You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@mesos.apache.org by Gaston Kleiman <ga...@mesosphere.io> on 2017/10/27 07:48:46 UTC

Review Request 63353: Don't clear the executor ID of non-command executors on re-registration.

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/
-----------------------------------------------------------

Review request for mesos, Benjamin Mahler and Greg Mann.


Bugs: MESOS-8135
    https://issues.apache.org/jira/browse/MESOS-8135


Repository: mesos


Description
-------

Previously the agent would sometimes clear the executor ID of
non-command executors before sending the `ReregisterSlaveMessage`
message.


Diffs
-----

  src/slave/slave.cpp d8477b4e364270f5ad61311aa8fe48f823bc7aac 
  src/tests/master_slave_reconciliation_tests.cpp d5eb7ba68c5308338236879e7cb1e970a01e48e6 


Diff: https://reviews.apache.org/r/63353/diff/1/


Testing
-------

Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.


Thanks,

Gaston Kleiman


Re: Review Request 63353: Don't clear the executor ID of non-command executors on re-registration.

Posted by Benjamin Mahler <bm...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/#review189861
-----------------------------------------------------------



Per offline discussion, looks like we don't actually need to be clearing it?

- Benjamin Mahler


On Nov. 1, 2017, 8:53 p.m., Gaston Kleiman wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/63353/
> -----------------------------------------------------------
> 
> (Updated Nov. 1, 2017, 8:53 p.m.)
> 
> 
> Review request for mesos, Benjamin Mahler, Greg Mann, and Vinod Kone.
> 
> 
> Bugs: MESOS-8135
>     https://issues.apache.org/jira/browse/MESOS-8135
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Previously the agent would sometimes clear the executor ID of
> non-command executors before sending the `ReregisterSlaveMessage`
> message.
> 
> 
> Diffs
> -----
> 
>   src/slave/slave.cpp 337083dbe60bba2d3773b785bdceeaf0b8fcd070 
> 
> 
> Diff: https://reviews.apache.org/r/63353/diff/2/
> 
> 
> Testing
> -------
> 
> Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.
> 
> 
> Thanks,
> 
> Gaston Kleiman
> 
>


Re: Review Request 63353: Don't clear the executor ID of executors on re-registration.

Posted by Benjamin Mahler <bm...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/#review189974
-----------------------------------------------------------


Ship it!




How about the following?

```
Fixed an issue where task executor IDs are missing in the master.

Due to a bug, the agent was erroneously clearing the executor ID
of non-command executor's tasks before sending the
`ReregisterSlaveMessage` message. This leads to the master having
tasks with missing executor IDs (see MESOS-8135).

Also, it turns out that the clearing of executor ID's is actually
unnecessary altogether, as command executor tasks already do not
have an executor ID set.

Review: https://reviews.apache.org/r/63353/
```

I think it's nice for bug fixes to clearly state what they're fixing in the summary (that's what people will see in the short log).

- Benjamin Mahler


On Nov. 2, 2017, 7:52 p.m., Gaston Kleiman wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/63353/
> -----------------------------------------------------------
> 
> (Updated Nov. 2, 2017, 7:52 p.m.)
> 
> 
> Review request for mesos, Benjamin Mahler, Greg Mann, and Vinod Kone.
> 
> 
> Bugs: MESOS-8135
>     https://issues.apache.org/jira/browse/MESOS-8135
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Previously the agent would sometimes clear the executor ID of
> non-command executors before sending the `ReregisterSlaveMessage`
> message.
> 
> 
> Diffs
> -----
> 
>   src/slave/slave.cpp 337083dbe60bba2d3773b785bdceeaf0b8fcd070 
> 
> 
> Diff: https://reviews.apache.org/r/63353/diff/4/
> 
> 
> Testing
> -------
> 
> Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.
> 
> 
> Thanks,
> 
> Gaston Kleiman
> 
>


Re: Review Request 63353: Don't clear the executor ID of executors on re-registration.

Posted by Gaston Kleiman <ga...@mesosphere.io>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/
-----------------------------------------------------------

(Updated Nov. 2, 2017, 12:52 p.m.)


Review request for mesos, Benjamin Mahler, Greg Mann, and Vinod Kone.


Summary (updated)
-----------------

Don't clear the executor ID of executors on re-registration.


