A starting point for automating cluster creation.
$ sudo pip install ansible (or setup your favorite virtualenv)
ensure you are running 2.4.1+ or later. Otherwise this fix is necessary: https://github.com/ansible/ansible/pull/30333
$ export AWS_ACCESS_KEY_ID=’AK123′
$ export AWS_SECRET_ACCESS_KEY=’abc123′$ cp ansible.cfg.default ansible.cfg && vim ansible.cfg
Set your pem pathBe sure to open up your aws dashboard and watch things work. If you see any other 'test-dev' instances in the ec2 dashboard you can terminiate them, but duplicates will not be made if you do not and all shoudl still be well.
$ ansible-playbook -i hosts ec2-cluster-create.ymlor with all debug output
$ ansible-playbook -i hosts ec2-cluster-create.yml -vvvOnce complete you should see something like this:
PLAY RECAP ****************************************************************************
172.18.100.103 : ok=3 changed=3 unreachable=0 failed=0
172.18.110.151 : ok=1 changed=1 unreachable=0 failed=0
172.18.110.173 : ok=1 changed=1 unreachable=0 failed=0
172.18.110.222 : ok=1 changed=1 unreachable=0 failed=0
172.18.70.40 : ok=3 changed=3 unreachable=0 failed=0
172.18.80.246 : ok=3 changed=3 unreachable=0 failed=0
172.18.90.96 : ok=3 changed=3 unreachable=0 failed=0
localhost : ok=19 changed=10 unreachable=0 failed=0
From best practices here
Ensure you are running 2.4.1+ or otherwise make this ansible core fix: ansible/ansible#30333
Managed by ec2.py
Read more here.
Configuration is kept at /ec2.ini.
Cmdline example to list all ec2 nodes.
$ ./ec2.py --listec2.py is set to return "test-dev" tagged instances only
Since there is no teardown yet be sure to delete you instances ( they are xl instances by default. Search for tag 'test-dev'