| Author |
Message
|
| jgray |
 Posted: Thu Mar 09, 2006 12:52 pm Post subject: PCFException - reason 2085 (object not found) |
 |
|
Newbie
Joined: 07 Sep 2005
Posts: 8
Location: nyc
|
i'm using support pac me01 with mq 5.3. from time to time i get a PCFException when our JMS service attempts to connect to one of our queues. the frustrating thing is, it's intermittent. it works most of the time, but fails to find & connect to the queue about 10 - 15 percent of the time. does anyone know why i'd be getting this intermittent behavior? the stack trace follows:
java.lang.NullPointerException
at com.ibm.mq.jms.context.MQContext.getAllQueues(MQContext.java:234)
at com.ibm.mq.jms.context.MQContext.list(MQContext.java:303)
at com.ibm.mq.jms.context.MQContext.list(MQContext.java:292)
at com.ibm.mq.jms.context.MQContext.lookup(MQContext.java:339)
at com.ibm.mq.jms.context.MQContext.lookup(MQContext.java:330)
at javax.naming.InitialContext.lookup(Unknown Source) |
|
| Back to top |
|
 |
| fjb_saper |
| Posted: Thu Mar 09, 2006 1:16 pm Post subject: Re: PCFException - reason 2085 (object not found) |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| jgray wrote: |
i'm using support pac me01 with mq 5.3. from time to time i get a PCFException when our JMS service attempts to connect to one of our queues. the frustrating thing is, it's intermittent. it works most of the time, but fails to find & connect to the queue about 10 - 15 percent of the time. does anyone know why i'd be getting this intermittent behavior? the stack trace follows:
java.lang.NullPointerException
at com.ibm.mq.jms.context.MQContext.getAllQueues(MQContext.java:234)
at com.ibm.mq.jms.context.MQContext.list(MQContext.java:303)
at com.ibm.mq.jms.context.MQContext.list(MQContext.java:292)
at com.ibm.mq.jms.context.MQContext.lookup(MQContext.java:339)
at com.ibm.mq.jms.context.MQContext.lookup(MQContext.java:330)
at javax.naming.InitialContext.lookup(Unknown Source) |
This looks to me like a JNDI process.
A little additional check might help.
Just verify that the object retrieved from JNDI is not null.
You might as well have to use a try/catch block for your lookup...
Enjoy 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| jgray |
| Posted: Thu Mar 09, 2006 1:33 pm Post subject: could do a JNDI check, but... |
|
|
Newbie
Joined: 07 Sep 2005
Posts: 8
Location: nyc
|
i still have the question about why looking up the queue object works most of the time, but fails intermittently. the queue is always there, on the same q manager, so a JNDI lookup should find it 100 times out of 100. since i'm looking up the queue objects directly on the MQ system admin queue (how support pac me01 works) is there something about how the sys admin queue responds to pcf commands that i should know that might help me explain this flaky behavior. thanks!
jon[/i] |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu Mar 09, 2006 3:26 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
It really depends on the internal works.
I would guess it may have to do something with the usage of the queue at time of enquiry and possibly with some internal time out?
Not having access while waiting on a synchronized operation and timing out ?
For JMS I would not look up queues dynamically. If you need dynamic queues have them created by the session...
The queues used by lookup should be static in the lookup. (You can tell I love the file context!)...
Enjoy 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Thu Mar 09, 2006 3:55 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Every time you do a lookup, it's possible that you create a connection to the queue manager. Certainly you have to *use* a connection.
It may be that you are occasionally using up all your available connections, and when that happens then the lookup fails.
See if there are errors in the queue manager error logs.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
|
|
