On Wed, May 27, 2009 at 3:54 PM, Alexander Schöcke <asc@xxxxxxxxxxxxxxxxxxxxxxx> wrote: > Hello everybody. > > I'm using a view (http://pgsql.tapoueh.org/site/html/news/20080131.bloat.html) to display the bloat (unused disk space) of the tables in a PostgreSQL database. Using this data I want to implement a database maintenance script automatically exectuting a VACUUM FULL on these tables. > Unfortunately I am finding a table to have bloat which can't be reclaimed. I have tried VACUUM, REINDEX, VACUUM FULL ANALYZE with REINDEX, and even dump and restore. The view always shows 375MB of bloat for the table. > > Is this normal? yes it is normal, if you do updates on it, and it doesn't use HOT (for instance if you update indexed column), now, in real life you shouldn't really ever need to vacuum full. Reindex - yes, but not vacuum full. Make sure, that you adjust fsm_* settings in your config, best thing is tu run vacuum analyze verbose, and see if it warns you about it being too small. -- GJ -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
