Dear Learning Equality Technical Team,
While working with Kolibri versions 0.19.0 to 0.19.3 on Linux, I encountered an issue where the manage benchmark command failed due to a configuration error. This command is important for assessing system performance and helps ensure that Kolibri remains efficient and responsive, particularly in low-resource environments where performance directly affects learners’ access to educational content.
Through debugging, I traced the issue to missing session support in the Django configuration. Specifically, "django.contrib.sessions" had not been included in the INSTALLED_APPS setting within the base.py file. Since session management is a core dependency for certain application processes, its absence caused the benchmark command to fail in these versions.
By adding "django.contrib.sessions" to the configuration, I restored the functionality of the benchmark command across the affected versions.
Hi @Abdulkarim_Ali_Patri,
While it may prevent the error you are seeing, this will not fix the benchmark command. It seems the fact that this was directly inspecting the Session objects was overlooked during the migration from file backed sessions to database backed sessions.
I will make sure the proper fix goes into our next Kolibri release - thanks for letting us know.
Kind Regards,
Richard
Hi Richard,
Thank you for the clarification and for digging deeper into the issue.
That makes sense. I understand now that the problem lies in how the benchmark command interacts with the session model rather than just the configuration. My initial fix addressed the immediate error, but I see why a more complete solution is needed.
I appreciate you logging the issue. I would be happy to help test the fix or explore contributing to a solution if that would be useful.
Kind regards,
Ali
Hi @Abdulkarim_Ali_Patri,
Thanks - we’ve merged a fix for this that will go into the next Kolibri patch release! Thanks for letting us know about this.
Kind Regards,
Richard