Hi all, We're experiencing an issue with Flink 1....
# troubleshooting
a
Hi all, We're experiencing an issue with Flink 1.20 where jobs are failing due to ZooKeeper connection problems. Here are the key details:
Copy code
- ZooKeeper configuration:
  - high-availability.zookeeper.client.session-timeout: 60s (default)
  - high-availability.zookeeper.client.tolerate-suspended-connections: true
Symptoms: - ZooKeeper connection state changes from SUSPENDED to LOST - JobMaster shuts down and stops leader election Whenever the JobManager reconnects to ZK, its not bale to recover. I am assuming although there was a LOST zk connection, it will attempt to restart the jobmanager. We are manually restarting the JobManager to recover it. is this the right way for recovering ? Any insights or suggestions would be greatly appreciated.