Open Stack

Hi there

We would like to write an update about the status of the InfraCloud ( if
you never heard of it, it's essentially a bunch of hardware donated by HPE
to the project to run an OpenStack cloud for CI testing. More info at
http://docs.openstack.org/infra/system-config/infra-cloud.html ).

We decided in the last infra mid-cycle at Fort Collins to split logically
the machines given (roughly 132 servers, with a mix of SL390s, SL230s and
SE1170s) in two different clouds: vanilla and chocolate, in order to have
each cloud running the current and previous OpenStack version.

The vanilla region, which is comprised of 48 SL390s nodes, has been
provisioned with Mitaka and this is how it looks like:

1 x Bifrost machine to provision the clouds control plane
1 x Controller
42 x Computes
4 x nodes currently being investigated due to faulty HW issues

This vanilla cloud has already been enabled in Nodepool:

https://review.openstack.org/#/c/363942/

So far, it's running quite well per the Logstash job logs collected from it.

Next step is to bump the max-servers to 50 servers (
https://review.openstack.org/#/c/364101/ ) to stress it a little bit more
to eventually unlock it to its full potential.
The SL390s report 24 cores, so this region can serve as quite a big
provider for CI runs.

We are looking forward to hear from you any feedback or problems you may
encounter.
We also plan to have a track on the infra mid-cycle in two weeks to plan
the rollout of the chocolate region and overall improvements.

Regards
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20160901/c13cd719/attachment.html>