Re: logical replication launcher crash on buildfarm

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

- Andres