Subject | Re: Slow Sweep speed on large database - Long post |
---|---|
Author | Eduardo A. Salgado |
Post date | 2005-08-04T21:08:40Z |
--- In firebird-support@yahoogroups.com, Alexandre Benson Smith
<iblist@t...> wrote:
CS and most would not. So, we keep one standard way.
Will look at this.
Good suggestions! Thanks.
BTW, that Sweep, it is still going (over 6 hours)!
<iblist@t...> wrote:
> In SMP machines CS performs better, did you tried it ?We have over 160 installations around the US. Some may benefit from
CS and most would not. So, we keep one standard way.
> Assuming that you have FK between the the larger and the very smallWill look at this.
> tables will anser a lot of your problems (sweep taking too long).
> Select count(*) will never be fast in a 4.1 million row tables,Read about it.
> Because of long duplicate chains of values in indices, run thestatis on
> indices are remove those indices that have a lot of duplicates (>10000)
Will look at this.
> limit that I will comment bellow (page buffers).Will correct down to 10000.
Good suggestions! Thanks.
BTW, that Sweep, it is still going (over 6 hours)!