Chris Cooper [:coop] (he/him) Reporter
	Description • 9 years ago

Short, likely incomplete list of systems that need to be made aware of the new 10.10 machines:

* buildbot-configs
* slavealloc
* graphserver
* slave health
* treeherder

	Justin Wood (:Callek)
	Comment 1 • 9 years ago

There is a lot of "prior art" on Bug 1191481 for what is required, including some handy SQL magic for graph server and slavealloc.

For treeherder I did: https://github.com/mozilla/treeherder/pull/846

	Kim Moir [:kmoir] ET
	Comment 2 • 9 years ago

Vlad and Alin, as discussed in today's meeting you can start writing patches for the components in comment #1 that require changes to stand up this new platform. As Amy suggested, we'll stand up this platform in parallel with the existing 10.10 machines so we can limit it to certain branches as we bring the machines online.

Here is a reference document for setting up a new platform. Some of the doc may need to be updated, so please update as you see errors or omissions
https://wiki.mozilla.org/ReferencePlatforms/HowToSetupNewPlatform

I don't think we will need new masters so you can start at this step
https://wiki.mozilla.org/ReferencePlatforms/HowToSetupNewPlatform#Open_a_bug_for_buildbot_and_puppet_changes

Let me know if you have any questions, I'm happy to help.

	Alin Selagea [:aselagea] Assignee
	Comment 3 • 9 years ago

Patch to add the new slaves in graph server.

	Alin Selagea [:aselagea] Assignee
	Comment 4 • 9 years ago

Patch to enable t-yosemite-r7 slaves in buildbot-configs.

	Alin Selagea [:aselagea] Assignee
	Comment 5 • 9 years ago

Also added the csv to add the slaves in DB.

	Kim Moir [:kmoir] ET
	Comment 6 • 9 years ago

Comment on attachment 8665950 [details] [diff] [review]
buildbot-configs.patch

You need to create a new platform pool for the r7 machines.  When we enable them we want to be able to distinguish between the existing r5 yosemite pool and the new r7 one.  We will only enable the new r7 machines on a few branches to start, which will allow us to disable some of the old r5 ones and use their cabinet space in the datacentre

So you need to have a new name here SLAVES['yosemite'] other than yosemite

Otherwise we'll just have a mixed pool of r5 and r7 machines

	Kim Moir [:kmoir] ET
	Comment 7 • 9 years ago

Comment on attachment 8665948 [details] [diff] [review]
graph_server.patch

Same issue here.  We need to add another platform pool to keep the r7 machines separate.

You need to add another platform here for the r7 machines

so see how their is MacOSX 10.10.2 defined here
insert into os_list values (55, "MacOSX 10.10");
insert into os_list values (57, "MacOSX 10.10 (e10s)");


You will need one something like
insert into os_list values (59, "MacOSX 10.10.5");
insert into os_list values (61, "MacOSX 10.10.5 (e10s)");

And then your corresponding database entries should reflect the values.

These are patches to add them to graph data base sql which is just sort of a a backup of the data.  To actually have them added to the production database you'll need to open a bug like this for the database team

https://bugzilla.mozilla.org/show_bug.cgi?id=1131072

You can open this bug after you get a r+ on your revised patches

I'll update the doc to reflect this, currently it's missing here
https://wiki.mozilla.org/ReferencePlatforms/HowToSetupNewPlatform#Open_bugs_for_graph_server_changes

	Kim Moir [:kmoir] ET
	Comment 11 • 9 years ago

Comment on attachment 8666009 [details] [diff] [review]
buildbot-configs.patch

Can you please make it yosemite_r7 instead to make it consistent with the other platforms?

	Alin Selagea [:aselagea] Assignee
	Comment 12 • 9 years ago

done!

	Kim Moir [:kmoir] ET
	Comment 13 • 9 years ago

Comment on attachment 8666033 [details] [diff] [review]
buildbot-configs.patch

thanks! great work!  You also need a patch to add the new platform to puppet.

	Alin Selagea [:aselagea] Assignee
	Comment 15 • 9 years ago

Added the patch for slave health. Note that /js/slave_health.js file contains a function "getSlavetypeForPendingJob(slaveclass, pending)" which determines the pool of slaves to address by searching after a certain pattern in the name of the pending job. However, in our case it is not possible to distinguish between t-yosemite-r5 and t-yosemite-r7 by using this criterion as the pool of jobs that run on the two types of slaves is identical. 
Talked to Coop and he suggested to leave it as it is at the moment (we'll probably change it when the transition from r5 to r7 is completed).

	Kim Moir [:kmoir] ET
	Comment 16 • 9 years ago

So I looked at the master capacity for macosx 10.10 tests.  Currently we have three masters for macosx tests (bm106, bm107, bm108) which serve the 106 10.10 hosts +  151 10.6 hosts.  The masters don't seem to be currently overloaded.

The plan is to enable 64 new machines but at the same time, disable the r5 builders on a branch/trunk. This will allow us to disable some r5 machines and move some r7 machines into their racks in the data centre. So I think we are okay for masters.

So once your trial runs work on your master, the next step is to write patches to enable the r7 machines in the buildbot configs. Then try these patches on your master and ensure the new builders appear as expected.  This doc has some good steps on how to verify that the new builders change as expected.

https://wiki.mozilla.org/ReleaseEngineering:TestingTechniques

One way forward would be to enable the r7 builders on try. (Currently 10.10 is disabled by default on try - you have to explicitly request it when you issue a try run).

See

http://hg.mozilla.org/build/buildbot-configs/file/9cf33da8ee95/mozilla-tests/config.py#l84

and note

try_by_default': False

Once you have verified that the new builders work on try in production you could enable on trunk and disable the associated r5 builders.  There are fewer 10.10 jobs on try than m-c for example.

	Alin Selagea [:aselagea] Assignee
	Comment 17 • 9 years ago

preliminary version, will need to update it.

	Kim Moir [:kmoir] ET
	Comment 18 • 9 years ago

Comment on attachment 8668925 [details] [diff] [review]
try_config.patch

This is a good start.  As I mentioned in our meeting, you will also need something like this to define the tests for yosemite r7

http://hg.mozilla.org/build/buildbot-configs/file/439b1791c08e/mozilla-tests/config.py#l1834

This is an example bug of where it was done before
bug 1122039

I think that is all you need, compare the builders before and after to see if it worked using builder_list.py

https://wiki.mozilla.org/ReleaseEngineering:TestingTechniques#builder_list.py_.2F_dump_master.py

	Alin Selagea [:aselagea] Assignee
	Comment 19 • 9 years ago

updated version

	Alin Selagea [:aselagea] Assignee
	Comment 20 • 9 years ago

Difference between the builders after applying the patch.

	Kim Moir [:kmoir] ET
	Comment 21 • 9 years ago

Comment on attachment 8669068 [details]
builder_diff.txt

The patch looks good. However, I'm wondering why there are new 10.10 and 10.6 builders added. The new builders should just be on 10.10.5.

	Alin Selagea [:aselagea] Assignee
	Comment 22 • 9 years ago

Added an updated version of the builders diff file. Things look good now.

	Alin Selagea [:aselagea] Assignee
	Comment 23 • 9 years ago

Uploaded patch to add trychooser syntax.

	Kim Moir [:kmoir] ET
	Comment 24 • 9 years ago

Comment on attachment 8669675 [details] [diff] [review]
trychooser.patch

You also need to update tools/trychooser/index.html to account for the new platform

	Alin Selagea [:aselagea] Assignee
	Comment 25 • 9 years ago

updated trychooser patch

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 26 • 9 years ago

The rest of yosemite-r7 slaves (8-64) have been added in slavealloc DB. The first 7 slaves that had been use for tests have been re-imaged.

