ERPNext Short Worker container (image:erpnext-python, command:worker)
ERPNext Long Worker container (image:erpnext-python, command:worker)
ERPNext Assets container (image:erpnext-assets)
Frappe SocketIO container (image:frappe-socketio)
sites folder is shared between all containers
assets folder is populated by erpnext-assets and shared with erpnext-python
nginx proxy pass the $http_host, (kubernetes ingress will manage domain names and certificates)
SocketIO, Chat, private public File upload works. Tell me what doesn’t work.
WIP: Helm Chart will have all of the above as deployments (independently scalable)
WIP: one helm release=one frappe-bench
WIP: ERPNext Assets and ERPNext python will be part of single pod (assets need to be shared)
WIP: Try out ReadWriteMany volumes and mount shared sites directory.
WIP: CronJob/Job kubernetes resource to migrate the bench
It does not use bench at all. Instead .py files are dropped in commands directory to do what bench does. Only basic bench functionality is needed in container environment.
Brief steps to get started (docker and docker-compose needed for testing)
I installed docker and docker-compose on a fresh CentOS 7 x64 VPS, following instructions up to Step Two of How To Install and Use Docker on CentOS 7, and and up to Step One of How To Install and Use Docker Compose on CentOS 7.
However, I ran into an error on this command:
$ git clone https://gitlab.com/revant.one/erpnext_kubernetes && cd erpnext_kubernetes
Cloning into 'erpnext_kubernetes'...
error: RPC failed; result=22, HTTP code = 404
fatal: The remote end hung up unexpectedly
Thanks @revant_one! I am going to try this out this weekend. I have had pretty poor performance (write latency) with s3fs style solutions previously, did csi-s3 work OK?
I have a schedule that builds version tagged images for v11, v12 and edge. Check updated post
It is ready for production deployment, checkout the gitlab repo.
I’ll write a separate post on
how to setup a dockerized erpnext in production with letsencrypt.
how to setup erpnext on kubernetes using helm chart
I am getting Site test.localhost already exists
I can, however go through with this when I choose a different site name. However I can not access my http://test2.localhost site afterwards
also in the terminal that is running the docker-compose ... up command I have noted
mariadb_1 | 2019-12-08 23:57:02 50 [Warning] Access denied for user '_fcadc07dbdbc21db'@'172.21.0.10' (using password: YES)
Have been trying to experiment with this but could not get it to successfully run. I’m up to a point where I can see the “Sorry! We will be back soon” page.
The erpnext_kubernetes_erpnext-python_1 container seems to have started fine based on the logs:
But I can not get past this. Any idea what I may be doing wrong?
Nginx will proxy to site available under bench/venv
name the sites with .localhost, e.g erp.localhost for trying out locally. That way you don’t have to edit /etc/hosts and browser treats localhost as secure
I tried IP address (container’s IP) and localhost, both did not work. Is there a way to get it working using port based?
I named the site as server1-docker.domain.local which is not the actual hostname of the server but something I made up, I also configured DNS to point this hostname to the actual server itself. Did I do something wrong here?