0

Im getting a 'degraded disk' error from an old PE 2850. I see the orange light blinking on the array so I know where the problem is. The system is running win2003 x64.

Im running OpenManage also and I'd like to add a hot-spare to the system so I can fail over the drive and yank it. When I go into OpenManage Im told there is a degraded 'virtual disk' and that its running RAID-5. Ok - this is fine.

When I click into the virtual disk Im told there are only two disks in the array. I know this isn't true (and just to make sure, I powered down the system and counted 4 disks).

Windows reports 136Gb of space which is consistent with 3 disks in the array (73Gb each).

So - OpenManage tells me I have two disks. Windows acts like it has 3 and the system truly has 4. Which is it?

ethrbunny
  • 2,369
  • 4
  • 41
  • 75

1 Answers1

1

I've seen that OpenManage will sometimes remove a failed physical disk from the listing in OMSA. Kind of a PITA, especially if you wanted to make it blink. I'm sure if you booted into the PERC BIOS, you'd see all of the physical disks.

mfinni
  • 36,144
  • 4
  • 53
  • 86
  • Im used to it being able to 'blink' disks for me. I've never seen it fail to show one, errors or not. – ethrbunny Apr 24 '12 at 15:36
  • I started at a new company in December, and we're running all Dell, mostly 1855 or 610/710 blades. I've seen at least a half-dozen drive failures on 1855 where OMSA no longer showed the failed physical disk, same symptom you're seeing. – mfinni Apr 24 '12 at 15:43
  • Ok - but it seems like OMSA is missing two disks. – ethrbunny Apr 24 '12 at 15:45
  • What does Windows disk management say it has for volumes and physical disks? – Bart Silverstrim Apr 24 '12 at 15:46
  • That's a little weirder, but just an extension of the same OMSA tomfoolery. Could mean you've got two drives dying. Run a DSET and look at the controller logs, if you can read SCSI diagnostics. Or boot into the controller BIOS and see if it tells you anything. Or just replace the drives with the blinking lights. I'm not sure I understand your question at this point, unless you're just asking "Does OMSA suck?" Yes it does. – mfinni Apr 24 '12 at 15:47
  • @Bart - Windows shouldn't see the physical disks at all. It should only see whatever virtual disks (RAID sets) that the PERC has created. – mfinni Apr 24 '12 at 15:48
  • It should, but I wondered what it actually says :-) – Bart Silverstrim Apr 24 '12 at 15:51
  • I only see one orange blinker - so I (naively) assume only one failing drive. I'd replace it but Im not entirely sure how to do this. Im spoiled and used to SANs with 'offline drive' and hot spares and easy regeneration. Im wondering how Win2003 will behave if I yank a drive. – ethrbunny Apr 24 '12 at 15:52
  • 1
    If it's on a PERC card you can just pull it and insert a new one and the array should rebuild. Unless there's something wonky with the PERC card or the config... – Bart Silverstrim Apr 24 '12 at 15:52
  • Since you're free to power down the server, do that. Go into the PERC BIOS and get the details, if you don't want to wade through the DSET/OMSA controller log. Replace anything actually failed while it's powered off. Let it rebuild. Then replace anything "failure predicted." – mfinni Apr 24 '12 at 15:53
  • Bart's right - replacing a failed drive in a server *should* have no impact to the host, assuming no other problems. Of course, if you replace the wrong one, you've caused a problem. And the rebuild will impact performance, if that matters. RAID failure may impact your overall business performance more, of course. – mfinni Apr 24 '12 at 15:54
  • So there is no need to 'offline' the disk first? – ethrbunny Apr 24 '12 at 16:52
  • Nope, not if it's failed. Even in the scenario if a healthy array, you *can* physically pull a disk without causing a problem. That's how the controller is supposed to deal with an accidental disk failure, right? – mfinni Apr 24 '12 at 17:15