	Kim Moir [:kmoir] ET
	Comment 27 • 9 years ago

key was missing in BuildSlaves.py.template

	Kim Moir [:kmoir] ET
	Comment 28 • 9 years ago

in production

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 29 • 9 years ago

:kmoir pushed some try tests on yosemite-r7, more details:
- https://treeherder.mozilla.org/#/jobs?repo=try&revision=a479174a8763
- https://treeherder.mozilla.org/#/jobs?repo=try&revision=0084ba8c2789

After the tests were completed, I made a comparison between them:
Failing jobs:
-> mozilla-central opt test gtest
-> mozilla-central debug test gtest
Warning jobs:
-> mozilla-central opt test mochitest-devtools-chrome-2
-> mozilla-central opt test mochitest-gl
-> mozilla-central opt test web-platform-tests-2
-> mozilla-central opt test web-platform-tests-3
-> mozilla-central opt test web-platform-tests-4
-> mozilla-central opt test web-platform-tests-reftests
-> mozilla-central debug test mochitest-2
-> mozilla-central debug test mochitest-devtools-chrome-3
-> mozilla-central debug test mochitest-gl
-> mozilla-central debug test web-platform-tests-2
-> mozilla-central debug test web-platform-tests-3

	Kim Moir [:kmoir] ET
	Comment 30 • 9 years ago

:VladC: it might be better to update bug 1210395 with this info as this is the bug that tracks the greening up of tests

	James Graham [:jgraham]
	Comment 31 • 9 years ago

So apparently this is the change that made pushing to try with [10.10] run on 10.10.5 rather than 10.10.2, thus causing unexpected breakage that doesn't reflect inbound. Can we fix the trychooser part so that doesn't happen?

	James Graham [:jgraham]
	Comment 32 • 9 years ago

Oh, right, just found that that's bug 1212887

	Kim Moir [:kmoir] ET
	Comment 33 • 9 years ago

Some hints on a patch to enable r7 on trunk and disable r5 on trunk

Here you disable r7 on every branch but try
http://hg.mozilla.org/build/buildbot-configs/file/c4ae6f2036cf/mozilla-tests/config.py#l2615

You could find the trunk branches like is done here
http://hg.mozilla.org/build/buildbot-configs/file/c4ae6f2036cf/mozilla-tests/config.py#l2662
to find the list of branches to enable r7 on trunk
and disable r5 on

compare the builder list before and after to confirm :-)

	Kim Moir [:kmoir] ET
	Comment 34 • 9 years ago

Vlad, here is a patch to enable yosemite_r7 tests on trunk as a hint.  The next step is to modify the patch so that it disables the yosemite tests on the same branches.

	Kim Moir [:kmoir] ET
	Comment 35 • 9 years ago

oops wrong patch with the last comment, here is a new one

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 36 • 9 years ago

Attached you can find the patch that enable yosmite_r7 on trunk and disable yosemite_r5

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 37 • 9 years ago

Attached you canfind the difference for the following patch: https://bug1203128.bmoattachments.org/attachment.cgi?id=8679507

	Kim Moir [:kmoir] ET
	Comment 38 • 9 years ago

Comment on attachment 8679507 [details] [diff] [review]
config_trunk.py-1203128.patch

we still want to keep the yosemite jobs running on the branches that are not trunk (i.e. not geck version 44).  For example branches like mozilla-beta, mozilla-release etc  So this line

delete_slave_platform(BRANCHES, PLATFORMS, {'macosx64': 'yosemite'}, branch_exclusions=[])

should include a list of those branches to keep running yosemite tests on

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 39 • 9 years ago

Updated the patch

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 40 • 9 years ago

Updated difference file

	Kim Moir [:kmoir] ET
	Comment 41 • 9 years ago

Comment on attachment 8679519 [details] [diff] [review]
config_trunk.py-1203128.patch

This looks like it is two patches concatenated together for the same file instead of just one patch.  Also, it doesn't leave the yosemite tests on the branches that aren't on the ride_trains_branches list

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 42 • 9 years ago

Attached the correct patch

	Kim Moir [:kmoir] ET
	Comment 43 • 9 years ago

Comment on attachment 8679987 [details] [diff] [review]
config_trunk.py-1203128.patch

You still need something to disable the corresponding tests on yosemite on the branches

see http://hg.mozilla.org/build/buildbot-configs/file/db1761df121c/mozilla-tests/config.py#l2621

for an example of this sort of change

but instead of disabling ubuntu64_vm_lnx_large slaves on the linux64 platform you will be disabling yosemite slaves on the macosx64 platform

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 44 • 9 years ago

Updated the patch

	Kim Moir [:kmoir] ET
	Comment 46 • 9 years ago

Comment on attachment 8680037 [details]
diff_enable_r7_trunk.txt

I didn't explain this well previously, I apologize.  The example I gave earlier was not a good one

You want to have results like this for an example job
+Rev5 MacOSX Yosemite 10.10.5 jamun talos g2 ScriptFactory
-Rev5 MacOSX Yosemite 10.10 jamun talos g2 ScriptFactory

So for every job 10.10.5 job enabled on trunk, disable the corresponding 10.10 job on trunk

	Alin Selagea [:aselagea] Assignee
	Comment 47 • 9 years ago

Added the patch the enable r7 on trunk and disable r5 on trunk.
R5 is still enabled on non-trunk branches though.

	Alin Selagea [:aselagea] Assignee
	Comment 48 • 9 years ago

Diff file.

	Kim Moir [:kmoir] ET
	Comment 49 • 9 years ago

Comment on attachment 8680651 [details] [diff] [review]
trunk_enable_r7_disable_r5.patch

Looks good, do you need this line

+PLATFORMS['macosx64']['talos_slave_platforms'] = ['yosemite_r7']

I think the calls to delete_slave_platform will take care of enabling and disabling talos for the correct platforms and branches

	Alin Selagea [:aselagea] Assignee
	Comment 50 • 9 years ago

Tested by leaving "PLATFORMS['macosx64']['talos_slave_platforms'] = ['yosemite', 'yosemite_r7']" as it is --> that won't remove talos jobs for yosemite, but it will add them for yosemite_r7.

If I completely remove that line --> we'll also have talos tests for yosemite and yosemite_r7, but some other talos tests for 10.6 will be added (see diff2_trunk.txt). 

My guess is that if no slave platform is specified, then talos jobs will be generated for all of them.

	Kim Moir [:kmoir] ET
	Comment 52 • 9 years ago

Comment on attachment 8680652 [details]
diff_trunk.txt

I think there is a remaining issue
-Rev5 MacOSX Yosemite 10.10 mozilla-beta talos g2 ScriptFactory
-Rev5 MacOSX Yosemite 10.10 mozilla-beta talos g1 ScriptFactory
-Rev5 MacOSX Yosemite 10.10 mozilla-beta talos svgr ScriptFactory
-Rev5 MacOSX Yosemite 10.10 mozilla-beta talos dromaeojs ScriptFactory
-Rev5 MacOSX Yosemite 10.10 mozilla-beta talos tp5o ScriptFactory
-Rev5 MacOSX Yosemite 10.10 mozilla-beta talos other ScriptFactory
-Rev5 MacOSX Yosemite 10.10 mozilla-beta talos chromez ScriptFactory
+Rev5 MacOSX Yosemite 10.10.5 mozilla-beta talos g2 ScriptFactory
+Rev5 MacOSX Yosemite 10.10.5 mozilla-beta talos g1 ScriptFactory
+Rev5 MacOSX Yosemite 10.10.5 mozilla-beta talos svgr ScriptFactory
+Rev5 MacOSX Yosemite 10.10.5 mozilla-beta talos dromaeojs ScriptFactory
+Rev5 MacOSX Yosemite 10.10.5 mozilla-beta talos tp5o ScriptFactory
+Rev5 MacOSX Yosemite 10.10.5 mozilla-beta talos other ScriptFactory
+Rev5 MacOSX Yosemite 10.10.5 mozilla-beta talos chromez ScriptFactory

