[
https://issues.apache.org/jira/browse/ZOOKEEPER-1892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13932824#comment-13932824
]
Dutch T. Meyer commented on ZOOKEEPER-1892:
-------------------------------------------
I opened what I think is the same issue a while back as ZOOKEEPER-1856
(https://issues.apache.org/jira/browse/ZOOKEEPER-1856.)
Since you've got a patch here, we should probably resolve that one as a
duplicate.
> addrvec_next gets called twice when failing over to the next server
> -------------------------------------------------------------------
>
> Key: ZOOKEEPER-1892
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-1892
> Project: ZooKeeper
> Issue Type: Bug
> Components: c client
> Reporter: Michi Mutsuzaki
> Fix For: 3.5.0
>
> Attachments: ZOOKEEPER-1892.patch
>
>
> zookeeper_interest() already calls zoo_cycle_next_server() when the socket is
> set to -1, so we shouldn't call addrvec_next in handle_error. This causes the
> next server to get skipped. Zookeeper_simpleSystem::testFirstServerDown fails
> unless the client gets connected to the server during the first round because
> the client keeps skipping the second server after the first round.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
