Hi Matt, Thanks for your reply ! Sure, I will use 5.18.1 version but I am not getting any exception like timeout for hung consumer in the consumer app code. Is prefetch limit also play role here ? it is default as of now. Should we try to reduce this limit as well ? Any suggestion/help is much appreciated. Thanks a lot in Advance !
Mit freundlichen Grüßen Gorav Singh externe Fachkraft von der Firma Coforge Limited im Auftrag der DB Systel GmbH SEZ Developer Unit Plot No. TZ-2 & 2A, Sector Tech Zone, Greater Noida, UP 201308 E-Mail: gorav.singh-ext...@deutschebahn.com -----Original Message----- From: Matt Pavlovich <mattr...@gmail.com> Sent: Friday, June 30, 2023 8:10 PM To: users@activemq.apache.org Subject: Re: Messages got stuck in ActiveMQ 5.18.0 until restart of server [You don't often get email from mattr...@gmail.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ] Hi Gorav- For sure, you'll need to use the activemq-client-jakarta 5.18.1 jar. The v5.18.0 activemq-client-jakarta had a minor packaging bug that caused it not to work at all, so I'm not sure how that client would be working in a happy path. Also, the PooledConnectionFactory does not have jakarta support yet (jakarta support is coming in v5.19.x), so there may be a couple of structural issues going on within the app to start. The scenario you are describing aligns most closely to an app coding bug where the application consumer is hung, and not acknowledging the messages. The broker dispatched messages to that consumer, so it is not passing them to another consumer. Restarting the POD causes the Consumer to go away, and that allows the broker to dispatch those messages to another consumer. Look for the queue metric 'InFlightCount'. If that value is about the number of messages not dequeuing, then most likely the issue is in the consumer app code. Thanks, Matt Pavlovich > On Jun 30, 2023, at 2:02 AM, Gorav Singh -Extern > <gorav.singh-ext...@deutschebahn.com> wrote: > > Hi Team, > > We are using activemq-client-jakarta 5.18.0, activemq-broker 5.18.0 with > spring boot 3.0 and jdk 17. We are consuming messages using @JmsListener > initially having 1-5 concurrency. > We are consuming mostly 2500-3000 messages from the queue as the part of some > campaigns. > > > Now it works fine most of the days but once or twice in a week, we found that > some of the messages got stuck in the queue which are not consumed but > messages which pushed later got consumed successfully. > We have to restart the pod every time in such cases to receive stuck messages. > > We tried some solution like not to use PooledConnectionFactory for consumers, > also we changed concurrency setting to 5-5 so that consumer could not be > killed but nothing worked. > > Could you please help me to find out the solution or is it a known > ActiveMQ issue ? (please note that in JMS configuration, we are using > ActiveMQConnectionFactory, DefaultJmsListenerContainerFactory, with > JmsTemplate) > > > Mit freundlichen Grüßen > > Gorav Singh > > externe Fachkraft von der Firma Coforge Limited im Auftrag der DB > Systel GmbH SEZ Developer Unit Plot No. TZ-2 & 2A, Sector Tech Zone, > Greater Noida, UP 201308 > E-Mail: > gorav.singh-ext...@deutschebahn.com<mailto:gorav.singh-extern@deutsche > bahn.com> > > > ________________________________ > > Pflichtangaben > anzeigen<https://www.deutschebahn.com/pflichtangaben/20230607> > > Nähere Informationen zur Datenverarbeitung im DB-Konzern finden Sie > hier: https://www.deutschebahn.com/de/konzern/datenschutz ________________________________ Pflichtangaben anzeigen<https://www.deutschebahn.com/pflichtangaben/20230607> Nähere Informationen zur Datenverarbeitung im DB-Konzern finden Sie hier: https://www.deutschebahn.com/de/konzern/datenschutz
