[Openstack] [Nova] Proposed removal of the PowerVM driver

Matt Riedemann mriedem at linux.vnet.ibm.com
Mon Apr 7 21:12:33 UTC 2014



On 3/31/2014 12:48 PM, Florian Chazal wrote:
> Hi Matt,
>
> Regarding your answer the driver seems to handle VLAN mode but I still
> don't see how it can without a specific neutron plugin. From what I see
> from the code there is no any call for IVM "mkdev -vlan" and "-tagid"
> parameters.
>
> Should I extend it myself or am I missing something ?
>
> Regards,
>
> Florian CHAZAL
>
>
>
>
> On 21 March 2014 18:52, Matt Riedemann <mriedem at linux.vnet.ibm.com
> <mailto:mriedem at linux.vnet.ibm.com>> wrote:
>
>
>
>     On 3/21/2014 6:45 PM, Florian Chazal wrote:
>
>         Hi All,
>
>         Better late that never:
>
>         Since few weeks we are testing IBM Power System R2 server and we
>         were
>         evaluating the capability to go on a larger scale.  We are
>         lacking of
>         information regarding PowerKVM status and PowerVC so we began to try
>         Havana + powerVM:
>         The launch of a VM went well (except an issue with glance which
>         does not
>         gzip the image ... troubleshoot in process) but how does it works
>         regarding the network ? What type of network configuration it
>         has been
>         tested with ?
>
>         Thank you in advance,
>
>         Florian CHAZAL
>
>
>         On 25 November 2013 08:26, Matt Riedemann
>         <mriedem at linux.vnet.ibm.com <mailto:mriedem at linux.vnet.ibm.com>
>         <mailto:mriedem at linux.vnet.__ibm.com
>         <mailto:mriedem at linux.vnet.ibm.com>>> wrote:
>
>
>
>              On Monday, November 25, 2013 10:14:00 AM, Anne Gentle wrote:
>
>
>
>
>                  On Mon, Nov 25, 2013 at 5:10 AM, Thierry Carrez
>                  <thierry at openstack.org <mailto:thierry at openstack.org>
>         <mailto:thierry at openstack.org <mailto:thierry at openstack.org>>
>                  <mailto:thierry at openstack.org
>         <mailto:thierry at openstack.org> <mailto:thierry at openstack.org
>         <mailto:thierry at openstack.org>>__>__>
>
>                  wrote:
>
>                       Chuck Short wrote:
>                       > On Fri, Nov 22, 2013 at 4:45 PM, Michael Still
>                       <mikal at stillhq.com <mailto:mikal at stillhq.com>
>         <mailto:mikal at stillhq.com <mailto:mikal at stillhq.com>>
>                  <mailto:mikal at stillhq.com <mailto:mikal at stillhq.com>
>         <mailto:mikal at stillhq.com <mailto:mikal at stillhq.com>>>
>                       > <mailto:mikal at stillhq.com
>         <mailto:mikal at stillhq.com> <mailto:mikal at stillhq.com
>         <mailto:mikal at stillhq.com>>
>                  <mailto:mikal at stillhq.com <mailto:mikal at stillhq.com>
>         <mailto:mikal at stillhq.com <mailto:mikal at stillhq.com>>>>> wrote:
>                       >
>                       >>     So, a few questions...
>                       >>
>                       >>      - If there are no users shall we remove it
>         from
>                  Havana and
>                       Grizzly as
>                       >>     well, or does that violate our stability
>         principles
>                  too much?
>                       >
>                       >
>                       > Speaking with my stable maintainers hat on I
>         would personally
>                       NACK this
>                       > since
>                       > it doesnt meet the standard of a stable maintenance
>                  requirement,
>                       imho.
>
>                       It definitely doesn't meet our stable maintenance
>         rules. We
>                  want
>                       people
>                       to be able to safely upgrade to the latest
>         stable/* when
>                  they are
>                       users
>                       of the release. We don't add new features,
>         removing them
>                  would be even
>                       worse.
>
>                       >>      - If we don't remove the code from stable,
>         what about
>                       removing all
>                       >>     references from the stable docs and putting
>         in a warning
>                       saying that
>                       >>     powervm is a dead end instead? I want to
>         minimise
>                  confusion
>                       on the
>                       >>     part of people deploying stable releases.
>                       >
>                       > This would be okay in my opinion.
>
>                       At the very minimum I would add the "dead-end"
>         comment to
>                  the Havana
>                       release notes. If Anne is fine by it, it could also be
>                  added to the
>                       stable/havana version of the docs.
>
>
>                  PowerVM was mostly documented in developer docs in the
>         nova repo. It
>                  also has entries in the Configuration Reference in the
>         nova.conf
>                  options tables from our autodoc process. Those will
>         just stay in the
>                  stable/havana repo, but I'm okay with backporting a note to
>                  stable/havana with a clear message as to the driver's
>         fate. We also
>                  have a mention of PowerVM in stable/grizzly where we
>         could also
>                  place
>                  a note. Release notes are a good place for this as
>         well, thanks
>                  Russell.
>
>                  Tracking with
>         https://bugs.launchpad.net/____openstack-manuals/+bug/1254780
