I'm currently working with Helm charts, where I have a "client" Helm chart that has a direct dependency on a "server" Helm chart. The server chart needs to be fully deployed and ready before the client chart can be deployed.
I want to ensure that Helm waits for the successful deployment of the server chart before starting the deployment of the client chart. Specifically, I need to know if there's a way to:
Wait for the server chart to be fully deployed, including all services and pods being in the "Running" state. Trigger the client chart deployment only after the server chart has been successfully deployed. I've looked into Helm hooks, but I'm not sure how to implement this dependency in a way that ensures the client doesn't start until the server is fully up and running.
What I’ve tried so far:
Helm doesn't have such functionality. Each helm chart that is getting installed is independent package. If you want to achieve such dependency between multiple helm charts, you have to do that in your installation environment using bash scripts or any other scripts. That's what we do in our organization - we use jenkins to deploy our helm packages to cluster using scripts we achieve this dependency between multiple helm packages.