I am setting up a production server to host Allura platform.
So I have some general questions before I start:
1. What is the expected system requirement/configuration for a production server to host Allura?
2. Are there any specific configurations that I need to do in Allura to get a good performance when the load is high (for >100 users)?
3. Any other points that I need to be aware about?
Do you recommend running MongoDB and Allura on separate servers?
Hi Vrinda,
The overall system requirements depend largely on how much traffic and how much data (repos, tickets, etc) will be in the system.
To configure Allura for good performance, see production-docker-example.ini. It is good to look through all of it, but lines 75-79 are particularly important for performance.
-Dave
Thank you very much Dave :)
I'll start with splitting MongoDB from the Allura server.
I looked around to find how I can set up MongoDB on a separate server and make it accessible from Allura server.
One hint that I got to make the MongoDB server accessible from Allura server is to change the bind_ip configuration in /etc/mongod.conf to point to 0.0.0.0 so as to allow any IP to access it
Is this all I have to do on the MongoDB server?
And in Allura's development.ini, change:
activitystream.master = mongodb://<mongodb ip="" server's="">:27017
and ming configurations to point to the MongoDB server.
Anything else that I need to change in Allura configurations?</mongodb>
Yep, activitystream and ming entries in the .ini file are the only settings you need to change for it.
Depending on your network/firewall setup, make sure that your mongodb server is not accessible to the outside world.
I sometimes see INFO messages 'Handling signal: winch' in the gunicorn terminal after which the application no longer responds. Why does this happen?