Proper way to setup Espo CRM with Git

[smaction]

Our company is doing a lot of customization of Espo. We test it on a staging server and then plan to move it to a production server via Git. Obviously things like users are not the same on staging and production. What files do we need to have in the repository? What files can be ignored? What type of customization is in datbase? how do you recommend merging those changes into production?

We have been experimenting but a lot of work has to be done manually after doing a pull from the branch.


TIA.

Scott

[yuri]

Hi Scott,

/data dir - config, cache, uploaded files, logs.
/custom dir - customizations done by administrator (like custom fields, custom layouts, labels, etc.)


There are no customization defenitions stored in database. All is stored in files (custom dir).

You can track these dirs in your repository:

/application
/custom
/client

How we do:
We have dev reposiory and there is a builder that builds an extension package. We upload the extension to the production (Administration > Extension). This allows us to upgrade w/o a hassle. Since usualy there are some scripts in upgrade packages, it can be a problem to upgrade crm if you use git to push changes to the production.