Post reply

Lazy Writer

  • June 9, 2011 at 9:32 pm

    #73805

    Comments posted to this topic are about the item Lazy Writer

  • June 9, 2011 at 9:33 pm

    #1336796

    Nice clear question - thanks

    If everything seems to be going well, you have obviously overlooked something.

    Ron

    Please help us, help you -before posting a question please read[/url]
    Before posting a performance problem please read[/url]

  • June 9, 2011 at 10:19 pm

    #1336801

    Good question, thanks.

    M&M

  • June 10, 2011 at 3:52 am

    #1336856

    Nice question Steve.

  • June 10, 2011 at 5:12 am

    #1336891

    Nice question...

    -----------------
    Gobikannan

  • June 10, 2011 at 6:20 am

    This was removed by the editor as SPAM

  • June 10, 2011 at 6:21 am

    #1336918

    Wow. I really had to dig for that one.

    Pretty confusing stuff. I'd like to know if anyone here utilizes this feature of splitting physical CPUs for the purpose of creating soft-NUMA nodes for multiple lazy writer threads.

    So with one processor, do you have 1 NUMA node by default?

    ______________________________________________________________________________________________
    Forum posting etiquette.[/url] Get your answers faster.

  • June 10, 2011 at 6:46 am

    #1336932

    Nice question! But I have questions.

    What is the best architecture in the real world? Soft-NUMA in SMP, NUMA hardware with soft-NUMA or NUMA hardware only? Is it work fine in any SQL Edition (2005/2008 - Standart, Enterprise, Workgroup, Express...)?

  • June 10, 2011 at 6:48 am

    #1336934

    Good question.

  • June 10, 2011 at 7:00 am

    #1336945

    Question is missing a response: "ten forum posts" 😛

    ---------------------------------------------------------
    How best to post your question[/url]
    How to post performance problems[/url]
    Tally Table:What it is and how it replaces a loop[/url]

    "stewsterl 80804 (10/16/2009)I guess when you stop and try to understand the solution provided you not only learn, but save yourself some headaches when you need to make any slight changes."

  • June 10, 2011 at 7:00 am

    #1336946

    calvo (6/10/2011)

    I'd like to know if anyone here utilizes this feature of splitting physical CPUs for the purpose of creating soft-NUMA nodes for multiple lazy writer threads.

    You get a lazy writer (and couple other things) per hard NUMA node, not soft. That reference is wrong.

    Let me see if I can find a better one....

    First the connect item for the incorrect BoL info: http://connect.microsoft.com/SQLServer/feedback/details/670328/books-online-incorrect-soft-numa-information

    And Bob Dorr's article on NUMA, this should be THE reference article on SQL and NUMA

    http://blogs.msdn.com/b/psssql/archive/2010/04/02/how-it-works-soft-numa-i-o-completion-thread-lazy-writer-workers-and-memory-nodes.aspx

    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

  • June 10, 2011 at 7:13 am

    #1336956

    GilaMonster (6/10/2011)

    calvo (6/10/2011)

    I'd like to know if anyone here utilizes this feature of splitting physical CPUs for the purpose of creating soft-NUMA nodes for multiple lazy writer threads.

    You get a lazy writer (and couple other things) per hard NUMA node, not soft. That reference is wrong.

    And Bob Dorr's article on NUMA, this should be THE reference article on SQL and NUMA

    http://blogs.msdn.com/b/psssql/archive/2010/04/02/how-it-works-soft-numa-i-o-completion-thread-lazy-writer-workers-and-memory-nodes.aspx

    Except that Bob's Article says

    The benefits of soft-NUMA include reducing I/O and lazy writer bottlenecks on computers with many CPUs and no hardware NUMA. There is a single I/O thread and a single lazy writer thread for each NUMA node. Depending on the usage of the database, these single threads may be a significant performance bottleneck. Configuring four soft-NUMA nodes provides four I/O threads and four lazy writer threads, which could increase performance.


    --Mark Tassin
    MCITP - SQL Server DBA
    Proud member of the Anti-RBAR alliance.
    For help with Performance click this link[/url]
    For tips on how to post your problems[/url]

  • June 10, 2011 at 7:18 am

    #1336962

    mtassin, I believe Bob is quoting BOL in that paragraph.

    Gail,

    I had found that link earlier and figured BOL would be the authoratative source. Thanks for clearing that up for us.

    Bob says "The lazy writer thread creations are tied to the SQL OS view of the physical NUMA memory nodes. So whatever the hardware presents as physical NUMA nodes will equate to the number of lazy writer threads that are created."

    So no matter how many soft NUMA nodes you have, the number of lazy writer threads is based on the presence of hardware NUMAs.

    ______________________________________________________________________________________________
    Forum posting etiquette.[/url] Get your answers faster.

  • June 10, 2011 at 7:31 am

    This was removed by the editor as SPAM

  • June 10, 2011 at 7:43 am

    #1336987

    jcrawf02 (6/10/2011)

    Question is missing a response: "ten forum posts" 😛

    Ha, ha! Thanks for the laugh!

Viewing 15 posts - 1 through 15 (of 37 total)

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