Sunday, May 11, 2008

Re: [HACKERS] XIDs and big boxes again ...

Hans-Juergen Schoenig wrote:

>> regards, tom lane
>>
>
>
> overhead is not an issue here - if i lose 10 or 15% i am totally fine as
> long as i can reduce vacuum overhead to an absolute minimum.
> overhead will vary with row sizes anyway - this is not the point.

I am not buying this argument. If you have a 5TB database, I am going to
assume you put it on enterprise class hardware. Enterprise class
hardware can handle the I/O required to appropriately run vacuum.

We have a customer that is constantly running 5 autovacuum workers on
only 28 spindles. We are in the process of upgrading them to 50 spindles
at which point I will likely try 10 autovacuum workers.

>
> the point is that you don't want to potentially vacuum a table when only
> a handful of records has been changed.

Right, generally speaking 20% is reasonable, although I tend to be much
more aggressive and try to keep it at 10%.

Sincerely,

Joshua D. Drake


--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

No comments: