[
https://issues.apache.org/jira/browse/ZOOKEEPER-2498?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Michael Han updated ZOOKEEPER-2498:
-----------------------------------
Attachment: ZOOKEEPER-2498.patch
> Potential resource leak in C client when processing unexpected / out of order
> response
> --------------------------------------------------------------------------------------
>
> Key: ZOOKEEPER-2498
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2498
> Project: ZooKeeper
> Issue Type: Bug
> Components: c client
> Affects Versions: 3.4.8, 3.5.2
> Reporter: Michael Han
> Assignee: Michael Han
> Fix For: 3.4.9, 3.5.3
>
> Attachments: ZOOKEEPER-2498.patch
>
>
> In C client, we use reference counting to decide if a given zh handle can be
> destroyed or not. This requires we always make sure to call api_prolog (which
> increment the counter) and api_epilog (which decrease the counter) in pairs,
> for a given call context.
> In zookeeper_process, there is a place where the code will return without
> invoking api_epilog, which would lead to potential zh resource leak.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
