[
https://issues.apache.org/jira/browse/ZOOKEEPER-2556?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15607790#comment-15607790
]
Alexander Shraer commented on ZOOKEEPER-2556:
---------------------------------------------
looks good, but I think the check that peerType is not in cfg shouldn't be in
the inner loop.
> peerType remains as "observer" in zoo.cfg even though we change the node from
> observer to participant runtime
> -------------------------------------------------------------------------------------------------------------
>
> Key: ZOOKEEPER-2556
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2556
> Project: ZooKeeper
> Issue Type: Bug
> Components: server
> Affects Versions: 3.5.1, 3.5.2
> Reporter: Rakesh Kumar Singh
> Assignee: Rakesh Kumar Singh
> Priority: Minor
> Attachments: ZOOKEEPER-2556.patch, ZOOKEEPER-2556.patch
>
>
> peerType remains as "observer" in zoo.cfg even though we change the node from
> observer to participant runtime
> Steps to reproduce:-
> 1. Start zookeeper in cluster with one node as observer by configuring
> peerType=observer in zoo.cfg and server.2=10.18.219.50:2888:3888:observer;2181
> 2. Start the cluster
> 3. start a client and change the node from observer to participant, the
> configuration related to peertype remained same though other things like
> clientport got from zoo.cfg
> >reconfig -remove 2 -add 2=10.18.219.50:2888:3888:participant;2181
> We should either remove this parameter or update with correct node type at
> run time
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
