Re: [HACKERS] can't stop autovacuum by HUP'ing the server

Dave Cramer wrote:

> AutoVacuumShmem->av_signal[AutoVacForkFailed] = false;
> pg_usleep(100000L); /* 100ms */
>
> SendPostmasterSignal(PMSIGNAL_START_AUTOVAC_WORKER);
> continue;
>
> Do these signals get cleaned up on a reload ?

Well, not on a reload specifically, but this signal is of prompt
response (i.e. the postmaster acts immediately on it). See
CheckPostmasterSignal. Also, note that this code starts a worker, not
the launcher which is what you're seeing.

The signal you're looking for is PMSIGNAL_START_AUTOVAC_LAUNCHER (see
the varsup.c code in the vicinity of what I posted earlier). The
postmaster response is to set start_autovac_launcher (see
sigusr1_handler in postmaster.c) and when this is seen set, the launcher
is started (see ServerLoop in postmaster.c). However the bit you're
interested in is *why* the signal is being sent, which is what the
freeze limits determine.

--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers