I currently have a php server (using yii) that uses a backend mysql database and i imported into gitlab as admin/test-ws.git. deployment today is NOT devops… ie i have a physical server containing apache/php connecting to another physical server with mysql.
this project has NO tests or composer.json in the root directory (because who did that 3 years ago)
i installed kubernetes in aws and have it configured in gitlab (12.10-ce) with runners inside of k8s. On every checkin “auto devops” runs and always fails at TEST and pretty much every other step. Googling this, it seems i need to setup environment variables and write my OWN .gitlab-ci.yml defeating the whole point of “auto” in autodevops. Docs are always confusing on this as it always switches from gitlab 8 (create yml) to gitlab 12 (auto) so i never know which applies to me. ok, fine, i have to manually write a 400 line yml file, so i use this template https://gitlab.com/gitlab-org/gitlab-ci-yml/blob/master/Auto-DevOps.gitlab-ci.yml but i still dont understand how it creates a container and then configures my php server to talk to mysql server which is a completely different project. and then actually deploy to the kubernetes cluster in production.
I wrote a docker-compose.yml file i use to deploy in a pod two containers, the php web server and the mysql webserver with volumes… however in “prod” i really want php and mysql to be separate and i have no idea where to start as “autodevops” only seems to work on “hello world” style applications.
I guess my question comes down to i must be missing something about how the system works because other people and internet implies it works. so what am i missing in that i must write a 400 line yaml file for every project?
also docs seem to imply that i would manage the deployment in production using helm, but again, i dont see where it creates the helm charts and how i would deploy/rollout new ones.
also the template has
but i dont use postgres so i am not sure what those values should be.