Re: Write performance

From: Scott Carey <scott(at)richrelevance(dot)com>
To: Janning <ml(at)planwerk6(dot)de>
Cc: "pgsql-performance(at)postgresql(dot)org" <pgsql-performance(at)postgresql(dot)org>
Subject: Re: Write performance
Date: 2010-06-25 17:30:06
Message-ID: 590C5896-DA69-451C-8138-831422140F59@richrelevance.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance


On Jun 24, 2010, at 6:16 AM, Janning wrote:

> On Thursday 24 June 2010 14:53:57 Matthew Wakeling wrote:
>> On Thu, 24 Jun 2010, Janning wrote:
>>> We have a 12 GB RAM machine with intel i7-975 and using
>>> 3 disks "Seagate Barracuda 7200.11, ST31500341AS (1.5 GB)"
>>
>> Those discs are 1.5TB, not 1.5GB.
>
> sorry, my fault.
>
>>> One disk for the system and WAL etc. and one SW RAID-0 with two disks for
>>> postgresql data. Our database is about 24GB.
>>
>> Beware of RAID-0 - make sure you can recover the data when (not if) a disc
>> fails.
>
> oh sorry again, its a raid-1 of course. shame on me.

If your WAL is not on RAID but your data is, you will lose data if the WAL log drive dies. You will then have a difficult time recovering data from the data drives even though they are RAID protected. Most likely indexes and some data will be corrupted since the last checkpoint. I have lost a WAL before, and the result was a lot of corrupted system indexes that had to be rebuilt in single user mode, and one system table (stats related) that had to be purged and regenerated from scratch. This was not fun. Most of the data was fine, but the cleanup is messy if you lose WAL, and there is no guarantee that your data is safe if you don't have the WAL available.

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Greg Smith 2010-06-25 18:48:07 Any recent AMD purchases?
Previous Message Greg Smith 2010-06-25 17:06:26 Re: sudden spurt in swap utilization (was:cpu bound postgresql setup.)