From: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> |
---|---|
To: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Using multi-row technique with COPY |
Date: | 2005-11-30 12:20:44 |
Message-ID: | 200511301220.jAUCKiE12701@candle.pha.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Simon Riggs wrote:
> On Tue, 2005-11-29 at 18:51 -0500, Bruce Momjian wrote:
> > Please let me back up and ask a more simplistic question. I understand
> > the idea of allowing COPY to insert rows with less locking, but I am
> > wondering about the NOLOGGING idea. On commit, we must guarantee that
> > all the rows are in the table, so what advantage is there to a NOLOGGING
> > option?
>
> We would need to flush all the blocks in the table out of cache at
> commit time, for that table only. (As with CTAS, CIDX).
>
>
> To allow a full discussion, I'll separate the various ideas:
> 1. COPY using bulk copy mode
What is "bulk copy mode"? Full page loading?
> 2. NOLOGGING
Means flush/fsync table pages on commit.
> 3. Created in this transaction
Reduces locking?
> 4. ERRORTABLES
> 5. Uniqueness violations
>
> Right now, I think you have reasonable objections/input to (2) that we
> should discuss more before agreeing a way forward. I would aim to do (1)
> first, then return with a full and much better explained proposal for
> (2) for us to discuss, since (2) depends upon (1) somewhat.
>
> (3) and (4) seem to have been generally accepted, but (5) seems not
> viable with present thinking.
--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073
From | Date | Subject | |
---|---|---|---|
Next Message | Pollard, Mike | 2005-11-30 13:54:49 | Re: ice-broker scan thread |
Previous Message | Muhamamd Irfan Azam | 2005-11-30 10:25:04 | LibPQ Error. |