You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@aurora.apache.org by Zameer Manji <zm...@twopensource.com> on 2014/10/01 02:18:00 UTC

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/
-----------------------------------------------------------

(Updated Sept. 30, 2014, 5:17 p.m.)


Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.


Changes
-------

Brian's Feedback.


Bugs: AURORA-728
    https://issues.apache.org/jira/browse/AURORA-728


Repository: aurora


Description
-------

Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.


Diffs (updated)
-----

  src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
  src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
  src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 

Diff: https://reviews.apache.org/r/25974/diff/


Testing (updated)
-------

./pants src/test/python/apache/aurora/executor:executor-small
./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh


Thanks,

Zameer Manji

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Kevin Sweeney <ke...@apache.org>.

> On Oct. 1, 2014, 4:34 p.m., Kevin Sweeney wrote:
> > src/main/python/apache/aurora/executor/common/announcer.py, line 228
> > <https://reviews.apache.org/r/25974/diff/6/?file=709964#file709964line228>
> >
> >     push this into `__start`, out of the constructor?
> >     
> >     At least on the Java side we try to avoid doing any I/O in constructors (as they're more for wiring) but instead delegate to explicit lifecycle methods like `start`.
> 
> Zameer Manji wrote:
>     I'd rather not. Keeping it in the constructor means it stays in the from_assigned_task method code path. Moving it to the start method moves it out of that code path and really changes when we do this sort of I/O.

is it common practice to do I/O in constructors elsewhere in this codebase? I'll defer to wickman here


- Kevin


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55161
-----------------------------------------------------------


On Sept. 30, 2014, 5:17 p.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Sept. 30, 2014, 5:17 p.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Zameer Manji <zm...@twopensource.com>.

> On Oct. 1, 2014, 4:34 p.m., Kevin Sweeney wrote:
> > src/main/python/apache/aurora/executor/common/announcer.py, line 228
> > <https://reviews.apache.org/r/25974/diff/6/?file=709964#file709964line228>
> >
> >     push this into `__start`, out of the constructor?
> >     
> >     At least on the Java side we try to avoid doing any I/O in constructors (as they're more for wiring) but instead delegate to explicit lifecycle methods like `start`.

I'd rather not. Keeping it in the constructor means it stays in the from_assigned_task method code path. Moving it to the start method moves it out of that code path and really changes when we do this sort of I/O.


- Zameer


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55161
-----------------------------------------------------------


On Sept. 30, 2014, 5:17 p.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Sept. 30, 2014, 5:17 p.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Kevin Sweeney <ke...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55161
-----------------------------------------------------------



src/main/python/apache/aurora/executor/aurora_executor.py
<https://reviews.apache.org/r/25974/#comment95545>

    superfluous return



src/main/python/apache/aurora/executor/common/announcer.py
<https://reviews.apache.org/r/25974/#comment95546>

    push this into `__start`, out of the constructor?
    
    At least on the Java side we try to avoid doing any I/O in constructors (as they're more for wiring) but instead delegate to explicit lifecycle methods like `start`.


- Kevin Sweeney


On Sept. 30, 2014, 5:17 p.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Sept. 30, 2014, 5:17 p.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Brian Wickman <wi...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55309
-----------------------------------------------------------



src/test/python/apache/aurora/executor/common/test_announcer.py
<https://reviews.apache.org/r/25974/#comment95688>

    I'm confused -- nobody calls client.connected anymore, right?  In theory, this test will pass, except it will take 30-60 seconds to run since client_connect_event.wait(timeout=...) will be called with the health check timeout.


- Brian Wickman


On Oct. 2, 2014, 11:20 p.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Oct. 2, 2014, 11:20 p.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Zameer Manji <zm...@twopensource.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/
-----------------------------------------------------------

(Updated Oct. 6, 2014, 11:53 a.m.)


Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.


Changes
-------

Fix import order and checkstyle issues.


Bugs: AURORA-728
    https://issues.apache.org/jira/browse/AURORA-728


Repository: aurora


Description
-------

Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.


Diffs (updated)
-----

  src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
  src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
  src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 

Diff: https://reviews.apache.org/r/25974/diff/


Testing
-------

./pants src/test/python/apache/aurora/executor:executor-small
./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh


Thanks,

Zameer Manji

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Zameer Manji <zm...@twopensource.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55402
-----------------------------------------------------------


Pinging Kevin for final ship it and commit.

- Zameer Manji


On Oct. 2, 2014, 6:44 p.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Oct. 2, 2014, 6:44 p.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Kevin Sweeney <ke...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55547
-----------------------------------------------------------

Ship it!


Ship It!

- Kevin Sweeney


On Oct. 2, 2014, 6:44 p.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Oct. 2, 2014, 6:44 p.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Brian Wickman <wi...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55313
-----------------------------------------------------------

Ship it!


