Re: Vacuum improvement - Mailing list pgsql-hackers

From Alvaro Herrera
Subject Re: Vacuum improvement
Date
Msg-id 20021016040423.GA15436@dcc.uchile.cl
Whole thread Raw
In response to Re: Vacuum improvement  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Tue, Oct 15, 2002 at 11:52:35PM -0400, Tom Lane wrote:
> Gavin Sherry <swm@linuxworld.com.au> writes:
> > have a parameter which specified how much of the table is vacuumed. That
> > is, you could specify:
> > VACUUM FULL test 20 precent;
> 
> Erm ... but which 20 percent?  In other words, how could you arrange for
> repeated applications of such a command to cover the whole table, and
> not just retrace an already-cleaned-out portion?

Maybe each relation block can have a last-vacuumed timestamp?  Somewhere
in the table there would have to be a linked list of least-recently
vacuumed blocks so the vacuum cleaner does not have to read every
block to know which one to clean.

Or maybe some system table can provide information about activity in
each block since last vacuum.  This forces the use of the stat
collector...

-- 
Alvaro Herrera (<alvherre[a]dcc.uchile.cl>)
"El sabio habla porque tiene algo que decir;
el tonto, porque tiene que decir algo" (Platon).


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Vacuum improvement
Next
From: Anuradha Ratnaweera
Date:
Subject: Postgresql and multithreading