57 lines
2.7 KiB
Plaintext
57 lines
2.7 KiB
Plaintext
Hello everyone!
|
|
|
|
I would like to announce my candidacy for the Winstackers PTL election.
|
|
|
|
I have been involved in OpenStack since Havana, and even longer with Hyper-V
|
|
and Windows related technologies, and I have been the PTL of Winstackers for 2
|
|
cycles.
|
|
|
|
In the past cycle, we've added another 2 projects to the list of consumers of
|
|
os-win: os-brick and designate, and we hope to add many more.
|
|
|
|
We have also added the pieces necessary for various technologies, which are to
|
|
be consumed in the upcomming cycles, such as Fibre Channel support and various utils for Hyper-V Shielded VMs.
|
|
|
|
And last, but not least, we have added networking-hyperv under Winstackers
|
|
governance.
|
|
|
|
In the new cycle, there are a couple of things that I'd like us to achieve:
|
|
|
|
* Expand the list of os-win consumers. At the moment, the projects that are
|
|
under consideration are monasca and freezer. New suggestions are welcome.
|
|
|
|
* Improve os-win's performance and how it consumens PyMI and other libraries.
|
|
PyMI provides a Python native module wrapper over the Windows Management
|
|
Infrastructure (MI) API, and it is used by os-win in order to interact with
|
|
Hyper-V and other Microsoft related technologies. Each cycle we improve PyMI
|
|
and how it is used, increasing the overall performance. These performance
|
|
increases are observed by running Rally scenarios. For example, comparing
|
|
Newton to Liberty, the amount of time needed to bind a Neutron port and its
|
|
security group rules to a Hyper-V vSwtich port has been reduced by at least
|
|
12 times. And there is still room for improvement.
|
|
|
|
* Improve how os-win is consumed in nova. It order to make os-win easier to
|
|
use and evolve, we need to introduce some os-win objects (for example,
|
|
VirtualMachine objects). Having such objects will also help reduce the amount
|
|
of WMI queries done (expensive operations) and improve the overall
|
|
performance especially for spawning instances.
|
|
|
|
* oslo Windows CI. There were a few several cases in the past when an oslo
|
|
library release would break Windows / Hyper-V support in some or most
|
|
projects. Having a voting Windows CI on oslo patches can minimize the chances
|
|
of this happening. We are already working on this, and we hope to see it up
|
|
and running soon.
|
|
|
|
* Add support for new emerging Windows technologies. At the moment, we are
|
|
looking into Windows Containers and the new Windows Server 2016 networking
|
|
stack, which will bring support for VXLAN network types. These new
|
|
technologies will have to fit perfectly into the OpenStack ecosystem.
|
|
|
|
These are the main objectives I have in mind that I would like to accomplish
|
|
during the Ocata cycle, among other minor goals. New goals may appear during
|
|
the cycle.
|
|
|
|
Thank you for your consideration,
|
|
|
|
Claudiu
|