Added new slaves and spec files for GPU FYI bots running Chromium recipe.

Added new slaves and spec files for GPU FYI bots running Chromium recipe.

Adding Linux and Mac builders and testers to start. More configurations
will be added once these are working.

BUG=542370
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=298105

[Ken Russell (switch to Gerrit), 2015-12-29 03:35:52]

kbr@chromium.org changed reviewers:
+ phajdan.jr@chromium.org, sergeyberezin@chromium.org, sergiyb@chromium.org,
tandrii@chromium.org

[Ken Russell (switch to Gerrit), 2015-12-29 03:35:52]

Anyone who happens to be around this week: could you please review this CL?
While most people are OOO, I'd like to deploy these new machines and debug as
many problems as possible.

This goes with the src-side change https://codereview.chromium.org/1547383002 .
Am I correct in thinking that that CL needs to land first?

https://codereview.chromium.org/1550223002/diff/1/masters/master.chromium.gpu...
File masters/master.chromium.gpu.fyi/slaves.cfg (right):

https://codereview.chromium.org/1550223002/diff/1/masters/master.chromium.gpu...
masters/master.chromium.gpu.fyi/slaves.cfg:318: # TODO(kbr): is this needed or
will it conflict with the Chromium
I'm 99% sure that in order for these new testers to work correctly, I need to
remove the "triggered_by" property. I added it specially for the
recipe_master_helper to be able to trigger the GPU testers from the builders.
However, it looks like the Chromium recipe uses recipe-based triggers, so the
testers should be added only to the buildbot "builders", and not the
"schedulers". Can anyone confirm whether that's correct?

(The chromium.gpu and chromium.gpu.fyi waterfalls are the only ones using the
recipe_master_helper -- it's very useful, and eliminates a lot of boilerplate
for the waterfalls. I studied the differences between master.chromium.gpu.fyi
and master.chromium.linux.)

[tandrii(chromium), 2015-12-29 11:19:30]

lg-tm but i'd rather get one from Sergey Berezin or Pawel.

https://codereview.chromium.org/1550223002/diff/1/masters/master.chromium.gpu...
File masters/master.chromium.gpu.fyi/slaves.cfg (right):

https://codereview.chromium.org/1550223002/diff/1/masters/master.chromium.gpu...
masters/master.chromium.gpu.fyi/slaves.cfg:318: # TODO(kbr): is this needed or
will it conflict with the Chromium
On 2015/12/29 03:35:52, Ken Russell wrote:
> I'm 99% sure that in order for these new testers to work correctly, I need to
> remove the "triggered_by" property. I added it specially for the
> recipe_master_helper to be able to trigger the GPU testers from the builders.
> However, it looks like the Chromium recipe uses recipe-based triggers, so the
> testers should be added only to the buildbot "builders", and not the
> "schedulers". Can anyone confirm whether that's correct?
> 
> (The chromium.gpu and chromium.gpu.fyi waterfalls are the only ones using the
> recipe_master_helper -- it's very useful, and eliminates a lot of boilerplate
> for the waterfalls. I studied the differences between master.chromium.gpu.fyi
> and master.chromium.linux.)

i'm with you here, but also for 99% - don't know for sure. Sergey - can you
verify this claim?

[Sergey Berezin, 2015-12-29 18:25:36]

LGTM, and I'm pretty sure you can remove 'triggered_by'. Though I'm also pretty
sure it wouldn't break anything, since nothing seems to read this flag.

https://codereview.chromium.org/1550223002/diff/1/masters/master.chromium.gpu...
File masters/master.chromium.gpu.fyi/slaves.cfg (right):

https://codereview.chromium.org/1550223002/diff/1/masters/master.chromium.gpu...
masters/master.chromium.gpu.fyi/slaves.cfg:318: # TODO(kbr): is this needed or
will it conflict with the Chromium
On 2015/12/29 11:19:30, tandrii(chromium) wrote:
> On 2015/12/29 03:35:52, Ken Russell wrote:
> > I'm 99% sure that in order for these new testers to work correctly, I need
to
> > remove the "triggered_by" property. I added it specially for the
> > recipe_master_helper to be able to trigger the GPU testers from the
builders.
> > However, it looks like the Chromium recipe uses recipe-based triggers, so
the
> > testers should be added only to the buildbot "builders", and not the
> > "schedulers". Can anyone confirm whether that's correct?
> > 
> > (The chromium.gpu and chromium.gpu.fyi waterfalls are the only ones using
the
> > recipe_master_helper -- it's very useful, and eliminates a lot of
boilerplate
> > for the waterfalls. I studied the differences between
master.chromium.gpu.fyi
> > and master.chromium.linux.)
> 
> i'm with you here, but also for 99% - don't know for sure. Sergey - can you
> verify this claim?

