github/kubernetes
1
0
Fork 0
mirror of https://github.com/k3s-io/kubernetes.git synced 2025-09-06 03:33:26 +00:00
Code Issues Projects Releases Wiki Activity

Run gendocs

Browse Source
This commit is contained in:
Tim Hockin
2015-07-17 15:35:41 -07:00
parent aacc4c864c
commit 33f1862830
210 changed files with 599 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
docs/getting-started-guides/rackspace.md
View File

@@ -58,23 +58,26 @@ The current cluster design is inspired by:
- [Angus Lees](https://github.com/anguslees/kube-openstack)
## Prerequisites
1. Python2.7
2. You need to have both `nova` and `swiftly` installed. It's recommended to use a python virtualenv to install these packages into.
3. Make sure you have the appropriate environment variables set to interact with the OpenStack APIs. See [Rackspace Documentation](http://docs.rackspace.com/servers/api/v2/cs-gettingstarted/content/section_gs_install_nova.html) for more details.
##Provider: Rackspace
## Provider: Rackspace
- To build your own released version from source use `export KUBERNETES_PROVIDER=rackspace` and run the `bash hack/dev-build-and-up.sh`
- Note: The get.k8s.io install method is not working yet for our scripts.
* To install the latest released version of kubernetes use `export KUBERNETES_PROVIDER=rackspace; wget -q -O - https://get.k8s.io | bash`
## Build
1. The kubernetes binaries will be built via the common build scripts in `build/`.
2. If you've set the ENV `KUBERNETES_PROVIDER=rackspace`, the scripts will upload `kubernetes-server-linux-amd64.tar.gz` to Cloud Files.
2. A cloud files container will be created via the `swiftly` CLI and a temp URL will be enabled on the object.
3. The built `kubernetes-server-linux-amd64.tar.gz` will be uploaded to this container and the URL will be passed to master/nodes when booted.
## Cluster
There is a specific `cluster/rackspace` directory with the scripts for the following steps:
1. A cloud network will be created and all instances will be attached to this network.
- flanneld uses this network for next hop routing. These routes allow the containers running on each node to communicate with one another on this private network.
@@ -83,6 +86,7 @@ There is a specific `cluster/rackspace` directory with the scripts for the follo
4. We then boot as many nodes as defined via `$NUM_MINIONS`.
## Some notes
- The scripts expect `eth2` to be the cloud network that the containers will communicate across.
- A number of the items in `config-default.sh` are overridable via environment variables.
- For older versions please either:
@@ -92,6 +96,7 @@ There is a specific `cluster/rackspace` directory with the scripts for the follo
* Download a [snapshot of `v0.3`](https://github.com/GoogleCloudPlatform/kubernetes/archive/v0.3.tar.gz)
## Network Design
- eth0 - Public Interface used for servers/containers to reach the internet
- eth1 - ServiceNet - Intra-cluster communication (k8s, etcd, etc) communicate via this interface. The `cloud-config` files use the special CoreOS identifier `$private_ipv4` to configure the services.
- eth2 - Cloud Network - Used for k8s pods to communicate with one another. The proxy service will pass traffic via this interface.