Here are the logs from that GCE instance: https://drive.google.com/open?id=1rg7auzLdwC-Wi-jpxkDqa713Xp8cV5RN

mmoroz@ could you provide more details on the setup of that GCE instance? Which Linux distro is it running?

Also, the output directory (the archive you provided) doesn't include stdout, which has some more information (e.g. at the beginning it logs some platform info, flags, etc.). Could you find the stdout?

I know the lifecycle management of wptserve could be improved. I tried last time but failed because I accidentally broke Windows (https://crrev.com/c/1030570). I can try again, but I'm not sure whether that'd solve the problem as I don't yet understand the root cause of your issue.

Ubuntu 16. More precisely:

Linux code-coverage-linux-0005 4.13.0-1015-gcp #19-Ubuntu SMP Fri May 4 08:43:58 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux


I'm afraid that stdout is lost at this point, but I'll copy that from another loop the bot will be running soon.

This might not be an issue after new options in
https://chrome-internal.googlesource.com/chrome/tools/code-coverage/+/master/scripts/run_layout_tests.bash#18

One report came out fine, will see if things remain stable throughout day and then close bug.

Re 5#: what were the new options? --disable-breakpad? That shouldn't affect wptserve. And as all process hanging issues, it might not be deterministic.

--skip-failing-tests
--no-show-results
lower --child-processes (16 instead of 32).

i got it from valgrind tests
https://cs.chromium.org/chromium/src/third_party/libyuv/tools_libyuv/valgrind/chrome_tests.py?rcl=a9626b9daf62a9b260737e9c2de821ad087b19a1&l=565

previously all five bots were getting stuck, so lets see how it goes today.

Works fine now. will continue to monitor.