suhas-mahajan
29th June 2011, 13:37
Hi Friends
During "Post integration transactions to finance transcations" (tfgld4201m000), we are facing session hung problem. During that, we see nobody will be able to run any session.
Can you please help in tracing?
Thanks and regards
-Suhas Mahajan
Neal Matthews
29th June 2011, 16:21
Hello,
There should be plenty of posts on tracing if you search the board. Is the session hanging in a job queue or hanging when run manually ?
Have you tried just running the session in small ranges such as different transaction types just to see if you can process any data at all.
Cheers Neal
suhas-mahajan
30th June 2011, 09:20
Hi Neal
Thanks for the info.
I was talking about internals not TT_SQL_TRACE or set DBSLOG=01770.
Its hanging when user runs manually and that too randomly. One more thing, I understand that when the system has huge load that time its frequently happens. May be users all forcibly killing logins as that time.
Any input please?
Thanks
-Suhas
Hitesh Shah
30th June 2011, 20:14
Somethings to check
1. Purge device que.
2. Check oracle table spaces and log file growth
3. Potential raid failures
4. Any inefficiently written programs run for long period by many people .
shah_bs
30th June 2011, 20:25
This may not be related, but check for an Application Lock entry using the Remove Application Lock session. I donot remember the id of the lock, but should not be difficult to figure out - should have tfgld in it somewhere. Once in a rare while, when two processes are trying to create the next Batch Number, and the Post Integration Transaction is also running, it can hang or cause the other process to hang. Once the lock is removed, it should be possible to continue.
But it is quite rare.
Also, it may be necessary to tweak the status of the Batch Status record (tfgld100) and the Batch Status by Transaction Type record (tfgld101) to 'Ready' - unfortunately have use General Table Maintenance for that, in case they show to be 'In use' - make sure nobody else is working the batch (manually creating, for example) before changing the status.
suhas-mahajan
30th June 2011, 21:40
Hi Shah (square),
Hitesh - 1. Device queue is not big before 1 month we have cleared.
2. Here were using Informix - but how growth helps not sure..because the issue is occuring randomly.
3. RAID failure - In that case it should not work..but this is not the reason for sure.
4. This could be the reason but how to trace its bit difficult.
BS - I agree with you..its rare..but application lock is not the issue.
Any other inputs?
mark_h
30th June 2011, 22:09
If you haven't contact infor support. They might be able to provide additional options on what the issue could be.
suhas-mahajan
1st July 2011, 06:52
Hi Mark
Thanks for info.
I did contact them but they are still "Analysing" not sure on what. Also, the standard message upgrade SP and PS.
mark_h
1st July 2011, 15:19
You know upgrading the portingset actually solved a session hanging in the past for us. The difference was our session always hung when you tried to run it.
suhas-mahajan
1st July 2011, 15:23
Mark
Thanks for info.
Installing SP and PS means huge project here. No other alternates?
Thanks
mark_h
1st July 2011, 16:23
I can't say for sure, plus there is no gurantee even that solves the problem. What we have done before is have infor give us all the objects in debug mode - then we would use that to debug the issue. The problem I see is that if it is random it might be impossible to tell where the issue really is.