Migrate docs to use new libnetworking UX

calico_containers/calico_ctl/__init__.py

@@ -1 +1 @@
-__version__ = "0.7.0.dev"
+__version__ = "0.8.0.dev"

docs/getting-started/default-networking/Demonstration.md

@@ -31,7 +31,7 @@ You should see output like this on each node
 
     vagrant@calico-01:~$ docker ps -a
     CONTAINER ID        IMAGE                    COMMAND                CREATED             STATUS              PORTS                                            NAMES
-    39de206f7499        calico/node:v0.7.0   "/sbin/my_init"        2 minutes ago       Up 2 minutes                                                         calico-node
+    39de206f7499        calico/node:v0.8.0   "/sbin/my_init"        2 minutes ago       Up 2 minutes                                                         calico-node
 
 ## Networking containers.
 
@@ -120,6 +120,6 @@ the host.
 
 For example:
 
-    calicoctl node --ip=172.17.8.101 --ip6=fd80:24e2:f998:72d7::1 --libnetwork
+    calicoctl node --ip=172.17.8.101 --ip6=fd80:24e2:f998:72d7::1
 
 See the [IPv6 demonstration](DemonstrationIPv6.md) for a worked example.

docs/getting-started/default-networking/EnvironmentSetup.md

@@ -70,7 +70,7 @@ Get the calicoctl binary onto each host.  You can download a specific
 [release][calico-releases] from github.  
 For example, to retrieve the latest v0.7.0 release, on each host run
 
-	wget https://github.com/projectcalico/calico-docker/releases/download/v0.7.0/calicoctl
+	wget https://github.com/projectcalico/calico-docker/releases/download/v0.8.0/calicoctl
 	chmod +x calicoctl
 
 This binary should be placed in your `$PATH` so it can be run from any
@@ -83,7 +83,7 @@ run `calicoctl node` the first time.  Select the same version of the Calico
 Docker image as you selected above.  For example, to pull the latest released 
 version
 
-    docker pull calico/node:v0.7.0
+    docker pull calico/node:v0.8.0
 
 ## Final checks
 
@@ -94,9 +94,9 @@ Check that the hosts have IP addresses assigned, and that your hosts can ping
 one another.
 
 You should also verify each host can access etcd.  The following will return 
-an error if etcd is not available.
+the current etcd version if etcd is available.
 
-    etcdctl ls /
+    curl -L http://127.0.0.1:4001/version
 
 [etcd]: https://coreos.com/etcd/docs/latest/
 [calico-releases]: https://github.com/projectcalico/calico-docker/releases/

docs/getting-started/default-networking/cloud-config/user-data-first

@@ -45,6 +45,6 @@ write_files:
   owner: root
   content: |
     #!/usr/bin/bash -e
-    wget -O /opt/bin/calicoctl https://github.com/projectcalico/calico-docker/releases/download/v0.7.0/calicoctl
+    wget -O /opt/bin/calicoctl https://github.com/projectcalico/calico-docker/releases/download/v0.8.0/calicoctl
     chmod +x /opt/bin/calicoctl
     /opt/bin/calicoctl checksystem --fix

docs/getting-started/default-networking/cloud-config/user-data-others

@@ -40,6 +40,6 @@ write_files:
   owner: root
   content: |
     #!/usr/bin/bash -e
-    wget -O /opt/bin/calicoctl https://github.com/projectcalico/calico-docker/releases/download/v0.7.0/calicoctl
+    wget -O /opt/bin/calicoctl https://github.com/projectcalico/calico-docker/releases/download/v0.8.0/calicoctl
     chmod +x /opt/bin/calicoctl
     /opt/bin/calicoctl checksystem --fix

docs/getting-started/default-networking/vagrant-coreos/Vagrantfile

@@ -8,7 +8,7 @@
 # image which slows down the install process, but speeds up the demonstration.
 # This version should match the version of calicotl installed in the cloud
 # config files.
-calico_docker_ver = "v0.7.0"
+calico_docker_ver = "v0.8.0"
 
 # Size of the cluster created by Vagrant
 num_instances=2

docs/getting-started/default-networking/vagrant-ubuntu/Vagrantfile

@@ -2,7 +2,7 @@
 # vi: set ft=ruby :
 
 # The version of calico to install
-calico_docker_ver = "v0.7.0"
+calico_docker_ver = "v0.8.0"
 
 # Size of the cluster created by Vagrant
 num_instances=2

docs/getting-started/libnetwork/Demonstration.md

@@ -1,5 +1,7 @@
 # Calico as a libnetwork plugin.
-This demonstration uses Docker's native [libnetwork network driver](https://github.com/docker/libnetwork), available in the Docker [experimental channel](https://github.com/docker/docker/tree/master/experimental) alongside the Docker 1.8 release. Docker's experimental channel is still moving fast and some of its features are not yet fully stable, so we recommend using a known, working version of the experimental docker binary from calico-docker release [0.5.3](https://github.com/projectcalico/calico-docker/releases/tag/v0.5.3) (July 21) or [0.5.4](https://github.com/projectcalico/calico-docker/releases/tag/v0.5.4) (August 5, optimized for 64-bit Ubuntu).
+This demonstration uses Docker's native 
+[libnetwork network driver](https://github.com/docker/libnetwork), available 
+from Docker 1.9 release and above.
 
