Array corrupting on 2 different raid controllers

Discussion in 'Storage & Backup' started by Vampire, Nov 11, 2010.

  1. Vampire

    Vampire Member

    Joined:
    Dec 3, 2004
    Messages:
    747
    Location:
    Hawkesbury
    Hey guys, having some annoying troubles with a RAID5 array I've been playing around with. Basically I built a raid5 array with a Promise Supertrak EX 8350 with 5 x 1.5tb seagate hdd's, installed Windows Server 2008 r2 on a separate disk for the OS. After 20 mins of use I would start to experience random file corruption only on the raid array. After a reconfiguring/recreating the array with different stripe sizes and what not, the exact same thing occurred. I thought it must have been due to the raid controller being faulty, so I gave up and removed it, then plugged all the drives into the motherboard to use onboard raid. I configured the drives again in raid 5 (with brand new cables if its worth mentioning) and fired up windows, within about 45 minutes I noticed further corruption occuring :(

    Im not pretty confused as to what the problem might be... a faulty drive maybe? If so, how would I even find out which one? The raid shows up as healthy using both the promise controller and onboard raid. The only other thing I could think of is possibly a ram issue... but that wouldnt really explain why I only experience corruption on the raid and not the standalone OS hard drive.

    Any thoughts on where I should be looking next? My only real thought is to run memtest and see what the results of that are.

    Cheers.
     
  2. saba

    saba Evil Vizier

    Joined:
    Aug 4, 2001
    Messages:
    2,710
    Location:
    Melbourne
    If the controller has memory, it could be dud. Also could be system memory, or maybe a dodgy power supply.
     
  3. OP
    OP
    Vampire

    Vampire Member

    Joined:
    Dec 3, 2004
    Messages:
    747
    Location:
    Hawkesbury
    But I tired a controller card AND onboard sata raid with the same result, so that would have to rule out a faulty controller. Ill do a memtest on it, as far as power supply goes.... it isnt a cheap PSU so I doubt it, not quite sure how to prove it though.
     
  4. young_einstein

    young_einstein Member

    Joined:
    Jan 19, 2006
    Messages:
    2,562
    Location:
    Caroline Springs, VIC
    I doubt it would be the power supply either, but it's pretty easy to 'prove' by just swapping it out with another one ...
     
  5. phrosty-boi

    phrosty-boi Member

    Joined:
    Jun 27, 2003
    Messages:
    1,091
    Location:
    Altona
    ummm run seatools on the drives maybe?
    (I assume seagates diag program is still called that?)
    have you checked the power settings of the drives / sleep time and such, obvious question but it never hurts to ask does it
    tried different cables by any chance, have had corrupt files on single drives before caused by bad data cables
     
  6. saba

    saba Evil Vizier

    Joined:
    Aug 4, 2001
    Messages:
    2,710
    Location:
    Melbourne
    Stop overclocking and run at stock for a while. Maybe even underclock if you've been running it for a while.
     
  7. OP
    OP
    Vampire

    Vampire Member

    Joined:
    Dec 3, 2004
    Messages:
    747
    Location:
    Hawkesbury
    Its running at totally stock speeds. This isnt a box in my sig, its a different one. I should have mentioned that. Running a memtest now, if it comes out good I'll run seatools on the drives.
     
  8. MetalPhreak

    MetalPhreak Member

    Joined:
    Apr 8, 2007
    Messages:
    1,763
    Location:
    Perth
    Perhaps its an issue with Server 2008 R2 and partitions over 2TB in size? Would explain why the OS drive isn't affected.
     
  9. OP
    OP
    Vampire

    Vampire Member

    Joined:
    Dec 3, 2004
    Messages:
    747
    Location:
    Hawkesbury
    I actually checked the computer this morning and it had bluescreened with the error "PFN_LIST_CORRUPT". I am running a memtest as we speak and Im already seeing a LOT of errors. Im guessing the memory is the culprit. I'll swap out the ram and see if memtest improves, and go from there :thumbup:
     
  10. OP
    OP
    Vampire

    Vampire Member

    Joined:
    Dec 3, 2004
    Messages:
    747
    Location:
    Hawkesbury
    I randomly took out 2 ram modules and booted with only 4gb of ram, memtest passed with flying colours. I have put the raid controller back, rebuilt the raid and I havent had any problems or corruption since then. :thumbup:
     
  11. T101

    T101 Member

    Joined:
    Sep 16, 2002
    Messages:
    549
    Location:
    Sydney

    How did you tell it was corrupting. Did u run stress tests on it or just transfer a whole bunch of files and did verifty on them after?

    I once had a problem like this.
    I kept replacing the memory thinking it was at fault (which it was) ..but for a short time. Ended up being a dodgy PSU from Antec.
     
  12. OP
    OP
    Vampire

    Vampire Member

    Joined:
    Dec 3, 2004
    Messages:
    747
    Location:
    Hawkesbury
    A couple of ways, downloading torrents (they kept corrupting and restarting) and also large transfers (~1tb). Still going well now running with only 2 ram modules and not 4 :)
     
  13. Zoltag

    Zoltag Member

    Joined:
    Aug 30, 2001
    Messages:
    703
    Location:
    London
    The last lot of RAM I had problems with, I ended up pulling it out, cleaning all the contacts and reseating it - Worked like a charm after that :)
     

Share This Page

Advertisement: