Horton

Re: SIGQUIT handling, redux

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Stephen Frost <sfrost@snowman.net>
Cc: pgsql-hackers@lists.postgresql.org
Date: 2020-09-09T14:55:23Z
Lists: pgsql-hackers 
I wrote:
> Stephen Frost <sfrost@snowman.net> writes:
>> As I mentioned over there, I agree that we should do this and we should
>> further have the statistics collector also do so, which currently sets
>> up SIGQUIT with ShutdownRequestPending() and in its loop decides it's
>> fine to write out the stats file (which we're going to remove during
>> recovery anyway...) and then call exit(0).

> I noticed that that was different from everything else, but it's not
> actually signal-unsafe, so it seems like a different topic from what
> I'm on about at the moment.  I don't mind if you or somebody else
> wants to change it, but I don't see it as a back-patchable bug fix.

Note also that the postmaster actually uses SIGQUIT to command normal
shutdown of the stats collector (cf reaper(), around line 3125 in HEAD).
So this needs a change in signaling conventions, not just internal
tweaks in the collector.  Not a big deal, but it reinforces my feeling
that it should be a HEAD-only change.

			regards, tom lane

Commits

  1. Log a message when resorting to SIGKILL during shutdown/crash recovery.

  2. Don't run atexit callbacks during signal exits from ProcessStartupPacket.

  3. Use _exit(2) for SIGQUIT during ProcessStartupPacket, too.

  4. Make archiver's SIGQUIT handler exit via _exit().