By default, `kubeadm` sets `100.64.0.0/12` as the subnet for services. This means when a service is created, its cluster IP, if not manually specified,
By default, `kubeadm` sets `100.64.0.0/12` as the subnet for services. This means when a service is created, its cluster IP, if not manually specified,
will be automatically assigned from the services subnet. If you would like to set a different one, use `--service-cidr`.
will be automatically assigned from the services subnet. If you would like to set a different one, use `--service-cidr`.
By default, `kubeadm` sets `cluster.local` as the cluster DNS domain. If you would like to set a different one, use `--service-dns-domain`.
By default, `kubeadm` sets `cluster.local` as the cluster DNS domain. If you would like to set a different one, use `--service-dns-domain`.
-`--schedule-workload=<bool>` (default: "false")
-`--schedule-workload=<bool>` (default: "false")
By default, `kubeadm` sets the master node kubelet as non-schedulable for workloads. This means the master node won't run your pods. If you want to change that,
By default, `kubeadm` sets the master node kubelet as non-schedulable for workloads. This means the master node won't run your pods. If you want to change that,
use `--schedule-workload=true`.
use `--schedule-workload=true`.
-`--cloud-provider=<cloud provider>`
-`--cloud-provider=<cloud provider>`
By default, `kubeadm` doesn't perform auto-detection of the current cloud provider. If you want to specify it, use `--cloud-provider`. Possible values are
By default, `kubeadm` doesn't perform auto-detection of the current cloud provider. If you want to specify it, use `--cloud-provider`. Possible values are
the ones supported by controller-manager, namely `"aws"`, `"azure"`, `"cloudstack"`, `"gce"`, `"mesos"`, `"openstack"`, `"ovirt"`, `"rackspace"`, `"vsphere"`.
the ones supported by controller-manager, namely `"aws"`, `"azure"`, `"cloudstack"`, `"gce"`, `"mesos"`, `"openstack"`, `"ovirt"`, `"rackspace"`, `"vsphere"`.
***TODO(phase1+)***
***TODO(phase1+)***
...
@@ -61,37 +61,3 @@ Here's an example on how to use it:
...
@@ -61,37 +61,3 @@ Here's an example on how to use it:
-`--token=<token>`
-`--token=<token>`
By default, when `kubeadm init` runs, a token is generated and revealed in the output. That's the token you should use here.
By default, when `kubeadm init` runs, a token is generated and revealed in the output. That's the token you should use here.
# User Experience Considerations
> ***TODO*** _Move this into the design document
a) `kube-apiserver` will listen on `0.0.0.0:443` and `127.0.0.1:8080`, which is not configurable right now and make things a bit easier for the MVP
b) there is `kube-discovery`, which will listen on `0.0.0.0:9898`
from the point of view of `kubeadm init`, we need to have
a) a primary IP address as will be seen by the nodes and needs to go into the cert and `kube-discovery` configuration secret
b) some other names and addresses API server may be known by (e.g. external DNS and/or LB and/or NAT)
from that perspective we don’t can assume default ports for now, but for all the address we really only care about two ports (i.e. 443 and 9898)
we should make ports configurable and expose some way of making API server bind to a specific address/interface
but I think for the MVP we need solve the issue with hardcode IPs and DNS names in the certs
so it sounds rather simple enough to introduce `--api-advertise-addr=<ip>` and `--api-external-dns-name=<domain>`, and allowing multiple of those sounds also simple enough
from the `kubeadm join` perspective, it cares about the two ports we mentioned, and we can make those configurable too
but for now it’s easier to pass just the IP address
plust it’s also require, so passing it without a named flag sounds convenient, and it’s something users are familiar with
that’s what Consul does, what what Weave does, and now Docker SwarmKit does the same thing also (edited)
flags will differ, as there are some Kubernetes-specifics aspects to what join does, but basic join semantics will remain _familiar_
if we do marry `kube-discovery` with the API, we might do `kubeadm join host:port`, as we’d end-up with a single port to care about
// TODO look into what this really means, scheduler prints it for some reason
//
//E0817 17:53:22.242658 1 event.go:258] Could not construct reference to: '&api.Endpoints{TypeMeta:unversioned.TypeMeta{Kind:"", APIVersion:""}, ObjectMeta:api.ObjectMeta{Name:"kube-scheduler", GenerateName:"", Namespace:"kube-system", SelfLink:"", UID:"", ResourceVersion:"", Generation:0, CreationTimestamp:unversioned.Time{Time:time.Time{sec:0, nsec:0, loc:(*time.Location)(nil)}}, DeletionTimestamp:(*unversioned.Time)(nil), DeletionGracePeriodSeconds:(*int64)(nil), Labels:map[string]string(nil), Annotations:map[string]string(nil), OwnerReferences:[]api.OwnerReference(nil), Finalizers:[]string(nil)}, Subsets:[]api.EndpointSubset(nil)}' due to: 'selfLink was empty, can't make reference'. Will not report event: 'Normal' '%v became leader' 'moby'
// WriteStaticPodManifests builds manifest objects based on user provided configuration and then dumps it to disk
// WriteStaticPodManifests builds manifest objects based on user provided configuration and then dumps it to disk
