From: | Petr Jelinek <petr(dot)jelinek(at)2ndquadrant(dot)com> |
---|---|
To: | Andres Freund <andres(at)anarazel(dot)de> |
Cc: | pgsql-hackers(at)postgresql(dot)org, Petr Jelinek <petr(at)2ndquadrant(dot)com>, Peter Eisentraut <peter_e(at)gmx(dot)net> |
Subject: | Re: logical replication launcher crash on buildfarm |
Date: | 2017-03-16 08:51:42 |
Message-ID: | 0337938f-4bda-3157-014e-c699d2d884f0@2ndquadrant.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 16/03/17 09:44, Andres Freund wrote:
> On 2017-03-16 09:40:48 +0100, Petr Jelinek wrote:
>> 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...
>
> I can give you a login to that machine, it doesn't do anything but run
> buildfarm animals... Will have to be my tomorrow however.
>
That would be helpful, thanks.
--
Petr Jelinek http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
From | Date | Subject | |
---|---|---|---|
Next Message | Andres Freund | 2017-03-16 08:53:22 | Re: logical replication launcher crash on buildfarm |
Previous Message | Andres Freund | 2017-03-16 08:45:53 | Quals not pushed down into lateral |