[ZOOKEEPER-1145] ObserverTest.testObserver fails at particular point after several runs of ant junt.run -Dtestcase=ObserverTest

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Duplicate
• Affects Version/s: 3.4.0
• Fix Version/s: 3.4.0
• Assignee: Vishal Kher
• Reporter: Eugene Joseph Koontz
• Created: 03/Aug/11 22:36
• Updated: 23/Nov/11 19:22
• Resolved: 15/Aug/11 01:02

Description

Use the attached repeat.sh to run ObserverTest repeatedly by doing:

src/repeat.sh ObserverTest

The test will will fail eventually after a few iterations; should be only a few minutes.

The line that fails in the test is:

zk = new ZooKeeper("127.0.0.1:" + CLIENT_PORT_OBS,
ClientBase.CONNECTION_TIMEOUT, this);

Attached as out.txt is the output showing a successful run, for comparison, followed by a failed run.

Note that in the seconds before the test fails, in the following lines, that there is a 24 second gap in time (between 22:13:02 and 22:13:26):

[junit] 2011-08-03 22:13:02,167 [myid:3] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:11229:ZooKeeperServer@833] - Client attempting to establish new session at /127.0.0.1:46929
[junit] 2011-08-03 22:13:26,003 [myid:2] - INFO [QuorumPeer[myid=2]/0:0:0:0:0:0:0:0:11228:Leader@419] - Shutting down
[junit] 2011-08-03 22:13:26,003 [myid:2] - INFO [QuorumPeer[myid=2]/0:0:0:0:0:0:0:0:11228:Leader@425] - Shutdown called
[junit] java.lang.Exception: shutdown Leader! reason: Only 0 followers, need 1

Issue Links

• relates to: ZOOKEEPER-1144 ZooKeeperServer not starting on leader due to a race condition