I will be moving to northern Idaho in less than a month and will need somewhere else to host homelab during the time it takes to do so. I already have an Amazon account, so I'll most likely spin up their smallest ec2 instance and upload it to that. I'm still going to need a Public IP to point cloudflare at and a few other things I'm sure. I'll go over all of that here. You can download the source code from github to follow along.
AWS
Defaults will normally work for this
- A default VPC
- Subnet within the VPC
- An ssh key pair generated to access the ec2 instance.
- An Elastic IP address linked to the instance to point DNS at.
- t2.micro ubuntu ec2 instance
- A security group
- Enable 22 while we provision (disable after finished)
- Enable port 80 for http traffic
EC2 Creation
I started up a t2.micro ubuntu 16.04 instance
- 1 Core vCPU (up to 3.3 GHz)
- 1 GiB Memory RAM
- 8 GB Storage
- FREE TIER ELIGIBLE
Download the pem file at the end of creation to access the instance.
chown it to restrict permission to the file and allow it to be added ssh-agent
chmod 600 homelab.pem
ssh-add homelab.pem
Next create an Elastic IP and link it to the instance from the 'Actions' drop down. With this newly acquired public IP we can now get to business with provisioning our new machine and deploying ghost to the cloud.
First lets test that we can login.
Be sure and use the default user [email protected]<ip_address> for a newly created ubuntu instance
ssh [email protected]
Welcome to Ubuntu 16.04.3 LTS (GNU/Linux 4.4.0-1047-aws x86_64)
...
[email protected]:~$
Ansible provisioning
Next I'll create an aws_inv.ini file for ansible to access this machine with a few key variables
- hostname
- username
- ip address
- network interface for swarm configs
- swarm availability mode
aws_inv.ini
[manager]
homelab
[manager:vars]
ansible_host=52.25.231.162
ansible_user=ubuntu
swarm_iface=eth0
swarm_availability=active
If we test ansible now, we'll see this error output.
ansible -i aws_inv.ini all -m ping
homelab | FAILED! => {
"changed": false,
"module_stderr": "Connection to 52.25.231.162 closed.\r\n",
"module_stdout": "/bin/sh: 1: /usr/bin/python: not found\r\n",
"msg": "MODULE FAILURE",
"rc": 0
}
Easy enough to fix this by installing python on the system first.
ssh [email protected]
sudo apt-get update && sudo apt-get install python
Now with python installed we should get a pong back.
ansible -i aws_inv.ini all -m ping
homelab | SUCCESS => {
"changed": false,
"ping": "pong"
}
Install the base
With that, we are ready to kick off a few scripts to get everything installed and running. First I need to update my site.yml file and take out the custom setup I have in there that I only needed for my laptop, like ignoring lid closure, etc...
Starter site.yml file to build on and add any dependencies needed.
site.yml
---
- hosts: all
become: true
become_method: sudo
tasks:
- name: apt-get update && install deps
apt:
update_cache: yes
name: "{{ item }}"
state: present
with_items:
- vim
- wget
- rsync
- make
tags:
- setup
- deps
Run this against the instance
ansible-playbook -i aws_inv.ini site.yml
PLAY [all] ********************************************************************************************************************
TASK [Gathering Facts] ********************************************************************************************************
ok: [homelab]
TASK [apt-get update && install deps] *****************************************************************************************
changed: [homelab] => (item=[u'vim', u'wget', u'rsync', u'make'])
PLAY RECAP ********************************************************************************************************************
homelab : ok=2 changed=1 unreachable=0 failed=0
Install docker and enable swarm mode
Next I will install docker and start a one node swarm cluster for hosting the ghost blog using an ansible playbook
ansible-playbook -i aws_inv.ini swarm.yml
After that finishes running I can log into the instance and see that docker is in-fact running and in swarm mode.
sudo docker node ls
ID HOSTNAME STATUS AVAILABILITY MANAGER STATUS
i1hjapxrz3tfkddva4nztnqhs * ip-172-31-0-177 Ready Active Leader
homelab backup
From my homelab I will tar up the docker volume in preparation to move it to AWS
tar -czvf content.tar.gz /data/ghost/content
mv content.tar.gz /home/jahrik/
Pull it to my local workstation
scp 192.168.2.106:/home/jahrik/content.tar.gz .
content.tar.gz 100% 19MB 7.3MB/s 00:02
SCP to AWS host
Move it up to the cloud
scp content.tar.gz [email protected]:/home/ubuntu/
content.tar.gz 100% 19MB 639.9KB/s 00:31
Log back in and Inflate it
sudo tar -xzvf content.tar.gz
Create the right directories to place it in.
sudo mkdir -p /data/ghost/content
And finally, move it into place
sudo cp -R data/ghost/content/* /data/ghost/content/
Start the docker service with a stack file
Clone the home_lab repository off github and start up the docker service.
git clone https://github.com/jahrik/home_lab.git
cd home_lab
# edit stack file and change:
# - '2368:2368'
# to
# - '80:2368'
vim ghost/ghost-stack.yml
Start it up
sudo docker stack deploy -c ghost/ghost-stack.yml ghost
Creating network ghost_default
Creating service ghost_blog
Tail the logs to it start up
sudo docker service logs -f ghost_blog
[email protected] | [2018-01-17 08:02:06] INFO Finished database migration!
[email protected] | [2018-01-17 08:02:08] INFO Ghost is running in production...
[email protected] | [2018-01-17 08:02:08] INFO Your blog is now available on http://homelab.business/
[email protected] | [2018-01-17 08:02:08] INFO Ctrl+C to shut down
[email protected] | [2018-01-17 08:02:08] INFO Ghost boot 2.106s
Browse to the public IP 52.25.231.162 and I can see it running.
Redirect Cloudflare to the new public IP.
