothing here is ever normal it seems :) – Travis May 24 '13 at 12:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.