Getting started with Google Cloud Platform(GCP)

Google provides the same cloud services as other cloud providers such as Amazon Web Services(AWS) and Microsoft (Azure). It refers it as Google Cloud Platform or GCP. You can easily get started by signing up for free – https://cloud.google.com/free/

List of all products provided in GCP – https://cloud.google.com/products/

Google provides several ways to interact with its services-

1. GCP console (web ui)
GCP console is a web user interface which lets you interact with GCP resources. You can view, create, update and delete cloud resources from this page.

How to create a Linux vm(instance) using the console – https://cloud.google.com/compute/docs/quickstart-linux

2. Command Line Interface (gcloud cli toolset)
Install gcloud : https://cloud.google.com/sdk/gcloud/

The gcloud toolkit is a command line interface tool to interact with GCP resources. Very useful in automating cloud tasks, with its command completion and help pages, it is almost a necessity to familiarize yourself with this tool.

How to create an instance using gcloud cli – https://cloud.google.com/sdk/gcloud/reference/compute/instances/create

3. Cloud deployment manager
GCP deployment manager allows you to create, delete and update GCP resources in parallel by declaring a set of templates written in jinja2 or python. Templates can be shared with other teams and can be re-used with little modification.

What deployment manager is and how it works – https://cloud.google.com/deployment-manager/

How to deploy an a GCP instance using deployment manager – https://cloud.google.com/deployment-manager/docs/quickstart

4. APIs
Google provides application programming interface(APIs) to interact with its GCP services. Google recommends using the client libraries over directly calling the RESTful apis.

a. Client libraries

List of client libraries for different programming languages – https://cloud.google.com/apis/docs/cloud-client-libraries

How to interact with Google Compute Engine(GCE) using the Python client library – https://cloud.google.com/compute/docs/tutorials/python-guide#addinganinstance

b. RESTful or raw APIs

API Reference – https://cloud.google.com/compute/docs/reference/beta/

Method for creating an instance – https://cloud.google.com/compute/docs/reference/beta/instances/insert

Ansible – How to run a portion of a playbook using tags.

If you have a large playbook it may become useful to be able to run a specific part of it or only a single task without running the whole playbook. Both plays and tasks support a “tags:” attribute for this reason.

In this specific scenario, I have a playbook which configures all productions servers from the moment the servers boot till they start taking traffic. While testing the plays in dev environment, I was debugging an issue on the parts which does dns configuration. This is where the “tags” attributes comes handy –

1. Tag the task –

...
- name: Configure resolv.conf
  template: src=resolv.conf.j2 dest=/etc/resolv.conf
  when: ansible_hostname != "ns1"
  tags:
    - dnsconfig
...

2. Run only the tasks tagged with a specific name –

root@linubuvma:/etc/ansible# ansible-playbook -i dc1/hosts dc1/site.yml --tags "dnsconfig" --check

PLAY [Setup data center 1 servers] *****************************************************

TASK: [common | Configure resolv.conf] ****************************************
skipping: [ns1]
changed: [docker]
ok: [ns2]
ok: [whitehat]
ok: [mail]
ok: [www]
ok: [ftp]

PLAY RECAP ********************************************************************
whitehat                   : ok=1    changed=0    unreachable=0    failed=0
docker                     : ok=1    changed=1    unreachable=0    failed=0
ns1                        : ok=0    changed=0    unreachable=0    failed=0
ns2                        : ok=1    changed=0    unreachable=0    failed=0
mail                        : ok=1    changed=0    unreachable=0    failed=0
www                   : ok=1    changed=0    unreachable=0    failed=0
ftp                   : ok=1    changed=0    unreachable=0    failed=0

Ansible playbook will run only the task with the specified tag, it will skip the rest of the tasks in the playbook. Use the ‘–list-tags’ flag to view all the tags.

References –

http://docs.ansible.com/playbooks_tags.html

https://www.percona.com/live/mysql-conference-2015/sites/default/files/slides/Ansible.pdf

Ansible – Enable logging

By default, Ansible logs the output of playbooks to the standard output only. In order to enable logging to a file for later review or auditing, it can be turned on by setting log_path to a path location where Ansible has a write access.

In my case, i have added the “log_path” setting in the ansible configuration file “/etc/ansible/ansible.cfg”

# grep log_path /etc/ansible/ansible.cfg
log_path = /var/log/ansible.log

Now I can view the log file to all the details on ansible runs –

root@linubuvma:/etc/ansible# ansible-playbook tasks/groupby.yml --check
PLAY [all:!swarm:!docker1] ****************************************************

TASK: [group_by key=os_{{ ansible_os_family }}] *******************************
changed: [ns2]
.....

root@linubuvma:/etc/ansible# ls -al /var/log/ansible.log
-rw-r--r-- 1 root root 4255 May 16 21:21 /var/log/ansible.log
root@linubuvma:/etc/ansible# head  /var/log/ansible.log
2015-05-16 21:21:43,732 p=22946 u=root |
2015-05-16 21:21:43,732 p=22946 u=root |  /usr/local/bin/ansible-playbook tasks/groupby.yml --check
2015-05-16 21:21:43,732 p=22946 u=root |
2015-05-16 21:21:43,734 p=22946 u=root |  ERROR: the playbook: tasks/groupby.yml could not be found
2015-05-16 21:21:48,575 p=22954 u=root |
2015-05-16 21:21:48,576 p=22954 u=root |  /usr/local/bin/ansible-playbook tasks/groupby.yml --check
2015-05-16 21:21:48,576 p=22954 u=root |
2015-05-16 21:21:48,594 p=22954 u=root |  PLAY [all:!swarm:!docker1] ****************************************************
2015-05-16 21:21:48,609 p=22954 u=root |  TASK: [group_by key=os_{{ ansible_os_family }}] *******************************
2015-05-16 21:21:48,641 p=22954 u=root |  changed: [ns2]

It logs dry-runs (–check) as well and it is smart enough not to log Password arguments.

References –

http://docs.ansible.com/ansible/latest/intro_configuration.html#log-path