I'm seeing some errors running tests on Mountain Lion. This bug will be used to track their resolution, whether it be test setup or otherwise.

Similar to issues reported here https://discussions.apple.com/thread/4143805

I think the that is a red herring. In any case, the next error message is that build opens and tries to access localhost/page_load_test/svg/gearflowers.svg and nothing happens. Armen looked at this with me and looked at the yaml file and suggested that I give jmaher access to the machine to ensure that the test framework works on Mountain Lion. John could you ping we when you get a chance and I'll give you vnc passwd to the machine?

I will be happy to look at this. I assume tomorrow would be a better day. Just ping me first thing in the morning.

jmaher, kmoir: Whats left to do here?

We resolved the issue in comment 2. I'm just leaving this bug open until I've run the tests again on my slaves in the staging environment, which I'm working on today.

Here are the remaining issues when running tests The tests run fine except for the fact that they can't write to the graphing database. See bug 785498. Rev5 MacOSX Mountain Lion 10.8 mozilla-central talos svgr Rev5 MacOSX Mountain Lion 10.8 mozilla-central talos other_mac Rev5 MacOSX Mountain Lion 10.8 mozilla-central talos dromaeo Rev5 MacOSX Mountain Lion 10.8 mozilla-central talos dirty Rev5 MacOSX Mountain Lion 10.8 mozilla-central talos tpn Rev5 MacOSX Mountain Lion 10.8 mozilla-central talos nochromer Rev5 MacOSX Mountain Lion 10.8 mozilla-central talos chromer I ran multiple sendchanges simultaneously, forced some builds by hand. So I think once the graphing server issue is resolved, they should be okay, but of course I'll rerun the tests to confirm. Ben mentioned that a sendchange for a 10.7 opt build should initiate all the tests that we need to run. I notice that there are a lot of other build types that aren't initiated by this type of sendchange. If I force a build they fail to download, I assume this is because the they need to be seeded somehow, is this correct? Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test mochitests-1/5 Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test mochitests-2/5 Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test mochitests-3/5 Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test mochitests-4/5 Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test mochitests-5/5 Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test mochitest-other Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test reftest Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test crashtest Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test xpcshell Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test jsreftest Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test jetpack Rev5 MacOSX Mountain Lion 10.8 mozilla-central opt test peptest

So I ran all the tests today with the correct sendchanges and they were all green except for the ones listed in bug 786084.

This can be closed, the test failure issues are being tracked in bug 786084