Closed Bug 1397829 Opened 8 years ago Closed 8 years ago

add buildernames for stylo-disabled tests which run on hardware, win7 and win10

Categories

(Infrastructure & Operations Graveyard :: CIDuty, task, P1)

Tracking

(firefox-esr52 unaffected, firefox55 unaffected, firefox56 unaffected, firefox58 fixed)

RESOLVED FIXED
Tracking Status
firefox-esr52 --- unaffected
firefox55 --- unaffected
firefox56 --- unaffected
firefox58 --- fixed

People

(Reporter: jmaher, Assigned: aselagea)

References

Details

Attachments

(12 files, 7 obsolete files)

7.28 KB, patch
Details | Diff | Splinter Review
134.17 KB, text/plain
jmaher
: feedback+
Details
11.30 KB, patch
kmoir
: review+
kmoir
: feedback+
aselagea
: checked-in+
Details | Diff | Splinter Review
61.86 KB, text/plain
jmaher
: feedback+
Details
1.38 KB, patch
jmaher
: review+
Details | Diff | Splinter Review
1.30 KB, patch
aobreja
: review+
aselagea
: checked-in+
Details | Diff | Splinter Review
6.82 KB, patch
aobreja
: review+
aselagea
: checked-in+
Details | Diff | Splinter Review
2.58 KB, patch
kmoir
: review+
aselagea
: checked-in+
Details | Diff | Splinter Review
2.39 KB, patch
kmoir
: review+
aselagea
: checked-in+
Details | Diff | Splinter Review
538 bytes, patch
aobreja
: review+
aselagea
: checked-in+
Details | Diff | Splinter Review
1.58 KB, patch
aobreja
: review+
aselagea
: checked-in+
Details | Diff | Splinter Review
4.10 KB, patch
aobreja
: review+
aselagea
: checked-in+
Details | Diff | Splinter Review
we noticed that the jobs we run on hardware are not running in parallel with -stylo-disabled, this is due to the buildernames not existing (and some small tweaks needed in taskcluster configs)
I am not sure if this is even close, :kmoir- I have bugged you quite frequently the last 2 weeks, can I bug you some more to help me get a builderdiff and polish up this patch?
Assignee: nobody → jmaher
Status: NEW → ASSIGNED
Attachment #8905589 - Flags: review?(kmoir)
I'll look at this this afternoon, have to address a few other bugs first.
:ting, I also see win10-asan having troubles with buildbot-bridge. Are win10-asan tests to be run on hardware or vm? Do all the tests currently run on it (or should they be)? I ask so we need to determine if win10-asan should be added to the buildbot changes and run via buildbot-bridge, or hacked around in taskcluster.
Flags: needinfo?(janus926)
Attached patch bug1397829-2.patch (obsolete) — Splinter Review
Attachment #8905589 - Attachment is obsolete: true
Attachment #8905589 - Flags: review?(kmoir)
Attached file bug1397829builder.diff (obsolete) —
builder diff
Attached patch bug1397829puppet.patch (obsolete) — Splinter Review
in this case we want to solve this for unittests, not talos tests; and stylo -> stylo-disabled or some variant. Is there a need to do the full copy like we did in the win10 patch last week?
Attached patch bug1397829tools.patch (obsolete) — Splinter Review
Attachment #8906059 - Flags: feedback?(jmaher)
Comment on attachment 8906059 [details] bug1397829builder.diff this is specifying a config of 'stylo' and it should be 'stylo-disabled', likewise this is just talos jobs and we need all the jobs (like mochitest, reftest)
Attachment #8906059 - Flags: feedback?(jmaher) → feedback-
This because the name you gave it here was "Windows 10 64-bit stylo" PLATFORMS['win64-stylo-disabled']['win64_hw_stylo'] = {'name': 'Windows 10 64-bit stylo', I changed it to 'Windows 10 64-bit stylo disabled' and the diff looks fine. I'll attach a new builder diff and patch
Attached patch bug1397829-3.patch (obsolete) — Splinter Review
Attachment #8906056 - Attachment is obsolete: true
does this also include unittests, not just talos? from looking over the patch it looks like just talos.
Sorry, I read hardware tests as talos tests. If this patch looks okay for talos I can modify it to include unit tests.
Attachment #8906059 - Attachment is obsolete: true
