hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tom White (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HADOOP-9212) Potential deadlock in FileSystem.Cache/IPC/UGI
Date Tue, 15 Jan 2013 10:40:13 GMT

     [ https://issues.apache.org/jira/browse/HADOOP-9212?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Tom White updated HADOOP-9212:

    Attachment: 1_jcarder_result_0.png

The scenario that leads to the potential deadlock:
* FS.Cache.closeAll(), which is holding the FS.Cache lock, calls DFS's close method which
calls close on the RPC proxy, which eventually calls ipc.Client.stop() and takes lock on Hashtable
of connections
* ipc.Client.getConnection(), which is holding lock on Hashtable of connections, takes lock
on UGI's class during some UGI setup that trigger's UGI.ensureInitialized()
* UGI.getCurrentUser(), which is holding UGI class lock, calls getLoginUser() which calls
Credentials.readTokenStorageFile, which uses FileSystem, taking a lock on FileSystem.Cache

> Potential deadlock in FileSystem.Cache/IPC/UGI
> ----------------------------------------------
>                 Key: HADOOP-9212
>                 URL: https://issues.apache.org/jira/browse/HADOOP-9212
>             Project: Hadoop Common
>          Issue Type: Bug
>          Components: fs
>    Affects Versions: 2.0.2-alpha
>            Reporter: Tom White
>            Assignee: Tom White
>         Attachments: 1_jcarder_result_0.png
> jcarder found a cycle which could lead to a potential deadlock.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

View raw message