giraph-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chuan Lei (JIRA)" <>
Subject [jira] [Commented] (GIRAPH-730) GiraphApplicationMaster race condition in resource loading
Date Fri, 23 Aug 2013 23:39:51 GMT


Chuan Lei commented on GIRAPH-730:

Hi Eli,

Thanks for your feedbacks. There could be two problems when 2 threads call the getTaskResourceMap

1. Two threads find that the LOCAL_RESOURCES is null, thus both of them try to initialize/compose

2. One thread (T1) finds that the LOCAL_RESOURCES is null, and starts initializing it. The
other thread (T2) finds the LOCAL_RESOURCES is not null (being initialized), and then directly
returns the LOCAL_RESOURCES, which has not been completely initialized by T1 yet.

Both cases above may cause the inconsistent state.

Synchronizing the block is fine. I will change it and resubmit a path shortly.

> GiraphApplicationMaster race condition in resource loading
> ----------------------------------------------------------
>                 Key: GIRAPH-730
>                 URL:
>             Project: Giraph
>          Issue Type: Bug
>    Affects Versions: 1.0.0
>         Environment: Giraph with Yarn
>            Reporter: Chuan Lei
>            Assignee: Chuan Lei
>         Attachments: GIRAPH-730.v1.patch
> In, getTaskResourceMap function is not multi-thread safe,
which causes the application master fail to distribute the resources (jar, configuration file,
etc.) to each container.

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:

View raw message