>         <https://bugs.launchpad.net/__openstack-manuals/+bug/1254780>
>
>
>         <https://bugs.launchpad.net/__openstack-manuals/+bug/1254780
>         <https://bugs.launchpad.net/openstack-manuals/+bug/1254780>__> --
>                  feel free
>                  to add the nova project as well to ensure the docs in
>         the nova repo
>                  are removed.
>
>                  I do want to point out that we don't have adequate
>         documentation of
>                  other drivers, like Hyper-V and Xen, just want to make
>         sure you all
>                  know there are gaps and documentation isn't the
>         indicator of
>                  "acceptance" of a hypervisor driver.
>                  Thanks -
>                  Anne
>
>                       --
>                       Thierry Carrez (ttx)
>
>                       ___________________________________________________
>                       Mailing list:
>         http://lists.openstack.org/____cgi-bin/mailman/listinfo/____openstack
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack>
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack
>         <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>>
>
>                       Post to     : openstack at lists.openstack.org
>         <mailto:openstack at lists.openstack.org>
>                  <mailto:openstack at lists.__openstack.org
>         <mailto:openstack at lists.openstack.org>>
>                       <mailto:openstack at lists.
>         <mailto:openstack at lists.>__open__stack.org <http://openstack.org>
>                  <mailto:openstack at lists.__openstack.org
>         <mailto:openstack at lists.openstack.org>>>
>                       Unsubscribe :
>         http://lists.openstack.org/____cgi-bin/mailman/listinfo/____openstack
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack>
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack
>         <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>>
>
>
>
>
>
>                  ___________________________________________________
>                  Mailing list:
>         http://lists.openstack.org/____cgi-bin/mailman/listinfo/____openstack
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack>
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack
>         <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>>
>
>                  Post to     : openstack at lists.openstack.org
>         <mailto:openstack at lists.openstack.org>
>                  <mailto:openstack at lists.__openstack.org
>         <mailto:openstack at lists.openstack.org>>
>                  Unsubscribe :
>         http://lists.openstack.org/____cgi-bin/mailman/listinfo/____openstack
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack>
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack
>         <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>>
>
>
>
>              Also to relay what was discussed in nova IRC this morning, this
>              topic is on the release meeting agenda for tomorrow:
>
>         https://wiki.openstack.org/____wiki/Meetings/ProjectMeeting#____Agenda_for_next_meeting
>         <https://wiki.openstack.org/__wiki/Meetings/ProjectMeeting#__Agenda_for_next_meeting>
>
>
>         <https://wiki.openstack.org/__wiki/Meetings/ProjectMeeting#__Agenda_for_next_meeting
>         <https://wiki.openstack.org/wiki/Meetings/ProjectMeeting#Agenda_for_next_meeting>>
>
>              --
>
>              Thanks,
>
>              Matt Riedemann
>
>
>
>              ___________________________________________________
>              Mailing list:
>         http://lists.openstack.org/____cgi-bin/mailman/listinfo/____openstack
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack>
>
>
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack
>         <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>>
>              Post to     : openstack at lists.openstack.org
>         <mailto:openstack at lists.openstack.org>
>              <mailto:openstack at lists.__openstack.org
>         <mailto:openstack at lists.openstack.org>>
>              Unsubscribe :
>         http://lists.openstack.org/____cgi-bin/mailman/listinfo/____openstack
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack>
>
>
>         <http://lists.openstack.org/__cgi-bin/mailman/listinfo/__openstack
>         <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>>
>
>
>
>
>         --
>         Florian Chazal
>
>
>     Hi Florian,
>
>     Some details on what testing we did internally with running Tempest:
>
>     - Controller/Compute Node: RHEL 6.3-6.5, Fedora 19 ppc64
>     - Hypervisor: PowerVM IVM 2.2.2.0+
>     - Platform: P7 and P7+ Racks, Blades, ITEs
>     - Networking:
>              - Neutron + Open vSwitch: Flat, VLAN - IPv4 only
>              - nova-network: Flat
>     - Cinder Storage Driver: Storwize/SVC 6.4.1/6.4.2 for FC
>
>     There are some teams here working on getting a PowerVC nova virt
>     driver into Stackforge.
>
>     --
>
>     Thanks,
>
>     Matt Riedemann
>
>
>
>
> --
> Florian Chazal

Hi Florian,

Looks like you're right, VLAN isn't supported in the PowerVM driver, 
only flat networking.  I was looking at some incorrect support charts 
which must have meant PowerVC instead of PowerVM.

By the way, I'm not sure if you've seen this but if you're looking for 
config drive support there is a patch here [1] that was being worked for 
Icehouse but was abandoned due to the driver being removed from nova 
upstream.  It went through decent review and was run internally for a 
couple of releases so it should be relatively stable.

[1] https://review.openstack.org/#/c/38716/

-- 

Thanks,

Matt Riedemann





More information about the Openstack mailing list