I'm feeling so stupid now. Indeed it works perfectly with the FQQN (which I should have tried as this is how our JAVA client accesses the queue).
On Sun, Apr 7, 2024 at 6:43 PM Justin Bertram <jbert...@apache.org> wrote: > Can you share the relevant code for each scenario? > > Have you tried using the fully qualified queue name (i.e. "ADDRESS::QUEUE") > [1]? > > > Justin > > [1] > > https://activemq.apache.org/components/artemis/documentation/latest/address-model.html#fully-qualified-queue-names > > On Fri, Apr 5, 2024 at 10:54 AM Florent TASSIN <tass...@gmail.com> wrote: > > > In the broker.xml, I have disabled auto-creation of addresses and > queues. I > > also have created an address (ADDRESS) and in the <multicast> a queue > > (QUEUE). > > <addresses> > > <address name="ADDRESS"> > > <multicast> > > <queue name="QUEUE" filter=""/> > > </multicast> > > </address> > > </addresses> > > > > This creates a managed-subscription. > > Now, I want my client to subscribe to this managed subscription using the > > qpid proton python api. > > > > However: > > - If I specify the queue name when calling create_receiver and the same > > queue name when specifying the container_id, it creates a queue named > > "QUEUE.QUEUE". > > - If I don't specify a container_id, it generates a queue named > > [some_uid].QUEUE > > - And if, on the other hand, I don't specify a name or an empty one when > > calling create_receiver, it generates a queue named "QUEUE.QUEUE-ADDRESS" > > > > In any case, I'm not able to subscribe to the existing queue (QUEUE) > > > > If someone has an idea, it will be deeply appreciated. > > > > Rgds, > > F. > > >