| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Josh Berkus <josh(at)agliodbs(dot)com> |
| Cc: | Kevin Grittner <kgrittn(at)ymail(dot)com>, Robert Haas <robertmhaas(at)gmail(dot)com>, Vik Fearing <vik(dot)fearing(at)dalibo(dot)com>, PG Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: idle_in_transaction_timeout |
| Date: | 2014-06-18 19:32:25 |
| Message-ID: | 27179.1403119945@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Josh Berkus <josh(at)agliodbs(dot)com> writes:
> There are plenty of badly-written applications which "auto-begin", that
> is, they issue a "BEGIN;" immediately after every "COMMIT;" whether or
> not there's any additional work to do. This is a major source of IIT
> and the timeout should not ignore it.
Nonsense. We explicitly don't do anything useful until the first actual
command arrives, precisely to avoid that problem.
It might be that we should slap such apps' wrists anyway, but given
that we've gone to the trouble of working around the behavior at the
system structural level, I'd be inclined to say not. What you'd be
doing is preventing people who have to deal with such apps from using
the timeout in any useful fashion.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andres Freund | 2014-06-18 19:32:43 | Re: pgsql: Reduce the number of semaphores used under --disable-spinlocks. |
| Previous Message | Tom Lane | 2014-06-18 19:28:46 | Re: pg_control is missing a field for LOBLKSIZE |