Re: TCP keepalive support for libpq

2010/2/15 Robert Haas <robertmhaas(at)gmail(dot)com>:
> On Mon, Feb 15, 2010 at 11:00 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> Euler Taveira de Oliveira <euler(at)timbira(dot)com> writes:
>>> Magnus Hagander escreveu:
>>>> If we want to do this, I'd be inclined to say we sneak this into 9.0..
>>>> It's small enough ;)
>>>>
>>> I'm afraid Robert will say a big NO. ;) I'm not against your idea; so if
>>> nobody objects go for it *now*.
>>
>> If Robert doesn't I will.  This was submitted *way* past the appropriate
>> deadline; and if it were so critical as all that, why'd we never hear
>> any complaints before?
>
> Agreed.
>
>> If this were actually a low-risk patch I might think it was okay to try
>> to shoehorn it in now; but IME nothing involving making new use of
>> system-dependent APIs is ever low-risk.  Look at Greg's current
>> embarrassment over fsync, a syscall I'm sure he thought he knew all
>> about.
>
> That's why I think we shouldn't change the default behavior, but
> exposing a new option that people can use or not as works for them
> seems OK.

Well, not changing the default will have us with a behaviour that's
half-way between what we have now and what we have on the server side.
That just seems ugly. Let's just punt the whole thing to 9.1 instead
and do it properly there.

--
Magnus Hagander
Me: http://www.hagander.net/
Work: http://www.redpill-linpro.com/