1) Hardly, I do not log timeout errors in my application, since I considered
them harmless :-)
2) Yes, I know that I have to enable notification again, I do this of course.
3) Those applications are not meant for load balancing, they are different
applications for different tasks, making up whole system. The applications
communicate through communication subsystem based on Message Queuing.
But I may decide to increase the interval - the shortest message timer we use is
60sec so I think even 30sec will be enough. Thanks for the idea.

Martin

________________________________

Od: msmq@yahoogroups.com za uživatele Keith Schaab
Odesláno: pá 7.10.2005 2:01
Komu: msmq@yahoogroups.com
Předmět: RE: [msmq] Possible problem with MQ?



Can you determine if you have ArrivedError events occurring with timeout errors
(0xc00e001b)? In the code below you are going to encounter these every time 10s
elapses without a message arriving in the queue you're receiving from.
Additionally you will need to call EnableNotification again when you receive the
ArrivedError event if you aren't already doing so. Is there a specific reason
you have chosen a 10s timeout period in your code? It may be that you will want
to increase this number and if 7 applications per server are there for load
balancing, I would say that you can also reduce the number of redundant
applications you have running at one time. Unfortunately I'm not sure what is
causing the hang you are seeing though.



________________________________

From: msmq@yahoogroups.com [mailto:msmq@yahoogroups.com] On Behalf Of Bodlák
Martin
Sent: Wednesday, October 05, 2005 11:43 PM
To: msmq@yahoogroups.com
Subject: RE: [msmq] Possible problem with MQ?



I use COM access to MQ functionality (callback interfaces and notification) and
I use

pQueue->EnableNotification(m_pEvent, NULL, &m_vTimeout); // 10000 msec ... 10
sec

There are about 7 applications on every server and they are timeouting every 10
seconds and it can be the error in log with the period about a second. Is it a
possible cause?

You wrote about timeout of 0, but I am not aware that there is some zero timeout
inside my application(s). Can this be some internal MQ/COM issue?



Martin



________________________________

Od: msmq@yahoogroups.com za uživatele Keith Schaab
Odesláno: čt 6.10.2005 0:44
Komu: msmq@yahoogroups.com
Předmět: RE: [msmq] Possible problem with MQ?

Usually this is caused by an attempt to retrieve a message with a timeout of 0.
This means MSMQ will fail with the error you describe if there isn't a message
in the queue.



________________________________

From: msmq@yahoogroups.com [mailto:msmq@yahoogroups.com] On Behalf Of Bodlák
Martin
Sent: Tuesday, October 04, 2005 4:57 AM
To: msmq@yahoogroups.com
Subject: [msmq] Possible problem with MQ?



Hello, I have a win2k domain with MQ installed and running and I have set a
debug mode for extended logging.

Now I have a msmq debug log with messages like



0xaf8> Wed Sep 21 02:41:42 2005: DRIVER Error: cancel/100, HR: 0xc00e001b, Data-
0x0
0xaf8> Wed Sep 21 02:41:42 2005: DRIVER Error: qbase/1330, HR: 0xc00e001b, Data-
0x0

nearly every second (we send many messages in our applications). I looked at the
error code and it said "The time specified for MQReceiveMessage to wait for the
message elapsed." I expected the respective system queue would be full of
undelivered messages, but there was only _one_ messages there!



What exactly does this error code mean? Can this be a reason (or a symptom
respectively) for hanging up the whole MQ service? (As it seems we had a problem
of that kind, but I can't find any clue of what exactly happened...)



BTW, Windows 2000 latest SP, MQ is used via COM interfaces. Application is
written in C++ 7.0 compiled in VisualStudio .NET.



Thanks in advance for any help!

Martin