thanks for the new builderdiff, the 'platform' name looks right now- the branches are good. to be safe we should do beta (or ensure that we migrate to beta when 57 goes there next week. Otherwise this is good and we should move forward with the unittest names! Thanks for your help :)
(In reply to Joel Maher ( :jmaher) (UTC-5) from comment #3) > :ting, I also see win10-asan having troubles with buildbot-bridge. Are > win10-asan tests to be run on hardware or vm? Do all the tests currently > run on it (or should they be)? I ask so we need to determine if win10-asan > should be added to the buildbot changes and run via buildbot-bridge, or > hacked around in taskcluster. IIRC, all the tests run on hardware. I am not sure why win64-asan is affected by buildbot changes, aren't the tests run by taskcluster? But anyway, please make win64-asan sync with linux64-asan if you can. Hopefully win64-asan will promote to tier-2 in next quarter.
Flags: needinfo?(janus926)
(In reply to Ting-Yu Chou [:ting] from comment #15) > IIRC, all the tests run on hardware. I was wrong. Just checked, win64-asan runs common-tests [1] and only mochitest-chrome [2], mochitest-clipboard [3] have virtualization "hardware". I am not sure, but I expect the other tests without defining virtualization run on VM. [1] http://searchfox.org/mozilla-central/rev/00fa5dacedb925022f53d025121f1a919508e7ce/taskcluster/ci/test/test-platforms.yml#250 [2] http://searchfox.org/mozilla-central/rev/00fa5dacedb925022f53d025121f1a919508e7ce/taskcluster/ci/test/tests.yml#665 [3] http://searchfox.org/mozilla-central/rev/00fa5dacedb925022f53d025121f1a919508e7ce/taskcluster/ci/test/tests.yml#725
we have only had win10 hardware support for the last few months and this was stood up only for Talos performance tests- so I will assume this needs the same vm || hardware as desktop unittests- right now desktop unittests don't run on windows 10 reliably, so it is sort of unrealistically to make win10-asan run our unittests and expect them to pass- although we do hope to have that all fixed in the next month or so. :kmoir, in the patch for unittests, can we have a win10-asan platform as well?
Flags: needinfo?(kmoir)
:kmoir, I understand you have been busy with the uplift, is there anyone else who can help here?
:jmaher I apologize for the delay. I was on course earlier this week and am on releaseduty so trying to catch up with other work. I'll look at it now while my merge scripts are running and if I can't get complete patches today I'll try to get someone from buildduty to look at it.
Flags: needinfo?(kmoir)
Mihai do you know if the buildduty folks might have time to look at this and help Joel? If not, I can try to look at it next week when releaseduty calms down.
Flags: needinfo?(mtabara)
I'll look at this today and see if I can help somehow.
Flags: needinfo?(mtabara)
Thanks Alin, let me know if you have questions!
Looked at the existing patches and tried a couple of solutions to enable the same unittests on "win64_hw_stylo" as the regular ones we have on Windows 10, but had little luck today :| If it's okay with you, I'll resume working on this on Monday when I return. Meanwhile, I'd like to clarify two things: a) is my approach above correct? do we need 'stylo disabled' variant for each Windows 10 test we run on hardware? Do we want this for Win 7 as well? b) what tests do we want for "win10-asan"?
Flags: needinfo?(jmaher)
For all tests that run on windows10-hw (including unittest and talos), we want the ability to schedule a |windows10-stylo-disabled| and a |windows10-asan| copy. For windows7, we also need a |windows7-stylo-disabled| copy, but not an asan copy. All changes should apply to version 57+, so this is mozilla-beta, try, mozilla-central, autoland, mozilla-inbound, and other trunk based branches.
Flags: needinfo?(jmaher)
