[
https://issues.apache.org/jira/browse/ZOOKEEPER-1219?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rakesh R updated ZOOKEEPER-1219:
--------------------------------
Assignee: (was: Rakesh R)
> LeaderElectionSupport recipe is unnecessarily dispatching the READY_START
> event even if the ELECTED node stopped/expired simultaneously.
> ----------------------------------------------------------------------------------------------------------------------------------------
>
> Key: ZOOKEEPER-1219
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-1219
> Project: ZooKeeper
> Issue Type: Improvement
> Components: recipes
> Affects Versions: 3.3.3
> Reporter: Rakesh R
> Fix For: 3.5.0
>
> Attachments: ZOOKEEPER-1219.patch
>
>
> Let's say node has determined as READY and has dispatched DETERMINE_COMPLETE
> event, at the same time the ELECTED node got stopped or expired . Still the
> f/w first dispatches the READY_START event to the node and then checks
> whether the ELECTED node exists() or not. Here it finds there is no 'Stat'
> corresponding to ELECTED and will again goes to leader determination phase.
> *Problem:*
> Unnecessarily the READY_START event is dispatching to the node and says node
> to be ready with the startup/init, even if there is no ELECTED node.
> *Proposal*
> Reverse the logic, first check whether ELECTED node exists() or not and then
> if success f/w can dispatch the READY_START event. Otherwise go to the leader
> determination phase.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
