• as2higpark (7/2/2011)


    Also wanted to add that we are seeing the CPUs (average of all 24) spike to around 50 - 70% sporatically throughout the work day. The spikes generally last 5 - 15 seconds.

    What is your max memory setting in the new instance? I have seen people forget to set that on a new instance (including me) and have CPU problems when SQL starts fighting with Windows over memory during peak usage times.

    Ignoring the new hardware scenario some random thoughts:

    > the articles Ninja's posted are a good place to look

    > it could also be a few bad plans might in play, or maybe some bad queries in general (scans or implicit data type conversions)

    > Parallelism could be set too high...although you said there are not a lot of CXPACKET waits...you have to capture those wait stats periodically and do a diff for them to be meaningful. You can clear them after you capture them too[/url]. They also get reset when the instance restarts.

    > Has any new code gone in since you've been on the new hardware?

    > Have you done any stats or index fragmentation maintenance since you got onto the new hardware?

    There are no special teachers of virtue, because virtue is taught by the whole community.
    --Plato