Re: [Openstack] Questions on nova code: networking live/offline migration

2011-09-12 Thread Alex Lyakas
Hello Kei, thank you for your attention. According to the code, this should reproduce every time, because of two reasons: 1) There is no code that ensures that the same backing file is present at the destination within the /var/lib/nova/instances/_base folder. At least I don't see such code,

Re: [Openstack] OpenStack Summit Session Ideas

2011-09-12 Thread Thierry Carrez
Brian Lamar wrote: I've been thinking about submitting a brainstorm session (or two) for the conference where we can go over some 'big picture' items I've been wrestling with lately. [...] Sounds like a great idea ! I'd say you need one session for #1 + #2 (Nova standard interface), one

[Openstack] Last day for registering on the waiting list for the Design Summit !

2011-09-12 Thread Thierry Carrez
Hi everyone, The waiting list for the Essex Design Summit is growing out of control so we'll close registrations tomorrow Tuesday, noon GMT ! If you missed the boat and still want to be considered for the Design Summit, you need to register on the waiting list before it's too late ! See

Re: [Openstack] Integration tests

2011-09-12 Thread Jay Pipes
Wow, another integration test framework emerges out of the woodwork :) Looking forward to getting all of these into a single place... and clearing up the lack of communication between teams on this subject! There's Kong, Stacktester, and Backfire. Any others out there we should know about?

Re: [Openstack] OpenStack Summit Session Ideas

2011-09-12 Thread Jay Pipes
On Fri, Sep 9, 2011 at 1:45 PM, Brian Lamar brian.la...@rackspace.com wrote: Topic #1 - Separating API projects from OpenStack Nova. I would push for separate projects for the EC2-compatible API, the Rackspace-compatible API, and the OpenStack Official API. Nova is not the place to stick

[Openstack] swift 1.4.3 (diablo) released

2011-09-12 Thread John Dickinson
We have released swift 1.4.3 today. The changelog is below. Baring something we missed in testing, this release is the swift version that will be released as Openstack Diablo. The launchpad landing page for this release is https://launchpad.net/swift/diablo/1.4.3. swift (1.4.3) *

[Openstack] Diablo 4 dashboard error missing openstackx extensions in nova.

2011-09-12 Thread shake chen
Hi I run Diablo 3 and dashboard, all the component is working as expect. (ubuntu 11.04) but when I update to Diablo 4, when I login the dashboard and check the instance, would report the error. Unable to get usage info: This error may be caused by missing openstackx extensions in nova. See the

Re: [Openstack] Proposal: URIs for X-Auth-Header Keystone tokens

2011-09-12 Thread Joe Savak
Mark Bryan, I haven't forgotten about this and included it on the keystone wiki so we won't lose track of it during Essex planning. http://wiki.openstack.org/keystone Thanks, Joe -Original Message- From: openstack-bounces+joe.savak=rackspace@lists.launchpad.net

Re: [Openstack] Diablo 4 dashboard error missing openstackx extensions in nova.

2011-09-12 Thread Anthony Young
Make sure that the openstackx/extensions contents are accessible via the flag FLAGS.osapi_extensions_path. nova-api uses that flag to dynamically load external extensions. Anthony On Mon, Sep 12, 2011 at 7:52 AM, shake chen shake.c...@gmail.com wrote: Hi I run Diablo 3 and dashboard, all

Re: [Openstack] Integration tests

2011-09-12 Thread Tim Simpson
I'm with the Reddwarf (Database as a Service) team at Rackspace. We employ a large set of integration / functional tests which run in a Vagrant controlled VM. Our tests use python-novaclient as well as a similar client for Reddwarf we've written ourselves. The motivation is to eat our own dog

[Openstack] Keystone/nova-api error

2011-09-12 Thread Marc Peiser
Hi all, I'm having issues with keystone install with nova/glance. I seem to be missing something as I'm getting this error from nova-api logs: 2011-09-12 19:15:16,558 DEBUG nova.api [-] ec2_listen : 0.0.0.0 from (pid=9026) module /usr/bin/nova-api:58 2011-09-12 19:15:16,558 DEBUG nova.api [-]

Re: [Openstack] Integration tests

2011-09-12 Thread Matt Dietz
Regarding novaclient, I was debating this myself last night. Our backfire suite currently uses it. However, I can see scenarios where nova has advanced but novaclient has yet to, and because all three projects are independently developed, we're stuck temporarily. I can see utility in tests

[Openstack] Screencasts ...

2011-09-12 Thread Sandy Walsh
Here's a wiki page of some screencasts I put together of recent Nova changes I've been working on. Please feel free to add your own. http://wiki.openstack.org/DemoVideos -S This email may include confidential information. If you received it in error, please delete it.

Re: [Openstack] Screencasts ...