Ship It!

- Brian Wickman


On Oct. 3, 2014, 1:44 a.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Oct. 3, 2014, 1:44 a.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Zameer Manji <zm...@twopensource.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/
-----------------------------------------------------------

(Updated Oct. 2, 2014, 6:44 p.m.)


Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.


Changes
-------

Brian's feedback.


Bugs: AURORA-728
    https://issues.apache.org/jira/browse/AURORA-728


Repository: aurora


Description
-------

Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.


Diffs (updated)
-----

  src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
  src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
  src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 

Diff: https://reviews.apache.org/r/25974/diff/


Testing
-------

./pants src/test/python/apache/aurora/executor:executor-small
./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh


Thanks,

Zameer Manji

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Zameer Manji <zm...@twopensource.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/
-----------------------------------------------------------

(Updated Oct. 2, 2014, 4:20 p.m.)


Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.


Changes
-------

Brian's feedback.


Bugs: AURORA-728
    https://issues.apache.org/jira/browse/AURORA-728


Repository: aurora


Description
-------

Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.


Diffs (updated)
-----

  src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
  src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
  src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 

Diff: https://reviews.apache.org/r/25974/diff/


Testing
-------

./pants src/test/python/apache/aurora/executor:executor-small
./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh


Thanks,

Zameer Manji

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Brian Wickman <wi...@apache.org>.

> On Oct. 1, 2014, 11:14 p.m., Brian Wickman wrote:
> >

btw overall this review looks great and is much cleaner than what i was envisioning.


- Brian


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55158
-----------------------------------------------------------


On Oct. 1, 2014, 12:17 a.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Oct. 1, 2014, 12:17 a.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Zameer Manji <zm...@twopensource.com>.

> On Oct. 1, 2014, 4:14 p.m., Brian Wickman wrote:
> > src/test/python/apache/aurora/executor/common/test_announcer.py, line 239
> > <https://reviews.apache.org/r/25974/diff/6/?file=709965#file709965line239>
> >
> >     out of scope for this review, but there's a new-ish pypi project called 'zake' that allows the kazoo client to be stubbed out with a mock zk tree (while preserving certain operations like create/delete.)  i'd love to explore using it here in the future.

It looks really nice for this sort of testing. It is definately out of the scope of this review. If we ever need to do more work on the Announcer/Service Discovery we should first migrate to zake.


- Zameer


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55158
-----------------------------------------------------------


On Sept. 30, 2014, 5:17 p.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Sept. 30, 2014, 5:17 p.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>

Re: Review Request 25974: Prevent initial ZK timeouts from killing the executor.

Posted by Brian Wickman <wi...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25974/#review55158
-----------------------------------------------------------



src/main/python/apache/aurora/executor/common/announcer.py
<https://reviews.apache.org/r/25974/#comment95541>

    i think "if self.__connect_event.is_set()" is more accurate -- it's possible for it to be connected and then race with disconnection.  we just want to validate that it managed to connect initially.
    
    this complicates your test a little bit since you do connected.set() right off the bat.  but you might be able to get around this by making a HealthCheckConfig that produces a tiny timeout.



src/test/python/apache/aurora/executor/common/test_announcer.py
<https://reviews.apache.org/r/25974/#comment95542>

    out of scope for this review, but there's a new-ish pypi project called 'zake' that allows the kazoo client to be stubbed out with a mock zk tree (while preserving certain operations like create/delete.)  i'd love to explore using it here in the future.


- Brian Wickman


On Oct. 1, 2014, 12:17 a.m., Zameer Manji wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25974/
> -----------------------------------------------------------
> 
> (Updated Oct. 1, 2014, 12:17 a.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney, Bill Farner, and Brian Wickman.
> 
> 
> Bugs: AURORA-728
>     https://issues.apache.org/jira/browse/AURORA-728
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Prevent initial ZK timeouts from killing the executor. In addition, prevent uncaught exceptions from killing the executor.
> 
> 
> Diffs
> -----
> 
>   src/main/python/apache/aurora/executor/aurora_executor.py 79a24855b2a68271b7478395dfdadab8755c3af2 
>   src/main/python/apache/aurora/executor/common/announcer.py c466da8d48bbc2aa227c2db157cab84665ad6602 
>   src/test/python/apache/aurora/executor/common/test_announcer.py 4f6e200ecb1a4ea7cb45acd466a57f19d5815326 
> 
> Diff: https://reviews.apache.org/r/25974/diff/
> 
> 
> Testing
> -------
> 
> ./pants src/test/python/apache/aurora/executor:executor-small
> ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh
> 
> 
> Thanks,
> 
> Zameer Manji
> 
>