Details
-
Bug
-
Status: Open
-
Major
-
Resolution: Unresolved
-
2.0.1
-
None
-
None
Description
After a network issue caused the last ISR to fail connecting to ZooKeeper, the attempt to reconnect failed with an ArrayIndexOutOfBoundsException.
2019-05-31 15:54:38,823 [zk-session-expiry-handler0-SendThread(zk2-1:2181)] WARN (org.apache.zookeeper.ClientCnxn) - Client session timed out, have not heard from server in 20010ms for sessionid 0x15000011b2b1042a 2019-05-31 15:54:38,823 [zk-session-expiry-handler0-SendThread(zk2-1:2181)] INFO (org.apache.zookeeper.ClientCnxn) - Client session timed out, have not heard from server in 20010ms for sessionid 0x15000011b2b1042a, closing socket connection and attempting reconnect 2019-05-31 15:54:39,702 [zk-session-expiry-handler0-SendThread(zk1-2:2181)] INFO (org.apache.zookeeper.client.ZooKeeperSaslClient) - Client will use DIGEST-MD5 as SASL mechanism. 2019-05-31 15:54:39,702 [zk-session-expiry-handler0-SendThread(zk1-2:2181)] ERROR (org.apache.zookeeper.client.ZooKeeperSaslClient) - Exception while trying to create SASL client: java.lang.ArrayIndexOutOfBoundsException: 0 2019-05-31 15:54:39,702 [zk-session-expiry-handler0-SendThread(zk1-2:2181)] INFO (org.apache.zookeeper.ClientCnxn) - Opening socket connection to server zk1-2/1.3.6.1:2181. Will attempt to SASL-authenticate using Login Context section 'Client' 2019-05-31 15:54:39,702 [zk-session-expiry-handler0-SendThread(zk1-2:2181)] INFO (org.apache.zookeeper.ClientCnxn) - Socket connection established to zk1-2/1.3.6.1:2181, initiating session 2019-05-31 15:54:39,703 [zk-session-expiry-handler0-SendThread(zk1-2:2181)] INFO (org.apache.zookeeper.ClientCnxn) - Session establishment complete on server zk1-2/1.3.6.1:2181, sessionid = 0x15000011b2b1042a, negotiated timeout = 30000 2019-05-31 15:54:39,703 [zk-session-expiry-handler0-SendThread(zk1-2:2181)] ERROR (org.apache.zookeeper.ClientCnxn) - SASL authentication with Zookeeper Quorum member failed: javax.security.sasl.SaslException: saslClient failed to initialize properly: it's null.
Kafka was "not live" in zookeeper and had to be manually restarted to recover from this error. It would be better if the last ISR could retry.
Attachments
Issue Links
- duplicates
-
KAFKA-7987 a broker's ZK session may die on transient auth failure
- Resolved