Re: Vacuum verbose output?

Robert

Thanks for that.

On 14/1/03 15:54, "Robert Treat" <xzilla(at)users(dot)sourceforge(dot)net> wrote:

> ISTM that at times your vacuuming regularly and actually recovering
> space, but then there are times when your space grows rather quickly.
> This is either an indication that you need to time your vacuums better
> (to coincide with large batch inserts or deletes) or you need to vacuum
> more often, to recover more space between batches.

We vacuum about every hour, immediately after a large batch delete. Inserts
are spread across the hour with a couple of peaks.

[FWIW tuple "life" (insert to subsequent delete) is about 3 hours on that
table, about 12-24 hours on other tables where performance is important, and
30 days on the rather larger but less performance critical large archive
tables.]

I've got limited data at the moment -- I need to collect more to see.

>> It grows steadily until a VACUUM FULL is performed, when it shrinks back.
>> If we never do a VACUUM FULL, will it just keep on growing, and does
>> performance suffer as it does so?
>>
>
> The thing with vacuum is at some point you should reach a condition
> where space on the disk increases only when the total number of inserts
> is greater than the total number of deletions over a given time frame.
> If it increases when inserts is less than deletes, your not vacuuming
> enough. If you can reach this level of "symbiosis" with regular vacuums,
> you should never have to do vacuum full.

OK, that's a helpful way of looking at it. Thanks.

Julian