Re: 8.4 open item: copy performance regression? - Mailing list pgsql-hackers

From Alan Li
Subject Re: 8.4 open item: copy performance regression?
Date
Msg-id 782056770906221729o573f9511s184b966fc7118e2@mail.gmail.com
Whole thread Raw
In response to Re: 8.4 open item: copy performance regression?  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Mon, Jun 22, 2009 at 7:16 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Alan Li <ali@truviso.com> writes:
> How much concern is there for the contention for use cases where the WAL
> can't be bypassed?

If you mean "is something going to be done about it in 8.4", the
answer is "no".  This is a pre-existing issue that there is no simple
fix for.

                       regards, tom lane


No no, I am certainly not implying anything for the 8.4 timeframe.

Moving forward, I imagine this being more of a problem for data warehouse applications, where bulk inserts occur on existing fact tables.  In this case, the WAL cannot be bypassed (unless the bulk insert occurs on a newly created partition).  And since COPY is cpu-bound, it would perhaps be advantageous to do parallel COPY's on the same table on multi-core systems, which won't work with WAL bypassing either.

Thanks, Alan

pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Changed error message for blocks by prepared transactions
Next
From: KaiGai Kohei
Date:
Subject: Re: security checks for largeobjects?