This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

DPA and processor Affinity

I believe DPA is monitoring all cpu's on the server, even though SQL is only using a subset of the processors.  So the metrics are getting skewed because of this.  Can anyone confirm this and does anyone now if there is a fix?

  • We will monitor two resource metrics.  One is O/S CPU utilization which is all processors available to the host.  The other metric is Instance CPU utilization (processing consumed by SQL Server).  You can choose whichever one makes more sense to focus on.  Is this what you are referring to or am I missing the mark here?

  • yes, but since those 2 metrics come out to be exactly the same, the instance metric maybe not working as it should??

  • Is this host dedicated to SQL Server?  I agree though, there *should* be some OS overhead, even if minimal.  Do you mind opening a support ticket?  We can share our calculations via a support case which may help determine why they are the same (we can go for the raw data directly from SQL Server).

  • Sure I can open up a support ticket.  But its pretty obvious that is is the same.  come to looking at it, all the graphs are skewed because its looking at what all cpus are doing, not what is assigned to sql...

    and to give you some insight on the cpu, i have 16 cpu's assigned, and 64 cpu's total.  so if the instance cpu utilization only tracks what sql is using, it should be pegged at 90-100% all the time, as i had to run a different tool to get the proper metrics...

    sql-affinity.PNG