Alin, how is it going on this bug? catlee asked me about it today, let me know if I can help in any way kmoir: do you / aselagea think you can get https://bugzilla.mozilla.org/show_bug.cgi?id=1397829 done in the next day or so? Bug 1397829 — ASSIGNED, jmaher@mozilla.com — add buildernames for stylo-disabled tests which run on hardware, win7 and win10 12:18 PM <kmoir> Kim Moir let me ask him the status of this work
Flags: needinfo?(aselagea)
I looked at Alin's dev-master but I wasn't sure of the state of his patches so I'll talk to him tomorrow about it and see how I can help if required.
Also, as a side note, catlee mentioned this bug has a script that might be useful for debugging buildbot bridge errors bug 1381597
Attached patch bug1397829-4.patch (obsolete) — Splinter Review
Note: this patch adds many builders, so I had to manually increase the builder limit in master/master.cfg on my master in order to test the patches. Also, we already have a "stylo-disabled" variant for the talos tests, so I wonder if we should only run that variant for the corresponding Win7 and Win10 platform where we want stylo disabled. We'll probably need to do some adjustments in TaskCluster too, didn't have the chance to take a look at that yet.
Flags: needinfo?(aselagea)
Attachment #8909759 - Flags: feedback?(kmoir)
Attached file builder_diff.txt
Builder diff for the patch above.
Comment on attachment 8909759 [details] [diff] [review] bug1397829-4.patch Note: this patch adds many builders, so I had to manually increase the builder limit in master/master.cfg on my master in order to test the patches. >> If we are hitting builder limits we may want reconfigure some windows masters to just run certain types of tests in production_masters.json otherwise the master will hit performance issues due to the number of builders. Before we do that though, I'd like to ask jmaher to review the builder list and see if it is correct. Perhaps we could just enable these tests on try first and see how the patches work and then we could deal with the number of builders issue after that. Also, we already have a "stylo-disabled" variant for the talos tests, so I wonder if we should only run that variant for the corresponding Win7 and Win10 platform where we want stylo disabled. >>This is a good approach We'll probably need to do some adjustments in TaskCluster too, didn't have the chance to take a look at that yet. Other feedback, for this bit starting here trunk_branches = [] +for name, branch in items_at_least(BRANCHES, 'gecko_version', 57): + trunk_branches.append(name) + +platforms = ['win32-stylo-disabled', 'win64-stylo-disabled', 'win64-asan'] +for branch in BRANCHES.keys(): + for platform in platforms: + if platform not in BRANCHES[branch]['platforms'].keys(): + continue + + # Windows 7 + if platform == 'win32-stylo-disabled': + BRANCHES[branch]['platforms'][platform]['talos_slave_platforms'] = ['win7_ix_stylo_disabled'] + # copy the regular tests we run on Windows 7 HW + BRANCHES[branch]['platforms'][platform]['win7_ix_stylo_disabled'] = deepcopy(BRANCHES[branch]['platforms']['win32']['win7_ix']) + # Windows 10 + if platform == 'win64-stylo-disabled': + BRANCHES[branch]['platforms'][platform]['talos_slave_platforms'] = ['win10_64_stylo_disabled'] + # copy the regular tests we run on Windows 10 HW + BRANCHES[branch]['platforms'][platform]['win10_64_stylo_disabled'] = deepcopy(BRANCHES[branch]['platforms']['win64']['win10_64']) + if platform == 'win64-asan': + BRANCHES[branch]['platforms']['win64-asan']['win10_64_asan'] = deepcopy(BRANCHES[branch]['platforms']['win64']['win10_64']) + + if branch not in trunk_branches: + del BRANCHES[branch]['platforms'][platform] Instead of defining the truck branches you could probably do something like platforms = ['win32-stylo-disabled', 'win64-stylo-disabled', 'win64-asan'] for name, branch in items_at_least(BRANCHES, 'gecko_version', 57): for platform in platforms: then the rest of your if statements, then you don't have the delete them from non-trunk branches
Attachment #8909759 - Flags: feedback?(kmoir) → feedback+
