You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Ben Mahler <be...@gmail.com> on 2013/04/09 00:31:20 UTC

Review Request: Fixed a deadlock in the Hadoop MesosScheduler by unsynchronizing resourceOffers().

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/10352/
-----------------------------------------------------------

Review request for mesos, Benjamin Hindman, Vinod Kone, and Brenden Matthews.


Description
-------

See my comments in the diff.

The benefit of this solution is that the only modification needed is to unsynchronize resourceOffers() and pull out the needed JobTracker data prior to synchronizing explicitly. Longer term we will want to consider something more robust against regressions.


This addresses bug MESOS-429.
    https://issues.apache.org/jira/browse/MESOS-429


Diffs
-----

  hadoop/mesos/src/java/org/apache/hadoop/mapred/MesosScheduler.java 50df625c914c158b368b71cf3e32428fbc637ad4 

Diff: https://reviews.apache.org/r/10352/diff/


Testing
-------

Ran the tutorial for both of the supported Hadoop distributions.


Thanks,

Ben Mahler


Re: Review Request: Fixed a deadlock in the Hadoop MesosScheduler by unsynchronizing resourceOffers().

Posted by Brenden Matthews <br...@diddyinc.com>.

> On April 8, 2013, 10:37 p.m., Brenden Matthews wrote:
> > This looks reasonable.  I'll try it out overnight and report back tomorrow.
> 
> Ben Mahler wrote:
>     Much appreciated!

It's been running with your patch for more than 24 hours now without any issues.  The patch gets my approval.


- Brenden


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/10352/#review18802
-----------------------------------------------------------


On April 8, 2013, 11:02 p.m., Ben Mahler wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/10352/
> -----------------------------------------------------------
> 
> (Updated April 8, 2013, 11:02 p.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Vinod Kone.
> 
> 
> Description
> -------
> 
> See my comments in the diff.
> 
> The benefit of this solution is that the only modification needed is to unsynchronize resourceOffers() and pull out the needed JobTracker data prior to synchronizing explicitly. Longer term we will want to consider something more robust against regressions.
> 
> 
> This addresses bug MESOS-429.
>     https://issues.apache.org/jira/browse/MESOS-429
> 
> 
> Diffs
> -----
> 
>   hadoop/mesos/src/java/org/apache/hadoop/mapred/MesosScheduler.java 50df625c914c158b368b71cf3e32428fbc637ad4 
> 
> Diff: https://reviews.apache.org/r/10352/diff/
> 
> 
> Testing
> -------
> 
> Ran the tutorial for both of the supported Hadoop distributions.
> 
> 
> Thanks,
> 
> Ben Mahler
> 
>


Re: Review Request: Fixed a deadlock in the Hadoop MesosScheduler by unsynchronizing resourceOffers().

Posted by Ben Mahler <be...@gmail.com>.

> On April 8, 2013, 10:37 p.m., Brenden Matthews wrote:
> > This looks reasonable.  I'll try it out overnight and report back tomorrow.

Much appreciated!


- Ben


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/10352/#review18802
-----------------------------------------------------------


On April 8, 2013, 11:02 p.m., Ben Mahler wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/10352/
> -----------------------------------------------------------
> 
> (Updated April 8, 2013, 11:02 p.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Vinod Kone.
> 
> 
> Description
> -------
> 
> See my comments in the diff.
> 
> The benefit of this solution is that the only modification needed is to unsynchronize resourceOffers() and pull out the needed JobTracker data prior to synchronizing explicitly. Longer term we will want to consider something more robust against regressions.
> 
> 
> This addresses bug MESOS-429.
>     https://issues.apache.org/jira/browse/MESOS-429
> 
> 
> Diffs
> -----
> 
>   hadoop/mesos/src/java/org/apache/hadoop/mapred/MesosScheduler.java 50df625c914c158b368b71cf3e32428fbc637ad4 
> 
> Diff: https://reviews.apache.org/r/10352/diff/
> 
> 
> Testing
> -------
> 
> Ran the tutorial for both of the supported Hadoop distributions.
> 
> 
> Thanks,
> 
> Ben Mahler
> 
>


Re: Review Request: Fixed a deadlock in the Hadoop MesosScheduler by unsynchronizing resourceOffers().

Posted by Ben Mahler <be...@gmail.com>.

> On April 8, 2013, 10:37 p.m., Brenden Matthews wrote:
> > This looks reasonable.  I'll try it out overnight and report back tomorrow.
> 
> Ben Mahler wrote:
>     Much appreciated!
> 
> Brenden Matthews wrote:
>     It's been running with your patch for more than 24 hours now without any issues.  The patch gets my approval.

Ok great, committing this now.


