Skip to content

Latest commit

 

History

History
125 lines (76 loc) · 3.29 KB

README.rst

File metadata and controls

125 lines (76 loc) · 3.29 KB
Raw

Overview

The ploy_ec2 plugin provides integration of Amazon EC2 with ploy.

Installation

ploy_ec2 is best installed with easy_install, pip or with zc.recipe.egg in a buildout.

Masters

To use ploy_ec2 you need an Amazon account and AWS keys.

Once you got your keys, you should put them in a secure location and reference them in your ploy.conf. Additionally you need to set the region of the master:

[ec2-master:ec2eu]
access-key-id = ~/.aws/ec2.id
secret-access-key = ~/.aws/ec2.key
region = eu-west-1

You can also set the AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables instead.

You need to define a master for each region you want to use.

Instances

Each instance has the following mandatory settings:

image
The Amazon Machine Image (AMI) that this instance will start up with.
keypair
The name of the SSH keypair to use.
placement
The availability zone in which to launch the instances.
securitygroups
The name of the Securitygroups this instance should be assigned to.

The following settings are optional:

instance_type

ip

startup_script
Path to a script which will be run right after creation and first start of the instance. This uses the User Data feature and needs to be supported by the AMI.

volumes

snapshots

device_map

delete-volumes-on-terminate

Securitygroups

description

connections
[ec2-securitygroup:app-server]
description = The production server
connections =
    tcp     22      22      0.0.0.0/0
    tcp     80      80      0.0.0.0/0

Volumes

You can define volumes via ec2-volume sections. The id of the section must not start with vol-. You can declare the size as a number of GB.

If the volume doesn't exist, it is automatically created.

[ec2-volume:a-volume-name]
size = 100

[ec2-instance:foo]
...
volumes = a-volume-name /dev/sdf

Macro expansion

For instances the ip and volumes options aren't copied when expanding macros.

Fingerprint verification

Automatic ssh fingerprint verification works by checking whether the fingerprint is in the console output of the instance.

After reboot or stop/start of an instance, the console output is refreshed. The problem with that is, that the fingerprint isn't included in the console anymore by default. To fix that you need to log the fingerprint on reboot somehow. One way to do that with Ubuntu is to add a script at /var/lib/cloud/scripts/per-boot/ssh-keys with this content:

#!/bin/sh
/usr/bin/ssh-keygen -l -f /etc/ssh/ssh_host_rsa_key.pub

Make sure that script is executable.