Re: backend hangs at immediate shutdown (Re: Back-branch update releases coming in a couple weeks)

From: "Robert Haas" <robertmhaas(at)gmail(dot)com>
On Thu, Jun 20, 2013 at 12:33 PM, Alvaro Herrera
> <alvherre(at)2ndquadrant(dot)com> wrote:
>> I will go with 5 seconds, then.
>
> I'm uncomfortable with this whole concept, and particularly with such
> a short timeout. On a very busy system, things can take a LOT longer
> than they think we should; it can take 30 seconds or more just to get
> a prompt back from a shell command. 5 seconds is the blink of an eye.

I'm comfortable with 5 seconds. We are talking about the interval between
sending SIGQUIT to the children and then sending SIGKILL to them. In most
situations, the backends should terminate immediately. However, as I said a
few months ago, ereport() call in quickdie() can deadlock indefinitely.
This is a PostgreSQL's bug. In addition, Tom san was concerned that some
PLs (PL/Perl or PL/Python?) block SIGQUIT while executing the UDF, so they
may not be able to respond to the immediate shutdown request.

What DBAs want from "pg_ctl stop -mi" is to shutdown the database server as
immediately as possible. So I think 5 second is reasonable.

Regards
MauMau