* You can choose where to install the master components. If you want your master node to act both as master (api,scheduler,controller) and node (e.g. accept workloads, create pods ...),
the server address has to be present on both groups 'kube-master' and 'kube-node'.
* Almost all kubernetes components are running into pods except *kubelet*. These pods are managed by kubelet which ensure they're always running
* For safety reasons, you should have at least two master nodes and 3 etcd servers
* Kube-proxy doesn't support multiple apiservers on startup ([Issue 18174](https://github.com/kubernetes/kubernetes/issues/18174)). An external loadbalancer needs to be configured.
@ -155,16 +152,6 @@ You can choose between 2 network plugins. Only one must be chosen.
The choice is defined with the variable '**kube_network_plugin**'
### Expose a service
There are several loadbalancing solutions.
The one i found suitable for kubernetes are [Vulcand](http://vulcand.io/) and [Haproxy](http://www.haproxy.org/)
My cluster is working with haproxy and kubernetes services are configured with the loadbalancing type '**nodePort**'.
eg: each node opens the same tcp port and forwards the traffic to the target pod wherever it is located.
Then Haproxy can be configured to request kubernetes's api in order to loadbalance on the proper tcp port on the nodes.
Please refer to the proper kubernetes documentation on [Services](https://github.com/kubernetes/kubernetes/blob/release-1.0/docs/user-guide/services.md)
### Check cluster status
@ -199,7 +186,7 @@ There are two ways of installing new apps
Additionnal apps can be installed with ```ansible-galaxy```.
ou'll need to edit the file '*requirements.yml*' in order to chose needed apps.
you'll need to edit the file '*requirements.yml*' in order to chose needed apps.
The list of available apps are available [there](https://github.com/ansibl8s)
For instance it is **strongly recommanded** to install a dns server which resolves kubernetes service names.
@ -268,6 +255,18 @@ calicoctl pool show
calicoctl endpoint show --detail
```
##### Optionnal : BGP Peering with border routers
In some cases you may want to route the pods subnet and so NAT is not needed on the nodes.
For instance if you have a cluster spread on different locations and you want your pods to talk each other no matter where they are located.
The following variables need to be set:
**peer_with_router** enable the peering with border router of the datacenter (default value: false).
you'll need to edit the inventory and add a and a hostvar **local_as** by node.
```
node1 ansible_ssh_host=95.54.0.12 local_as=xxxxxx
```
#### Flannel
* Flannel configuration file should have been created there
Antoine Legrand
8 years ago