szkoludasebastian commented on issue #23908: URL: https://github.com/apache/pulsar/issues/23908#issuecomment-2710416567
> [compose.yaml.txt](https://github.com/user-attachments/files/18877610/compose.yaml.txt) > > I tried to reproduce this issue by the above test setting, but I couldn't. Can you help to reproduce this issue by modifying the above setup script? > > Also, for this repro step: > > ``` > 1. Create partitioned topic A with 100 partitions > 2. Connect 4 consumers to topic A (Subscription type does not matter, we noticed this problem for shared and key_shared subscriptions) > 3. Start sending data to topic A with a throughput of about 5,000 messages per second > 4. Restart bookies (in our case we had like 6 bookies and 6 brokers) - we restarted it by killing k8s pods > 5. After all bookies restarted and are up and running wait for 30 seconds and stop sending data > 6. Check pulsar admin topic stats that backlog is only growing. We were observing backlog for every partition in grafana so we knew for which partition check stats-internal. In our stats-internal we usually see such situation: > ``` > > it appears that the Step 5 kills the producer. After step 5, how do we see backlog increase without producer? We use pulsar library for java so I think that producer is automatically recreated or reconnected. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@pulsar.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
