After SQL Server 2000 cluster failover DATA folder disappears from Shared Drive

  • Hi ,

    I am trying to setup SQL Server 2000 Active/Passive two node cluster on Windows 2003 Enterprise edition.I am using following editions.

    SQL Server 2000 Enterprise Edition SP4

    Win2003 Enterprise Edition SP2

    The active/passive cluster has been installed successfully however when I try to failover the DATA folder which is under shared drive M:\Program Files\Sql Server.... disappears completely. Other folder remains unaffected on the M drive.

    This takes SQL Server 2000 resource offline as all master databases disappear.

    Could anyone help why only DATA folder disappear from the shared disks ?

    Many thanks,

    V.

  • Hi V,

    Dodgy indeed. Could it be that M: on one node is actually not the same LUN on the SAN as M: on the other node? (assuming the DATA directory "comes back" when you fail back to the first node).

    /Elisabeth

    elisabeth@sqlserverland.com
    MCITP | MCT
    http://sqlblog.com/blogs/elisabeth_redei/
    http://linkedin.com/in/elisabethredei

  • Hi Elizabeth,

    Thank you for your reply.

    It is very strange why DATA folder disappeared.

    Fortunately I have resolved it now and nothing much I have done !

    I have shutdown cleanly all servers including DNS Server.

    Restarted one by one servers and it is working fine. Tested failover and DATA folder doesn't disappear.

    But going through all this pain I want to investigate why it was happening.

    I am checking Eventlogs and Cluster.log & SQL Server Error logs files.

    Could you suggest any other log files , locations I should check to investigate.

    Many thanks,

    V.

  • Hi,

    Good to hear! Eventlog and clusterlogs is where you want to look but I suspect anything of interest has been overwritten in the cluster.log. Aood practice is to make a copy of it ASAP when anything happens in the cluster and you might also want to increase the size of it:

    http://support.microsoft.com/kb/168801.

    The article also shows you how to increase the loglevel, you might want to do that now until you feel confident about your disks.

    Depending on what SAN you have, there are filter drivers involved and depending on the driver, you may or may not see anything of interest in the eventlog (these days they do usually raise errors in the eventlog when appropriate).

    Check with your "SAN guys" if they can turn on some addtional logging for a while as well (or if they already have it; to inspect those logfiles).

    To summarize; cluster.log and eventlog are the logs you want to inspect. Someone should also inspect the logfiles from the SAN provider. Keep an extra eye on your disk IO subsystem for a while to ensure this disappearing DATA did not indicate anything serious.

    As DBA, your worst nightmare is an unreliable disk subsystem.

    Good luck!

    /Elisabeth

    elisabeth@sqlserverland.com
    MCITP | MCT
    http://sqlblog.com/blogs/elisabeth_redei/
    http://linkedin.com/in/elisabethredei

  • Thanks Elisabeth.

    Your help is really appreciated.

    Thanks for Microsoft articale which is very helpful.

    Regards,

    V

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

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