high CXPACKET and LATCH_EX

  • Users claims the slow performance on the database. I collected top waits for the server as below. How can I find out what queries that have high CXPACKET and LATCH_EX? Thanks.

    wait_typewait_time_spctrunning_pct

    CXPACKET83931185.2964.9264.92

    LATCH_EX18154049.1914.0478.97

    BROKER_RECEIVE_WAITFOR6752891.585.2284.19

    BROKER_EVENTHANDLER6749025.915.2289.41

    CMEMTHREAD4391072.633.4092.81

    ASYNC_NETWORK_IO1701190.661.3294.12

    PAGEIOLATCH_SH1467659.301.1495.26

    SOS_SCHEDULER_YIELD1145887.640.8996.15

    CXROWSET_SYNC967398.980.7596.89

    TRACEWRITE946843.310.7397.63

    WRITELOG765721.930.5998.22

    PAGELATCH_UP515339.540.4098.62

    LCK_M_S239391.560.1998.80

    ASYNC_IO_COMPLETION196326.980.1598.95

    LATCH_SH167410.820.1399.08

  • And?

    Gail Shaw
    Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
    SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability

    We walk in the dark places no others will enter
    We stand on the bridge and no one may pass
  • Hi Mr. Gila,

    I have same first 3 issues. Can you please give suggestions?

  • Jonathan Kehayias has a query on this blog post[/url] that identifies parallel plans in the cache which can help you start tuning the SQL that might be causing the CXPACKET waits.

  • smtzac (5/12/2014)


    Hi Mr. Gila,

    I have same first 3 issues. Can you please give suggestions?

    I just spent a long time replying to this and then an error occurred when I submitted my post! :angry:

    I started off by saying that it's Mrs (or Ms) Gila - or Shaw to be correct. I then wrote a bit about CXPACKET waits, but I've since come across this from Paul Randall:

    CXPACKET: this wait type always accrues when parallelism happens, as the control thread in a parallel operation waits until all threads have completed. However, when parallel threads are given unbalanced amounts of work to do, the threads that finish early also accrue this wait type, leading to it maybe becoming the most prevalent. So this one could be benign, as the workload has lots of good parallelism, but could be malignant if there’s unwanted parallelism or problems causing skewed distribution of work among parallel threads.

    You can affect parallelism through the use of Max Degreee of Parallelism (MAXDOP) and Cost Threshold for Parallelism, BUT, I would strongly advise against changing those settings on a Production server with testing in a Test environment first.

    Regards

    Lempster

Viewing 5 posts - 1 through 4 (of 4 total)

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