our current linux64 browser-chrome tests run on really slow single core VM instances- this takes a lot of time up in getting results and is not very realistic. When moving over, we see common failures in the above two tests. Here is a try push highlighting it: https://treeherder.mozilla.org/#/jobs?repo=try&revision=eb311273537cfcdca2cd373774658c75e9918730&selectedJob=137079382&filter-searchStr=debug%20%20x64 the log is pretty long, but bc2 on linux64, and bc10 on linux64-stylo-disabled you can see the failures.

:standard8- would you have any ideas on how to work around the timeout and failures in the above mentioned tests?

Simple bit first: browser_library_commands.js looks like fallout from browser_library_batch_delete.js failing. The info it is complaining about not being defined is in promiseBookmarksNotification(), which would imply that the observer hasn't been cleaned up. Since the previous test timed out waiting for that observer to give the right notification, that isn't a surprise. So browser_library_batch_delete.js is the one to think about here - I'm still thinking about what that could be.

I don't think that bug 1409771 will have fixed this, so I'm now running with latest central, and some debug on try to see if I can work out what's happening here: https://treeherder.mozilla.org/#/jobs?repo=try&revision=2b45da30cb57f4282d93cd0430864756f36d83a5

Joel, I've just done another try push without my debug, and both the one before, and this one appear to be all green: https://treeherder.mozilla.org/#/jobs?repo=try&revision=cffbaf1312ff30b448d10fd81989d6436aa560ac Please can you check what I've pushed, hence, if I've been testing the right thing? I guess it is possible that bug 1409771 or something else has fixed this in the meantime, I'm just a bit surprised.

:standard8- this looks to be working just fine; I am confused why you are not getting any failures given that it was "perma" fail for me before. I did double check that bc9 ran the test and that it was on the correct instance type. Maybe we can close this bug as worksforme

Ok, lets go WFM then, if it comes back we can always re-open.