More Mac CQ builder adjusting.

More Mac CQ builder adjusting.

This CL takes mac_chromium_gyp_rel out of the CQ altogether,
and drops mac_chromium_10.10_rel_ng to a 50% experiment to
help it get caught up.

TBR=kbr@chromium.org
NOTRY=true
BUG=626492
Committed: https://crrev.com/12b791371788f449ff3782d4707992f3a017a2f7
Cr-Commit-Position: refs/heads/master@{#404361}

[Dirk Pranke, 2016-07-08 03:09:26]

The CQ bit was checked by dpranke@chromium.org

[commit-bot: I haz the power, 2016-07-08 03:09:54]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-08 14:14:55]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-07-08 14:17:55]

Description was changed from

==========
More Mac CQ builder adjusting.

This CL takes mac_chromium_gyp_rel out of the CQ altogether,
and drops mac_chromium_10.10_rel_ng to a 50% experiment to
help it get caught up.

TBR=kbr@chromium.org
NOTRY=true
BUG=626492
==========

to

==========
More Mac CQ builder adjusting.

This CL takes mac_chromium_gyp_rel out of the CQ altogether,
and drops mac_chromium_10.10_rel_ng to a 50% experiment to
help it get caught up.

TBR=kbr@chromium.org
NOTRY=true
BUG=626492

Committed: https://crrev.com/12b791371788f449ff3782d4707992f3a017a2f7
Cr-Commit-Position: refs/heads/master@{#404361}
==========

[commit-bot: I haz the power, 2016-07-08 14:17:56]

Patchset 1 (id:??) landed as
https://crrev.com/12b791371788f449ff3782d4707992f3a017a2f7
Cr-Commit-Position: refs/heads/master@{#404361}

[Ken Russell (switch to Gerrit), 2016-07-08 22:33:17]

LGTM, but the new slaves in https://codereview.chromium.org/2133863003 are
needed. Currently this pool has 5 slaves; the main pool has 90. Any experiment
over 5% in my opinion runs the risk of swamping the optional slaves.

[Dirk Pranke, 2016-07-09 00:37:57]

On 2016/07/08 22:33:17, Ken Russell wrote:
> LGTM, but the new slaves in https://codereview.chromium.org/2133863003 are
> needed. Currently this pool has 5 slaves; the main pool has 90. Any experiment
> over 5% in my opinion runs the risk of swamping the optional slaves.

I'm not sure how this is related to the other CL. In this case, this takes the
builder
that uses the optional pool out of the CQ entirely, and the 10.10 builder uses
a completely different pool.

I actually don't think we should put any experimental builder into an optional
pool
at all; I think builders should be moved to the main pools before they are added
as experiments (and it was a mistake that I didn't do this for
mac_chromium_gyp_rel).

[Ken Russell (switch to Gerrit), 2016-07-09 01:35:10]

On 2016/07/09 00:37:57, Dirk Pranke wrote:
> On 2016/07/08 22:33:17, Ken Russell wrote:
> > LGTM, but the new slaves in https://codereview.chromium.org/2133863003 are
> > needed. Currently this pool has 5 slaves; the main pool has 90. Any
experiment
> > over 5% in my opinion runs the risk of swamping the optional slaves.
> 
> I'm not sure how this is related to the other CL. In this case, this takes the
> builder
> that uses the optional pool out of the CQ entirely, and the 10.10 builder uses
> a completely different pool.

Ah, sorry, I thought they were in the same pool. Agreed, they're unrelated.

> I actually don't think we should put any experimental builder into an optional
> pool
> at all; I think builders should be moved to the main pools before they are
added
> as experiments (and it was a mistake that I didn't do this for
> mac_chromium_gyp_rel).

Agreed.