| From: | Andrew Sullivan <ajs(at)crankycanuck(dot)ca> |
|---|---|
| To: | John Smith <sodgodofall(at)gmail(dot)com> |
| Cc: | pgsql-general(at)postgresql(dot)org |
| Subject: | Re: slow, long-running 'commit prepared' |
| Date: | 2008-11-26 04:47:56 |
| Message-ID: | 20081126044756.GC11225@shinkuro.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Tue, Nov 25, 2008 at 08:05:36PM -0800, John Smith wrote:
> Thanks for the quick response, Andrew. How could the 'commit prepared'
> be I/O bound? Isn't all the I/O accomplished during 'prepare
> transaction', leaving only a tiny bit of data to write at commit?
No. When you do PREPARE TRANSACTION you have to save a whole lot of
state. And when you then do COMMIT PREPARED, you say, "Lose that
state in favour of this new state." All that work has to be done
again, in the other direction.
It's not for nothing that people think 2PC is a heavyweight and
expensive system. :-(
A
--
Andrew Sullivan
ajs(at)crankycanuck(dot)ca
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Scott Marlowe | 2008-11-26 05:04:13 | Re: Postgres 8.3 only uses seq scan |
| Previous Message | John Smith | 2008-11-26 04:05:36 | Re: slow, long-running 'commit prepared' |