docker
/
docs
mirror of https://github.com/docker/docs.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

update roadmap 

Signed-off-by: Evan Hazlett <ejhazlett@gmail.com>
This commit is contained in:
Evan Hazlett 2015-03-29 01:02:23 -04:00
parent ee7617c2e7
commit 64de4e80b7
No known key found for this signature in database
GPG Key ID: A519480096146526
1 changed files with 8 additions and 40 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
ROADMAP.md
View File

@ -13,46 +13,14 @@ provisioning**.
For what is coming in specific releases, see our [upcoming For what is coming in specific releases, see our [upcoming
milestones](https://github.com/docker/machine/milestones).) milestones](https://github.com/docker/machine/milestones).)
## Client/server architecture ### Docker Engine / Swarm Configuration
Currently there are only a few things that can be configured in the Docker Engine and Swarm. This will enable more operations such as Engine labels and Swarm strategies.
Machine currently only works with a single user. Any machines you create are ### Boot2Docker Migration Support
stored on your local filesystem, meaning only you can control them. This means Currently both Machine and Boot2Docker provider similar functionality. This will enable users to migrate from boot2docker to machine.
you can't use it in teams or for large deployments.
Machines should instead have some kind of central storage. This could be done by ### Expand Provisioner
having a client/server architecture, where a central server keeps track of what Machine currently supports running Boot2Docker for "local" providers and Ubuntu for "remote" providers. This will expand the provisioning capabilities to include other base operating systems such as RedHat like distributions and possibly other "just enough" operating systems.
machines exist and the credentials used to access them. Clients could then
connect to this server to create and manage machines.
Once we have a server, a number of other things become possible: monitoring the ### Windows Experience
health of running hosts, access control, automatically scaling them, etc. Currently, the Machine on Windows experience is not as good as the Mac / Linux. There is no "recommended" path to use Machine and there are several inconsistencies on Windows such as logging and output formatting.
Here are some of the pieces of work being done to make this happen:
- [**Internal refactoring**](https://github.com/docker/machine/issues/553):
Some of Machine's internals need to be reorganised to make an external API
possible.
- **Add a REST API for internal API**: The first step towards a
server is adding a REST API for the stuff that Machine does already.
## Swarm integration
For production, Machine should be able to create and manage Swarm clusters.
This probably means there is a new high-level object in Machine called a Swarm,
which abstracted away the machines inside of it. You would be able to create and
configure a Swarm on a provider, then scale it up and down.
## Flexible provisioning
Currently, the operating systems and methods that Machine uses to provision
hosts are fixed. In production, you may want to be able to customise this
depending on your environment and requirements.
- **Customize the provisioning**: It should be possible to do custom
provisioning of your server. This will probably be done by [setting up
machines with cloudinit](https://github.com/docker/machine/issues/124).
- **Customize the operating system**: It should be possible to deploy operating
systems which aren't boot2docker or Ubuntu.
- **Customize the Docker Engine options**: It should be possible to pass
options to the Docker daemon inside the machine.