zookeeper-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "gopalakrishna (JIRA)" <j...@apache.org>
Subject [jira] [Created] (ZOOKEEPER-2702) zookeeper ensemble took 20 minutes to come back up after leader failed
Date Wed, 22 Feb 2017 12:42:44 GMT
gopalakrishna created ZOOKEEPER-2702:
----------------------------------------

             Summary: zookeeper ensemble took 20 minutes to come back up after leader failed
                 Key: ZOOKEEPER-2702
                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2702
             Project: ZooKeeper
          Issue Type: Bug
    Affects Versions: 3.4.9
         Environment: OS version is ubuntu 14.04(trusty)
            Reporter: gopalakrishna


Zookeeper version : 3.4.9

OS version is ubuntu 14.04(trusty)

Default configuration of zoo.cfg 
tickTime=2000
initLimit=10
syncLimit=5

I have setup the zookeeper ensemble with three servers zk1.com, zk2.com, zk3.com.

Initial State:

ZK1(FOLLOWER)---ZK2(LEADER)-------ZK3(FOLLOWER)


This morning, ZK2(LEADER) went down and it became a FOLLOWER with in fraction of seconds.
It took 20 minutes for new LEADER to be decided for the ensemble. ZK3 was the new LEADER.

New State:
ZK(FOLLOWER)----ZK2(FOLLOWER)-----ZK3(LEADER) (after 20 minutes).


Can somone help me to debug what happened? 

Zookeeper is managing the solr cloud 2shards, 4 nodes. 



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message