On Mon, 26 Jul 2004 , Arne Georg Gleditsch wrote:
* Robert L Mathews
I have also discovered several other trivially fixable issues that dramatically slow down OTRS with PostgreSQL (many of the SQL queries are made in a way that prevents PostgreSQL from using bigint indexes); I'll post more about them separately and try to provide a patch that fixes it once I'm confident I've found them all.
I've posted about this before, but there was no response at the time: http://article.gmane.org/gmane.comp.otrs.user/2898 . My suggestion is still to ditch the ridiculous bigint-serials. 32 bits are going to be enough for everyone.
(Yes, these are "famous last words", but if anyone, ever, should need more than 32 bits (which I have a hard time imagining), they'll certainly be dimensioned to afford special-case hacking of the schema and source code for their own installation. The sheer volume of data in the postgres database implied by needing bigint-serials is not something either hardware or software is going to cope with without serious herding anyway.)
As a late followup to this, I have submitted a patch that solves this problem, in OTRS bugzilla: http://bugs.otrs.org/show_bug.cgi?id=518 It makes a huge difference to OTRS performance on PostgreSQL when you have more than a few thousand tickets. -- Robert L Mathews, Tiger Technologies http://www.tigertech.net/ "Ignorance more frequently begets confidence than does knowledge." -- Darwin