| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Slow Model Queue on Windows System |
« View previous topic :: View next topic » |
| Author |
Message
|
| insel |
 Posted: Mon Mar 04, 2002 5:29 am Post subject: |
 |
|
Newbie
Joined: 03 Mar 2002
Posts: 3
|
Hi all,
We have a Queuemanager version 5.2 under Windows 2000.
At this shows itself the phenomenon, that local Queues written up more quickly will can were produced than Queues them by way of a Model Queue. The same holds good for that read from the Queue.
As example 1024 records were written a 1024 byte.
On a local queue the program needed 1. 7 seconds.
On a model queue (them already existed DEFTYPE=PERMDYN) then program needed 19.8 seconds.
If somebody knows why that in such a manner is.
Thanx
Insel
|
|
| Back to top |
|
 |
| mqonnet |
| Posted: Mon Mar 04, 2002 6:05 am Post subject: |
|
|
Grand Master
Joined: 18 Feb 2002
Posts: 1114
Location: Boston, Ma, Usa.
|
The reason for this time consumption is because even though you specified Perm dynamic queues, they are non-existent entities at the time of app start. And they need to be created. This happens on an Mqopen of the model queue. So, there is always an additional task involved here. Opening a model queue, take the template def from it and create a dyna queue, open this dyna queue and then pass it's handle back. All this process on the Open only. And then it tries to do the put/get. And thus it has this overhead and takes more time than a direct open,put/get of a local queue.
Cheers.
Kumar
_________________
IBM Certified WebSphere MQ V5.3 Developer
IBM Certified WebSphere MQ V5.3 Solution Designer
IBM Certified WebSphere MQ V5.3 System Administrator |
|
| Back to top |
|
|
| insel |
| Posted: Wed Mar 06, 2002 12:01 am Post subject: |
|
|
Newbie
Joined: 03 Mar 2002
Posts: 3
|
Hi Kumar,
thank you for your answer, but indeed goes these at the theme past.
For even when the new preparing claims the Queue much stretch of time, it should not have these drastic consequences.
Which as well in the UNIX (Sun) or OS /390 associated area is not the case. But as well as all the time, I prepare the Queue not at all newly! !!! If one separate use them some day times via "Model Queue" was produced and have then these "STRETCHES OF TIME"!
Detlef |
|
| Back to top |
|
|
| mqonnet |
| Posted: Wed Mar 06, 2002 4:35 am Post subject: |
|
|
Grand Master
Joined: 18 Feb 2002
Posts: 1114
Location: Boston, Ma, Usa.
|
So, am i right to understand that your dynamic queues are defined already and exit before you start your program. If yes, then is this the same program which is being used to test the performance of Predefined and dyna queues. Or do you have different programs to do this. The reason i am asking this is, persistent Messages take more time than non-persistent. Also if you have uow, it might also be a consideration. By the way, i assume that both these queues, predefined as well the permanent dynamic queue are local to the QM.
Also, did you use the same programs to test on Solaris and other boxes or you used different programs. What was the result on those platforms.
Cheers.
Kumar
_________________
IBM Certified WebSphere MQ V5.3 Developer
IBM Certified WebSphere MQ V5.3 Solution Designer
IBM Certified WebSphere MQ V5.3 System Administrator |
|
| Back to top |
|
|
| insel |
| Posted: Thu Mar 07, 2002 5:53 am Post subject: |
|
|
Newbie
Joined: 03 Mar 2002
Posts: 3
|
OK,
1.) I'm create with program A via ModelQueue one new Queue (PERMDYN).
2.) With program B i test the Performance (NON-PERSISTENT in both Local Queue and the new Queue from Model).
3.) For the test i use on all systems the same program
|
|
| Back to top |
|
|
| mqonnet |
| Posted: Thu Mar 07, 2002 6:09 am Post subject: |
|
|
Grand Master
Joined: 18 Feb 2002
Posts: 1114
Location: Boston, Ma, Usa.
|
If you say that you have are testing on two queues, one predefined and the other a permanent dynamic queue(whose name is known to you). And you do puts on these queues using the same application, i would think that the results on both should more or less be the same.
Not quite sure about your app. But i would first try modifying the standard amqsput sample to put 100 messages and try to note the time difference. If it is almost near, then would investigate your program. Else, i would go deeper into this one.
Cheers.
Kumar
_________________
IBM Certified WebSphere MQ V5.3 Developer
IBM Certified WebSphere MQ V5.3 Solution Designer
IBM Certified WebSphere MQ V5.3 System Administrator |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
