Clustering CoreOS Docker Hosts Using Fleet

Once the CoreOS Docker Hosts are Clustered we will be able to manage the Docker Hosts from a single server.

Getting the new Discovery URL.

curl -w "\n" https://discovery.etcd.io/new

We will get somthing like

https://discovery.etcd.io/16043bf6be5ecf5c42a0bcc0d9237954

make sure we use the new Discovery URL in the Config-core.yaml

Configure the new Config File with the URL
>>cat config-core.yaml
===============================================
#cloud-config
coreos:
  etcd:
    # generate a new token for each unique cluster from https://discovery.etcd.io/new
    #discovery: https://discovery.etcd.io/<token>
    discovery: https://discovery.etcd.io/a41bcad1e117d272d47eb938e060e6c8
    # multi-region and multi-cloud deployments need to use $public_ipv4
    addr: $private_ipv4:4001
    peer-addr: $private_ipv4:7001
  units:
    - name: etcd.service
      command: start
    - name: fleet.service
      command: start
write_files:
  - path: /etc/resolv.conf
    permissions: 0644
    owner: root
    content: |
      nameserver 8.8.8.8
ssh_authorized_keys:
  # include one or more SSH public keys
  - ssh-rsa AA7dasdlakdjlksajdlkjasa654d6s5a4d6sa5d465df46sdg4rdfghfdhdg74wefg4sd32f13f468e Generated-by-Nova
===============================================
If we are using Openstack make sure that the neutron metadata service is running fine and working properly because the private_ipv4 attribute works only if the instance get the meta data properly.

Use the Above Config File to start the CoreOs VM's. and Once the CoreOs VM are running Then we can use the Following Commands to check the fleet status.

List the machines in the Fleet
>> fleetctl list-machines
MACHINE         IP              METADATA
0a1cad1d...     192.36.0.65      -
220f3e64...     192.36.0.67      -
31dbd5ca...     192.36.0.66      -

If we need to add a new server into the same fleet we can get the Discovery URL by

grep DISCOVERY /run/systemd/system/etcd.service.d/20-cloudinit.conf

Using the above command we can get the Discovery URL from a running machine in the Fleet.

Once we get the URL start the new instance with the same config-core.yaml.