 ## Environment
 This demonstration makes some assumptions about the environment you have. See 
@@ -27,60 +29,86 @@ This will start a container on each host. Check they are running
 
 You should see output like this on each node
 
-    vagrant@calico-01:~$ docker ps -a
-    CONTAINER ID        IMAGE                    COMMAND                CREATED             STATUS              PORTS                                            NAMES
-    39de206f7499        calico/node-libnetwork   "/sbin/my_init"        2 minutes ago       Up 2 minutes                                                         calico-node
-
+    vagrant@calico-01:~$ docker ps
+    CONTAINER ID        IMAGE                           COMMAND                  CREATED             STATUS              PORTS               NAMES
+    eec9ebbfb486        calico/node-libnetwork:v0.4.0   "./start.sh"             21 seconds ago      Up 19 seconds                           calico-libnetwork
+    ffe6cb403e9b        calico/node:v0.8.0              "/sbin/my_init"          21 seconds ago      Up 20 seconds                           calico-node
 
 ## Creating networked endpoints
 
-The experimental channel version of Docker introduces a new flag to 
-`docker run` to network containers:  `--publish-service <service>.<network>.<driver>`.
+As of Docker 1.9, the integration of Docker with libnetwork introduces a new
+mechanism to provide networking for Docker containers.
+
+The new command `docker network` can be used to create a logical network.  
+A new flag is introduced to `docker run` to join a container to a particular 
+network:  `--net <network>`.
 
- * `<service>` is the name by which you want the container to be known on the network.
- * `<network>` is the name of the network to join.  Containers on different networks cannot communicate with each other.
- * `<driver>` is the name of the network driver to use.  Calico's driver is called `calico`.
+So let's go ahead and create some networks and start a few containers 
+on each host spread between these networks.
 
-So let's go ahead and start a few containers on each host.
+On either host, create three networks:
+
+    docker network create --driver=calico --subnet=192.168.0.0/24 net1
+    docker network create --driver=calico --subnet=192.168.1.0/24 net2
+    docker network create --driver=calico --subnet=192.168.2.0/24 net3
+
+Note that we use the Calico driver `calico`.  This driver is run within 
+the calico-node container.  We explicitly choose an IP Pool for each network
+rather than using the default selections - this is to avoid potential conflicts
+with the default NAT IP assignment used by VirtualBox.  Depending on your
+specific environment, you may need to choose different IP Pool CIDRs.
 
 On calico-01
 
-    docker run --publish-service srvA.net1.calico --name workload-A -tid busybox
-    docker run --publish-service srvB.net2.calico --name workload-B -tid busybox
-    docker run --publish-service srvC.net1.calico --name workload-C -tid busybox
+    docker run --net net1 --name workload-A -tid busybox
+    docker run --net net2 --name workload-B -tid busybox
+    docker run --net net1 --name workload-C -tid busybox
 
 On calico-02
 
-    docker run --publish-service srvD.net3.calico --name workload-D -tid busybox
-    docker run --publish-service srvE.net1.calico --name workload-E -tid busybox
+    docker run --net net3 --name workload-D -tid busybox
+    docker run --net net1 --name workload-E -tid busybox
 
 By default, networks are configured so that their members can communicate with 
 one another, but workloads in other networks cannot reach them.  A, C and E are
- all in the same network so should be able to ping each other.  B and D are in 
- their own networks so shouldn't be able to ping anyone else.
+all in the same network so should be able to ping each other.  B and D are in 
+their own networks so shouldn't be able to ping anyone else.
 
-On calico-01 check that A can ping C and E.
+On calico-01 check that A can ping C and E.  You can use either the IP address
+or the name for containers within the same network.
 
-    docker exec workload-A ping -c 4 srvC
-    docker exec workload-A ping -c 4 srvE
+    docker exec workload-A ping -c 4 `docker inspect --format "{{ .NetworkSettings.IPAddress }}" workload-C`
+    docker exec workload-A ping -c 4 workload-E.net1
 
-Also check that A cannot ping B or D
+Also check that A cannot ping B or D.  This is slightly trickier because the
+hostnames for different networks will not be added to the host configuration of
+the container - so we need to determine the IP addresses assigned to containers
+B and D.
 
-    docker exec workload-A ping -c 4 srvB
-    docker exec workload-A ping -c 4 srvD
+On calico-01 run
 
-To see the list of networks, use
+    docker inspect --format "{{ .NetworkSettings.IPAddress }}" workload-B
+
+this returns the IP address of workload-B.
+
+On calico-02 run
 
-    docker network ls
+    docker inspect --format "{{ .NetworkSettings.IPAddress }}" workload-D
+
+This returns the IP address of workload-D.
 
-## IPv6 (Optional)
+On calico-01 run
 
-IPv6 networking is also supported.  If you are using IPv6 address spaces as
-well, start your Calico node passing in both the IPv4 and IPv6 addresses of
-the host.
+    docker exec workload-A ping -c 4 <IP address of B>
+    docker exec workload-A ping -c 4 <IP address of D>
 
-For example:
+replacing the `<...>` with the appropriate IP address for B and D.  These pings
+will fail.
+
+To see the list of networks, use
 
-    calicoctl node --ip=172.17.8.101 --ip6=fd80:24e2:f998:72d7::1 --libnetwork
+    docker network ls
+
+## IPv6 (Optional)
 
-See the [IPv6 demonstration](DemonstrationIPv6.md) for a worked example.
+IPv6 networking is not yet supported for Calico networking with libnetwork.

docs/getting-started/libnetwork/EnvironmentSetup.md

@@ -37,9 +37,9 @@ The servers also need:
 - A specific Docker release to be running - since the Calico agent is packaged
 as a Docker container, and the libnetwork features required are currently
 only available in an experimental release.
-- A consul server used for clustering Docker.
+- A consul server used for clustering Docker
 - An Etcd cluster - which Calico uses for coordinating state between the nodes.
-- The `calicoctl` to be placed in the `$PATH`.
+- The `calicoctl` binary to be placed in the system `$PATH`.
 
 ## Requirements
 
@@ -50,35 +50,16 @@ We recommend configuring the hosts with the hostname `calico-01` and
 `calico-02`.  The demonstration will refer to these hostnames.
 
 They must have the following software installed:
-- The experimental release of [Docker](#experimental-docker)
+- [Docker 1.9 or greater](#Docker)
 - etcd installed and available on each node: [etcd documentation][etcd]
 - `ipset`, `iptables`, and `ip6tables` kernel modules.
-- A [consul server](#consul) running on calico-01
 
-### Consul
+### Docker
 
-To install consul, download and unzip the consul binary and give it executable
-permissions.  For example:
-
-    wget https://dl.bintray.com/mitchellh/consul/0.5.2_linux_amd64.zip -O consul.zip
-    unzip -o consul.zip
-    rm consul.zip
-    chmod +x consul
-
-You can start consul using the following:
-
-    ./consul agent -server -bootstrap-expect 1 -data-dir /tmp/consul -client <IPV4>
-
-where <IPV4> is replaced with your appropriate IPv4 address.  This address 
-should be accessible by both servers.
-
-### Experimental Docker
-
-Follow the instructions for installing the 
-[experimental channel of Docker][experimental-docker-git].
+Follow the instructions for installing
+[Docker][docker].
 
-Docker's experimental channel is still moving fast and some of its 
-features are not yet fully stable.
+A version of 1.9 or greater is required.
 
 ### Docker permissions
 
@@ -98,7 +79,7 @@ Get the calicoctl binary onto each host.  You can download a specific
 [release][calico-releases] from github.  
 For example, to retrieve the latest v0.7.0 release, on each host run
 
-	wget https://github.com/projectcalico/calico-docker/releases/download/v0.7.0/calicoctl
+	wget https://github.com/projectcalico/calico-docker/releases/download/v0.8.0/calicoctl
 	chmod +x calicoctl
 
 This binary should be placed in your `$PATH` so it can be run from any
@@ -110,7 +91,7 @@ You can optionally preload this image to avoid the delay when you run
 `calicoctl node --libnetwork` the first time.  For example, to pull the latest 
 released version, run
 
-    docker pull calico/node-libnetwork:v0.3.0
+    docker pull calico/node-libnetwork:v0.4.0
 
 ## Final checks
 
@@ -120,17 +101,17 @@ you'll need to adjust the demonstration instructions accordingly.
 Check that the hosts have IP addresses assigned, and that your hosts can ping
 one another.
 
-Check that you are running with the experimental version of Docker.
+Check that you are running with a suitable version of Docker.
 
     docker version
 
-It should indicate a version of 1.8.0 and experimental.
+It should indicate a version of 1.9 or greater.
 
 You should also verify each host can access etcd.  The following will return 
-an error if etcd is not available.
+the current etcd version if etcd is available.
 
-    etcdctl ls /
+    curl -L http://127.0.0.1:4001/version
 
 [etcd]: https://coreos.com/etcd/docs/latest/
 [calico-releases]: https://github.com/projectcalico/calico-docker/releases/
-[experimental-docker-git]: https://github.com/docker/docker/tree/master/experimental
+[docker]: https://docs.docker.com/installation/