Moodle 2.5 release notes/Performance tests

Revision as of 10:05, 9 May 2013 by David Monllaó Olivé (talk | contribs) (Test results)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

The comparison tests are between a 2.4.3+ site (Build: 20130425) and the same site upgraded to 2.5beta+ (2013050800.00, 66d6221f9ac9319b9f106643afc571930bade41e to be more specific) with the default new settings values, including badges enabled by default.

Test results

Database reads Database writes Time used Files included KiloBytes Session size getstring() calls
Moodle 2.4.3+ site (Build: 20130425) 1832.8 28.4 69.7 2810 3434.9 375.1 6681.5
Moodle 2.5beta+ (Build: 20130502) 1748 28.4 64.6 2734 3467.5 406.2 6708.4

Site info

  • All of Moodle's advanced features enabled but web services
  • Debug level set to NONE
  • Default cache stores
  • Sessions stored in the database
  • The frontpage settings are:
    • Without being logged: List of courses
    • Being logged: Enroled courses (same behaviour than 'List of courses' in 2.4, showing only the ones where the user is enroled)
  • Site numbers: 
    • 1.000.000 total users
    • 67 courses
    • 25000 role assignments
  • Tested course numbers:
    • 8761 role assignments
    • 126 course modules

Test plan

  • JMeter 2.3.4
  • 100 threads (using a CSV file with the users login data)
  • 10 ramp up period
  • A web server (Apache 2) and a database server (MySQL) in the same internal network

Adding to being enroled to the tested course, the 100 users (enroled as teachers) used for the tests, were also enroled to 10 other random courses.

  • View home page
  • Login
  • View site calendar
  • View next month of the site calendar
  • Browse a course
  • View a forum
  • View a discussion
  • View a course calendar
  • Browse the course again
  • Browse the forum again
  • Add a discussion
  • Return to the forum
  • View forum discussion
  • Delete the created forum post
  • Logout