[
https://issues.apache.org/jira/browse/KAFKA-5049?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15966978#comment-15966978
]
Jun Rao commented on KAFKA-5049:
--------------------------------
[~ijuma], yes, that does seem to be a problem,
[~anukin], if you want to work on thi, you just need to create a new instance
of ZkPath by passing in an instance of ZkClient and store it in ZkUtils. Then,
in ZkUtils, use the stored ZkPath instance instead of the singleton.
> Chroot check should be done for each ZkUtils instance
> -----------------------------------------------------
>
> Key: KAFKA-5049
> URL: https://issues.apache.org/jira/browse/KAFKA-5049
> Project: Kafka
> Issue Type: Bug
> Reporter: Ismael Juma
> Labels: newbie
> Fix For: 0.11.0.0
>
>
> In KAFKA-1994, the check for ZK chroot was moved to ZkPath. However, ZkPath
> is a JVM singleton and we may use multiple ZkClient instances with multiple
> ZooKeeper ensembles in the same JVM (for cluster info, authorizer and
> pluggable code provided by users).
> The right way to do this is to make ZkPath an instance variable in ZkUtils so
> that we do the check once per ZkUtils instance.
> cc [~gwenshap] [~junrao], who reviewed KAFKA-1994, in case I am missing
> something.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
