Re: bug in SignalSomeChildren

Robert Haas <robertmhaas@gmail.com>

From: Robert Haas <robertmhaas@gmail.com>

To: Alvaro Herrera <alvherre@commandprompt.com>

Cc: Tom Lane <tgl@sss.pgh.pa.us>, Fujii Masao <masao.fujii@gmail.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>

Date: 2010-12-20T16:27:47Z

Lists: pgsql-hackers

On Mon, Dec 20, 2010 at 8:20 AM, Alvaro Herrera
<alvherre@commandprompt.com> wrote:
> Excerpts from Robert Haas's message of vie dic 17 14:08:04 -0300 2010:
>
>> I'm having a bit of trouble confirming this on MacOS X, though.
>> Attaching gdb to either the startup process or a WAL sender causes
>> PostmasterIsAlive to return false, resulting in a near-immediate exit.
>>  Seems pretty stupid for attaching gdb to change the return value of
>> getppid() but it seems like that must be what's happening.
>
> Yeah, this problem has been known for some time and causes quite some
> pain.  We have an open problem report on autovacuum failing to run after
> some time, and we haven't been able to get a backtrace or strace because
> of this issue -- trying to attach to it causes a full system restart
> (PostmasterIsAlive returns false, autovac launcher dies, this death
> causes postmaster to shut everything down and restart).  This is on
> FreeBSD.

Can we add a develop option to force use of the kill(0) method?

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company