Hanging Portal Sessions

Hi Greg,
I know I'm replying to a post from March... but... we have a lot of performance related issues in both Portal and LID since going to LSF 9. I have seen Portal hang like this and/or display a blank screen temporarily. I've also seen it hang while scrolling through records (inquire) for up to 40 seconds.
We've had LPS here twice and opened 5 separate cases with GSC. No resolution to this point. We're getting ready to bring LPS back on site to continue evaluating.

Have you had any success resolving this issue?

We have similar portal and LID performance issue. Have you figured out what is the problem?

Can you describe some of your symptoms? Which forms? Add? Change? Inquire?

We've experienced hanging/slowness on PO screens mostly.  Sometimes we've seen it on Matching screens, but these would tie back to PO data.  It is online forms only.  Users have random hanging or delays in Inquire/Change/dropdowns/page up or down.  Sometimes the form hangs 5-10 seconds before the data refreshes.  We go through spurts of slowness for 10-30 minutes followed by more "normal" speeds, which are almost instantaneously. 

We've had numerous cases with GSC.  We have had LPS on site 3 times now.  No resolution or progress after 4 months.  Still working with LPS and trying to devise some better methods of tracking trends.

FYI... prior to April 08, we ran 8.0x environment on a Unix platform with the same network infrastructure and had no performance issues at all.  We're now LSF 9.0.0.5 on Windows 2003 server.

Are you experiencing these problems in LID, Portal, or both? Do you have any logging files growing at high rates in %LAWDIR%/system/* or Websphere? You might have trace turned on in Websphere or LSF9 somewhere.

Also, what database are you running, and is it on the same server?

Mostly in Portal (80% of our users are Portal) but also in LID to a lesser degree. What I've seen in LID is more like this... you page down like 4 times, it's instant, the 5th time it hangs for couple seconds, then it's fast again... and so on.

Nothing going crazy in %LAWDIR%/system/. In fact LPS/GSC says there's nothing remarkable in any of our log files, including WAS and the Windows event viewer.

We're Oracle 10g. The DB sits on its own server.

Currently LPS is to provide a "slowness form" to help us track slowness trends. It seems like we've looked at everything, though, from Lawson config to WAS config to database to network to individual user PC settings.

Try looking elsewhere - I do not beleive the issue is with your online screens - although that is where it is manifested.

My take on this is the batch jobs you are running at the same time as the online screens are locking records that you need to get at, when the batch report/update releases the locks - your screen moves on just fine.

You can adjust this on a per-batch-job basis by changing the MAX-OPS_IN_TRAN value for the job in jobdef ( .....from the default (50 transactions) down to a less lock-hungry 5 transactions - all that will happen is that your batch jobs will commit to the DB more frequently, with no apparent downside I have ever observed, and your online screens should suffer less....This issue is more prevalent in LSF9 than 8.0.x because LSF9 screams thru batch jobs far faster than 8.0.x ever could - more activity - more locks being held...

Is there a way to get a list of job's run per person based on a particular time frame? I would like to try what scooby has suggested as we have been having this problem for months! Our end users are getting very frustrated with Lawson! Any further suggestions regarding performance would be greatly appreciated...our situation is very similar to khartman's. We are LSF9.0 sp4, apps 8.0.3, windows...

This seems to be a very common problem. We experience it here as well.
We have Windows 2003 with SQL 2005 back end, when it runs fine, it fly, other times, its really slow and the HW is hardly utilized.

We are in the process of fine-tunning the performance, will be sharing it soon with the Guru user community.

Where do you do this?  I've looked in the jobdef for the particular batch job and I don't see that setting anywhere.  Am I missing something (LSF9.0.4 / 8.0.3 Apps / Windows SQL) Thanks!

-Shari

<!-- Converted from text/plain format -->

You have to use jobdef. Position on a step in the job. Then press F6 and select "Logical Transaction Size".

Any updates from the community on this? We are now implementing LFS 9 on windows and this post seems very helpful, describing our situation perfectly.