Re: Interpreting vacuum verbosity - Mailing list pgsql-general

From Tom Lane
Subject Re: Interpreting vacuum verbosity
Date
Msg-id 15289.1083958067@sss.pgh.pa.us
Whole thread Raw
In response to Re: Interpreting vacuum verbosity  (Jeff Boes <jboes@nexcerpt.com>)
List pgsql-general
Jeff Boes <jboes@nexcerpt.com> writes:
> At some point, someone was going to write a "white paper" detailing how one
> might go about setting these parameters.

In 7.4, it's relatively easy to check on whether your settings are
reasonable: just do a VACUUM VERBOSE (database-wide) and check the
FSM requirements indicated at the end of the tediously chatty output.

All I have handy to illustrate with is a test server that has only the
regression test database loaded in it, so these numbers are very small,
but what I see is:

INFO:  free space map: 280 relations, 520 pages stored; 4720 total pages needed
DETAIL:  Allocated FSM size: 1000 relations + 20000 pages = 178 kB shared memory.

This says that what I actually need to keep track of the present free
space in the database is 280 FSM relation slots and 4720 FSM page slots.
So the allocated space is plenty comfy here.  If the "pages needed"
number is significantly larger than your max_fsm_pages setting, then you
have a problem.

            regards, tom lane

pgsql-general by date:

Previous
From: Greg Stark
Date:
Subject: Re: Storing a file hash as primary key
Next
From: Joe Conway
Date:
Subject: Re: Storing a file hash as primary key