sowravi / ansible-spark-1 Goto Github PK

1. To install cloudmesh client for the first run 'playbook-cloudmesh-first-time-install.yml' with following command:

ansible-playbook playbook-cloudmesh-first-time-install.yml --ask-sudo-pass -vvvv

2.1 After doing step 1, Open, ~/.cloudmesh/cloudmesh.yaml and edit the following section,the entry with < > should be customized as per your credentials:

profile:
     firstname: <first name>
     lastname: <last name>
     email: <email id>
     user: <chameleon/jetsream/other cloud username>

2.2. Change the entry of active cloud to use your favorite cloud,for example chamelon shown here:


    active:
      - chameleon
    clouds:
      ...

2.3.Under your favourite cloud(chameleon/jestream/..) change the following entry, the entry with < > should be customized as per your credentials:

credentials:
OS_PASSWORD: <enter your chameleon cloud password here>
OS_TENANT_NAME: CH-818664
OS_TENANT_ID: CH-818664
OS_PROJECT_NAME: CH-818664
OS_USERNAME: <username>


2.4. Also change, the type of OS and flavor you want for hadoop/spark cluster(recommended) under your favorite cloud:

default:
        flavor: m1.medium
        image: CC-Ubuntu14.04

3. Edit the file hadoop-spark-playbook.yml under the area: '-name:Preparing cloudmesh- setting default user', the entry with < > should be customized as per user credentials: 

     - name: Preparing cloudmesh- setting default user
       become_user: "{{ lookup('env','USER') }}"
       shell: cm default user=<chameleon/jetstream user>

4. Now run the following command to deploy hadoop/spark cluster which will run the hadoop-spark-playbook.yml. It will again upgrade cloudmesh client, so that we are up to date with any changes:

ansible-playbook  hadoop-spark-playbook.yml --ask-sudo-pass -vvvv