On 16/03/17 04:42, Andres Freund wrote: > On 2017-03-15 20:28:33 -0700, Andres Freund wrote: >> Hi, >> >> I just unstuck a bunch of my buildfarm animals. That triggered some >> spurious failures (on piculet, calliphoridae, mylodon), but also one >> that doesn't really look like that: >> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=culicidae&dt=2017-03-16%2002%3A40%3A03 >> >> with the pertinent point being: >> >> ================== stack trace: >> pgsql.build/src/test/regress/tmp_check/data/core ================== >> [New LWP 1894] >> [Thread debugging using libthread_db enabled] >> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". >> Core was generated by `postgres: bgworker: logical replication launcher >> '. >> Program terminated with signal SIGSEGV, Segmentation fault. >> #0 0x000055e265bff5e3 in ?? () >> #0 0x000055e265bff5e3 in ?? () >> #1 0x000055d3ccabed0d in StartBackgroundWorker () at >> /home/andres/build/buildfarm-culicidae/HEAD/pgsql.build/../pgsql/src/backend/postmaster/bgworker.c:792 >> #2 0x000055d3ccacf4fc in SubPostmasterMain (argc=3, argv=0x55d3cdbb71c0) at >> /home/andres/build/buildfarm-culicidae/HEAD/pgsql.build/../pgsql/src/backend/postmaster/postmaster.c:4878 >> #3 0x000055d3cca443ea in main (argc=3, argv=0x55d3cdbb71c0) at >> /home/andres/build/buildfarm-culicidae/HEAD/pgsql.build/../pgsql/src/backend/main/main.c:205 >> >> it's possible that me killing things and upgrading caused this, but >> given this is a backend running EXEC_BACKEND, I'm a bit suspicous that >> it's more than that. The machine is a bit backed up at the moment, so >> it'll probably be a while till it's at that animal/branch again, >> otherwise I'd not have mentioned this. > > For some reason it ran again pretty soon. And I'm afraid it's indeed an > issue: > https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=culicidae&dt=2017-03-16%2003%3A30%3A02 >
Hmm, I tried with EXEC_BACKEND (and with --disable-spinlocks) and it seems to work fine on my two machines. I don't see anything else different on culicidae though. Sadly the backtrace is not that informative either. I'll try to investigate more but it will take time... -- Petr Jelinek http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training & Services -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers