DB migration

Collapse
This topic is closed.
X
X
 
  • Time
  • Show
Clear All
new posts
  • michael

    DB migration

    I just moved a SQL 2000 database to a new server (backup/restore).
    This server is much beefier with respect to hardware.

    The problem is that the system doesn't seem to be very fast at all. It
    seems to be having issues with thread count. For example, I saw some
    errors like this in my client apps:

    "The query processor could not start the necessary thread resources
    for parallel query execution."

    Also, connections from SSMS occasionally timeout and running this
    "query": "PRINT GETUTCDATE()" takes as long as 2 minutes to return the
    current date.

    What should I be looking for? What things should I check?

    Is it possible the db is just taking IO hits like crazy as it spins up
    (it's been up for about an hour now)?

    Thanks!
  • michael

    #2
    Re: DB migration

    On Jun 17, 12:22 pm, michael <mha...@gmail.c omwrote:
    I just moved a SQL 2000 database to a new server (backup/restore).
    This server is much beefier with respect to hardware.
    >
    The problem is that the system doesn't seem to be very fast at all. It
    seems to be having issues with thread count. For example, I saw some
    errors like this in my client apps:
    >
    "The query processor could not start the necessary thread resources
    for parallel query execution."
    >
    Also, connections from SSMS occasionally timeout and running this
    "query": "PRINT GETUTCDATE()" takes as long as 2 minutes to return the
    current date.
    >
    What should I be looking for? What things should I check?
    >
    Is it possible the db is just taking IO hits like crazy as it spins up
    (it's been up for about an hour now)?
    >
    Thanks!
    It turns out that this was caused by the max degree of parallelism
    threshold being set to 0. I guess the optimizer was parallelizing
    everything, even the simplest of queries. I changed it back to 5 (the
    default) and it's running much, much better now.

    Comment

    • Erland Sommarskog

      #3
      Re: DB migration

      michael (mharen@gmail.c om) writes:
      It turns out that this was caused by the max degree of parallelism
      threshold being set to 0. I guess the optimizer was parallelizing
      everything, even the simplest of queries. I changed it back to 5 (the
      default) and it's running much, much better now.
      The default is actually 0, but it's good to hear that your issue was
      resolved.



      --
      Erland Sommarskog, SQL Server MVP, esquel@sommarsk og.se

      Books Online for SQL Server 2005 at

      Books Online for SQL Server 2000 at

      Comment

      • Dan Guzman

        #4
        Re: DB migration

        To add to Erland's response, consider examining the execution plans of the
        parallel queries to see if index/query tuning is warranted. SQL Server will
        sometimes throw parallelism at a query to compensate for lack of useful
        indexes. I have seen cases where tuning eliminates parallelism and improves
        performance.

        --
        Hope this helps.

        Dan Guzman
        SQL Server MVP


        "michael" <mharen@gmail.c omwrote in message
        news:944ac9e6-6751-43b6-81dc-54f025770485@i7 6g2000hsf.googl egroups.com...
        >I just moved a SQL 2000 database to a new server (backup/restore).
        This server is much beefier with respect to hardware.
        >
        The problem is that the system doesn't seem to be very fast at all. It
        seems to be having issues with thread count. For example, I saw some
        errors like this in my client apps:
        >
        "The query processor could not start the necessary thread resources
        for parallel query execution."
        >
        Also, connections from SSMS occasionally timeout and running this
        "query": "PRINT GETUTCDATE()" takes as long as 2 minutes to return the
        current date.
        >
        What should I be looking for? What things should I check?
        >
        Is it possible the db is just taking IO hits like crazy as it spins up
        (it's been up for about an hour now)?
        >
        Thanks!

        Comment

        Working...