Server is Slow When All Performance Numbers are Good

  • Hi All

    We have an HP ML570 G3 windows server 2003 Enterprise edition with SQL 2000 Enterprise Edition SP4. It has 4 Xeon 3.33 Ghz processors running Hyper-Threading so we have total 8 processors and 3.5 GB of RAM. All drivers and firmwares are up-to-date. O/S and SQL logs are on RAID 1 and SQL data files are on RAID 5; plenty of unused space. The only problem we have is the server is very slow. This is a back end box for a web-based application running on another box. The previous version of the application were running fast, everyone was happy with it. When we installed the new version we also migrated it to a newer server. However the app was starting to run slower. I monitored the server performance and all the numbers are perfect but the app is still slow. On the SQL the average CPU usage is nerver go higher than 10% at anytime so I just wonder if SQL is misconfigured. Can you guys help look for anything wrong here? Thanks.

    Below is result from sp_configure. Sorry the numbers are not aligned.

    affinity mask-21474836482147483647620

    allow updates0100

    awe enabled0111

    c2 audit mode0100

    cost threshold for parallelism03276722

    Cross DB Ownership Chaining0100

    cursor threshold-12147483647-1-1

    default full-text language0214748364710331033

    default language0999900

    fill factor (%)010000

    index create memory (KB)704214748364700

    lightweight pooling0100

    locks5000214748364700

    max degree of parallelism03200

    max server memory (MB)4214748364725602560

    max text repl size (B)021474836476553665536

    max worker threads3232767255255

    media retention036500

    min memory per query (KB)512214748364710241024

    min server memory (MB)0214748364700

    nested triggers0111

    network packet size (B)5123276740964096

    open objects0214748364700

    priority boost0100

    query governor cost limit0214748364700

    query wait (s)-12147483647-1-1

    recovery interval (min)03276700

    remote access0111

    remote login timeout (s)021474836472020

    remote proc trans0111

    remote query timeout (s)02147483647600600

    scan for startup procs0100

    set working set size0100

    show advanced options0111

    two digit year cutoff1753999920492049

    user connections03276700

    user options03276714041404

  • Why did you concentrate on the SQL server parameters only?

    What about your network traffic? Do you have a single network card on this box? What is the network speed?

  • Glen (11/14/2008)


    Why did you concentrate on the SQL server parameters only?

    What about your network traffic? Do you have a single network card on this box? What is the network speed?

    I got to look on the SQL part first because I am DBA and it seems everyone is going to blame others first so I must make sure I am not doing anything wrong on my part 🙂

    We are on a 100 MB Cisco switch network. The NIC's was at 100 MB full. I changed to Auto, nothing improved.

    Thanks.

  • What is actually slow?

    If the application is responding slowly then have you tried running the queries ran by application manually under query analyzer? What is the Index structure and index fragmentation(are they getting

    rebuilt or defragmented regularly)and check the execution plan of queries if they run too slow when ran manually.

    Also, post these four counters output from performance monitor:

    SQLServer:Buffer Manager\Buffer cache hit ratio

    SQLServer:Buffer Manager\Checkpoint pages/sec

    SQLServer:Buffer Manager\Lazy writes/sec

    SQLServer:Buffer Manager\Page life expectancy

    MJ

  • MANU (11/14/2008)


    What is actually slow?

    If the application is responding slowly then have you tried running the queries ran by application manually under query analyzer? What is the Index structure and index fragmentation(are they getting

    rebuilt or defragmented regularly)and check the execution plan of queries if they run too slow when ran manually.

    Also, post these four counters output from performance monitor:

    SQLServer:Buffer Manager\Buffer cache hit ratio

    SQLServer:Buffer Manager\Checkpoint pages/sec

    SQLServer:Buffer Manager\Lazy writes/sec

    SQLServer:Buffer Manager\Page life expectancy

    MJ

    That is the hard part: the web app and SQL database are from a vendor. I don't know what it does in each stored procedure because they are encrypted. Monitoring a web app is impossible to tell what stored procedures are called when a web user click on a button on the page because in profiler I see hundreds of stored procedures running by only one login for the IIS front end.

    Nightly the database is checked, optimized, and reindexed so I am sure fragment is not a problem.

    SQLServer:Buffer Manager\Buffer cache hit ratio: 99.8

    SQLServer:Buffer Manager\Checkpoint pages/sec: 0 most of the time. It peaks for less than 5 seconds then drop to 0 again

    SQLServer:Buffer Manager\Lazy writes/sec: same as Checkpoint pages/sec

    SQLServer:Buffer Manager\Page life expectancy: always greater than 300 (expected number) unless a checkpoint occurs then it drops below 300 but only for couple seconds

    Thanks.

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

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