Re: crash-safe visibility map, take three - Mailing list pgsql-hackers

From Heikki Linnakangas
Subject Re: crash-safe visibility map, take three
Date
Msg-id 4CF525F2.3030806@enterprisedb.com
Whole thread Raw
In response to Re: crash-safe visibility map, take three  (Robert Haas <robertmhaas@gmail.com>)
List pgsql-hackers
On 30.11.2010 18:22, Robert Haas wrote:
> On Tue, Nov 30, 2010 at 11:16 AM, Tom Lane<tgl@sss.pgh.pa.us>  wrote:
>> How much is "quite a lot"?  Do we have any real reason to think that
>> this solution is unacceptable performance-wise?
>
> Well, let's imagine a 1GB insert-only table.  It has 128K pages.  If
> you XLOG setting the bit on each page, you'll need to write 128K WAL
> records, each containing a 12-byte relfilenode and a 4-byte block
> offset, for a total of 16 bytes of WAL per page, thus 2MB of WAL.

Plus WAL headers, I think it's something like 32 or 40 bytes of WAL per 
page.

> But you did just dirty a gigabyte of data.

Good point.

--   Heikki Linnakangas  EnterpriseDB   http://www.enterprisedb.com


pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: crash-safe visibility map, take three
Next
From: Tom Lane
Date:
Subject: Re: Another proposal for table synonyms