Helm deployment architecture

The official helm deployment of ERPnext is designed for a production environment. I’ve found the new-site chart, which allows for creation of new sites. What I’m not finding is the intended dev->stage->Prod deployment architecture.

Are we supposed to deploy the new site via chart in Prod and then simply migrate an existing folder structure into the Prod /sites directory? This would overwrite the data created by the chart (and probably break things).

How was movement of data between sites intended to be accomplished?


I added this “Job” yesterday, new-site is not a chart.
Not really useful for production if you’re installing more than 1 site.
This is a job that will create 1st site and if you enable ingress it’ll also create first ingress.
I’m planning to use this feature to deploy and tear-down helm releases for staging

Install the same chart as 3 separate helm releases. Customize and manage 3 separate values.yaml

new-site job will create site with name set in values in createSite.siteName. If that is unique name it’ll not conflict. If there is already existing site with same name it’ll fail with “site already exists” error.

Which data? use ReST API?
Files data? to copy files data, do it manually using the PVC and mounting it in Pods.
Database? it resides in separate service. Back it up and restore in mariadb.

Kubernetes based installation of ERPNext gives you variety of options that are available in cloud native setups.
Helm Chart is just another Helm Chart for any environment (prod/dev/staging/local) you plan to install it on.