CPU 99% after 2000 to 2005 migration

  • Sunday, moved 50 databases (total 800 G) from 2000 sp4 to 2005 sp3 onto a VMWare server. This particular SQL server is alone on the VM box, so no contention with other servers.

    Virtual "processors" and "CPU" are about the same as the old physical 2000 server

    I ran dbcc updateusage ('DB_Name) with count_rows

    and

    sp_updatestats

    for all databases. I have not reindexed yet. I know I should, but will indexes neessarily be horrible from migrating ?

    Our CPU is almost always near 100 %. Jobs are just hanging with WAIT_TYPE = CXPACKET.

    I read that changing max_parallelism to 1 or 2, instead of 0 might help. I changed it an hour ago, but no difference right now. I assume it won't help yet if some process has already grabbed all 4 processors ?

    That's about the limit of my knowledge in this area .... any thoughts or things I should look at ?

  • homebrew01 (3/18/2009)


    Sunday, moved 50 databases (total 800 G) from 2000 sp4 to 2005 sp3 onto a VMWare server. This particular SQL server is alone on the VM box, so no contention with other servers.

    Virtual "processors" and "CPU" are about the same as the old physical 2000 server

    I ran dbcc updateusage ('DB_Name) with count_rows

    and

    sp_updatestats

    for all databases. I have not reindexed yet. I know I should, but will indexes neessarily be horrible from migrating ?

    Our CPU is almost always near 100 %. Jobs are just hanging with WAIT_TYPE = CXPACKET.

    I read that changing max_parallelism to 1 or 2, instead of 0 might help. I changed it an hour ago, but no difference right now. I assume it won't help yet if some process has already grabbed all 4 processors ?

    That's about the limit of my knowledge in this area .... any thoughts or things I should look at ?

    sp_updatestats != UPDATE STATISTICS {tbl} WITH FULLSCAN,ALL

    Full scan is important!


    * Noel

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

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