[BOOKKEEPER-116] Bookkeeper server cannot re-connect to the zookeeper until zookeeper recognizes

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Component/s: hedwig-client

Environment

CentOS 5.5 for all servers and workstations (however zookeeper, bookies, and hubs are all built in Ubuntu 11);
OpenJDK Runtime Environment (IcedTea6 1.9.10) (rhel-1.23.1.9.10.el5_7-i386);
OpenJDK Client VM (build 19.0-b09, mixed mode);

• Assignee: Unassigned
• Reporter: Daniel Kim
• Created: 22/Nov/11 21:48
• Updated: 11/Sep/13 04:28
• Resolved: 11/Sep/13 04:28

Description

When a bookie is disconnected from the zk server, zk does not know about this immediately. As a result, if I try to restart the bookie, meanwhile, it will not be able to. Zk server believes that the server trying to reconnect already has a connection. Instead of throwing KeeperException for existing node, zk server should wait until the pre-existing connection times out and tries to reconnect. Or zk server should ignore the old metadata about this node, and re-connect immediately (but I guess this solution may cause security issues).

Issue Links

• duplicates: BOOKKEEPER-294 Not able to start the bookkeeper before the ZK session timeout.