[mojo-edk] Re-enable ping pong multiprocess tests

[mojo-edk] Re-enable ping pong multiprocess tests

These were timing out for potentially two reasons:

  1. 50000 iterations takes a long time to run in Debug builds. We
     should investigate why this slows down so much over time.
     Bug 585909 is filed for that.

  2. If sending a message failed in a non-fatal way (e.g. EAGAIN) we
     were dropping the handles. Since this can happen flakily,
     the tests could block indefinitely waiting for handles that
     would never arrive.

This CL fixes the handle dropping problem and reduces the test to run
500 iterations of 50 handles each.

BUG=585784
TBR=sammc@chromium.org

Committed: https://crrev.com/878a6845b81cfdeff147bfb6653469dd883aebef
Cr-Commit-Position: refs/heads/master@{#374811}

[Ken Rockot(use gerrit already), 2016-02-10 20:37:53]

PTAL. Note this only does 25k handles per test now instead of 50k. We should be
able to bump it back up if bug 585909 gets sorted out.

[Sam McNally, 2016-02-10 23:37:10]

On 2016/02/10 20:37:53, Ken Rockot wrote:
> PTAL. Note this only does 25k handles per test now instead of 50k. We should
be
> able to bump it back up if bug 585909 gets sorted out.

25k isn't enough to catch the fd leak on my desktop; the default fd limit there
is 32k. Perhaps we can set the limit lower in the child process to catch the
leak without a slow test.

[Ken Rockot(use gerrit already), 2016-02-10 23:38:44]

Note that if I make this leak locally I still don't catch it at 50k...

Limiting the child process SGTM. Can we still land this to turn the tests
back on and make sure they don't flake out for other reasons?

On Wed, Feb 10, 2016 at 3:37 PM, <sammc@chromium.org> wrote:

> On 2016/02/10 20:37:53, Ken Rockot wrote:
> > PTAL. Note this only does 25k handles per test now instead of 50k. We
should
> be
> > able to bump it back up if bug 585909 gets sorted out.
>
> 25k isn't enough to catch the fd leak on my desktop; the default fd limit
there
> is 32k. Perhaps we can set the limit lower in the child process to catch the
> leak without a slow test.
> https://codereview.chromium.org/1686243002/
>
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[Sam McNally, 2016-02-10 23:40:23]

LGTM

On 2016/02/10 23:38:44, Ken Rockot wrote:
> Note that if I make this leak locally I still don't catch it at 50k...
> 
> Limiting the child process SGTM. Can we still land this to turn the tests
> back on and make sure they don't flake out for other reasons?
> 
> On Wed, Feb 10, 2016 at 3:37 PM, <mailto:sammc@chromium.org> wrote:
> 
> > On 2016/02/10 20:37:53, Ken Rockot wrote:
> > > PTAL. Note this only does 25k handles per test now instead of 50k. We
> should
> > be
> > > able to bump it back up if bug 585909 gets sorted out.
> >
> > 25k isn't enough to catch the fd leak on my desktop; the default fd limit
> there
> > is 32k. Perhaps we can set the limit lower in the child process to catch the
> > leak without a slow test.
> > https://codereview.chromium.org/1686243002/
> >
> >
> 
> -- 
> You received this message because you are subscribed to the Google Groups
> "Chromium-reviews" group.
> To unsubscribe from this group and stop receiving emails from it, send an
email
> to mailto:chromium-reviews+unsubscribe@chromium.org.

[Ken Rockot(use gerrit already), 2016-02-10 23:50:20]

The CQ bit was checked by rockot@chromium.org

[commit-bot: I haz the power, 2016-02-10 23:53:16]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1686243002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1686243002/1

[commit-bot: I haz the power, 2016-02-11 00:56:46]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-02-16 22:33:24]

Description was changed from

==========
[mojo-edk] Re-enable ping pong multiprocess tests

These were timing out for potentially two reasons:

  1. 50000 iterations takes a long time to run in Debug builds. We
     should investigate why this slows down so much over time.
     Bug 585909 is filed for that.

  2. If sending a message failed in a non-fatal way (e.g. EAGAIN) we
     were dropping the handles. Since this can happen flakily,
     the tests could block indefinitely waiting for handles that
     would never arrive.

This CL fixes the handle dropping problem and reduces the test to run
500 iterations of 50 handles each.

BUG=585784
TBR=sammc@chromium.org
==========

to

==========
[mojo-edk] Re-enable ping pong multiprocess tests

These were timing out for potentially two reasons:

  1. 50000 iterations takes a long time to run in Debug builds. We
     should investigate why this slows down so much over time.
     Bug 585909 is filed for that.

  2. If sending a message failed in a non-fatal way (e.g. EAGAIN) we
     were dropping the handles. Since this can happen flakily,
     the tests could block indefinitely waiting for handles that
     would never arrive.

This CL fixes the handle dropping problem and reduces the test to run
500 iterations of 50 handles each.

BUG=585784
TBR=sammc@chromium.org

Committed: https://crrev.com/878a6845b81cfdeff147bfb6653469dd883aebef
Cr-Commit-Position: refs/heads/master@{#374811}
==========

[commit-bot: I haz the power, 2016-02-16 22:33:26]

Patchset 1 (id:??) landed as
https://crrev.com/878a6845b81cfdeff147bfb6653469dd883aebef
Cr-Commit-Position: refs/heads/master@{#374811}