Open Stack

On 2 May 2014 09:06, Aaron Rosen <aaronorosen at gmail.com> wrote:

> In all seriousness though, this was discussed in lengths on openstack-dev.
> I do feel your pain as an operator if you find that support for a feature
> you were using was removed. I think this is the only way to go though in
> order to continue raising the openstack bar.
>

When looking at which of the hypervisors to implement, one of the criteria
for me is certainly to consider how well tested it is. If it's not well
tested, then if you implement it into production you run the risk of
hitting odd bugs and therefore take on some responsibility to be more
heavily involved in the development process for that driver.

Essentially this means that if your drivers in not in Group A [1], then be
prepared to accept that you'll have to take on more responsibility in the
code base. A strategy you could adopt would be to get the driver from Group
B/C to group A by committing resources into the development of the tests
required to do so.

Doing so would make it much more likely for the driver to get into the main
codebase and also ensure that the distro's package the code along with the
other drivers.

In short, if you want the LXC/Docker driver to be a higher grade driver,
then commit some resources to help with the development and testing
maturity. If you want to ensure that it is never removed from the codebase,
do the same. If this is not an option for you, then don't implement the
driver in production.

[1] https://wiki.openstack.org/wiki/HypervisorSupportMatrix
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20140502/6e84f7aa/attachment.html>