Re: "Interrupt requested" in postgresql-DAY.log

On May 27, 2014, at 10:52 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Louis Battuello <louis(dot)battuello(at)etasseo(dot)com> writes:
>> I have a PostgreSQL 9.3.4 database running on CentOS 6 with PostGIS 2.1.2. Each night, I run a cron job to dump (pg_dump) a few schemas for development snapshots. Everything runs without error.
>
>> However, it seems that the dump process consistently results in a single line in the postgresql-*.log file. Oddly enough, this entry seems to ignore the log_line_prefix configuration parameter.
>
>> postgres(at)db:/var/lib/pgsql/9.3/data $ cat pg_log/postgresql-Fri.log
>> Interrupt requested
>
> That probably represents some non-Postgres bit of code deciding to bleat
> to stdout or stderr. The logging-collector mechanism is designed to catch
> such output coming from a backend process, but it can't stick a
> log_line_prefix on it.
>
> What exactly is bleating, I can't say. A quick grep confirms that there
> is no such string in the Postgres sources, but I dunno about PostGIS.
> If you've got any code in plperl, plpython, etc, the culprit might lurk
> somewhere there. glibc might even be to blame, though I don't think it
> ordinarily prints error messages.
>
> regards, tom lane
>
>
> --
> Sent via pgsql-general mailing list (pgsql-general(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-general
>

Thanks, Tom.

This does in fact look like something new bubbling up from PostGIS 2.1.2 and 2.1.3.

https://github.com/postgis/postgis/blob/svn-2.1/postgis/postgis_module.c

I don’t see it triggered in 2.1.1, so I’ll take my observation over to the GIS folks.

Regards,
Louis