Comment on attachment 8909760 [details] builder_diff.txt Jmaher is this the builder list you are expecting? Alin is running into some builder limits so we wanted to ensure that this is correct. Would it make sense to enable these builders on try only first to see if they tests run as expected and then enable on other branches?
Attachment #8909760 - Flags: feedback?(jmaher)
Comment on attachment 8909760 [details] builder_diff.txt a few optimizations: 1) I am not sure about the oak branch - do we need this there? 2) win asan, could be try and mozilla-central only if we need to save builders 3) .*web-platform-tests.* can be ignored, this is VM only 4) .*devtools-chrome.* can be ignored, this is VM only we don't need to optimize, but those are the quick wins- I could nitpick, but that gets complicated. Thanks for getting this so far!
Attachment #8909760 - Flags: feedback?(jmaher) → feedback+
Alin do you have any more questions on this bug?
Flags: needinfo?(aselagea)
Sorry for the late response, I was hitting some issues while working on this patch since it turns out it's a bit more complicated than it seems. What I tried to do in this patch is to adjust the builder list to match Joel's optimizations mentioned above.
Flags: needinfo?(aselagea)
Attachment #8910859 - Flags: review?(kmoir)
Attached file builder_diff2.txt
(In reply to Kim Moir [:kmoir] ET from comment #30) > Comment on attachment 8909759 [details] [diff] [review] > bug1397829-4.patch > Also, we already have a "stylo-disabled" variant for the talos tests, so I > wonder if we should only run that variant for the corresponding Win7 and > Win10 platform where we want stylo disabled. > > >>This is a good approach Tried to do this, but it seems it's a bit more difficult to accomplish that due to the way these talos tests are defined. There were 54 such tests added with my latest patch, so I preferred not to touch them for now. > Instead of defining the truck branches you could probably do something like > platforms = ['win32-stylo-disabled', 'win64-stylo-disabled', 'win64-asan'] > for name, branch in items_at_least(BRANCHES, 'gecko_version', 57): > for platform in platforms: > > then the rest of your if statements, then you don't have the delete them > from non-trunk branches That was my intention too, but we define the new platforms at the beginning of the script and they are then updated in various places in the code (so it's not like I'm adding then deleting something in the same snippet). Rather than touching various parts of the code to prevent adding builders on these platforms, I preferred to simply remove what's not needed.
Comment on attachment 8910860 [details] builder_diff2.txt Could you verify these builder changes, you probably have more context than me on the changes required
Attachment #8910860 - Flags: feedback?(jmaher)
Comment on attachment 8910859 [details] [diff] [review] bug1397829-5.patch I think this look good to me, but this is dependent on Joel's verification that the builder list looks good. Did you have the adjust the number of builders or did the optimizations reduce it sufficiently?
Attachment #8910859 - Flags: feedback+
Comment on attachment 8910860 [details] builder_diff2.txt this is looking great; possibly we are not at builderlimits with this list? Looking forward to this landing.
Attachment #8910860 - Flags: feedback?(jmaher) → feedback+
Comment on attachment 8910859 [details] [diff] [review] bug1397829-5.patch since Jmaher approved the builder list, I will r+. I think the tools and puppet patches need to be updated as well. Thank you so much Alin! Much appreciated!
Attachment #8910859 - Flags: review?(kmoir) → review+
(In reply to Joel Maher ( :jmaher) (UTC-5) from comment #39) > Comment on attachment 8910860 [details] > builder_diff2.txt > > this is looking great; possibly we are not at builderlimits with this list? > > Looking forward to this landing. We are not hitting any builder limit with this patch. I think we should also check the TC configs before landing this though.
