| Author |
Message
|
| arju |
 Posted: Sat Jun 21, 2014 5:52 am Post subject: Why is trigger set off ? |
 |
|
Novice
Joined: 13 Apr 2014
Posts: 22
|
We are at z/OS 1.13 and MQ 7.0.1.0. We have a cics transaction on a triggered queue which picks up messages on the queue when they arrive on the queue. But when cics has troubles with transaction, the trigger is set to off automatically. We can not understand this behavior.
Any explanation from gurus here? |
|
| Back to top |
|
 |
| bruce2359 |
| Posted: Sat Jun 21, 2014 7:06 am Post subject: Re: Why is trigger set off ? |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| arju wrote: |
...when cics has troubles with transaction, the trigger is set to off automatically. We can not understand this behavior.
Any explanation from gurus here? |
Has troubles? You mean the transaction abended? Or, something else?
One of the requirements for triggering is that the application executes successfully - that is: it exists, can be loaded, does not abend, etc.. The triggering mechanism is smart enough to realize that a failed app should not be triggered again, and again, and again...
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| arju |
| Posted: Sat Jun 21, 2014 8:04 am Post subject: |
|
|
Novice
Joined: 13 Apr 2014
Posts: 22
|
| yes, the cics transaction is disabled. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sat Jun 21, 2014 8:09 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| arju wrote: |
| yes, the cics transaction is disabled. |
Did the transaction abnormally terminate (abend)? What abend code?
| arju wrote: |
| the trigger is set to off automatically. |
Do you mean that the triggered application queue is set to TRIGTYPE(NONE) or NOTRIGGER? Or, that the transaction is disabled?. Or both.
Please be more technically precise in your posts.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| arju |
| Posted: Sat Jun 21, 2014 8:36 am Post subject: |
|
|
Novice
Joined: 13 Apr 2014
Posts: 22
|
| The trigger is set to off when the transaction abends with an internal abend and rollsback. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sat Jun 21, 2014 8:40 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| arju wrote: |
| The trigger is set to off when the transaction abends with an internal abend and rollsback. |
WMQ has no OFF trigger setting attribute value. Where do you see OFF?
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| arju |
| Posted: Sat Jun 21, 2014 8:46 am Post subject: |
|
|
Novice
Joined: 13 Apr 2014
Posts: 22
|
| It is set to N. And trigger is defined as First. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sat Jun 21, 2014 9:48 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| arju wrote: |
| It is set to N. And trigger is defined as First. |
Ah. You are using the ISPF panel interface to display queue attributes and values.
Y/N are the possible values: N=no, Y=yes. N=notrigger, Y=trigger. No OFF value.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| arju |
| Posted: Sun Jun 22, 2014 3:31 am Post subject: |
|
|
Novice
Joined: 13 Apr 2014
Posts: 22
|
Thanks for your patience. Somehow the trigger is set to n and we have to set it to y on every monday.. then the consumer consumes everything on the queue.
But we would like to know who turns the trigger from y to n. Is it qmgr? Is it cics?
We think CICS is only an application for QMGR so it has no business to change any internal object of the qmgr. It's sole purpose is to get the messages from the queue and in case of failure, depending on failure either to commit or rollback. As CICS is doing a normal two phase commit here.
Please help us the solve the mystery !! |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sun Jun 22, 2014 5:54 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
This only occurs on Mondays? What else occurs in the same timeframe? Perhaps some batch job? WMQ has no 'if it's Monday, then ...' capability; but your automation software likely does.
What other useful information can you provide us?
Is this a new problem? If so, what else has recently been changed? Programs recompiled? New batch jobs?
Look at the 'last altered' date on the triggered queue to find out exactly when triggering was set to notrigger. Look at syslog to see what else was running.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sun Jun 22, 2014 6:27 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| arju wrote: |
| We think CICS is only an application for QMGR so it has no business to change any internal object of the qmgr. It's sole purpose is to get the messages from the queue and in case of failure, depending on failure either to commit or rollback. As CICS is doing a normal two phase commit here. |
CICS is far more than 'only an application for QMGR'. CICS is a transaction-manager capable of managing the execution of thousands of concurrent transactions.
You need to contact (work with) your CICS systems programmer to identify the real source of your problem. Much depends on exactly and precisely how your consumer is triggered, what options have been set for your consumer CICS transaction, what automation might be involved, and so on.
Apparently, more than just 'a normal two-phase commit' is taking place; and you (and your sysprog) need to identify the root cause.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sun Jun 22, 2014 8:54 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| arju wrote: |
| The trigger is set to off when the transaction abends with an internal abend and rollsback. |
Be a bit more precise, please.
What do you mean by 'internal abend?' Is the application written to terminate itself? If so, how does it do so?
What abend code or abend message does the application encounter?
Does the transaction abend with a Unnn abend code - like U123? An Snnn abend code - like a S0c7? Or something else?
Is the abend code found in the CICS Messages and Codes manual? Or user documentation?
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jul 21, 2014 6:55 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
Any progress on this issue?
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
|
|