Bugs: MESOS-8135
    https://issues.apache.org/jira/browse/MESOS-8135


Repository: mesos


Description
-------

Previously the agent would sometimes clear the executor ID of
non-command executors before sending the `ReregisterSlaveMessage`
message.


Diffs
-----

  src/slave/slave.cpp 337083dbe60bba2d3773b785bdceeaf0b8fcd070 


Diff: https://reviews.apache.org/r/63353/diff/3/


Testing
-------

Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.


Thanks,

Gaston Kleiman


Re: Review Request 63353: Don't clear the executor ID of non-command executors on re-registration.

Posted by Gaston Kleiman <ga...@mesosphere.io>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/
-----------------------------------------------------------

(Updated Nov. 2, 2017, 12:52 p.m.)


Review request for mesos, Benjamin Mahler, Greg Mann, and Vinod Kone.


Changes
-------

We found out that there's actually no need to clear executor IDs at any point.


Bugs: MESOS-8135
    https://issues.apache.org/jira/browse/MESOS-8135


Repository: mesos


Description
-------

Previously the agent would sometimes clear the executor ID of
non-command executors before sending the `ReregisterSlaveMessage`
message.


Diffs (updated)
-----

  src/slave/slave.cpp 337083dbe60bba2d3773b785bdceeaf0b8fcd070 


Diff: https://reviews.apache.org/r/63353/diff/3/

Changes: https://reviews.apache.org/r/63353/diff/2-3/


Testing
-------

Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.


Thanks,

Gaston Kleiman


Re: Review Request 63353: Don't clear the executor ID of non-command executors on re-registration.

Posted by Gaston Kleiman <ga...@mesosphere.io>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/
-----------------------------------------------------------

(Updated Nov. 1, 2017, 1:53 p.m.)


Review request for mesos, Benjamin Mahler, Greg Mann, and Vinod Kone.


Changes
-------

Moved the test to a new patch.


Bugs: MESOS-8135
    https://issues.apache.org/jira/browse/MESOS-8135


Repository: mesos


Description
-------

Previously the agent would sometimes clear the executor ID of
non-command executors before sending the `ReregisterSlaveMessage`
message.


Diffs (updated)
-----

  src/slave/slave.cpp 337083dbe60bba2d3773b785bdceeaf0b8fcd070 


Diff: https://reviews.apache.org/r/63353/diff/2/

Changes: https://reviews.apache.org/r/63353/diff/1-2/


Testing
-------

Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.


Thanks,

Gaston Kleiman


Re: Review Request 63353: Don't clear the executor ID of non-command executors on re-registration.

Posted by Mesos Reviewbot <re...@mesos.apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/#review189410
-----------------------------------------------------------



Patch looks great!

Reviews applied: [63353]

Passed command: export OS='ubuntu:14.04' BUILDTOOL='autotools' COMPILER='gcc' CONFIGURATION='--verbose --disable-libtool-wrappers' ENVIRONMENT='GLOG_v=1 MESOS_VERBOSE=1'; ./support/docker-build.sh

- Mesos Reviewbot


On Oct. 27, 2017, 7:48 a.m., Gaston Kleiman wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/63353/
> -----------------------------------------------------------
> 
> (Updated Oct. 27, 2017, 7:48 a.m.)
> 
> 
> Review request for mesos, Benjamin Mahler and Greg Mann.
> 
> 
> Bugs: MESOS-8135
>     https://issues.apache.org/jira/browse/MESOS-8135
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Previously the agent would sometimes clear the executor ID of
> non-command executors before sending the `ReregisterSlaveMessage`
> message.
> 
> 
> Diffs
> -----
> 
>   src/slave/slave.cpp d8477b4e364270f5ad61311aa8fe48f823bc7aac 
>   src/tests/master_slave_reconciliation_tests.cpp d5eb7ba68c5308338236879e7cb1e970a01e48e6 
> 
> 
> Diff: https://reviews.apache.org/r/63353/diff/1/
> 
> 
> Testing
> -------
> 
> Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.
> 
> 
> Thanks,
> 
> Gaston Kleiman
> 
>


Re: Review Request 63353: Don't clear the executor ID of non-command executors on re-registration.

Posted by Gaston Kleiman <ga...@mesosphere.io>.

> On Oct. 31, 2017, 3:43 p.m., Benjamin Mahler wrote:
> > src/tests/master_slave_reconciliation_tests.cpp
> > Lines 721 (patched)
> > <https://reviews.apache.org/r/63353/diff/1/?file=1870005#file1870005line721>
> >
> >     TaskExecutorIds?