2011-09-12 Thread Trey Morris
very nice! On Mon, Sep 12, 2011 at 12:55 PM, Sandy Walsh sandy.wa...@rackspace.comwrote: Here's a wiki page of some screencasts I put together of recent Nova changes I've been working on. Please feel free to add your own. http://wiki.openstack.org/DemoVideos -S This email may include

Re: [Openstack] Keystone/nova-api error

2011-09-12 Thread Vishvananda Ishaya
Looks like the version of nova that is installed in pymodules is old. You somehow have a new api-paste.ini but old actual nova code. Vish On Sep 12, 2011, at 10:27 AM, Marc Peiser wrote: Hi all, I'm having issues with keystone install with nova/glance. I seem to be missing something as

Re: [Openstack] Integration tests

2011-09-12 Thread Jay Pipes
On Mon, Sep 12, 2011 at 1:22 PM, Matt Dietz matt.di...@rackspace.com wrote: Regarding novaclient, I was debating this myself last night. Our backfire suite currently uses it. However, I can see scenarios where nova has advanced but novaclient has yet to, and because all three projects are

Re: [Openstack] Diablo 4 dashboard error missing openstackx extensions in nova.

2011-09-12 Thread Roman Sokolkov
I ve see this error too.In httpd error logs I have the same traceback. Additional my nova-api.log : 2011-09-12 17:12:03,732 DEBUG routes.middleware [-] No route matched for GET /admin/services from (pid=4312) __call__ /usr/li b/python2.7/site-packages/routes/middleware.py:97 2011-09-12

Re: [Openstack] Integration tests

2011-09-12 Thread Sandy Walsh
From: Jay Pipes [jaypi...@gmail.com] Can we discuss pulling novaclient into Nova's source tree at the design summit? +1 This email may include confidential information. If you received it in error, please delete it.

Re: [Openstack] Integration tests

2011-09-12 Thread Tim Simpson
It would be advantageous for the tests to only use one client, whatever it turns out to be. I think it would be most practical to use a bleeding edge version of Nova client and add any features necessary to it in its own repo directly or by somehow mixing it in with the test code. As for

Re: [Openstack] Integration tests

2011-09-12 Thread Matt Dietz
Ditto On 9/12/11 2:10 PM, Sandy Walsh sandy.wa...@rackspace.com wrote: From: Jay Pipes [jaypi...@gmail.com] Can we discuss pulling novaclient into Nova's source tree at the design summit? +1 This email may include confidential information. If you

Re: [Openstack] Integration tests

2011-09-12 Thread Gabe Westmaas
On 9/12/11 2:54 PM, Tim Simpson tim.simp...@rackspace.com wrote: It would be advantageous for the tests to only use one client, whatever it turns out to be. I think it would be most practical to use a bleeding edge version of Nova client and add any features necessary to it in its own repo

Re: [Openstack] Integration tests

2011-09-12 Thread Matt Dietz
Actually to extend this, Novaclient is used for zone communication inside of nova. It not being included is silly, at this point On 9/12/11 1:37 PM, Jay Pipes jaypi...@gmail.com wrote: On Mon, Sep 12, 2011 at 1:22 PM, Matt Dietz matt.di...@rackspace.com wrote: Regarding novaclient, I was

Re: [Openstack] Integration tests

2011-09-12 Thread Tim Simpson
One solution is to make a second version of the Nova client with the same interface, so that the test config could use one or the other on different test runs. We could make this client communicate via XML to ensure we hit the XML serialization code. To me the biggest benefit of using the

Re: [Openstack] Integration tests

2011-09-12 Thread Ed Leafe
On Sep 12, 2011, at 3:16 PM, Gabe Westmaas wrote: It would be advantageous for the tests to only use one client, whatever it turns out to be. I think it would be most practical to use a bleeding edge version of Nova client and add any features necessary to it in its own repo directly or by

Re: [Openstack] Diablo 4 dashboard error missing openstackx extensions in nova.

2011-09-12 Thread Anthony Young
Another thing to check - the os1.1 keystone endpoints have been modified in d4 to include a tenant prefix. If you are upgrading an existing install, you may need to update your keystone service catalog accordingly: keystone-manage $* endpointTemplates add RegionOne nova http://%= api_vip

Re: [Openstack-poc] making ppb meetings better (?)

2011-09-12 Thread Thierry Carrez
John Dickinson wrote: I have found the PPB meetings to be very disjointed and hard to follow. A classic example is from our last meeting where 3 votes were proposed and only 2 were voted on. At times we have several conversations going on at once. Sometimes we vote, someone continues the

Re: [Openstack-poc] making ppb meetings better (?)

2011-09-12 Thread Jonathan Bryce
Thanks John for bringing this up and everyone for the input. This is good feedback. Here are a couple of my thoughts on it: 1) I agree that the meetings get rather chaotic at times. I would say that most of the time that situation is less related to non-PPB users and more related to those of

Re: [Openstack-poc] making ppb meetings better (?)

2011-09-12 Thread John Dickinson
On Sep 12, 2011, at 11:14 AM, Jonathan Bryce wrote: Thanks John for bringing this up and everyone for the input. This is good feedback. Here are a couple of my thoughts on it: 1) I agree that the meetings get rather chaotic at times. I would say that most of the time that situation is

Re: [Openstack-poc] making ppb meetings better (?)

2011-09-12 Thread Jonathan Bryce
3) Along the lines of getting community feedback, do you all think that we should move to a two-part process where we discuss, send to a summary to the community requesting comments and then vote? Should we just try to gather feedback before we ever discuss? I'd suggest that for most