Query regarding CICS performance.



Support for CICS/ESA, CICS/TS & Transaction Gateway, CICS Configuration Manager and CICS Performance Analyzer

Query regarding CICS performance.

Postby Viswanathchandru » Tue Feb 08, 2011 12:01 pm

Hi,
We have a sandbox system where in the system hangs when the CEDF command is invoked and the system resumes only after the task is purged. What could be the reason for this. And how this can be solved. Can anyone please guide me on this. Apologies if m wrong.


Regards,
V
Viswanathchandru
 
Posts: 271
Joined: Mon Oct 25, 2010 2:24 pm
Has thanked: 25 times
Been thanked: 0 time

Re: Query regarding CICS performance.

 

Re: Query regarding CICS performance.

Postby dick scherrer » Wed Feb 09, 2011 12:39 am

Hello,

How do you "purge the task" if the system is hung. . . :?
Hope this helps,
d.sch.
User avatar
dick scherrer
Global moderator
 
Posts: 6308
Joined: Sat Jun 09, 2007 8:58 am
Has thanked: 3 times
Been thanked: 90 times

Re: Query regarding CICS performance.

Postby Robert Sample » Wed Feb 09, 2011 4:33 am

If the LPAR is running on a single CPU, and the task is in a "hard hang" where no CICS commands can be executed, the most likely way to successfully recover is to IPL the LPAR.

If the LPAR is running on multiple CPUs with a "hard hang", you should be able to get in and purge (forcepurge) the region to recover.

If there truly is a task in a loop with no CICS commands, I cannot think of any ways in CICS to end the task since CICS is devoting all its time to that one task.
Robert Sample
Global moderator
 
Posts: 3130
Joined: Sat Dec 19, 2009 8:32 pm
Location: Bellevue, Iowa
Has thanked: 1 time
Been thanked: 185 times

Re: Query regarding CICS performance.

Postby Viswanathchandru » Fri Feb 11, 2011 3:25 pm

Robert: thanks a lot for your post!! really helped me...
Dick: the task was cancelled through a different RACF ID( i mean from another sys prog) :D
Viswanathchandru
 
Posts: 271
Joined: Mon Oct 25, 2010 2:24 pm
Has thanked: 25 times
Been thanked: 0 time

Re: Query regarding CICS performance.

Postby dick scherrer » Sat Feb 12, 2011 2:10 am

Hello,

So the cics region was not "hung" - only the problem transaction?

If this is a test cics, suggest you have the cics support people reduce the amount of time a transaction is permitted. Then it will time-out and use no more resources.
Hope this helps,
d.sch.
User avatar
dick scherrer
Global moderator
 
Posts: 6308
Joined: Sat Jun 09, 2007 8:58 am
Has thanked: 3 times
Been thanked: 90 times

Re: Query regarding CICS performance.

Postby David Jackson » Tue Feb 15, 2011 5:58 am

If the CICS system is in a true looping condition due to task, the only (and even then it is not assured) is to KILL the Transaction cross memory from outside CICS.
Omegamon provided such a means.
User avatar
David Jackson
 
Posts: 10
Joined: Wed Jan 26, 2011 5:09 am
Location: California
Has thanked: 0 time
Been thanked: 0 time

Re: Query regarding CICS performance.

Postby stevexff » Mon Feb 21, 2011 9:08 pm

Yes. CICS has a CPU limit that allows a task to be purged with abend AICA when it is exceeded; the only problem is that this count gets reset when the program does an EXEC CICS command, so if that is part of your loop you are a bit stuck. Omegamon, Mainview and other CICS monitoring tools can kill tasks automatically when they hit thresholds of CPU, storage allocation, and I/O calls - your sysprogs should have these set already in your test regions to stop runaway tasks, and if they haven't you could ask them to set them up.
Steve
stevexff
 
Posts: 56
Joined: Wed Nov 10, 2010 7:48 pm
Has thanked: 0 time
Been thanked: 0 time


Return to CICS

 


  • Related topics
    Replies
    Views
    Last post