In OneDev a build is designed to be fully reproducible. For this purpose, it should only relies on files in git repository and artifacts retrieved via project/job dependencies (can be declared in job definition).

So cache in OneDev is merely used to increase speed. In case cache is not available the build should also work, just slower.

Cache can be defined in more settings of a job, it expects a key, and a relative path to job workspace. This relative path is initially empty, and your build may store intermediate files into this path. Next time when your build starts, the cache will be mounted with intermediate files generated in previous build.

Thanks Robin, this is exactly what I was looking for. You might want to create an entry for that feature in the documentation.

Really, OneDev ROCKS!!!

The previous recommendation worked perfectly but I want to go a step further. I am building docker images that depend on a "base" image, vsr-bas. So:

• I created a project for one of the "child" images, vsr-geo
• I defined its parent as vsr-bas
• I pull the vsr-bas:latest image from my registry
• and build the new image from there

Knowing that vsr-geo is a child project of vsr-bas and that vsr-bas has a cache defined, how do I access the same cache from vsr-bas?

Cache with same key will be shared even if they belongs to different project or job . You may use same cache key in vsr-geo as defined in vsr-bas.

On the other hand, if you want project/job specific cache, just enclose project path/job name variable in the cache key.