Re: Index vacuum improvements in 8.2

From: Wes <wespvp(at)msg(dot)bt(dot)com>
To: pgsql general <pgsql-general(at)postgresql(dot)org>
Subject: Re: Index vacuum improvements in 8.2
Date: 2007-01-04 22:23:45
Message-ID: C1C2DA91.3A6C2%wespvp@msg.bt.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

From 8.2 release notes:
Speed up vacuuming of B-Tree indexes (Heikki Linnakangas, Tom)

From <http://archives.postgresql.org/pgsql-patches/2006-05/msg00007.php>

"2. Vacuum is now done in one phase, scanning the index in physical order.
That significantly speeds up index vacuums of large indexes that don't fit
into memory. However, btbulkdelete doesn't know if the vacuum is a full or
lazy one. The patch just assumes it's a lazy vacuum, but the API really
needs to be changed to pass that information earlier than at
vacuum_cleanup."

Are these talking about the same thing - 8.2 now vacuums indexes in disk
order instead of index sort order? If so, that will make a tremendous
difference in vacuum speed for us (when we can afford the downtime to do a
dump/upgrade/reload). For us, vacuum starts at about 4 hours after a
reindex, then goes downhill. We reindex when it starts taking 24 hours to
vacuum.

The above referenced posting is from May 1, 2006. Is pgsql 8.2 the first
release to have this code?

Wes

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Thomas F. O'Connell 2007-01-05 00:18:40 Interrupted pg_dump / pg_restore Upgrade
Previous Message Angva 2007-01-04 20:42:48 upgrading and pg_restore versions