We should still have talos tests on 10.10 on mozilla-beta and aurora.  All the changes should be on trunk branches, not on older versions.  Also, as a side note, the merge happened today so you may have to change checking for 44 to 45 

http://whattrainisitnow.com/

	Alin Selagea [:aselagea] Assignee
	Comment 54 • 9 years ago

We only keep the non-trunk talos jobs on 10.10:

trunk_yosemite)[alin.selagea@dev-master2.bb.releng.use1.mozilla.com trunk_yosemite]$ cat new | grep '10.10 ' | grep talos
Rev5 MacOSX Yosemite 10.10 mozilla-aurora talos g2 ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-aurora talos g1 ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-aurora talos svgr ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-aurora talos dromaeojs ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-aurora talos tp5o ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-aurora talos other ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-aurora talos chromez ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-beta talos g2 ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-beta talos g1 ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-beta talos svgr ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-beta talos dromaeojs ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-beta talos tp5o ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-beta talos other ScriptFactory
Rev5 MacOSX Yosemite 10.10 mozilla-beta talos chromez ScriptFactory

	Kim Moir [:kmoir] ET
	Comment 55 • 9 years ago

Comment on attachment 8681952 [details] [diff] [review]
trunk_enable_r7_disable_r5.patch

Looks good.  We'll wait to until the two remaining bugs that are tracking test failures that block on bug 1210395 are resolved before landing this patch

	Kim Moir [:kmoir] ET
	Comment 56 • 9 years ago

Looks like the last patch is bitrotten, need to fix it so it will land cleanly on the current code base.

	Alin Selagea [:aselagea] Assignee
	Comment 57 • 9 years ago

I added the updated version of the patch. 

This may not be the final one too, as /buildbot-configs/file/tip/mozilla-tests/config.py file can suffer some changes until this patch will be landed (e.g. bug 1223072)

	Joel Maher ( :jmaher ) (UTC -8)
	Comment 58 • 9 years ago

the last issue on the tests is close to having a patch ready.  I assume by Monday we will be all green on tests.  Will we be ready with the configs and machines to make this happen starting early next week?

	Chris Cooper [:coop] (he/him) Reporter
	Comment 59 • 9 years ago

(In reply to Joel Maher (:jmaher) from comment #58)
> the last issue on the tests is close to having a patch ready.  I assume by
> Monday we will be all green on tests.  Will we be ready with the configs and
> machines to make this happen starting early next week?

Yes, our plan is to push forward with deployment on Monday.

	Joel Maher ( :jmaher ) (UTC -8)
	Comment 60 • 9 years ago

we might not be ready Monday- but ideally we will have the last issues resolved by then.

	Chris Cooper [:coop] (he/him) Reporter
	Comment 61 • 9 years ago

Joel: is there a reason why we wouldn't just flip the switch and hide the broken test? The longer we wait, the higher the chance that something else breaks before we flip the switch.

	Joel Maher ( :jmaher ) (UTC -8)
	Comment 62 • 9 years ago

it isn't a single test, it is an entire directory of tests we would need to disable.  We can do that, but need to get it landed prior to the switch.  If we would have done that a week earlier we could have avoided this trap- so I am on board with solving this, I just want to make sure we have a realistic plan.

	Kim Moir [:kmoir] ET
	Comment 63 • 9 years ago

Comment on attachment 8687133 [details] [diff] [review]
trunk_enable_r7_disable_r5.patch

Looks good, did you do a builder diff to ensure that that the builders still look sane?

	Alin Selagea [:aselagea] Assignee
	Comment 64 • 9 years ago

(In reply to Kim Moir [:kmoir] from comment #63)
> Comment on attachment 8687133 [details] [diff] [review]
> trunk_enable_r7_disable_r5.patch
> 
> Looks good, did you do a builder diff to ensure that that the builders still
> look sane?

Yeah, I did, things look fine.

	Chris Cooper [:coop] (he/him) Reporter
	Comment 65 • 9 years ago

Comment on attachment 8687133 [details] [diff] [review]
trunk_enable_r7_disable_r5.patch

https://hg.mozilla.org/build/buildbot-configs/rev/2a66ca6c29a9

	Release Engineering SlaveAPI Service
	Comment 66 • 9 years ago

In production: https://hg.mozilla.org/build/buildbot-configs/rev/2a66ca6c29a9

	Joel Maher ( :jmaher ) (UTC -8)
	Comment 67 • 9 years ago

is there anything left to do here?

