[Openstack-operators] Upgrade Strategy from Essex to Folsom+

Christian Parpart trapni at gmail.com
Mon Jan 7 16:45:14 UTC 2013


Hey,
this is great news for the first step, however, since every guide I found
so far is exclusevely talking about quantum, what implications would I
expect in upgrading w/o taking the whole cluster offline to the internet ?

Thanks,
Christian.

On Mon, Jan 7, 2013 at 5:43 PM, heut2008 <heut2008 at gmail.com> wrote:

> It isn't a must to use quantum and openvswitch,also change from
> nova-network to quantum is really a rough work.
> you can use  nova-network in folsom as well.  the folsom split
>  nova-volume to cinder, so cinder is a must  to use.
>
>
> 2013/1/8 Christian Parpart <trapni at gmail.com>
>
>> Hi all,
>>
>> I am having the following Essex Architecture, all on Ubuntu 12.04 LTS:
>>
>> - central nova-network service (2 HA'd nodes)
>> - central controller node containing scheduler, nova-volume, mysql,
>> rabbitmq,. ...
>> - 30 compute nodes.
>>
>> NOW, I'd like to upgrade from Essex to Folsom, and lucky me, I found a
>> rather
>> small intro on how to get the folsom packages by Ubuntu directly into
>> this LTS,
>> but that's it.
>>
>> I now know, that Cinder is nothing else than nova-volume split up into
>> its own project,
>> so I don't think i'll have issues upgrading this part, but I am really
>> very very concerned
>> about the networking node, Quantum is also a splitup of nova-network into
>> its own module
>> but it now contains open-vswitch as integral part, which is a cool thing,
>> I'm not against it,
>> but I wonder whether it is possible to to a live upgrade, that is, not
>> taking down any service
>> except the node that I am currently upgrading by the service (process)
>> starts I am to have.
>>
>> Is it also possible to upgrade compute nodes just like this? open-vswitch
>> is incompatible to
>>  native linux bridges - the bridge kernel module must not be loaded when
>> using
>> open-vswitch with kernel module acceleration enabled.
>>
>> Is it possible to partially upgrade? That is, upgrading one node after
>> the other,
>> and do the compute instances require a restart/reboot?
>>
>> Many questions, but all tightly bound to how to upgrade from Essex to
>> Folsom, and I'd be happy if
>> at least some of them could be answered. :)
>>
>> Many thanks in advance,
>> Christian.
>>
>>
>> _______________________________________________
>> OpenStack-operators mailing list
>> OpenStack-operators at lists.openstack.org
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
>>
>>
>
>
> --
> Yaguang Tang
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20130107/96065db7/attachment.html>


More information about the OpenStack-operators mailing list