| Author |
Message
|
| PeterPotkay |
 Posted: Sat Feb 21, 2004 9:19 am Post subject: New Queue Attribute in CSD06 |
 |
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
The MQExplorer does not show it, but runmqsc does.
I wonder if there is any performance hit? If not, why not set it to HIGH always? Unless you are the type to bounce QMs solely to get rid of Persistent messages?
From the read me.....
| Quote: |
New Queue Attribute
This WebSphere MQ V5.3 update introduces a new queue attribute
NPMCLASS which applies to local and model queues. NPMCLASS can
take one of two values:
1.NPMCLASS(NORMAL) - This is the default value and indicates
that non-persistent messages on this queue are only lost following
a failure, or a queue manager shutdown. These messages will be
discarded in the event of a queue manager restart.
2.NPMCLASS(HIGH) - This setting enables non-persistent messages on
this queue to be retained across a queue manager restart.
Non-persistent messages may still be lost in the event of a
failure.
The WebSphere MQ V5.3 JMS classes are extended to allow exploitation
of the NPMCLASS queue attribute. This is done by adding a new value,
JMSC.MQJMS_PER_NPHIGH, to the list of p persistence modes accepted by
the Destination.setPersistence(int persistence) method. If this
value is set and the underlying queue(s) supports it, then persistent
JMS messages will be mapped to non-persistent (MQPER_NOT_PERSISTENT)
MQ messages. If the underlying queue(s) do not specify
NPMCLASS(HIGH), then persistent JMS messages will map to persistent
(MQPER_PERSISTENT) MQ messages.
In order to determine if the underlying queue supports NPMCLASS then
when JMSC.MQJMS_PER_NPHIGH is specified the JMS classes will issue
an MQINQ against the base queue. JMSC.MQJMS_PER_NPHIGH should not
be specified when the JMS application is not authorized to inquire
upon the base queue.
It is the administrator's responsibility to ensure that all
appropriate queues specify NPMCLASS(HIGH). Note that for
Publish/Subscribe applications, both the publisher and subscriber
queues must be specify NPMCLASS(HIGH).
For maximum reliability it is advised that the JMS Persistent
attribute is mapped to WebSphere MQ persistence rather than use of
JMSC.MQJMS_PER_NPHIGH.
|
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
 |
| Michael Dag |
| Posted: Sat Feb 21, 2004 11:31 am Post subject: |
|
|
Jedi Knight
Joined: 13 Jun 2002
Posts: 2607
Location: The Netherlands (Amsterdam)
|
Interesting new feature, I hope it doesn't introduce new defects aswell 
The only reason for the default setting I can see is: not all admins read the readme and
would probably go nuts finding a lot of non-persistent messages after a restart... 
Michael |
|
| Back to top |
|
|
| JasonE |
| Posted: Sun Feb 22, 2004 2:26 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003
Posts: 1220
Location: Hursley
|
Its not the default for lots of reasons, including the fact that it would never change mid release. It makes the difference between NP and P less obvious as well, as the term 'event of a failure' is very vague. I believe its basically there to give a performance boost if you want to process messages but dont mind if they get list in the 'event of a failure' but want them if possible across a restart. NP vs P is just a question of what gets logged - less i/o's == faster messages...
As to introducing new defects - I hope not too  |
|
| Back to top |
|
|
| pgorak |
| Posted: Mon Feb 23, 2004 2:39 am Post subject: |
|
|
Disciple
Joined: 15 Jul 2002
Posts: 158
Location: Cracow, Poland
|
| Quote: |
| It makes the difference between NP and P less obvious as well, as the term 'event of a failure' is very vague. |
I bet it'll cause a lot of confusion. Just wait for new posts 
Piotr
_________________
***
IBM Certified Solution Developer WebSphere MQ 5.3 |
|
| Back to top |
|
|
| JasonE |
| Posted: Mon Feb 23, 2004 4:02 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003
Posts: 1220
Location: Hursley
|
I wonder if it changes the answers to any of the certification questions?  |
|
| Back to top |
|
|
| Michael Dag |
| Posted: Mon Feb 23, 2004 4:53 am Post subject: |
|
|
Jedi Knight
Joined: 13 Jun 2002
Posts: 2607
Location: The Netherlands (Amsterdam)
|
uhm...
Which statement is TRUE about non-persisent messages?
a) ..
b) ..
c) non-persistent messages do not survive a queuemanager restart
d) ..
This is obviously not TRUE anymore, but like Piotr says, what about the less obvious situations...
Michael |
|
| Back to top |
|
|
|
|
