[rt-users] RT related Postgres question

[Vivek Khera]

On Nov 30, 2007, at 2:31 PM, Kenneth Marshall wrote:

> Here is a list of indexes that we use here. Check your schema
> to see if you are missing any. You should not be getting a
> sequential scan.

sequential scans are not automatically bad.  if your result set is  
more than a certain percentage of the whole table, it is more  
efficient to scan the table than to do the extra disk seeks of going  
back and forth to the index file too.

having too many indexes is bad, as you have just added a *lot* of  
overhead for every insert/update/delete in that you must update all  
the extra indexes.

there is some fine balance in between, which takes a lot of experience  
and analysis to find.