We should see which suites are failing and which tests within them (few if I recall correctly) and paste the URLs to the logs. The failing jobs can be found in here: https://tbpl.mozilla.org/?tree=Cedar&jobname=Rev3%20WINNT%206.2 We should then find the code of each failing test and file a bug for each one of them and CC the owner/reviewers of the test.

I'm not sure what those mochitest-2 failures are, so I haven't filed anything on them. For the ref tests, there are multiple failures in different kinds of tests, so we'll probably have to break those out under bug 847893. Most though seem to have to do with border radius tests.

jimm is actually finding owners for the test failures or fixing them himself. Here's our current list of failures (ignore the talos jobs): https://tbpl.mozilla.org/?tree=Cedar&jobname=Rev3%20WINNT%206.2 M-1: * https://tbpl.mozilla.org/php/getParsedLog.php?id=20534287&tree=Cedar&full=1 * bug 847879 - some of them can be fixed by IT by deploying a package M-2 debug: * https://tbpl.mozilla.org/php/getParsedLog.php?id=20560698&tree=Cedar&full=1 Reftests: * bug 847893 * https://tbpl.mozilla.org/php/getParsedLog.php?id=20534932&tree=Cedar&full=1 Reftests unaccelerated: * 4 failures

I would like to enable win8 on m-i, m-c and try to produce load on the machines. I will hide those jobs that are perma-oranges. I would like to do it by tomorrow. Please let me know if you have any objections.

I've merged mc to Cedar to see where we are with tests. I will also be enabling tests on the mentioned branches in the next hour (and hide oranges).

Jobs are running on mozilla-central: https://tbpl.mozilla.org/?jobname=winnt%206.2 A few orange, but may be intermittent - have retriggered before I hide.

(In reply to Armen Zambrano G. [:armenzg] from comment #5) > I would like to enable win8 on m-i, m-c and try to produce load on the > machines. > I will hide those jobs that are perma-oranges. > > I would like to do it by tomorrow. Please let me know if you have any > objections. What should we use as opts on try to get these to run?

(In reply to Ed Morley [:edmorley UTC+0] from comment #7) > Jobs are running on mozilla-central: > https://tbpl.mozilla.org/?jobname=winnt%206.2 > > A few orange, but may be intermittent - have retriggered before I hide. M2 and reftests should be hidden for now. (In reply to Jim Mathies [:jimm] from comment #8) > (In reply to Armen Zambrano G. [:armenzg] from comment #5) > > I would like to enable win8 on m-i, m-c and try to produce load on the > > machines. > > I will hide those jobs that are perma-oranges. > > > > I would like to do it by tomorrow. Please let me know if you have any > > objections. > > What should we use as opts on try to get these to run? try: -b do -p win32 -u all -t all I think we can only choose on which platform to build but not limit which platforms does the build trigger.

(In reply to Armen Zambrano G. [:armenzg] from comment #9) > I think we can only choose on which platform to build but not limit which > platforms does the build trigger. In other words, I don't even need to make changes to trychooser IIUC.

(In reply to Armen Zambrano G. [:armenzg] from comment #9) > (In reply to Ed Morley [:edmorley UTC+0] from comment #7) > > Jobs are running on mozilla-central: > > https://tbpl.mozilla.org/?jobname=winnt%206.2 > > > > A few orange, but may be intermittent - have retriggered before I hide. > > M2 and reftests should be hidden for now. I've also hidden debug and pgo reftests (just opt were hidden), since they are permaorange too.

Also on inbound.

Please can someone file/investigate: https://tbpl.mozilla.org/php/getParsedLog.php?id=20649616&tree=Firefox https://tbpl.mozilla.org/php/getParsedLog.php?id=20649432&tree=Firefox

Also: https://tbpl.mozilla.org/php/getParsedLog.php?id=20645880&tree=Firefox

jimm, do you have any idea why this is coming up? http://cl.ly/NYjF (In reply to Ed Morley [:edmorley UTC+0] from comment #13) > Please can someone file/investigate: > https://tbpl.mozilla.org/php/getParsedLog.php?id=20649432&tree=Firefox > I disabled t-w864-ix-003 since it is burning more jobs than necessary.

Ah great - thank you :-)

bug 851203 - I have disabled machines 1 to 16 as requested by Q. He says that those machines are missing .net.

We're currently seeing a lot talos timeouts on these (bug 798219).

Sorry, we've made Win8 not shown by default on m-c and inbound. The intermittent failure rate is too high :(.

(In reply to Ryan VanderMeulen [:RyanVM] from comment #19) > Sorry, we've made Win8 not shown by default on m-c and inbound. The > intermittent failure rate is too high :(. Aside from the remaining reftest problems (one bug left!), what seems to be the problem. Looking across inbound the tests appear pretty stable to me.

For whoever wants to see them: https://tbpl.mozilla.org/?tree=Mozilla-Inbound&jobname=WINNT 6.2&showall=1 There are intermittent timeouts, leaks and crashes. Unfortunately the screenshots don't show anything. Debug m2 and debug m-other are most repeating intermittent offenders. All other test suites are quite green and very low ratio of intermittent oranges. Could we please re-enable all other suites except: * debug m2 * debug m-o * opt/deb reftests?

* debug m2 * debug m-o These don't appear to be any more failure prone than Win7 tests, imho.

I would like to enable everything that is green and leave the intermittent suites and the oranges to be debated. Objections?

wfm :-)

(we can always iterate more later)

Done: https://tbpl.mozilla.org/?tree=Mozilla-Inbound&jobname=WINNT%206.2 https://tbpl.mozilla.org/?jobname=WINNT%206.2 Please if there are other intermittent oranges hide only those and report in here.

The last two ref test fixes should land as soon as inbound is open. AFAIC that's all the work we'll do here, unless tree managers feel some other test suite still needs work? Everything remaining that blocks this bug is intermittent on other platforms as well.

This looks new for win8 (had similar on other platforms but not for 12-18 months): https://tbpl.mozilla.org/php/getParsedLog.php?id=20773068&tree=Firefox { 03:01:20 WARNING - TEST-UNEXPECTED-FAIL | chrome://mochitests/content/browser/browser/components/privatebrowsing/test/browser/perwindow/browser_privatebrowsing_placestitle.js | Test timed out }

Bug 686828 also came back to life on Win8 recently.

Win8 pgo/opt m-2 and mochitest-other were both hidden on inbound but visible on m-c; have just unhidden them on inbound also. Note: _debug_ mochitest-other and m-2 are still hidden on both. Looking at m-oth there are new win8 specific failures afaict.

All deps closed, no activity, closing.