Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/cgwalters/min-cloud-agent

Minimal client implementation of AWS/OpenStack metadata API
https://github.com/cgwalters/min-cloud-agent

Last synced: 26 days ago
JSON representation

Minimal client implementation of AWS/OpenStack metadata API

Awesome Lists containing this project

README 

        
Currently not actively developed

--------------------------------



Note: this repository is not under active development; the author

is working instead on https://cloudinit.readthedocs.org/en/latest/



min-cloud-agent

===============



In contrast to cloud-init, the vision of this project is to be an

ultra-small bootstrapping stage that one can use to start other

management tools.



At present, it implements a very small subset of the AWS EC2 metadata

API (which is also provided by the OpenStack hypervisor).



The list of requests implemented is:



 * `/2009-04-04/user-data`:



    Execute arbitrary script (must start with #!, e.g. #!/bin/bash).  The

    best practice is for this script to be idempotent, but min-cloud-agent

    will make a best effort to ensure this script runs exactly once when

    the system is first booted.



 * `/2009-04-04/meta-data/public-keys/0/openssh-key`:



    Provision a single ssh key to the `root` account's `.ssh/authorized_keys` file.

    If the file already exists, it will not be modified.



In contrast to cloud-init, the user-data must be an interpreted

program; it can not be the YAML cloud-config, have includes, or

anything like that.



For example, suppose you want to add a user to the system.  Rather

than using cloud init's "users" attribute in the cloud-config YAML,

you should simply invoke the operating system native tool,

e.g. `/usr/sbin/useradd` directly in a shell script.



It's strongly recommended that all operating system vendors shipping

min-cloud-agent include the `curl` utility.  This will allow user-data

script authors to download further code from a URL, or perform a POST

to registration/state server announcing that the system is online.



Another very useful thing to do in user-data is to call into the

system software management (e.g. yum/apt-get/rpm-ostree) to perform

upgrades or install further applications (such as Docker) or higher

level configuration management tools (such as Puppet).



min-cloud-agent will never grow an abstraction layer for any of these

things; if you are scripting multiple operating systems or versions,

your userdata scripts must carry the abstractions.



Dependencies

------------



 * GLib: https://git.gnome.org/browse/glib/

 * libgsystem: https://git.gnome.org/browse/libgsystem/



Planned features

----------------



 * user-data

 * configurable account for ssh keys

 * hostname



Why create this instead of just using `cloud-init`?

---------------------------------------------------



cloud-init is a very featureful program - it has many data providers

and backends.  This leads to a fuzzy line between what it does and

what other operating system management and configuration tools like

Puppet should do.  It means cloud-init is likely to keep growing.



Beyond simply being a large program in itself, cloud-init is written

Python, and this agent is targeted at small host operating systems

that don't necessarily include Python by default.



Contributing

------------



Use Github pull requests.  You can discuss this program on the Fedora

cloud list: https://admin.fedoraproject.org/mailman/listinfo/cloud



License

-------



min-cloud-agent is licensed under the LGPLv2+.  See the COPYING

file for more information.