Message-ID: <60032.10.10.6.20.1236289562.squirrel@mail.autorevenue.com>
In-Reply-To: <41416.10.10.6.20.1236287704.squirrel@mail.autorevenue.com>
References: <41416.10.10.6.20.1236287704.squirrel@mail.autorevenue.com>
Date: Thu, 5 Mar 2009 16:46:02 -0500 (EST)
Subject: Re: another PITR recovery question
From: msteben@autorevenue.com
To: pgsql-admin@postgresql.org
User-Agent: SquirrelMail/1.4.6-1
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Importance: Normal

May be helpful to add on some memory information to my question:
 shared_buffers - 250MB
 work_mem  - 75 MB
 maintenance_work_mem - 100MB
 max_fsm_pages        - 1500000
 wal_buffers     -       96
 checkpoint_segments  -  32

And some information on the machine (VMWARE)

 ipcs -l

------ Shared Memory Limits --------
max number of segments = 4096
max seg size (kbytes) = 4194303
max total shared memory (kbytes) = 1073741824
min seg size (bytes) = 1

This is the only process running on this virtual machine.

Thanks,
Mark
>
> I'm running postgres 8.2.5 and I'm running a PITR restore on a
> remote machine.
> The archived logs are being restored at a rate of about 5 every 3 minutes
> I was reading in compressed logs and uncompressing them as they are being
> restored. I am now reading in uncompressed logs straight into the restore.
>
> The restore rate only improved slightly - to the rate of 6 every 3
> minutes.
>
> Is there something I can do in postgresql.conf or in the shared memory to
> get this to move faster?  At this rate my recovery will run for days as
> I'm
> copying WAL logs from my production machine.
>
>
> Thanks
>
>
>
> Mark Steben
>
> Autorevenue
>
>
>
>