| Author |
Message
|
| mqjeff |
 Posted: Thu Jan 15, 2009 3:37 am Post subject: |
 |
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| Vitor wrote: |
| (Come, Igor! The storm reaches it's height!) |
That's *my* line. |
|
| Back to top |
|
 |
| Vitor |
| Posted: Thu Jan 15, 2009 3:54 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| mqjeff wrote: |
| Vitor wrote: |
| (Come, Igor! The storm reaches it's height!) |
That's *my* line. |
Not now I've had the opperation. My hunch back is much better now and I'm setting up on my own! No more "Yes, Master" and stealing brains in jars.
Though truth be told, that really had nothing to do with my duties and was just a hobby......
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| sebastia |
| Posted: Thu Jan 15, 2009 4:03 am Post subject: |
|
|
Grand Master
Joined: 07 Oct 2004
Posts: 1003
|
Exactly :
| Quote: |
| It's good for what it's for. |
It is good to have a copy of all the trigger messages !
Which can be veru useful in some installation,
in a primary stage of the project,
when few things seem not to work ...
Cheers. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Thu Jan 15, 2009 4:14 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| Vitor wrote: |
| mqjeff wrote: |
| Vitor wrote: |
| (Come, Igor! The storm reaches it's height!) |
That's *my* line. |
Not now I've had the opperation. My hunch back is much better now and I'm setting up on my own! No more "Yes, Master" and stealing brains in jars. |
Then I believe it is exerk's line. |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Jan 15, 2009 5:19 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| sebastia wrote: |
| It is good to have a copy of all the trigger messages ! |
Why?
The trigger message is a transitory thing, which says nothing about the message that caused it to be raised.
| sebastia wrote: |
Which can be veru useful in some installation,
in a primary stage of the project,
when few things seem not to work ...
|
Define a scenario.
All the trigger message tells you is that the queue manager reacted to an event. If "a few things seem not to work" where does this get you? If the application is not responding properly to the trigger or message data, nothing in the message will tell you this.
Indeed, this could lead you up a false path. Consider this scenario: a message is sent to a triggered queue and disappears. Not on the queue, database not updated with the message contents, nothing. You check and discover there's no trigger message, cry "ah ha!" and look to see why the message never arrived on the target queue, which you know because the trigger never fired on it's arrival.
Some period of time later, you discover there's a bug in the application preventing the database update. There was no trigger message on the TRIGGER(FIRST) queue because the application had a long wait period & had the queue open when the message arrived.
Your scenario please.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| sebastia |
| Posted: Thu Jan 15, 2009 7:21 am Post subject: |
|
|
Grand Master
Joined: 07 Oct 2004
Posts: 1003
|
My scenario ?
An APP writes into a queue, and that msg
must trigger the consumer to send the answer.
The data msg stays in queue, the trigger msg is gone somehow,
and the response application is not awakened at all.
I think that if I am able to demonstrate
that the trigger message was there,
then half of the probable causes can be discarded.
In other words - it can help and it doesn't hurt.
Cheers. |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Jan 15, 2009 7:27 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| sebastia wrote: |
I think that if I am able to demonstrate
that the trigger message was there,
then half of the probable causes can be discarded.
|
And you can obtain exactly this by either a) using RFHUtil as a trigger monitor or b) running the trigger monitor in foreground. No need for exits or exotic mirroring strategies. Both of these methods have the added advantage of removing the need to tie message times together to see if the trigger message you've recorded matches the input message.
But if you want to hit yourself over the head with a stick while crawling over broken glass, I defend your right to make that choice. 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| sebastia |
| Posted: Thu Jan 15, 2009 7:43 am Post subject: |
|
|
Grand Master
Joined: 07 Oct 2004
Posts: 1003
|
Yes, RFHUTIL is ok, and foreground Trigger Monitor is ok,
but to copy the message can be done at midnight,
and look into the messages the next morning ...
Yes, I do crawl around broken glasses,
and someone else tries to hit me,
but I've learned lots of tricks to survive
those and worse situations ...
