Hi Folks I am debugging an issue where the c client library takes a while to reconnect to Zookeeper server (version 3.5.1-alpha), any help is understanding what the problem is highly appreciated.
I tried 2 cases: 1) The server is restarted with no change in server configuration. The client enters connecting state when the socket to zookeeper server closes and within a few seconds enter connected state (reconnects within 5 seconds) 2) The server is restarted with some config changes (we go from 2 zookeeper server to 1) The client enters connecting state when the socket to zookeeper server closes but enters connected state after a while ( I have seen up to 2/3*session_timeout). The session timeout I use is 180 seconds. I am not able to understand why the client takes a long time to enter connected state in case 2. Is there a way to nudge zookeeper client library to reconnect faster? Regards, Pramod https://zookeeper.apache.org/doc/r3.5.1-alpha/zookeeperProgrammers.html When a client (session) becomes partitioned from the ZK serving cluster it will begin searching the list of servers that were specified during session creation. Eventually, when connectivity between the client and at least one of the servers is re-established, the session will either again transition to the "connected" state (if reconnected within the session timeout value) or it will transition to the "expired" state (if reconnected after the session timeout). It is not advisable to create a new session object (a new ZooKeeper.class or zookeeper handle in the c binding) for disconnection. The ZK client library will handle reconnect for you. In particular we have heuristics built into the client library to handle things like "herd effect", etc... Only create a new session when you are notified of session expiration (mandatory).
