UNSOLVED

remcohn

updated

20 years ago

R

remcohn

4 Posts

0

3308

September 9th, 2006 13:00

2850 raid errors

Hi,
i have a 2850 running since april now with 4x146gig seagate. but since i got the machine the 4rd drive gave 3 errors, resulting in a degraded array. I fixed it again by rebuilding the disk.

A other machine (2800 with the same disks) had exacly the same problem once.

is this a known problem? and can anyone tell me if the problem is in the raid controller (PERC4e/Di) or in the disks?

remco
  • NumberNine

    17 Posts

    203

    0

    Posted September 13th, 2006 13:00

    Based on my recent experience with "rebuilding" a failed drive member of an array (or a drive reporting errors in Array Manager, but still seeming to work) this is asking for trouble. We ended up getting disk corruption on a RAID5 array on a machine that was an Active Directory domain controller, and the files that got lost/scrambled during the next reboot's CHKDISK operation were... you guessed it, the active directory database itself ;-) Yikes!. To make matters worse, the backup domain controller machine (an old IBM box) had lost its raid controller a few days prior so we were temporarily running that domain on a single server awaiting replacement parts. When it rains, it pours... so thank goodness for good backup tapes!

    In short, when a hard drive that's a member of an redundant array (raid1, 5, 0+1, or 1+0, etc) begins to report errors, pull it and replace with a known good replacement drive ASAP!!! The downtime and headaches will be minimized that way. My humble opinions anyway...
  • pcmeiners

    6 Operator

    •

    1771 Posts

    203

    0

    Posted September 13th, 2006 20:00

    Agree with number9, I would get a replacement disk

     

  • remcohn

    4 Posts

    203

    0

    Posted September 14th, 2006 05:00

    I just called dell, and sounds like its a known problem with the seagate 146gig's revision D703.
    It was not needed to replace all 4 drives, but i will get 1 replacement tomorow

    remco
  • pcmeiners

    6 Operator

    •

    1771 Posts

    203

    0

    Posted September 14th, 2006 17:00

    I would schedule a "consistency check" at least bi monthly, this will help pickup errors before they create nasty problems.