I think you might have to do add the new platforms for talos here so they go through bb bridge https://dxr.mozilla.org/mozilla-central/source/taskcluster/taskgraph/transforms/job/mozharness_test.py?q=mozharness_test.py&redirect_type=direct#476 If you get the buildbot changes landed and the builders enabled jmaher can probably run a try run for the tc changes to ensure they work. The buildbot changes won't schedule the tests, just enable the builders.
This should adjust the TC buildernames for talos tests so they match the ones existing in BB.
Attachment #8911722 - Flags: review?(jmaher)
I still believe that we should stick to either "stylo disabled" platforms or "stylo disabled" tests variants. What we have now is a mix of them, which may be confusing.
Updated patch for puppet.
Attachment #8906060 - Attachment is obsolete: true
Attachment #8906086 - Attachment is obsolete: true
Attachment #8909759 - Attachment is obsolete: true
Attachment #8911749 - Flags: review?(aobreja)
Updated patch for tools.
Attachment #8906069 - Attachment is obsolete: true
Attachment #8911750 - Flags: review?(aobreja)
Comment on attachment 8911749 [details] [diff] [review] bug1397829_puppet.patch Looks good.
Attachment #8911749 - Flags: review?(aobreja) → review+
Comment on attachment 8911750 [details] [diff] [review] bug1397829_tools.patch Looks good.
Attachment #8911750 - Flags: review?(aobreja) → review+
Comment on attachment 8911722 [details] [diff] [review] adjust_tc_buildernames.patch Review of attachment 8911722 [details] [diff] [review]: ----------------------------------------------------------------- this looks nice and simple
Attachment #8911722 - Flags: review?(jmaher) → review+
Assignee: jmaher → aselagea
Pushed by aselagea@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/23494177c6b2 add buildernames for stylo-disabled tests which run on hardware, win7 and win10, r=jmaher
Status: ASSIGNED → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
This work is not yet done, reopening.
Status: RESOLVED → REOPENED
Keywords: leave-open
Resolution: FIXED → ---
Few updates here: Status: - created and landed patches to add these buildernames in Buildbot configs and various other places - created a patch to adjust the buildernames in TaskCluster => no new jobs in TH yet Next steps: 1. Talos: - we have "stylo-disabled" talos tests that run on the regular platforms - we also have dedicated "stylo-disabled" platforms, but the above mentioned "stylo-disabled" talos tests don't run there [1] Should we try to move everything that's "stylo-disabled" to dedicated platforms? 2. do we actually want the new unittests to run via BBB and appear under those dedicated platforms? [1] https://treeherder.mozilla.org/#/jobs?repo=mozilla-central&filter-searchStr=stylo-disabled
Component: General → Buildduty
Priority: -- → P1
QA Contact: catlee
Had a meeting with :jmaher where we discussed about the next steps required on this bug. a) at this point, we have two sets of Windows unittests running on BB - one of them should be "stylo-disabled", but it's being displayed under the wrong platform. e.g. https://treeherder.mozilla.org/#/jobs?repo=mozilla-central&revision=35fbf14b96a633c3f66ea13c1a163a3f3a4219b9&filter-searchStr=windows%20x64%20reftest Even though we added the required buildernames above, some pieces are still missing and I will focus on solving that. b) there's also some confusion regarding the way "stylo-disabled" talos tests are being run: - they should be displayed under the appropriate platform - we shouldn't run regular talos tests for "stylo-disabled" platforms Will file a new bug to address that.
Blocks: 1403600