	Kim Moir [:kmoir] ET
	Comment 68 • 9 years ago

We need to land patches to add the remaining batch of yosemite machines to slavealloc and the range of machines in production config.py (machines > 64)

Alin or Vlad could you work on these patches?

	Joel Maher ( :jmaher ) (UTC -8)
	Comment 70 • 9 years ago

Comment on attachment 8689529 [details] [diff] [review]
bug1203128seta.patch

Review of attachment 8689529 [details] [diff] [review]:
-----------------------------------------------------------------

this looks good, I am not 100% certain if this will do the trick, but it does seem like it.

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 71 • 9 years ago

Attached you can find the production_config.py where yosemite range has been increased to 200

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 72 • 9 years ago

Attached the correct patch

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 73 • 9 years ago

Added the csv file to add them also in slavealloc DB.

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 74 • 9 years ago

Attached you can find the inserts for data.sql

	Vlad Ciobancai [:vladC] [:vciobancai]
	Comment 75 • 9 years ago

Attached the correct patch, there was some mistakes when I created the inserts

	Alin Selagea [:aselagea] Assignee
	Comment 76 • 9 years ago

Added the new slaves to graph DBs (both prod and staging).

	Release Engineering SlaveAPI Service
	Comment 77 • 9 years ago

In production: https://hg.mozilla.org/build/buildbot-configs/rev/e5d571f51809

	Release Engineering SlaveAPI Service
	Comment 78 • 9 years ago

In production: https://hg.mozilla.org/build/buildbot-configs/rev/b27086167ddd

	Kim Moir [:kmoir] ET
	Comment 79 • 9 years ago

fix to address reconfig bustage

	Kim Moir [:kmoir] ET
	Comment 80 • 9 years ago

added machines 65-200 in slavealloc.  Saw that machines 64-74 had puppetized so enabled them

	Kim Moir [:kmoir] ET
	Comment 81 • 9 years ago

also enabled 75-86 as I see they are up too

	Amy Rich [:arr] [:arich]
	Comment 82 • 9 years ago

All of the r7s up through t-yosemite-r7-0112.test.releng.scl3.mozilla.com are now enabled.

	Amy Rich [:arr] [:arich]
	Comment 83 • 9 years ago

The following have now been enabled as well:

t-yosemite-r7-0113.test.releng.scl3.mozilla.com
t-yosemite-r7-0114.test.releng.scl3.mozilla.com
t-yosemite-r7-0115.test.releng.scl3.mozilla.com
t-yosemite-r7-0116.test.releng.scl3.mozilla.com
t-yosemite-r7-0117.test.releng.scl3.mozilla.com
t-yosemite-r7-0118.test.releng.scl3.mozilla.com
t-yosemite-r7-0119.test.releng.scl3.mozilla.com
t-yosemite-r7-0120.test.releng.scl3.mozilla.com
t-yosemite-r7-0121.test.releng.scl3.mozilla.com
t-yosemite-r7-0122.test.releng.scl3.mozilla.com
t-yosemite-r7-0123.test.releng.scl3.mozilla.com
t-yosemite-r7-0124.test.releng.scl3.mozilla.com
t-yosemite-r7-0125.test.releng.scl3.mozilla.com
t-yosemite-r7-0126.test.releng.scl3.mozilla.com
t-yosemite-r7-0127.test.releng.scl3.mozilla.com
t-yosemite-r7-0128.test.releng.scl3.mozilla.com
t-yosemite-r7-0129.test.releng.scl3.mozilla.com
t-yosemite-r7-0130.test.releng.scl3.mozilla.com
t-yosemite-r7-0131.test.releng.scl3.mozilla.com
t-yosemite-r7-0132.test.releng.scl3.mozilla.com
t-yosemite-r7-0133.test.releng.scl3.mozilla.com
t-yosemite-r7-0134.test.releng.scl3.mozilla.com
t-yosemite-r7-0135.test.releng.scl3.mozilla.com
t-yosemite-r7-0136.test.releng.scl3.mozilla.com

