mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ilya Pronin <ipro...@twopensource.com>
Subject Re: Review Request 62241: Added SchedulerHttpApiTest.UpdateHttpToPidSchedulerAndBack test.
Date Mon, 16 Oct 2017 18:17:18 GMT


> On Oct. 13, 2017, 6:40 p.m., Anand Mazumdar wrote:
> > src/tests/scheduler_http_api_tests.cpp
> > Lines 674 (patched)
> > <https://reviews.apache.org/r/62241/diff/1/?file=1820104#file1820104line674>
> >
> >     s/an/a

I'm pretty sure it's "an HTTP" :) https://en.wikipedia.org/wiki/Hypertext_Transfer_Protocol


> On Oct. 13, 2017, 6:40 p.m., Anand Mazumdar wrote:
> > src/tests/scheduler_http_api_tests.cpp
> > Lines 712-716 (patched)
> > <https://reviews.apache.org/r/62241/diff/1/?file=1820104#file1820104line712>
> >
> >     Do you care about a `HEARTBEAT` even for this test?

No. I copied that part from the other test :) I think we can can remove this part and failover
right after we get a `SUBSCRIBED` event. Ran the updated test with `--gtest_repeat=1000 --gtest_break_on_failure`
to verify that I don't introduce any flakiness and it seems fine.


> On Oct. 13, 2017, 6:40 p.m., Anand Mazumdar wrote:
> > src/tests/scheduler_http_api_tests.cpp
> > Lines 737-747 (patched)
> > <https://reviews.apache.org/r/62241/diff/1/?file=1820104#file1820104line737>
> >
> >     Kill this? We don't need this to test if the downgrade worked.

The crash that this test reproduces occurs when the master handles a `Message` from the scheduler.
We can remove the reconciliation part and the crash would still happen when the scheduler
driver sends a `TEARDOWN` call message from `SchedulerProcess::stop()`. But I think it is
better to be explicit here and not leave exercising the code that we were fixing to the cleanup
logic.


- Ilya


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/62241/#review188053
-----------------------------------------------------------


On Oct. 16, 2017, 11:17 a.m., Ilya Pronin wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/62241/
> -----------------------------------------------------------
> 
> (Updated Oct. 16, 2017, 11:17 a.m.)
> 
> 
> Review request for mesos and Anand Mazumdar.
> 
> 
> Bugs: MESOS-7867
>     https://issues.apache.org/jira/browse/MESOS-7867
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This test verifies that we are able to upgrade from a PID based
> framework to an HTTP framework and then downgrade back without
> restarting the master.
> 
> 
> Diffs
> -----
> 
>   src/tests/scheduler_http_api_tests.cpp cc03be0d2110f08023bf0ce181da37658737bd7c 
> 
> 
> Diff: https://reviews.apache.org/r/62241/diff/2/
> 
> 
> Testing
> -------
> 
> Ran `make check`. Verified that the test fails without the previous patch.
> 
> 
> Thanks,
> 
> Ilya Pronin
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message