View previous topic :: View next topic
|
Author |
Message |
polymathtarun
New User
Joined: 01 May 2016 Posts: 10 Location: India
|
|
|
|
There have been instances in in our environment when CICS region runs out of active task limit due to excessive concurrent users hits. At times this creates connectivity loss thus impacting downstream applications and further business outrage. Can someone please advise what alternate solutions are available to ensure better system availability in this situation? |
|
Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
|
|
|
|
There are a number of limits set for a CICS region, usually through changes to the SIT (system initialization table). You absolutely MUST contact your site support group and work with them to change these values. Modifying them can impact the system as a whole, but if you have adequate business reasons for changing them then your site support group can make the changes. |
|
Back to top |
|
|
polymathtarun
New User
Joined: 01 May 2016 Posts: 10 Location: India
|
|
|
|
Thank you so much. |
|
Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010 Posts: 3051 Location: NYC,USA
|
|
|
|
Act as per what Robert has recommended but also did you think on TCP/IP OR Network balancing ? |
|
Back to top |
|
|
polymathtarun
New User
Joined: 01 May 2016 Posts: 10 Location: India
|
|
|
|
Rohit- There is no network issue. Can you please explain me how network balancing is going to help here? |
|
Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
|
|
|
|
If the issue is CICS max tasks being reached, or CICS storage limits being reached, then I don't see how TCP/IP or Network balancing could possibly help. If max tasks is reached, for example, it doesn't matter where in the network the terminal is coming from it still cannot run a transaction until the max tasks condition is resolved. And, actually, if network balancing allows more terminals access to the region than are currently getting there, network balancing could actually make the problem WORSE, not better. |
|
Back to top |
|
|
polymathtarun
New User
Joined: 01 May 2016 Posts: 10 Location: India
|
|
|
|
Awesome and thank you for an explanation, Robert Sir. |
|
Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010 Posts: 3051 Location: NYC,USA
|
|
|
|
I thought the transactions can be re routed when one region gets exhausted it's limit.
Upon further readings it should probably be good to combo transaction CLasses (TCLASS) with MXT Parameter to make it more scalable. |
|
Back to top |
|
|
polymathtarun
New User
Joined: 01 May 2016 Posts: 10 Location: India
|
|
|
|
Robert and Rohit
Can you please advise which specific parameters of SIT table must be fine tuned to get rid of this problem. Thank You. |
|
Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
|
|
|
|
Have you done (or had your site support group do) a STAT and reviewed it? That can help pinpoint problem areas.
Most likely, based on the little you've said, the parameter to change will be MXT. This directly controls how many concurrent tasks are allowed in a CICS region. The default is 250 but few, if any, sites use the default. If the value of MXT gets increased a lot, I'd say to review DSALIM and EDSALIM since more transactions means more memory in use by those transactions. |
|
Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010 Posts: 3051 Location: NYC,USA
|
|
|
|
How more specific do you need ? with those above keywords in bold take a look in the IBM Manuals and then you should get close to the solution and understand it. |
|
Back to top |
|
|
|