Il 20/03/2015 12:22, Karanbir Singh ha scritto:
> On 03/20/2015 10:42 AM, Sandro Bonazzola wrote:
>>>>>>> For opening the discussion I suggest:
>>>>>>>
>>>>>>> - virt${release}-common : packages not related to xen or kvm, used by at least 2 projects
>>>>>>> - virt${release}-xen : xen hypervisor related packages
>>>>>>> - virt${release}-kvm : kvm hypervisor related packages (like qemu-kvm-ev)
>>>>>>> - virt${release}-ovirt : ovirt packages and dependencies, not required by other projects
>>>>>>> - virt${release}-docker : docker packages and dependencies, not required by other projects
>>>>>>
>>> note: the proposal here is going to result in ~ 24 different branches
>>> for each target ( into git.centos.org )
>>>
>>
>> We can reduce it a bit, just keeping virt7-ovirt instead of virt${release}-ovirt; no plan to get it on el6 right now.
>
> Can you detail the ovirt plan for the Virt SIG - so far, I dont think
> we've seen any ovirt content at all. Apologies if I've missed a
> conversation in the past around this, feel free to point me at that
> instead if its easier.
Packages built and in queue:
http://wiki.centos.org/SpecialInterestGroup/Virtualization/Roadmap
How to test:
http://wiki.centos.org/HowTos/oVirt
So far, with the packages in the CentOS Virt SIG you can provision the hypervisors / gluster hosts with oVirt 3.5.1.1 packages for VDSM and latest
qemu-kvm-ev.
The manager is not yet packaged for CentOS, missing several required dependencies that in ovirt repositories are provided by binary wrapper rpm packages.
>
>> We can also drop common if we decide to provide the same rpm in 2 or more different projects repo.
>> I'm not sure it's wise to do the same for the kvm target.
>
> The key thing here to keep in mind is that we want the auth and acl
> process to be simple for the entire pipeline. This starts from Git all
> the way to having released content on the mirrors. Take a look at this
> as an example of what the entire pipeline looks like :
> ref:
> http://lists.centos.org/pipermail/centos-devel/attachments/20150316/e6b4b15d/attachment.png
>
>
> Given that SIG subslipts are going to be an issue ( and are for Storage
> and Cloud infra SIG, in addition to the Virt SIG already ), we might
> need to rethink that a bit, hence my question on how this mapping will work.
>
--
Sandro Bonazzola
Better technology. Faster innovation. Powered by community collaboration.
See how it works at redhat.com
