Start with a clean machine,
- Install a bench
frappe-bench and setup erpnext on it.
- Install another bench
frappe-bench2 (ie use
frappe init frappe-bench2) and get erpnext running on it
After this, setup supervisor and edit one of the supervisor config to
run on another port (say 5000). Adjust this port corresponding nginx
In case of error, do you see any tracebacks in frappe-bench/logs/web.error.log?
I managed to configure this by changing not just the port name but also all the program names in supervisor.conf for the second bench. Then I also changed redis conf files for the second bench (port, pid, database) and socketio.js to reflect what was changed in redis files.
I have two benches (one running two sites, the other one running one site) working and I see no problems when working in ERPNext in any of them.
However, when I run supervisorctl status, I see that “node-socketio” and what I call “second-node-socketio” keep restarting every time somone uses any site on these benches. All the other processes (e.g. frappe-web, second-frappe-web…) keep running for days, but node-socketio restarts as soon as you log in to a site.
How can I fix this?
I don’t really understand what these processes are used for… just managed to guess how to change the config files … do I even need “second-node-socketio”, “second-redis-cache” , etc.?
You should look at the process logs (stdout/stderr you specified in the config). I am also convinced that supervisor config process names should not be hardcoded.