You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Jie Yu <yu...@gmail.com> on 2012/07/17 03:48:42 UTC
Review Request: Fixed a bug in SocketManager::next(int s) that will cause an
assertion failure in slave.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/5995/
-----------------------------------------------------------
Review request for mesos, Benjamin Hindman and Vinod Kone.
Description
-------
The bug is reported at here:
https://issues.apache.org/jira/browse/MESOS-220
The cause of this bug is that we cannot assume the Socket handle s exists in SocketManager::next. Because it is likely that s has already been closed by socket_manager->close(s). For example, process::recv_data will get called asynchronously if the remotely linked process exits. In that case, since the physical socket is broken, process::recv_data will remove the Socket handle from sockets in SocketManager.
Diffs
-----
third_party/libprocess/src/process.cpp b560b7c
Diff: https://reviews.apache.org/r/5995/diff/
Testing
-------
make check.
Thanks,
Jie Yu
Re: Review Request: Fixed a bug in SocketManager::next(int s) that will
cause an assertion failure in slave.
Posted by Benjamin Hindman <be...@berkeley.edu>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/5995/#review9307
-----------------------------------------------------------
Ship it!
Ship It!
- Benjamin Hindman
On July 17, 2012, 1:48 a.m., Jie Yu wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/5995/
> -----------------------------------------------------------
>
> (Updated July 17, 2012, 1:48 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Vinod Kone.
>
>
> Description
> -------
>
> The bug is reported at here:
> https://issues.apache.org/jira/browse/MESOS-220
>
> The cause of this bug is that we cannot assume the Socket handle s exists in SocketManager::next. Because it is likely that s has already been closed by socket_manager->close(s). For example, process::recv_data will get called asynchronously if the remotely linked process exits. In that case, since the physical socket is broken, process::recv_data will remove the Socket handle from sockets in SocketManager.
>
>
> Diffs
> -----
>
> third_party/libprocess/src/process.cpp b560b7c
>
> Diff: https://reviews.apache.org/r/5995/diff/
>
>
> Testing
> -------
>
> make check.
>
>
> Thanks,
>
> Jie Yu
>
>