As noted in https://bugzilla.mozilla.org/show_bug.cgi?id=1204281#c282: > something odd happened here on November 8th/9th (I think the 9th) and > bc3/bc4 on linux64 asan e10s is taking longer. Possibly investigating times > before/after the 9th and seeing if we are right at the limit for 3600 > seconds. Either more chunks or longer timeout- I prefer the more chunks > approach. I looked back as far as Nov 1 on m-c and found that asan bc3 and bc4 were long-running even then. There is intermittent variation in the job time, but it looks like that variation had existed well prior to Nov 8: I think we just need more capacity, and I also favor more chunks.

https://treeherder.mozilla.org/#/jobs?repo=try&revision=1c5e1ee33ae4a356dfac99af648afa3b73950501 ...unfortunately based on a rev that closed m-c (bug 1317296).

This increases mochitest-bc test chunks for linux64-asan only, from 7 to 10. Most chunks run in 20 to 30 minutes, but bc3 on e10s still takes about 45 minutes -- should be good enough to avoid timeouts. Is it okay to only adjust linux64-asan? That's the only platform hitting (or close to) timeouts, but I sometimes wonder if we should keep chunks the same across platforms. https://treeherder.mozilla.org/#/jobs?repo=try&revision=e10ec92fae78441bd23a85899c96ec7841b0cd29

Comment on attachment 8810624 [details] [diff] [review] increase mochitest-bc chunks to 10 for linux64-asan only Review of attachment 8810624 [details] [diff] [review]: ----------------------------------------------------------------- while it is nice to have parity between types, I don't think that is a requirement- we won't have parity between linux and osx/win for a long time and we haven't for the last 6 months or longer.

Pushed by gbrown@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/89700defd669 Increase test chunks for linux64-asan browser-chrome; r=jmaher