In the process of setting up a #Docker based data processing service, to be deployed in a #PaaS system (#OKD)
Containers ran fine on our local machines, so we pushed it...
Instant errors on OKD, despite using the exact same image as our own.
The culprit?
OKD forcibly runs the process as some random UID, differing from docker behaviour...
So what's the point of containerisation if we still find ways to cause "But it works on my machine!"
Correct us if we're wrong, but if your docker image requires the user to trigger a download of python packages after spinning up a new container...
You're not using docker right, right?
The python deps. should be in the image, right?
Anyhow, we're experimenting with
#InvenTree https://inventree.org/
It looks like a far more modern version of #PartKeepr, but its setup could use a bit of polish :P
@xaseiresh hi there! No one from #inventree was very active here so we did not see you post.
The update command checks if the dependencies are up to date as docker is not the only deployment method and the update steps are the same across all distribution methods but this should be optional. The image you can pull from dockerhub is fully self-contained and should just work.
Is this still an issue for you? Happy to help!
Xasin, Neira & Mesh
Matthias Mair