- Ben


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/10352/#review18802
-----------------------------------------------------------


On April 8, 2013, 11:02 p.m., Ben Mahler wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/10352/
> -----------------------------------------------------------
> 
> (Updated April 8, 2013, 11:02 p.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Vinod Kone.
> 
> 
> Description
> -------
> 
> See my comments in the diff.
> 
> The benefit of this solution is that the only modification needed is to unsynchronize resourceOffers() and pull out the needed JobTracker data prior to synchronizing explicitly. Longer term we will want to consider something more robust against regressions.
> 
> 
> This addresses bug MESOS-429.
>     https://issues.apache.org/jira/browse/MESOS-429
> 
> 
> Diffs
> -----
> 
>   hadoop/mesos/src/java/org/apache/hadoop/mapred/MesosScheduler.java 50df625c914c158b368b71cf3e32428fbc637ad4 
> 
> Diff: https://reviews.apache.org/r/10352/diff/
> 
> 
> Testing
> -------
> 
> Ran the tutorial for both of the supported Hadoop distributions.
> 
> 
> Thanks,
> 
> Ben Mahler
> 
>


Re: Review Request: Fixed a deadlock in the Hadoop MesosScheduler by unsynchronizing resourceOffers().

Posted by Brenden Matthews <br...@diddyinc.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/10352/#review18802
-----------------------------------------------------------


This looks reasonable.  I'll try it out overnight and report back tomorrow.

- Brenden Matthews


On April 8, 2013, 10:31 p.m., Ben Mahler wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/10352/
> -----------------------------------------------------------
> 
> (Updated April 8, 2013, 10:31 p.m.)
> 
> 
> Review request for mesos, Benjamin Hindman, Vinod Kone, and Brenden Matthews.
> 
> 
> Description
> -------
> 
> See my comments in the diff.
> 
> The benefit of this solution is that the only modification needed is to unsynchronize resourceOffers() and pull out the needed JobTracker data prior to synchronizing explicitly. Longer term we will want to consider something more robust against regressions.
> 
> 
> This addresses bug MESOS-429.
>     https://issues.apache.org/jira/browse/MESOS-429
> 
> 
> Diffs
> -----
> 
>   hadoop/mesos/src/java/org/apache/hadoop/mapred/MesosScheduler.java 50df625c914c158b368b71cf3e32428fbc637ad4 
> 
> Diff: https://reviews.apache.org/r/10352/diff/
> 
> 
> Testing
> -------
> 
> Ran the tutorial for both of the supported Hadoop distributions.
> 
> 
> Thanks,
> 
> Ben Mahler
> 
>


Re: Review Request: Fixed a deadlock in the Hadoop MesosScheduler by unsynchronizing resourceOffers().

Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/10352/
-----------------------------------------------------------

(Updated April 8, 2013, 11:02 p.m.)


Review request for mesos, Benjamin Hindman and Vinod Kone.


Changes
-------

Vinod's comment.


Description
-------

See my comments in the diff.

The benefit of this solution is that the only modification needed is to unsynchronize resourceOffers() and pull out the needed JobTracker data prior to synchronizing explicitly. Longer term we will want to consider something more robust against regressions.


This addresses bug MESOS-429.
    https://issues.apache.org/jira/browse/MESOS-429


Diffs (updated)
-----

  hadoop/mesos/src/java/org/apache/hadoop/mapred/MesosScheduler.java 50df625c914c158b368b71cf3e32428fbc637ad4 

Diff: https://reviews.apache.org/r/10352/diff/


Testing
-------

Ran the tutorial for both of the supported Hadoop distributions.


Thanks,

Ben Mahler


Re: Review Request: Fixed a deadlock in the Hadoop MesosScheduler by unsynchronizing resourceOffers().

Posted by Vinod Kone <vi...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/10352/#review18803
-----------------------------------------------------------

Ship it!



hadoop/mesos/src/java/org/apache/hadoop/mapred/MesosScheduler.java
<https://reviews.apache.org/r/10352/#comment39255>

    s/requires/require


- Vinod Kone


On April 8, 2013, 10:31 p.m., Ben Mahler wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/10352/
> -----------------------------------------------------------
> 
> (Updated April 8, 2013, 10:31 p.m.)
> 
> 
> Review request for mesos, Benjamin Hindman, Vinod Kone, and Brenden Matthews.
> 
> 
> Description
> -------
> 
> See my comments in the diff.
> 
> The benefit of this solution is that the only modification needed is to unsynchronize resourceOffers() and pull out the needed JobTracker data prior to synchronizing explicitly. Longer term we will want to consider something more robust against regressions.
> 
> 
> This addresses bug MESOS-429.
>     https://issues.apache.org/jira/browse/MESOS-429
> 
> 
> Diffs
> -----
> 
>   hadoop/mesos/src/java/org/apache/hadoop/mapred/MesosScheduler.java 50df625c914c158b368b71cf3e32428fbc637ad4 
> 
> Diff: https://reviews.apache.org/r/10352/diff/
> 
> 
> Testing
> -------
> 
> Ran the tutorial for both of the supported Hadoop distributions.
> 
> 
> Thanks,
> 
> Ben Mahler
> 
>