zookeeper-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Norbert Kalmar (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ZOOKEEPER-3445) Concurrency issue in ReferenceCountedACLCache
Date Fri, 28 Jun 2019 10:20:00 GMT

    [ https://issues.apache.org/jira/browse/ZOOKEEPER-3445?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16874833#comment-16874833

Norbert Kalmar commented on ZOOKEEPER-3445:

Thanks for reporting this.

As discussed on the PR, aclIndex is guarded by a lock. Only synchronized blocks access the

> Concurrency issue in ReferenceCountedACLCache
> ---------------------------------------------
>                 Key: ZOOKEEPER-3445
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3445
>             Project: ZooKeeper
>          Issue Type: Bug
>          Components: server
>    Affects Versions: 3.5.5, 3.4.14
>            Reporter: Jonathan Halterman
>            Priority: Critical
>              Labels: pull-request-available
>          Time Spent: 1h 40m
>  Remaining Estimate: 0h
> While debugging some unexpected "ACL not available for long" exceptions we were seeing,
I noticed that ReferenceCountedACLCache does not mark aclIndex as volatile, which it should
since it appears to be read from multiple threads. This may or may not be the cause of the
behavior we're seeing, but should be fixed regardless.

This message was sent by Atlassian JIRA

View raw message