I looked at it together with dnj@, and apparently 'triggered_by' will not be
used at all. Instead, what you did with in the config specifying the parent
builder is apparently enough to set up all the triggers. Disclaimer - I didn't
actually use it myself, so the same 99% :-)

[Ken Russell (switch to Gerrit), 2015-12-30 02:05:22]

I'm going to CQ this and request a master restart in order to start debugging
these new bots.

https://codereview.chromium.org/1550223002/diff/1/masters/master.chromium.gpu...
File masters/master.chromium.gpu.fyi/slaves.cfg (right):

https://codereview.chromium.org/1550223002/diff/1/masters/master.chromium.gpu...
masters/master.chromium.gpu.fyi/slaves.cfg:318: # TODO(kbr): is this needed or
will it conflict with the Chromium
On 2015/12/29 18:25:36, Sergey Berezin wrote:
> On 2015/12/29 11:19:30, tandrii(chromium) wrote:
> > On 2015/12/29 03:35:52, Ken Russell wrote:
> > > I'm 99% sure that in order for these new testers to work correctly, I need
> to
> > > remove the "triggered_by" property. I added it specially for the
> > > recipe_master_helper to be able to trigger the GPU testers from the
> builders.
> > > However, it looks like the Chromium recipe uses recipe-based triggers, so
> the
> > > testers should be added only to the buildbot "builders", and not the
> > > "schedulers". Can anyone confirm whether that's correct?
> > > 
> > > (The chromium.gpu and chromium.gpu.fyi waterfalls are the only ones using
> the
> > > recipe_master_helper -- it's very useful, and eliminates a lot of
> boilerplate
> > > for the waterfalls. I studied the differences between
> master.chromium.gpu.fyi
> > > and master.chromium.linux.)
> > 
> > i'm with you here, but also for 99% - don't know for sure. Sergey - can you
> > verify this claim?
> 
> I looked at it together with dnj@, and apparently 'triggered_by' will not be
> used at all. Instead, what you did with in the config specifying the parent
> builder is apparently enough to set up all the triggers. Disclaimer - I didn't
> actually use it myself, so the same 99% :-)

recipe_master_helper.py, which is used to set up this waterfall, pays attention
to the "triggered_by" property. I'm pretty sure I need to modify it to avoid
setting up a scheduler and Triggerable for these testers.

[Ken Russell (switch to Gerrit), 2015-12-30 02:05:42]

The CQ bit was checked by kbr@chromium.org

[Ken Russell (switch to Gerrit), 2015-12-30 02:05:42]

The patchset sent to the CQ was uploaded after l-g-t-m from
sergeyberezin@chromium.org
Link to the patchset: https://codereview.chromium.org/1550223002/#ps20001
(title: "Added sentinel for bots using recipe-side triggers.")

[commit-bot: I haz the power, 2015-12-30 02:05:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1550223002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1550223002/20001

[commit-bot: I haz the power, 2015-12-30 02:09:02]

Description was changed from

==========
Added new slaves and spec files for GPU FYI bots running Chromium recipe.

Adding Linux and Mac builders and testers to start. More configurations
will be added once these are working.

BUG=542370
==========

to

==========
Added new slaves and spec files for GPU FYI bots running Chromium recipe.

Adding Linux and Mac builders and testers to start. More configurations
will be added once these are working.

BUG=542370

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=298105
==========

[commit-bot: I haz the power, 2015-12-30 02:09:03]

Committed patchset #2 (id:20001) as
http://src.chromium.org/viewvc/chrome?view=rev&revision=298105