DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=mime-version:in-reply-to:references:date:message-id:subject:from:to
	:content-type:content-transfer-encoding;
	b=vxEiON2PGKydUsQZFSkiQN7Dfc/G4/b8d7CTIUSygg+Gr2F+Gr4EO8wN13cKNzCLww
	H6WJ0Thj/rxfwgZQgQ0cb3Cgaqjm5A2vSAHWXfzvBbnoImS/ZdnJxKVEBB5k9oirrOUI
	UAYgGeeKzSkhdytuA18XtcAlRQ/cu/zmb4Tq0=
MIME-Version: 1.0
In-Reply-To: <4BCA2DE0.5090602@pinpointresearch.com>
References: <g2z649cb4eb1004170931o39755d41k66e5be957fc89591@mail.gmail.com>
	<4BC9F464.20901@pinpointresearch.com>
	<v2m649cb4eb1004171308ve81536f2qf125a7722447c4ed@mail.gmail.com>
	<4BCA2DE0.5090602@pinpointresearch.com>
Date: Sat, 17 Apr 2010 18:09:24 -0400
Message-ID: <l2w649cb4eb1004171509p1c20174eg9940313652c2bd07@mail.gmail.com>
Subject: Re: pg_restore error
From: Bryan White <nicktook@gmail.com>
To: pgsql-admin@postgresql.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

On Sat, Apr 17, 2010 at 5:53 PM, Steve Crawford
<scrawford@pinpointresearch.com> wrote:
> Bryan White wrote:
>>
>> ...
>> The new server is a dual socket Nahalem. =A08 cores, 16 threads, 48 GB
>> ram, 2 WAL drives in RAID1 and 12 database drives in RAID10.
>>
>
> I would try setting -j higher - in your case try 8 for starters. Also tur=
n
> off fsync and autovacuum (and turn them back on !!!). See these and more
> tips at:
> http://it.toolbox.com/blogs/database-soup/using-84-parallel-restore-with-=
your-83-or-82-database-31575
>

Success. The -j6 run completed.  The time to restore went from 152
minutes to 45 minutes.  I was already doing some of the things on the
above page.  I plan on implementing others and than do a few test runs
to find the optimal number of jobs.

Thanks!

--=20
Bryan White