Chiao, |
|
| Back to top |
|
|
| mqjeff |
| Posted: Thu Jan 15, 2009 8:26 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
The contents of the trigger messages will only change when you change the process definition.
The only value for mirroring trigger messages is knowing that a message occured... and you can instead do something like change the process definition to point to a shell script that logs that it started up and then runs the real process. |
|
| Back to top |
|
|
| sebastia |
| Posted: Thu Jan 15, 2009 9:09 am Post subject: |
|
|
Grand Master
Joined: 07 Oct 2004
Posts: 1003
|
I think I got the idea : to do the log in the responding process.
But my premise was "the responding program is not started" ...
??? |
|
| Back to top |
|
|
| mqjeff |
| Posted: Thu Jan 15, 2009 9:43 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
So, there are three reasons the responding process did not get started.
Either the trigger monitor had an error starting it - which the default runmqtrm will log to it's stdout.
Or the process setup was malformed or the shell environment the trigger monitor is running in, in which case the responding process logging will show that error.
Or there was no trigger message. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Thu Jan 15, 2009 12:07 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
| Vitor wrote: |
Moving slightly off-topic, may I enquire why? "Bored" is a valid response here, as is "spirit of adventure and enquiry". |
I've always maintained that the manuals do an excellent job of explaining why a Trigger Message is generated. Exactly why and when a Trigger Monitor consumes a trigger message is not that obvious and one maybe surpised what they see.
Playing around like this under the covers is the best way to learn. If you understand why MQ works the way it does, all of the sudden a lot of the pieces come together and you start to understand other areas of MQ more quickly, and your problem resolution skills rocket up. One can never, ever gain this type of insight from studying the manuals, even if you remember 1000% of what you read.
Our z/OS QMs are always up. Nothing ever goes wrong with them. They were built before I came. Consequently I am weak with z/OS MQ compared to, say, um, MQ on Windows, MSCS, triggering, clustering, etc, all of which needed a lot of care and feeding to be correctly tuned here. Now things run pretty good across the board, if I do say so myself. Just gotta play like this with those z/OS QMs some more. Darn z/OS MQ guy on my team set things up to good.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Jan 15, 2009 12:15 pm Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| PeterPotkay wrote: |
| Playing around like this under the covers is the best way to learn. If you understand why MQ works the way it does, all of the sudden a lot of the pieces come together and you start to understand other areas of MQ more quickly, and your problem resolution skills rocket up. One can never, ever gain this type of insight from studying the manuals, even if you remember 1000% of what you read. |
It's for exactly the reasons you laid out that I emphasised that "spirit of adventure and enquiry" was a valid answer. I'm banging on in this thread for 2 principle reasons:
- to prevent an apparently inexperienced person writing a new trigger monitor in the belief they needed to;
- to prevent an apparently inexperienced person potentially blowing his queue manager up with an exit in the belief they needed to.
Doing these because you want to, or as a training exercise is quite valid. Though who'd want to use an exit is a question in itself!
| PeterPotkay wrote: |
| Our z/OS QMs are always up. Nothing ever goes wrong with them. |
Naturally.
| PeterPotkay wrote: |
| Darn z/OS MQ guy on my team set things up to good. |
A quality platform, administered by quality people. 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| exerk |
| Posted: Thu Jan 15, 2009 2:27 pm Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mqjeff wrote: |
| Vitor wrote: |
| mqjeff wrote: |
| Vitor wrote: |
| (Come, Igor! The storm reaches it's height!) |
That's *my* line. |
Not now I've had the opperation. My hunch back is much better now and I'm setting up on my own! No more "Yes, Master" and stealing brains in jars. |
Then I believe it is exerk's line. |
Much as I would like the credit I can't take it  , though I may have used 'throw the switch Igor!' somewhere in here.
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Thu Jan 15, 2009 2:38 pm Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
Actually, I got that wrong.
I meant to imply that you were Vitor's Igor these days, as a response to his claims that he was no longer mine. |
|
| Back to top |
|
|
|
|
