Artemis PullMessage issues

[Svetlana Undalova <un...@gmail.com>]

Hi,

I'm trying to reach you as I ran into an issue trying to use Apache Artemis
(2.16).
I have slow consumers and went with prefetch size = 0 which according to
docs results in pulling messages by the client. All works well if consumer
starts after messages are already in the queue, but if it becomes idle
after a while and manages to call consumer.Receive() N times then next time
when messages appear in the queue I can see that all N are sent to the
client (via checking UnconsumedMessageCount)

I checked the same set up with ActiveMQ (5.16) and there
both scenarios work correctly. I think that issue may happen because
Artemis broker ignores PullMessage timeout.

I can provide you with a simple test to reproduce the issue.

Please let me know what is the best way to reach you (maybe via slack)

Thanks,
Svetlana

Re: Artemis PullMessage issues

[Justin Bertram <jb...@apache.org>]

This mailing list is probably the best way to reach the community. Please
open a Jira [1] and attach your test to reproduce the issue. Thanks!


Justin

[1] https://issues.apache.org/jira/browse/ARTEMIS

On Mon, Feb 8, 2021 at 7:20 AM Svetlana Undalova <un...@gmail.com>
wrote:

> Hi,
>
> I'm trying to reach you as I ran into an issue trying to use Apache Artemis
> (2.16).
> I have slow consumers and went with prefetch size = 0 which according to
> docs results in pulling messages by the client. All works well if consumer
> starts after messages are already in the queue, but if it becomes idle
> after a while and manages to call consumer.Receive() N times then next time
> when messages appear in the queue I can see that all N are sent to the
> client (via checking UnconsumedMessageCount)
>
> I checked the same set up with ActiveMQ (5.16) and there
> both scenarios work correctly. I think that issue may happen because
> Artemis broker ignores PullMessage timeout.
>
> I can provide you with a simple test to reproduce the issue.
>
> Please let me know what is the best way to reach you (maybe via slack)
>
> Thanks,
> Svetlana
>