We’ve recently setup a CD pipeline that deploys review, staging, and (eventually) production versions of an application using Rancher. we use the rancher CLI to perform most of the deployment jobs and It works reasonably well, but one thing I’ve noticed is that while changes to our docker-compose.yml are always picked up in during a “rancher up -u” command, anything in the rancher-compose.yml is just ignored after the initial deployment.
I don’t understand why this would be the case. In our situation, we attach our application to an existing rancher load balancer using a selector label in the application’s docker-compose.yml file and then in the rancher-compose.yml we specify the port_rules for that connection; most importantly, a “hostname”. This hostname is pretty dynamic and specified in the CD pipeline, but we’re kind of stuck in that once the application is deployed, trying to upgrade it with changes in rancher-compose file is pointless, even with a --force-upgrade option on the command line. Any changes int hat file are just ignored.
Am I missing something important here? Why can’t we “upgrade” the rancher-compose.yml like we can the docker-compose.yml?