GitHub user InAnimaTe opened a pull request:
https://github.com/apache/libcloud/pull/876
New Rancher Container Driver
## Rancher Container Driver
### Description
This is a new Container Driver for [Rancher](http://rancher.com/), the
fantastic container orchestration platform support the likes of Kubernetes,
Docker Swarm, and Mesos frameworks. It provides a fantastic abstraction layer
for working with these frameworks on any infrastructure you'd like! Plus it's
[open-source](https://github.com/rancher/rancher)!
Utilizing the v1 API and extensively working with the
[docs](http://docs.rancher.com/rancher/v1.2/en/api/), I've managed to get an
initial version of a fairly reliable ContainerDriver for Libcloud. It follows
the normal Container conventions as closely as possible and provides extensive
docstrings for excellent interaction, even without a massive understanding of
Rancher itself.
The three main top-level Rancher items are implemented:
* Containers
* Services (`ex_*`)
* Environments (`ex_*`)
Although not every possible action is available, the necessities like
creation and deletion plus things like activation, get, and list are included.
> As you review this, please ask questions if you aren't sure about how
something is implemented!
#### Background
With [Arroyo Networks'](https://arroyonetworks.com/) upcoming NaaS
platform, [Inflow](https://arroyonetworks.com/#services), I've been hard at
work establishing our Infrastructure assets. Rancher has seemed a clear choice
to help us maintain a nimble platform not tied to any specific back-end
Infrastructure provider. Hence, much of our Python code needs a way to interact
with our clusters, so writing a Libcloud plugin was pretty much a no-brainer.
#### Example
```
from libcloud.container.providers import get_driver
from libcloud.container.types import Provider
driver = get_driver(Provider.RANCHER)
connection = driver(key="ACCESS_KEY_HERE", secret="SECRET_KEY_HERE",
host="172.30.0.100", port=8080)
image = ContainerImage("hastebin", "hastebin", "rlister/hastebin",
"latest", driver=None)
newcontainer = connection.deploy_container("myawesomepastebin", image,
environment={"STORAGE_TYPE": "file"})
```
### Status
- done, ready for review
I have tested this works appropriately with Rancher Server v1.1.3 (API v1).
Of course, I want persons looking at this to pay attention to the whole PR.
However, there are a few key areas I want some focus on:
* Unnecessary sections (do I really need `parse_body`?)
* Inefficient handling (is there a better way to do Image parsing or
TLS/SSL support?)
* Error Handling looks/works sanely.
* Convention adherence (I'm a newer programmer; I may not know well
established `*ism's`)
I encourage reviewers to pull my branch and play with this ;)
#### Notes
* Does not support user based API credentials, only Environment API
credentials (one env, no cluster support)
* Does not support images other than docker formatted images. `docker:`
prefix is forced!
* Images follow a standardized format. See `deploy_container` docstring!
* Does not support filters for searching/listing. (no `image` or `all`
support in `list_containers`)
* The `launchConfig` for `ex_deploy_service` is actually all defined at the
top level. This is because I easily pass it through `_build_payload` (**kwargs)
since the options are nearly identical to those of a normal container. No
reason to duplicate work. This also makes it easier since most people only
generally want to deploy a service with only one container anyways. This may
need to change later when we want >1 container for a service.
* Passing your own cert/key for SSL/TLS is not presently supported.
* For SSL/TLS (https) support with newer versions of OpenSSL, the following
is necessary (in your own code):
import ssl
import libcloud.security
libcloud.security.SSL_VERSION = ssl.PROTOCOL_TLSv1_2
### Checklist (tick everything that applies)
- [x] [Code
linting](http://libcloud.readthedocs.org/en/latest/development.html#code-style-guide)
(required, can be done after the PR checks)
- [x] Documentation - See Docstrings
- [ ] [Tests](http://libcloud.readthedocs.org/en/latest/testing.html) -
*Not yet...*
- [ ]
[ICLA](http://libcloud.readthedocs.org/en/latest/development.html#contributing-bigger-changes)
(required for bigger changes) - *Should be done in the next day/two.*
### Thanks
I have worked closely with [Anthony Shaw](https://github.com/tonybaloney),
a core contributor to Libcloud (especially the Container stuffs) as well as
[Vincent Fiduccia](https://github.com/vincent99), software architect at Rancher
Labs. The have been invaluable in helping me establish a solid driver for
Rancher.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/ArroyoNetworks/libcloud rancher_driver
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/libcloud/pull/876.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #876
----
commit 3143283f5805b2339489fee05569d2563ca48d39
Author: Mario Loria <ma...@arroyonetworks.com>
Date: 2016-09-28T03:12:46Z
initial push
commit 675cb6b8f77ec466e107eb06c72bbf7e50f70993
Author: Mario Loria <ma...@arroyonetworks.com>
Date: 2016-09-29T01:57:39Z
docstrings, flake8, error handling
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