(In reply to Alin Selagea [:aselagea][:buildduty] from comment #57) > a) at this point, we have two sets of Windows unittests running on BB - one > of them should be "stylo-disabled", but it's being displayed under the wrong > platform. > e.g. > https://treeherder.mozilla.org/#/jobs?repo=mozilla- > central&revision=35fbf14b96a633c3f66ea13c1a163a3f3a4219b9&filter- > searchStr=windows%20x64%20reftest > Even though we added the required buildernames above, some pieces are still > missing and I will focus on solving that. > > b) there's also some confusion regarding the way "stylo-disabled" talos > tests are being run: > - they should be displayed under the appropriate platform > - we shouldn't run regular talos tests for "stylo-disabled" platforms > Will file a new bug to address that. Now that we have bug 1403600 for the stylo-disabled talos tests and platform names, can this bug be closed?
(In reply to Chris Peterson [:cpeterson] from comment #59) > Now that we have bug 1403600 for the stylo-disabled talos tests and platform > names, can this bug be closed? No, this is still needed since we also want the regular unittests to run on the corresponding "stylo-disabled" platform.
Spent some time investigating this and it seems there are a couple of problems we need to fix before getting this work done. First of them is a builder limit on BB: "AssertionError: t-w1064-ix-156 has 4099 builders; limit is 4084". Not sure why we hit it again (maybe due to further pushes in bb-configs since my last push), but since we'll also need to deal with bug 1403600 soon, we could take care of it first so that we don't run regular talos tests on "stylo-disabled" platforms. That will save us some builders and get us past the builder limit.
No longer blocks: 1403600
Depends on: 1403600
The test platforms defined in TC [1], have "stylo-disabled" in their name. When we retrieve the test variant, we use this method [2]. Since we initially specified "stylo disabled" as variant, we won't get any change in the builder name as that variant can't be found on the platform name. I also updated the configs to account for the changes required when computing the builder names for the unit tests as well. [1] https://dxr.mozilla.org/mozilla-central/source/taskcluster/ci/test/test-platforms.yml#85-132 [2] https://dxr.mozilla.org/mozilla-central/source/taskcluster/taskgraph/transforms/job/mozharness_test.py#70-74 @Kim: I'll address this review request to you. Please feel free to redirect it if you consider it necessary/you're too busy with release duty.
Attachment #8913629 - Flags: review?(kmoir)
Buildbot will need these small changes as well. NOTE: these last two patches will need to land *before* the patches that will come up in bug 1403600.
Attachment #8913630 - Flags: review?(aobreja)
Depends on: 1404303
Attachment #8913630 - Flags: review?(aobreja) → review+
Comment on attachment 8913629 [details] [diff] [review] bug1397829_fix_tc_buildernames.patch Thanks you both for working on these patches. We really appreciate all your work to test and implement this functionality. I also appreciate you picking up this work while I was on releaseduty.
Attachment #8913629 - Flags: review?(kmoir) → review+
Pushed by aselagea@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/349550665585 Fix bb buildernames for 'stylo-disabled' tests, r=kmoir
Attachment #8913629 - Flags: checked-in+
It seems I omitted to update the variant name in one place, here's a small fix for that.
Attachment #8914666 - Flags: review?(aobreja)
Comment on attachment 8914666 [details] [diff] [review] fix_test_variant.patch Looks good
Attachment #8914666 - Flags: review?(aobreja) → review+
Pushed by aselagea@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/a938d79a7ef8 fix variant name for 'stylo-disabled', r=aobreja
Attachment #8914666 - Flags: checked-in+
Comment on attachment 8914666 [details] [diff] [review] fix_test_variant.patch I had to back out this change since the "stylo-disabled" tests scheduled in TC were not starting in BB. Will investigate what else is needed and eventually test on try before landing this again.
Attachment #8914666 - Flags: checked-in+ → checked-in-
