A reproducer is the best way for us to help diagnose/fix the issue. Failing that, are you able to profile the broker to see which objects are getting leaked?
There are a couple of MQTT specific things to check such as retained messages and session state. The life cycle of these is controlled by the client and will they essentially hang around forever unless the client clears them. There are a coupe of other gotchas that might cause the broker to start running out of memory (providing you haven't configured a PAGE/BLOCK policy) i.e. having a slow consumer with an address with lots of traffic e.g."/" If you can't reproduce it in a test, could you provide more info on the use case and include your broker config. Thanks On Mon, Jun 5, 2017 at 2:18 AM, Clebert Suconic <clebert.suco...@gmail.com> wrote: > Look especially for how you ack your messages, how you subscribe.. if > you're not leaking consumers... etc... > > On Sun, Jun 4, 2017 at 4:15 PM, hwaastad <he...@waastad.org> wrote: > > Hi, > > i'll see if i can produce some kind of test of the scenario i'm using. > > > > /hw > > > > Sendt fra min iPhone > > > >> Den 4. jun. 2017 kl. 15.37 skrev clebertsuconic [via ActiveMQ] < > ml+s2283324n4726996...@n4.nabble.com>: > >> > >> Any way you could replicate it on an example? > >> > >> Or test? > >> > >> > >> > >> On Sun, Jun 4, 2017 at 9:38 AM hwaastad <[hidden email]> wrote: > >> > >> > Hi, > >> > ran 2.2.0-snapshot from yesterday until today. > >> > > >> > To see if I could get the issue earlier I added: -Xms1024M -Xmx1024M. > >> > > >> > So after ~24hours, the memory leak happens again. and disk is full. > >> > > >> > /hw > >> > > >> > > >> > > >> > -- > >> > View this message in context: > >> > http://activemq.2283324.n4.nabble.com/artemis-2-1-0-logging-disaster- > tp4726922p4726995.html > >> > Sent from the ActiveMQ - User mailing list archive at Nabble.com. > >> > > >> -- > >> Clebert Suconic > >> > >> > >> If you reply to this email, your message will be added to the > discussion below: > >> http://activemq.2283324.n4.nabble.com/artemis-2-1-0-logging-disaster- > tp4726922p4726996.html > >> To unsubscribe from artemis 2.1.0 logging disaster, click here. > >> NAML > > > > > > > > > > -- > > View this message in context: http://activemq.2283324.n4. > nabble.com/artemis-2-1-0-logging-disaster-tp4726922p4727000.html > > Sent from the ActiveMQ - User mailing list archive at Nabble.com. > > > > -- > Clebert Suconic >
