Re: BUG #16817: kill process cause postmaster hang

commit 82233ce7ea42d6ba519aaec63008aff49da6c7af
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Date:   Fri Jun 28 17:20:53 2013 -0400

    Send SIGKILL to children if they don't die quickly in immediate shutdown

    On immediate shutdown, or during a restart-after-crash sequence,
    postmaster used to send SIGQUIT (and then abandon ship if shutdown); but
    this is not a good strategy if backends don't die because of that
    signal.  (This might happen, for example, if a backend gets tangled
    trying to malloc() due to gettext(), as in an example illustrated by
    MauMau.)  This causes problems when later trying to restart the server,
    because some processes are still attached to the shared memory segment.

    Instead of just abandoning such backends to their fates, we now have
    postmaster hang around for a little while longer, send a SIGKILL after
    some reasonable waiting period, and then exit.  This makes immediate
    shutdown more reliable.

    There is disagreement on whether it's best for postmaster to exit after
    sending SIGKILL, or to stick around until all children have reported
    death.  If this controversy is resolved differently than what this patch
    implements, it's an easy change to make.

    Bug reported by MauMau in message 20DAEA8949EC4E2289C6E8E58560DEC0@maumau

    MauMau and Álvaro Herrera