|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
 |
|
Message Expiry - Implementation Info |
« View previous topic :: View next topic » |
Author |
Message
|
IntegratorWMB |
Posted: Mon Aug 18, 2008 5:53 am Post subject: Message Expiry - Implementation Info |
|
|
Apprentice
Joined: 06 Dec 2007 Posts: 27
|
All,
I am using WMB 6.1 and this question pertains to the use of Message Expiry of MQ.
Currently we are developing monitoring of messages at the application level. Till it goes for production, we have a stop gap solution for monitoring the message by making a copy of the messages processed. One of the copy queues grows by 40000 messages a day. The messages needs to be in the copy queue for atleast 48 hours. I had set the MQMD Expiry property accordingly and it works.
My question is to do with the Implementation of the Expiry. I see the message expiry time for each message in the copy queue gets modified for every milliseconds. With 80,000 messages on a copy queue over 2 days is it wise to use the Expiry property. I need to know the innards of Expiry implementation. Is it using Bucket Timer algorithmn (or) Will there be a performance hit in using the Expiry property in a scenario like this ?
Thanks for any pointers. |
|
Back to top |
|
 |
belchman |
Posted: Mon Aug 18, 2008 5:56 am Post subject: |
|
|
Partisan
Joined: 31 Mar 2006 Posts: 386 Location: Ohio, USA
|
For one... it is never wise to use MQ as a database... _________________ Make three correct guesses consecutively and you will establish a reputation as an expert. ~ Laurence J. Peter |
|
Back to top |
|
 |
SAFraser |
Posted: Mon Aug 18, 2008 12:12 pm Post subject: |
|
|
 Shaman
Joined: 22 Oct 2003 Posts: 742 Location: Austin, Texas, USA
|
I do not know the answer to your question about expiry and how the headers are updated.
I do, however, live in a paranoid environment where WMB is used to "just make sure" MQ got the message. We popped in a trace node and we write certain fields of the input message into a text file. These types of trace nodes have been running in production (I know, it's bad) for a number of years, with average message volumes of 80,000/day and a peak of 400,000/day with no measureable impact on performance.
The bigger issue, really, is all the disk space the logs require. |
|
Back to top |
|
 |
gbaddeley |
Posted: Mon Aug 18, 2008 7:24 pm Post subject: |
|
|
 Jedi Knight
Joined: 25 Mar 2003 Posts: 2538 Location: Melbourne, Australia
|
The MQMD.Expiry on queued messages is only updated when the message is accessed by MQ at a point in time when it can expire. eg. Doing MQGET destructive or browse, Message Channel Agents, or the z/OS MQ expiry agent runs (see EXPIRYINT qmgr attribute).
There is no efficiency penalty for using Expiry. _________________ Glenn |
|
Back to top |
|
 |
PeterPotkay |
Posted: Tue Aug 19, 2008 5:09 am Post subject: |
|
|
 Poobah
Joined: 15 May 2001 Posts: 7722
|
gbaddeley wrote: |
...or the z/OS MQ expiry agent runs (see EXPIRYINT qmgr attribute) |
And as of MQ 6.0 there is also a built in behind the scenes Expiry message scavenger on distributed QMs. But it only looks at queues that are open, even if its just for MQINQ. Unlike on z/OS, there is no QM attribute to determine how often it runs. In fact, there is not much written about it at all, although if you look at the QM Error logs at start up time you will see it starting. _________________ Peter Potkay
Keep Calm and MQ On |
|
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
|
|
|
|