Initial slowness after 9.0.1.11 upgrade

 2 Replies
 0 Subscribed to this topic
 27 Subscribed to this forum
Sort:
Author
Messages
Gary P
New Member
Posts: 1
New Member
    If only there was a head scratcher category!  We upgraded our test environment from 9.0.1.5 to 9.0.1.11 and portal to match.  In addition we upgraded AIX from 5.3 to 7.1, WebSphere from 6 to 7, Oracle from 10G to 11G, and moved ADAM from 2003 server to 2008 R2.  Anyway, after the upgrade we are encountering this odd issue.  After a reboot, or the first time you access another persons information - say like time entry detail from summary approval, the reponse time is horribly slow.  After that first access, it is much quicker.  A couple examples.  I go into period time entry - it takes 10 seconds for the screen to come up.  I go out, come back in (or close browser as well), it takes 2 seconds.  I go to payment modeling, it takes 9 seconds to come up.  I go out, come back in, 2 or 3 seconds.  I go to manager summary approval.  Select detail for first report - takes 10 seconds.  Select second report - takes 10 seconds..go back to first report - 1 second, go to second report again 1 second..go to third report - 10 seconds.  Clear out IOS cache, still a good response time.  If I go to a form, it takes a bit long the first time, but subsequent other forms seem to be quicker/normal.  If someone else goes in, they experience the same type of performance. 
    Greg Moeller
    Veteran Member
    Posts: 1498
    Veteran Member
      We've noticed the same thing, but I don't think it was just with the upgrade... it started long before that. I think that it has something to do with caching, but have no data to back that up. We've pretty much just accepted it as a WAD (working as designed).... It may have stared with the update to WAS 7, which we did before the update to ESP 11. But, no!! I can remember it back when everyone was using LID as well. Hmmmmm!!!
      Alex Tsekhansky
      Veteran Member
      Posts: 92
      Veteran Member
        It's not quite WAD, but close to it. If it bugs you, you should evaluate several items before you do more drastic steps:
        1. Check WS logs and make sure you do not have hung threads. WAS would tell you if you do. In general with WAS7 on AIX multiple people using LSA will often cause hung threads, which will slow down certain things, and may even crash WAS if the number of such threads exceeds certain threshold (usually - between 20 and 40 hung threads)
        2. Make sure that Oracle is 11.2 (and not 11.1). Some patch levels of 11.1 have really weird issues with LSF, depending on the products you have.
        3. If you have more than 500 named users defined in LSF, make sure you modified ADAM parameters, so it can return more results than 500 in a single query.
        4. Make sure that Lawson's hostname is resolved by the Windows box that runs ADAM quickly, and vise versa (the easiest way - clear DNS cache, and try nslookup)

        If none of the items above presents a problem, post back into this thread, and we will deal with less obvious items.