On Mon, Jun 17, 2019 at 5:01 PM Victor Stinner <vstin...@redhat.com> wrote: > > It is the first for a very long time that I see a buildbot stuck in a > loop and don't produce any output. The main test runner process should > write the progress every 30 seconds. But here the job ran for 3 days. > Maybe it is a regression un test_asyncio, but it doesn't explain why > the main process fails to report update. > > 0:00:06 load avg: 0.58 [ 8/423] test_pyexpat passed > 0:00:07 load avg: 0.58 [ 9/423] test_aifc passed > running: test_asyncio (37 sec 198 ms), test_html (30 sec 59 ms) > remoteFailed: [Failure instance: Traceback (failure with no frames): > <class 'twisted.internet.error.ConnectionLost'>: Connection to the > other side was lost in a non-clean fashion. > ] >
Yeah, I wasn't able to see anything in there either. The final disconnect was probably as I powered down the VM itself - the process was, apparently, unable to cleanly terminate. I'm not sure whether a TCP RESET was sent during shutdown, or if the connection stayed open until the build master next tried to ping it, but I suspect the former. In any case, there was no application-level shutdown. > CIs are great. Every day, a new surprise! Every day I open my PyBB email > folder, I am excited, what will we get today?? You don't say! That's like my reason for subscribing to the tz-announce list - what new timezone or DST insanity will we see next? Tz-announce happens maybe ten times a year, but it's such gloriously ridiculous stuff that it's worth it. (Egypt, I shall never forget the day you reminded me that there were still worse things that could be done with DST.) ChrisA _______________________________________________ Python-Buildbots mailing list Python-Buildbots@python.org https://mail.python.org/mailman/listinfo/python-buildbots
