CLUSTER trashing indexes - Mailing list pgsql-novice

From Jason Davis
Subject CLUSTER trashing indexes
Date
Msg-id 200010220758.SAA09128@zeus.tassie.net.au
Whole thread Raw
Responses Re: CLUSTER trashing indexes  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-novice
Hi all

I have a rather large table, around 7 million rows. I created 3 indexes on
this table, which resulted in rather large indexes. As there will be no
more rows added to this table, I thought optimizing the indexes with the
CLUSTER command might be a good idea to speed up lookups. After the first
index had finished being processed by the cluster command, my remaining
indexes diappeared. The actual index files in the ~/data/base/dbname
directory had completely disappeared. Therefore I have to go and re-create
the indexes. Have I missed something in the release notes or known bugs
list, or is there a size limitation or problem when dealing with indexes
and/or clustering them? The average physical index file size being around
200MB each.

My version btw is 7.0.2 running on FreeBSD 3.4.

cheers,

Jason Davis
DB Admin/Programmer
Tas Access
www.tassie.net.au


pgsql-novice by date:

Previous
From: Frank Hilliard
Date:
Subject: Using KDE to start Postgres
Next
From: Tom Lane
Date:
Subject: Re: CLUSTER trashing indexes