From: | Robert Treat <xzilla(at)users(dot)sourceforge(dot)net> |
---|---|
To: | Peter Childs <blue(dot)dragon(at)blueyonder(dot)co(dot)uk> |
Cc: | pgsql-admin(at)postgresql(dot)org |
Subject: | Re: interpret vacuum verbose output |
Date: | 2003-06-05 18:06:40 |
Message-ID: | 1054836400.29943.157.camel@camel |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
On Thu, 2003-06-05 at 04:54, Peter Childs wrote:
> On Wed, 4 Jun 2003, Shankar K wrote:
>
> > hi all,
> >
> > I'm trying to evaluate the frequecy to run vacuum
> > analyze on key tables. so if anyone could help me to
> > interpret the output of vacuum analyze verbose output
> > that would be great. below is the output of one of our
> > major indexes.
> >
> UnUsed: Number of Records In Table not in use currently. (I think vacuum
> full should remove these)
yes
> Changed: Number of Records Changed (Not sure)
Number of Pages Changed
> Keep: Number of Records to Keep (Not Sure)
Correct. Usually these are tuples that are currently being accessed
within a transaction.
> If deleted gets too big you should be vacuuming more often
>
not really, unused is far more significant for that. You might say if
deleted is too small, you might want to vacuum less frequently.
> If unused gets too big try and increase your fsm size.
>
if it gets too big, it's a sign you're not vacuuming enough.
if it gets way too big, you need to do a vacuum full.
note recommendations on deleted and unused both assume you have
increased your fsm settings appropriately.
also note that in pre-7.4 releases, vacuum cannot remove dead pages
within indexes, so if your indexes grow to large you will need to
REINDEX
Robert Treat
--
Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL
From | Date | Subject | |
---|---|---|---|
Next Message | Andrew Biagioni | 2003-06-05 18:21:48 | Re: defined max length |
Previous Message | Robert Treat | 2003-06-05 17:57:01 | Re: Urgent Help Needed !!!!!!! |