Switch goes down = DB corruption

  • Greetings,

    SQL Server 2005 latest SP

    Windows 2003 latest SP

    SAN appliance

    Some of the DBs that are in trouble have write/read activity, some other have just read activity or even not at all (They are there just in case).

    It happened a few times now. When the main switch goes down, a "random" number of databases go in corruption mode on some or all (random too) of our SQL Server 2005. The behaviour can be of few type:

    - The + sign on the left of the DB is gone

    - You can use the + sign, but dbcc checkdb will show errors and any query on the DB will generate an error msg.

    - The + sign is there, but when you click on it, nothing happens.

    Is it the switch, the SAN or SQL Server 2005 fault? I am puzzled to how to make possible a switch failure possible without having our DBs corrupting.

    MBA

    MCSE, MCDBA, MCSD, MCITP, IBM DB2 Expert, I-Net+, CIW

    Proud member of the NRA

    -Anti-gun laws prevent law abiding citizens to buy guns and defend themselves against bad guys who do not care about the law and get their gun illegally.

    - Democracy is 2 wolves and one sheep talking about their next dinner. Freedom is 2 wolves and one armed sheep with a .357 magnum talking about their next dinner.

  • I presume that it's not a SQL Server fault. However it's not clear are you using MSCS or it is a just single server with single path on SAN?

    Or you have multiple SAN switches, and paths to LUNs?

    If it's true you shuold check versions of drivers, firmware's on HBA and most of all proper functioning of multipath software.

    Maybe to post some more info on HW level config?

    rgds

    Sinisa

Viewing 2 posts - 1 through 1 (of 1 total)

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