Skip to main content
Archived

Version Control in FME Server: Next Steps

Related products:FME Flow
  • August 23, 2017
  • 26 replies
  • 216 views

Show first post
This post is closed to further activity.
It may be an old question, an answered question, an implemented idea, or a notification-only post.
Please check post dates before relying on any information in a question or answer.
For follow-up or related questions, please post a new question or idea.
If there is a genuine update to be made, please contact us and request that the post is reopened.

26 replies

smol
Contributor
  • Contributor
  • August 17, 2021

To give some feedback about this and share my ideas on this.

 

We have been looking at supporting release cycles with version control in the last year. For us, the above feature did not fully align with our setup. We use FME Workbenches as a service for our public-facing services, which means the workbenches are very much coupled to these consuming services. This means it is quite important for us to keep in sync with the release cycles of our applications.

 

Our applications use git with CI/CD to manage the release cycles, and thus makes it is quite important to let the coupled FME workbenches follow this same principle. Luckily this was possible for us by using the FME API to update the workbenches in a release pipeline. This works perfectly as we can also let all the workbenches stay within the same application repository and falls in line with our idea of "horizontal repositories" (repo containing all its dependencies) with embedded FME team members. But I can imagine this can become better out of the box.

 

So this works quite okay for us for the Workbenches and even the FME Transformers. There are however many more settings/entities that are not defined as code at all as they are not related to FME Desktop. Think about App Servers, Schedule definitions, Automations, etc. Many of these definitions are also dependent on a git-based release cycle and it is currently very hard to keep these in sync. A possibility is to trigger (API) FME Projects migrations through a release pipeline, but this can be improved to have a git-first approach.

 

The above post is an FME -> Git approach, but I would advocate to (also) have a Git -> FME approach mainly so things are explicitly peer-reviewed first and fall under the Git version control.