Sounds like a bug. Please go ahead and open a Jira and include steps, code, configuration, etc. to reproduce the issue as simply as possible (i.e. a minimal reproducible example). Thanks!
Justin On Thu, Jun 9, 2022 at 11:13 AM Max Sidnin <msid...@gmail.com> wrote: > Hi, > It seems there is a bug related to "default-delay-before-dispatch" and > "default-consumers-before-dispatch". The scenario is: > > 1. Configure some queue via address-setting: > <address-setting match="*.something"> > > <default-consumers-before-dispatch>50</default-consumers-before-dispatch> > <default-delay-before-dispatch>5000</default-delay-before-dispatch> > > > <default-group-first-key>JMSXGroupFirstForConsumer</default-group-first-key> > <default-group-rebalance>true</default-group-rebalance> > > > <default-group-rebalance-pause-dispatch>true</default-group-rebalance-pause-dispatch> > <default-group-buckets>512</default-group-buckets> > <page-size-bytes>52428800</page-size-bytes> > <default-consumer-window-size>0</default-consumer-window-size> > </address-setting> > > 2. Run a producer, post 100-200 messages and disconnect. There should be no > active producers. > > 3. After some delay (~10 secs) start 10 consumers (less than our > "default-consumers-before-dispatch"). > > Expected: after 5 seconds consumers should start to consume messages > Actual: nothing happens until something touches the queue. For example: a > new message arrives, redistribution happens, the "resume()" operation is > called. After these actions everything wakes up and starts to work. > > There is no issue if a producer is alive and a queue is active. Artemis > version is 2.22.0 > > Should I report a bug? >
