[Openstack] Metadata and File Injection (code summit session?)

Steven Dake sdake at redhat.com
Tue Apr 10 23:28:39 UTC 2012


On 04/10/2012 03:04 PM, Justin Santa Barbara wrote:
>     Having the ability to read config data from a runtime changeable
>     metadata server (rather then a config file on an injected disk) serves a
>     use case I am interested in.  The only problem is horizontal scalability
>     of the metadata server in this model which may not be a problem with
>     some tinkering.
> 
> 
> Can you please share that use case?  I'm especially interested in
> finding use cases that would not better be better served by e.g. SSH or
> Zookeeper or Corosync..
> 

Sure

The exact use case is to support updating configurations of running
systems without installing a specific SSH key or requiring SSH at all.

Here is how it would work:

developer wants to tell the vm about 3 pieces of updated information

1) developer uses metadata set api to set key/value pairs for 3 pieces
of information

2) VM can then read these metadata blobs at its leisure

There are a few alternatives for runtime configurable data that are simple:
1) use ssh - this is a good model, but has a few constraints.  First a
key has to be loaded into the VM that the orchestration system knows
about.  Second, ssh has to be used over the network, which creates some
problems for some security people for pragmatic reasons (how do we get
SSH over VPNs), and less rationale reasoning (ie: ssh is evil, etc).

or

2) each service ends up implementing its own dynamic metadata server

You mentioned zookeeper or corosync.  those are big deps to pull in
simply to get runtime configurable metadata.  If they have other uses,
that may make sense but simply extending the metadata server seems
fairly  simple.

Option 2 would work but why have 10 of something when a common component
would do the job.

Cheers
-steve




More information about the Openstack mailing list