Skip to Content
Menu
This question has been flagged
5 Replies
21623 Views

I have installed openerp v7 on a Intel dual core machine

  • RAM: 4G
  • OS: FreeBSD 9.1
  • PostgreSQL: 9.2

It's a machine for testing OpenERP so there's only 2~3 users.

The reaction is kinda slow even there's only one concurrent user.

The machine's kinda old so I'm trying to squeeze the juice out of it... So I'm wondering does anyone have a tuning guide of tips?

Avatar
Discard
Best Answer

You don't mention the version of PostgreSQL you have installed (9.2 is faster than 9.1), or how much RAM you have, or how many total users you support, how many concurrent users you support, or what they do, or how performance is now - what about those details?

These are the pieces that affect performance of OpenERP:

  • Disk hardware - very little you can do
  • Operating System - no experience with FreeBSD
  • Database - PostgreSQL - see below
  • Application - Python - Unless you want to start profiling code and re-programming there isn't anything here you can do
  • Python Interpreter - Doubful there is anything here you can do
  • Web Server - CherryPy - A single Pentium M 1.6 Ghz with 1GB can support over 500 requests a second
  • CPU - nothing you can do apart from upgrade
  • Memory - nothing you can do apart from upgrade/increase

So your only real option is PostgreSQL.

See http://www.slideshare.net/openobject/tips-on-how-to-improve-the-performance-of-your-custom-modules-for-high-volumes-deployment-olivier-dony-open-erp

Avatar
Discard
Author

Thank you!

Author

So actually the only "tuning" we can do is all about Postgre?

Additionnaly you could find this one : http://fr.slideshare.net/openobject/performance2014-35689113

Best Answer

I know this post is a bit old, but this actually helped me this time.

We had an issue with slow POS orders sync. The config recommendations helped, and especially ensuring AUTOVACUUM is set-up. It was turned off for a while and the DB eventually bloated. 2000% decrease in processing time after completing the process. 

Tip: as we are at a "bloated" state, doing a VACUUM takes a lot of time. Our workaround was to export the whole DB, and re-import. 

Avatar
Discard
Related Posts Replies Views Activity
4
Dec 24
14784
1
Jun 24
623
1
May 24
323
1
Nov 23
682
2
Sep 23
7019