On a closer look, it seems those "stylo-disabled" unittests won't start since they're not on the bb masters - that is because the reconfigs fail with the following error: cd master && /builds/buildbot/tests1-windows/bin/buildbot checkconfig Traceback (most recent call last): File "/builds/buildbot/tests1-windows/lib/python2.7/site-packages/buildbot-0.8.2_hg_d769fbdcf8ee_production_0.8-py2.7.egg/buildbot/scripts/runner.py", line 1042, in doCheckConfig ConfigLoader(configFileName=configFileName) File "/builds/buildbot/tests1-windows/lib/python2.7/site-packages/buildbot-0.8.2_hg_d769fbdcf8ee_production_0.8-py2.7.egg/buildbot/scripts/checkconfig.py", line 31, in __init__ self.loadConfig(configFile, check_synchronously_only=True) File "/builds/buildbot/tests1-windows/lib/python2.7/site-packages/buildbot-0.8.2_hg_d769fbdcf8ee_production_0.8-py2.7.egg/buildbot/master.py", line 652, in loadConfig exec f in localDict File "/builds/buildbot/tests1-windows/master/master.cfg", line 94, in <module> c['slaves'].append(BuildSlave(name, SlavePasswords[slave_platform], max_builds=1, KeyError: 'win7_ix_stylo_disabled' make: *** [checkconfig] Error 1 This patch adds an entry for 'win7_ix_stylo_disabled' in Puppet.
Attachment #8915040 - Flags: review?(aobreja)
Attachment #8915040 - Flags: review?(aobreja) → review+
Comment on attachment 8915040 [details] [diff] [review] bug1397829_2_puppet.patch Landed this patch, checked that the Windows bb masters reconfigured successfully and can confirm these tests are now present on the said masters. Did a try push [1] and I can see "stylo-disabled" unittests being run via BBB and having the correct builder names. They will be moved to the correct platform in TH once bug is fixed. [1] https://treeherder.mozilla.org/#/jobs?repo=try&revision=efb95a9cacb2e0198a67066e57ad1b88e08adfce&filter-searchStr=stylo-disabled
Attachment #8915040 - Flags: checked-in+
According to [1], we can leave the "stylo-disabled" talos tests to run as they do now and not move them to separate platforms. Regular talos tests on "stylo-disabled" platforms are also not needed, so let's just remove them from BB to reduce the builder count. [1] https://bugzilla.mozilla.org/show_bug.cgi?id=1403600#c17
Attachment #8915610 - Flags: review?(aobreja)
Once we have the TH changes (bug 1404303) and the patch above landed, we can go on and re-land https://bugzilla.mozilla.org/attachment.cgi?id=8914666 above to have these unittests run under on the corresponding platform.
Comment on attachment 8915610 [details] [diff] [review] bug1397829_remove_talos.patch cat old | grep stylo-disabled | wc -l 1315 cat new | grep stylo-disabled | wc -l 959
Attachment #8915610 - Flags: review?(aobreja) → review+
Pushed by aselagea@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/e6050fffa0c5 add buildernames for stylo-disabled tests which run on hardware, r=aobreja
Attachment #8914666 - Flags: checked-in- → checked-in+
Okay, so all Treeherder, BB and TC changes mentioned above are now landed. Here's a push to m-i showing the "stylo-disabled" tests running at this point: https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=f232c635ecd51d1557683b0457c926e7e2f23128&filter-searchStr=stylo%20disabled Is there anything else needed here? As for the Windows 10 asan jobs, I checked the list of tasks and noticed we don't run any asan job for Windows. If those are still required to be enabled, we could probably treat that in a separate bug.
Flags: needinfo?(jmaher)
this is great, win-asan is try only and can be handled with a different bug.
Status: REOPENED → RESOLVED
Closed: 8 years ago8 years ago
Flags: needinfo?(jmaher)
Resolution: --- → FIXED
Removing leave-open keyword from resolved bugs, per :sylvestre.
Keywords: leave-open
Product: Release Engineering → Infrastructure & Operations
Product: Infrastructure & Operations → Infrastructure & Operations Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: