ansible-flink

An Ansible way to deploy Apache Flink.

I will be reverifying this repo and adding new features in this one soon enough

Edit the hosts_sample to add the hostnames or IP of your nodes Setup Flink using
$ ansible-playbook -i hosts_sample roles/configure.yaml

Run Flink using
$ ansible-playbook -i hosts_sample roles/run_flink.yaml

Stop Flink using
$ ansible-playbook -i hosts_sample roles/stop_flink.yaml

Download and configure Hadoop using $ ansible-playbook -i hosts_sample roles/configure-hadoop.yaml

You will need to modify the variables specified in the ansible-flink/group_vars/all/vars.yaml file. There are two variable defined there:

1. base_path: is the directory name where flink will be downloaded and setup
2. user_name: is the user name that you have on the remote machines.

If /home is mounted on NFS If you have a home directory /homemounted on a NFS. The root user might not be able to access that directory. Since by default ansible creates .ansible directory in/home/username, might run into some troubles for the commands that require sudo. In order to avoid this problem you will need to direct ansible to create the .ansbile directory in a place that is accessible by the root user, perhaps /tmp. Change the path for .ansible in ansible_sample.cfg and then $ cp ansible_sample.cfg ansible.cfg

TODO:

• Add the ability to provision servers on AWS before running flink playbook.
• Add support for other Cloud providers