	Kim Moir [:kmoir] ET
	Comment 84 • 9 years ago

enable 10.10.5 on aurora and disable corresponding 10.10 now that jmaher has uplifted patched to aurora

	Kim Moir [:kmoir] ET
	Comment 85 • 9 years ago

builder diff to enable 10.10.5 on aurora

	Kim Moir [:kmoir] ET
	Comment 86 • 9 years ago

enable 10.10.5 on try by default and disable 10.6 once we have enough machine capacity

	Joel Maher ( :jmaher ) (UTC -8)
	Comment 87 • 9 years ago

Comment on attachment 8695862 [details] [diff] [review]
bug1203128aurora.patch

Review of attachment 8695862 [details] [diff] [review]:
-----------------------------------------------------------------

I could be missing something here- please tell me if my comments are wrong and I can r+.

::: mozilla-tests/config.py
@@ +2603,5 @@
>  BRANCHES['try']['platforms']['win32']['win7-ix']['opt_unittest_suites'] += WEB_PLATFORM_TESTS_CHUNKED_E10S
>  
>  ride_trains_branches = []
>  for name, branch in items_at_least(BRANCHES, 'gecko_version', 45):
>      ride_trains_branches.append(name)

can we get rid of ride_trains_branches and just use the new clause for at_least(version, 44) ?  I don't see us using ride_trains_branches anymore in this small context.

@@ +2608,5 @@
>  
> +ride_trains_branches_aurora = []
> +for name, branch in items_at_least(BRANCHES, 'gecko_version', 44):
> +    ride_trains_branches_aurora.append(name)
> +    

nit: extra whitespace on this blank line.

	Kim Moir [:kmoir] ET
	Comment 88 • 9 years ago

builder diff is the same

	Joel Maher ( :jmaher ) (UTC -8)
	Comment 89 • 9 years ago

Comment on attachment 8696029 [details] [diff] [review]
bug1203128aurora-2.patch

Review of attachment 8696029 [details] [diff] [review]:
-----------------------------------------------------------------

I looked over the code and this looks good.  The new variable name is much better.

::: mozilla-tests/config.py
@@ +2725,5 @@
>  
> +r7_inactive_branches = []
> +for name, branch in items_before(BRANCHES, 'gecko_version', 44):
> +    r7_inactive_branches.append(name)
> + 

nit- one small space on this blank line remains!

	Kim Moir [:kmoir] ET
	Comment 90 • 9 years ago

Comment on attachment 8696029 [details] [diff] [review]
bug1203128aurora-2.patch

fixed whitspace thanks jmaher!

	Kim Moir [:kmoir] ET
	Comment 91 • 9 years ago

mozilla-aurora now has 10.10.5 jobs in running in production, reconfig completed

	Kim Moir [:kmoir] ET
	Comment 92 • 9 years ago

patch to enable 10.10.5 on release

	Kim Moir [:kmoir] ET
	Comment 93 • 9 years ago

builder diff to enable on release

	Joel Maher ( :jmaher ) (UTC -8)
	Comment 94 • 9 years ago

Comment on attachment 8699477 [details] [diff] [review]
bug1203128release.patch

Review of attachment 8699477 [details] [diff] [review]:
-----------------------------------------------------------------

oh nice!  are there special exceptions for mozilla-release?  it seems like we could get this out and assume it will work.

	Kim Moir [:kmoir] ET
	Comment 95 • 8 years ago

fix builder name to reflect correct hardware rev

	Alin Selagea [:aselagea] Assignee
	Comment 96 • 8 years ago

Comment on attachment 8700186 [details] [diff] [review]
bug1203128buildername.patch

Review of attachment 8700186 [details] [diff] [review]:
-----------------------------------------------------------------

The patch looks good. 
I think we should do the same change for the Thunderbird builders (in thunderbird_config.py) to reflect the new hardware revision.

	Alin Selagea [:aselagea] Assignee
	Comment 97 • 8 years ago

Added the patch to update the Thunderbird builder names as well.

	Kim Moir [:kmoir] ET
	Comment 98 • 8 years ago

This can probably be closed