On Tue, Jun 26, 2012 at 1:58 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> On Thu, Mar 29, 2012 at 3:04 AM, Daniel Farina <daniel(at)heroku(dot)com> wrote:
>> On Mon, Mar 26, 2012 at 12:14 AM, Jeff Davis <pgsql(at)j-davis(dot)com> wrote:
>>> On Tue, 2012-03-20 at 01:38 -0700, Daniel Farina wrote:
>>>> On Thu, Mar 15, 2012 at 9:39 PM, Fujii Masao <masao(dot)fujii(at)gmail(dot)com> wrote:
>>>> > On Fri, Mar 16, 2012 at 8:14 AM, Daniel Farina <daniel(at)heroku(dot)com> wrote:
>>>> >> Parallel to pg_cancel_backend, it'd be nice to allow the user to just
>>>> >> outright kill a backend that they own (politely, with a SIGTERM),
>>>> >> aborting any transactions in progress, including the idle transaction,
>>>> >> and closing the socket.
>>>> > +1
>>>> Here's a patch implementing the simple version, with no more guards
>>>> against signal racing than have been seen previously. The more
>>>> elaborate variants to close those races is being discussed in a
>>>> parallel thread, but I thought I'd get this simple version out there.
>>> After reading through the threads, it looks like there was no real
>>> objection to this approach -- pid recycling is not something we're
>>> concerned about.
>>> I think you're missing a doc update though, in func.sgml:
>>> "For the less restrictive <function>pg_cancel_backend</>, the role of an
>>> active backend can be found from
>>> the <structfield>usename</structfield> column of the
>>> <structname>pg_stat_activity</structname> view."
>>> Also, high-availability.sgml makes reference to pg_cancel_backend(), and
>>> it might be worthwhile to add an "...and pg_terminate_backend()" there.
>>> Other than that, it looks good to me.
>> Good comments. Patch attached to address the doc issues. The only
>> iffy thing is that the paragraph "For the less restrictive..." I have
>> opted to remove in its entirely. I think the documents are already
>> pretty clear about the same-user rule, and I'm not sure if this is the
>> right place for a crash-course on attributes in pg_stat_activity (but
>> maybe it is).
>> "...and pg_terminate_backend" seems exactly right.
>> And I think now that the system post-patch doesn't have such a strange
>> contrast between the ability to send SIGINT vs. SIGTERM, such a
>> contrast may not be necessary.
>> I'm also not sure what the policy is about filling paragraphs in the
>> manual. I filled one, which increases the sgml churn a bit. git
>> (show|diff) --word-diff helps clean it up.
> I went ahead and committed this.
> I kinda think we should back-patch this into 9.2. It doesn't involve
> a catalog change, and would make the behavior consistent between the
> two releases, instead of changing in 9.1 and then changing again in
> 9.2. Thoughts?
I think that would be good.
It is at the level of pain whereas I would backpatch from day one, but
I think it would be a welcome change to people who couldn't justify
gnashing their teeth and distributing a tweaked Postgres like I would.
It saves me some effort, too.
In response to
pgsql-hackers by date
|Next:||From: Robert Haas||Date: 2012-06-26 21:53:26|
|Subject: Re: Posix Shared Mem patch|
|Previous:||From: Martijn van Oosterhout||Date: 2012-06-26 21:44:42|
|Subject: Re: why roll-your-own s_lock? / improving scalability|