Lanayx opened a new issue #6154: Brokers constantly crash in the k8s cluster URL: https://github.com/apache/pulsar/issues/6154 **Describe the bug** Brokers constantly failing in a cluster with a _CrashLoopBackOff_ . They work for several hours and then fail. Tried installing clean cluster 3 days in a raw and the behavior persisted. **To Reproduce** Steps to reproduce the behavior: 1. Use default helm chart values-mini.yaml on GKE 3 nodes 2. Use following values: ``` # Pulsar proxy proxy: yes # Bookkeeper auto-recovery autoRecovery: no # Pulsar dashboard # Deprecated # Replace pulsar-dashboard with pulsar-manager dashboard: yes # pulsar manager pulsar_manager: no # Bastion pod for administrative commands bastion: yes # Monitoring stack (prometheus and grafana) monitoring: no ``` 3. Wait for some time (4+ hours) 4. See error **Expected behavior** Brokers should fail in the CrashLoopBackOff **Screenshots**  **Additional context** Adding logs from 2 failing brokers and 3 bookies. Also adding last logs from one of the evicted pods. [bookeeper-1.log](https://github.com/apache/pulsar/files/4121328/bookeeper-1.log) [bookeeper-2.log](https://github.com/apache/pulsar/files/4121329/bookeeper-2.log) [bookeeper-3.log](https://github.com/apache/pulsar/files/4121331/bookeeper-3.log) [broker-1.log](https://github.com/apache/pulsar/files/4121332/broker-1.log) [broker-2.log](https://github.com/apache/pulsar/files/4121333/broker-2.log) [evicted_broker.txt](https://github.com/apache/pulsar/files/4121424/evicted_broker.txt)
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
