Re:Kubernetes 1.10 released
by outworlder on Wednesday March 28, 2018 @04:39PM
(#16701703)
> Is there an easy way to get a single-node production-ready Kubernetes instance?
Not really. There are plenty of ways of getting a single node instance. None of them will give you a "production-ready" one, because they don't define it that way (and I happen to agree). You can of course do whatever you want.
Since you are using VMWare anyway, why can't you spin up more VMs (maybe smaller)? You can vmotion them away to different nodes when you are ready to actually make the cluster HA. It is a really really good idea to keep master and workers separate, even if you run a single node for each.
Failure of the worker will of course bring down your applications. When you recover it or spin up another one, K8s will recover your apps for you. Failure of the master will not adversely affect your systems only the cluster's ability to manage itself and its self-healing capabilities (which will affect uptime at some point).
Failure of the combined master/worker/etcd node should be recoverable, but frankly, at this point, should you care? I would just shoot it in the head and add some automation to provision a brand new cluster and deploy those applications again. Since you are not worried about HA and just want a place to deploy the containers, just make the k8s single-node-cluster cattle.
Not really. There are plenty of ways of getting a single node instance. None of them will give you a "production-ready" one, because they don't define it that way (and I happen to agree). You can of course do whatever you want.
Since you are using VMWare anyway, why can't you spin up more VMs (maybe smaller)? You can vmotion them away to different nodes when you are ready to actually make the cluster HA. It is a really really good idea to keep master and workers separate, even if you run a single node for each.
Failure of the worker will of course bring down your applications. When you recover it or spin up another one, K8s will recover your apps for you. Failure of the master will not adversely affect your systems only the cluster's ability to manage itself and its self-healing capabilities (which will affect uptime at some point).
Failure of the combined master/worker/etcd node should be recoverable, but frankly, at this point, should you care? I would just shoot it in the head and add some automation to provision a brand new cluster and deploy those applications again. Since you are not worried about HA and just want a place to deploy the containers, just make the k8s single-node-cluster cattle.