Since you are not using HA supervisor and referencing a specific HA core container version I guess you won't be able to use the update mechanism from within HA itself? How are you dealing with that aspect?
Updating to new versions of HA? I have an update entity in HA to let me know when a new version comes out, and when it does, I just edit that line in the deployment and "kubectl apply -f" it. A little more work, I suppose, but since I like to read through the detailed release notes to figure out what might break with each version anyway, not all that much.
There is one change worth making to the above files (I really need to make a post on everything I've learned since that last post, so I'll try and get that one out in the next couple of weeks) to make this go more fluidly - in the spec for the Home Assistant deployment, between the "selector:" and "template:" clauses, you should add:
strategy:
type: Recreate
That's because by default Kubernetes will try to start up the upgraded pod before it terminates the old one, and HA can be a mite cranky about two of it trying to run at once. Setting the strategy to "Recreate" means that it won't do this, and will terminate before restarting. No muss, no fuss.
I have the master nodes on the x64 machines (even though I need to add another one to get things working properly); I didn't think it would be a great idea to hammer the Pis with etcd.
Since you are not using HA supervisor and referencing a specific HA core container version I guess you won't be able to use the update mechanism from within HA itself? How are you dealing with that aspect?
Updating to new versions of HA? I have an update entity in HA to let me know when a new version comes out, and when it does, I just edit that line in the deployment and "kubectl apply -f" it. A little more work, I suppose, but since I like to read through the detailed release notes to figure out what might break with each version anyway, not all that much.
There is one change worth making to the above files (I really need to make a post on everything I've learned since that last post, so I'll try and get that one out in the next couple of weeks) to make this go more fluidly - in the spec for the Home Assistant deployment, between the "selector:" and "template:" clauses, you should add:
strategy:
type: Recreate
That's because by default Kubernetes will try to start up the upgraded pod before it terminates the old one, and HA can be a mite cranky about two of it trying to run at once. Setting the strategy to "Recreate" means that it won't do this, and will terminate before restarting. No muss, no fuss.
I curious a bout your kubernetes setup. Are you control nodes on the pis? Or are you just running one control node?
I have the master nodes on the x64 machines (even though I need to add another one to get things working properly); I didn't think it would be a great idea to hammer the Pis with etcd.
Wow ! Thank you! I got a happy 400 but hey. Need to deploy w/o that helm guy.
Added PV / PVC and adapted to Nginx ingress