Re: Segmentation fault mayhem

Hi,

On 2017-09-25 20:19:45 +0200, Sebastian Gröbler wrote:
> 2017-09-25 10:27:24 UTC [101978-1] dms(at)booking LOG:  starting logical decoding for slot "uhbeasbguomldspa_00016400_846fa332_0d73_4e40_bef4_da6c7a70c868"
> 2017-09-25 10:27:24 UTC [101978-2] dms(at)booking DETAIL:  streaming transactions committing after B96/1FD4BB78, reading WAL from B96/1FD241C8
> 2017-09-25 10:27:24 UTC [101978-3] dms(at)booking LOG:  logical decoding found consistent point at B96/1FD241C8
> 2017-09-25 10:27:24 UTC [101978-4] dms(at)booking DETAIL:  There are no running transactions.
> 2017-09-25 10:27:25 UTC [94708-8] LOG:  server process (PID 101978) was terminated by signal 11: Segmentation fault
> 2017-09-25 10:27:25 UTC [94708-9] LOG:  terminating any other active server processes
>
> Core was generated by `postgres: 9.6/main: wal sender process dms x.x.x.x(xxxx) idle           '.
> Program terminated with signal SIGSEGV, Segmentation fault.
> #0  0x00007fd3190ccad6 in ReorderBufferCommit ()
> (gdb) bt
> #0  0x00007fd3190ccad6 in ReorderBufferCommit ()
> #1  0x00007fd3190c5e00 in LogicalDecodingProcessRecord ()
> #2  0x00007fd3190d08a9 in ?? ()
> #3  0x00007fd3190d2433 in ?? ()
> #4  0x00007fd3190d3282 in exec_replication_command ()
> #5  0x00007fd319116b35 in PostgresMain ()
> #6  0x00007fd318e930a8 in ?? ()
> #7  0x00007fd3190b52b0 in PostmasterMain ()
> #8  0x00007fd318e94267 in main ()

Any chance to get a backtrace with debug symbols intsalled? And perhaps
from a newer point release? There've been a number of fixes since
9.6.1. Without more information this is essentially impossible to pin down.

Greetings,

Andres Freund