| From: | "Phoenix Kiula" <phoenix(dot)kiula(at)gmail(dot)com> |
|---|---|
| To: | "PG-General Mailing List" <pgsql-general(at)postgresql(dot)org> |
| Subject: | Vacuum Full is *hideously* slow! |
| Date: | 2008-11-15 21:36:32 |
| Message-ID: | e373d31e0811151336n9bea510lbe0dff1f531b63df@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Hi.
Per this thread:
http://archives.postgresql.org/pgsql-general/2008-11/msg00608.php -- I
think I understood that the time had come for my db to have a VACUUM
FULL. (Regular autovacuum etc is working)
I know a full vacuum is slow. A simple google search had suggested so.
But I had no idea it would take HOURS! I started the process against a
5GB database (about 8.5 million rows in all) and it went on for more
than an hour. I had to kill the process.
I am now reindexing just to be sure.
Is this normal? If a vacuum full takes hours or even days then what's
the point?
I read here - http://archives.postgresql.org/pgsql-performance/2005-07/msg00375.php
- that it's better to drop the indices, then vacuum, and then recreate
the indices. Is this true? This is also a bad decision for production
servers, but would this be better?
Thanks!
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Serge Fonville | 2008-11-15 21:45:10 | PostgreSQL Windows x64 Build |
| Previous Message | Pavel Stehule | 2008-11-15 19:36:37 | Re: Seek within Large Object, within PL/* function? |