Post reply

RAID rebuild - shutdown SQL Server?

  • August 12, 2004 at 9:10 am

    #85153

    I had a drive fail and need to rebuild the array (RAID 5).  The last time I had database corruption, I had Microsoft review the corrupt database and O/S events, etc., and they said it was related to the array rebuild that I had performed the week previous.  They said to shutdown SQL when performing array rebuilds.

    Anyone else had similiar experience or comment? 

    HP Proliant, with SQL 7 and single named instance of SQL 2K

    Thanks

    Jay

  • August 12, 2004 at 9:43 am

    #518711

    I have had the opportunity to do 6 rebuilds. We did 4 offline and 2 online. No issues either time with database. We did however on the online tries lost those drives within a few weeks again. Hot swap altough works seems to have an issue in some machines with killing the drive as we found on several forums when we ran into this.

  • August 13, 2004 at 8:05 am

    #518892

    Never done this with raid 5 but done several raid 10's.  I never had any problems and "popping" disks in and out of raid 1 mirrors is an accepted method of backup etc. - esp for o/s before sp's.

    I swapped out a single disk on a raid 10 on a fully operational system - there was no noticable performance difference. and didn't have any problems after.

    replacing a disk in an operational array is a good test anyway - but be careful.

    [font="Comic Sans MS"]The GrumpyOldDBA[/font]
    www.grumpyolddba.co.uk
    http://sqlblogcasts.com/blogs/grumpyolddba/

  • August 13, 2004 at 12:23 pm

    #518950

    Thanks for the comments.  After much P&M'ing I have given 'permission' to take down the server to rebuild the array.  I haven't had too many problems with the drives themselves after a hot swap, it's just that MS was trying to lay blame when one of my databases corrupted (150GB database) on a RAID rebuild.  I didn't want to go thru those hoops again if I can avoid it.

    Just proof positive that hard drives will eventually fail after you force them to run in 125-130F ambient - yes, the A/C failed, and no, we weren't allowed to take down the machines.  Thankfully the servers shut themselves off when internal temp reached 150F (55C).

    Much obliged

  • August 16, 2004 at 1:43 am

    #519039

    I've air con probs in a server room with disk boxes, found that the practice of not leaving gaps between equipment can be unhelpful, I managed to put single U spaces between my array boxes and then pointed a couple of large fans at them .. kept them cooler.

    [font="Comic Sans MS"]The GrumpyOldDBA[/font]
    www.grumpyolddba.co.uk
    http://sqlblogcasts.com/blogs/grumpyolddba/

Viewing 5 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic. Login to reply