Renamed in RR 63474.


- Gaston


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/#review189761
-----------------------------------------------------------


On Nov. 1, 2017, 1:53 p.m., Gaston Kleiman wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/63353/
> -----------------------------------------------------------
> 
> (Updated Nov. 1, 2017, 1:53 p.m.)
> 
> 
> Review request for mesos, Benjamin Mahler, Greg Mann, and Vinod Kone.
> 
> 
> Bugs: MESOS-8135
>     https://issues.apache.org/jira/browse/MESOS-8135
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Previously the agent would sometimes clear the executor ID of
> non-command executors before sending the `ReregisterSlaveMessage`
> message.
> 
> 
> Diffs
> -----
> 
>   src/slave/slave.cpp 337083dbe60bba2d3773b785bdceeaf0b8fcd070 
> 
> 
> Diff: https://reviews.apache.org/r/63353/diff/2/
> 
> 
> Testing
> -------
> 
> Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.
> 
> 
> Thanks,
> 
> Gaston Kleiman
> 
>


Re: Review Request 63353: Don't clear the executor ID of non-command executors on re-registration.

Posted by Benjamin Mahler <bm...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/#review189761
-----------------------------------------------------------


Fix it, then Ship it!




Can you split the fix and the test? We don't usually cherry pick back the tests.


src/tests/master_slave_reconciliation_tests.cpp
Lines 721 (patched)
<https://reviews.apache.org/r/63353/#comment266992>

    TaskExecutorIds?



src/tests/master_slave_reconciliation_tests.cpp
Lines 849-851 (patched)
<https://reviews.apache.org/r/63353/#comment266993>

    Put the output on the next line here?


- Benjamin Mahler


On Oct. 27, 2017, 7:48 a.m., Gaston Kleiman wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/63353/
> -----------------------------------------------------------
> 
> (Updated Oct. 27, 2017, 7:48 a.m.)
> 
> 
> Review request for mesos, Benjamin Mahler, Greg Mann, and Vinod Kone.
> 
> 
> Bugs: MESOS-8135
>     https://issues.apache.org/jira/browse/MESOS-8135
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Previously the agent would sometimes clear the executor ID of
> non-command executors before sending the `ReregisterSlaveMessage`
> message.
> 
> 
> Diffs
> -----
> 
>   src/slave/slave.cpp d8477b4e364270f5ad61311aa8fe48f823bc7aac 
>   src/tests/master_slave_reconciliation_tests.cpp d5eb7ba68c5308338236879e7cb1e970a01e48e6 
> 
> 
> Diff: https://reviews.apache.org/r/63353/diff/1/
> 
> 
> Testing
> -------
> 
> Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.
> 
> 
> Thanks,
> 
> Gaston Kleiman
> 
>


Re: Review Request 63353: Don't clear the executor ID of non-command executors on re-registration.

Posted by Mesos Reviewbot Windows <re...@mesos.apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63353/#review189780
-----------------------------------------------------------



PASS: Mesos patch 63353 was successfully built and tested.

Reviews applied: `['63353']`

All the build artifacts available at: http://dcos-win.westus.cloudapp.azure.com/mesos-build/review/63353

- Mesos Reviewbot Windows


On Oct. 27, 2017, 7:48 a.m., Gaston Kleiman wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/63353/
> -----------------------------------------------------------
> 
> (Updated Oct. 27, 2017, 7:48 a.m.)
> 
> 
> Review request for mesos, Benjamin Mahler, Greg Mann, and Vinod Kone.
> 
> 
> Bugs: MESOS-8135
>     https://issues.apache.org/jira/browse/MESOS-8135
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Previously the agent would sometimes clear the executor ID of
> non-command executors before sending the `ReregisterSlaveMessage`
> message.
> 
> 
> Diffs
> -----
> 
>   src/slave/slave.cpp d8477b4e364270f5ad61311aa8fe48f823bc7aac 
>   src/tests/master_slave_reconciliation_tests.cpp d5eb7ba68c5308338236879e7cb1e970a01e48e6 
> 
> 
> Diff: https://reviews.apache.org/r/63353/diff/1/
> 
> 
> Testing
> -------
> 
> Verified that the new test fails on GNU/Linux without the rest of the patch, but passes with it.
> 
> 
> Thanks,
> 
> Gaston Kleiman
> 
>