Re-build RAID array after failure

Servers

Servers
Information and ideas on Dell PowerEdge rack, tower and blade server solutions.

Re-build RAID array after failure

This question is answered

I have a PE2500 with a single RAID 5 array (PERC 3/Di) consisting of three Seagate ST336754LC disks. Recently one of the disks failed and i replaced it with a ST33660LC which was the closest I could get. The disk appears in the array and I've initialised it but the array does not attempt to rebuild. I have tried assigning it as a dedicated spare but the disk always shows the space (33.9G) as unallocated.

I have been using Dell OMSA but have even tried uninstalling and using the older Array Manager console.

The new disk is slighly smaller that the original (34.18GB) but the old disks have 286MB Unallocated so there should be enough room - could this be the problem?

All three disks flash the left hand led at about 0.5s intervals

Any ideas would be much appreciated - the server runs MS 2003 Server with all latest patches, BIOS, Drivers, etc.

Thanks in advance

 

Tim Ray

Verified Answer
  • Ok now I've fixed it. I found a slightly newer version of the Perc firmware (2.8-1 (Bld 7692) and updated it. I then restarted with the faulty drive inserted and a dedicated spare (the 33.6Gb drive).

    The re-build failed after 20 minutes or so and it immediately started re-building onto the spare.

All Replies
  • To rebuild a raid volume the replacement disk must be the same size of bigger. So your new disk should have at least 34.18gb. As it is smaller it won't rebuild, even if there is free space. You may have to use a 73gb if you can't get the exact size.

  • Thanks for the advice but I thought that might be the case so this afternoon I tried a 73GB disk as well but it does exactly the same thing.

    I have configured the 73GB drive as a hot spare but it won't seem to use it - the virtual disk shows 2 x disks on-line and the spare Ready but not in use.

    I have a second identical server (same updates, drivers, etc) and when I run OMSA on that I get an option to re-configure virtual disks but that option is not available on the server with the faulty disk. Both servers are using OMSA 4.5.

    Any other ideas?

     

     

     

  • remove it as a hot spare so that it's just 'ready' in OMSA right click on the drive and see what options you get. There may be an option to rebuild. If not pull the disk, wait a minute or so and re-insert it. Might force a rebuild. Also try the drive in a different slot.

  • Thanks for the advice - I have tried your suggestions but without success. OMSA now says that the array is degraded and has two 34G drives on line and one 73G drive 'ready'.

    On all three drives the power leds flash every half second or so but there is no activity on the replacement drive. Any other suggestions to save a complete re-build.

    Thanks in advance

     

  • Ok now I've fixed it. I found a slightly newer version of the Perc firmware (2.8-1 (Bld 7692) and updated it. I then restarted with the faulty drive inserted and a dedicated spare (the 33.6Gb drive).

    The re-build failed after 20 minutes or so and it immediately started re-building onto the spare.