You are viewing a plain text version of this content. The canonical link for it is here.

Posted to common-issues@hadoop.apache.org by "Colin Patrick McCabe (JIRA)" <ji...@apache.org> on 2015/02/10 01:18:36 UTC

[jira] [Updated] (HADOOP-10953) NetworkTopology#add calls NetworkTopology#toString without holding the netlock

     [ https://issues.apache.org/jira/browse/HADOOP-10953?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Colin Patrick McCabe updated HADOOP-10953:
------------------------------------------
    Summary: NetworkTopology#add calls NetworkTopology#toString without holding the netlock  (was: a minor concurrent bug inside NetworkTopology)

> NetworkTopology#add calls NetworkTopology#toString without holding the netlock
> ------------------------------------------------------------------------------
>
>                 Key: HADOOP-10953
>                 URL: https://issues.apache.org/jira/browse/HADOOP-10953
>             Project: Hadoop Common
>          Issue Type: Bug
>          Components: net
>    Affects Versions: 3.0.0
>            Reporter: Liang Xie
>            Assignee: Liang Xie
>            Priority: Minor
>         Attachments: HADOOP-10953.txt
>
>
> Found this issue while reading the related code. In NetworkTopology.toString() method, there is no thread safety guarantee directly, it's called by add/remove, and inside add/remove, most of this.toString() calls are protected by rwlock, except a couple of error handling codes, one possible fix is that moving them into lock as well, due to not heavy operations, so no obvious downgration should be observed per my current knowledge.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)