08:06:21 <joehuang> #startmeeting multisite
08:06:21 <collabot`> Meeting started Thu Nov 17 08:06:21 2016 UTC.  The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:06:21 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:06:21 <collabot`> The meeting name has been set to 'multisite'
08:06:31 <joehuang> hello
08:07:07 <joehuang> short feature implementation review
08:07:20 <joehuang> Ashish, how about keypair part
08:08:21 <SAshish> Yes, It is as per the discussion happend in the document.
08:09:11 <joehuang> Good, will wait for your patches
08:09:47 <joehuang> Is there any topic about the keypair need to discuss?
08:10:16 <SAshish> right now, i think let us get the keypair thing working
08:10:26 <joehuang> understand
08:10:28 <SAshish> Goutham will do it once he is done with kbclint
08:10:37 <goutham_> i will do that :)\
08:11:04 <joehuang> so Goutham will work on the keypair
08:11:15 <goutham_> yes
08:11:42 <joehuang> how about flavor/volume type replication?
08:12:30 <SAshish> for flavor, few may have compute specific dependencies
08:12:35 <SAshish> with extra specs
08:13:13 <SAshish> so leaving those type of flavors, we should pick others
08:13:31 <SAshish> though have not thought of it completely right now
08:13:32 <SAshish> :)
08:13:36 <joehuang> agree, but I think when Adminitrator decide to replicate, we assume that the administrator understand compute specific
08:14:00 <joehuang> no need to leave those type of flavors
08:14:32 <SAshish> If it is tagged with compute hostID
08:14:38 <SAshish> then it will be problem right
08:14:39 <joehuang> we may make assumption that Admin will take the risk and he has the knowledge about that
08:15:11 <joehuang> no hostID, but extra spec key-value
08:15:16 <SAshish> can there be a case where a flavor is tagged with particular compute
08:15:44 <joehuang> it's on host aggregate level, not on single hot level
08:16:03 <joehuang> it's identitified with host aggregate extra spec too
08:16:11 <joehuang> one second, I'll find the link
08:16:48 <SAshish> okay, I will look into it once, right now I dnt have much understanding of this. Will read about it and then I can discuss
08:16:55 <SAshish> sure
08:18:50 <joehuang> "Configure scheduler to support host aggregates" section in http://docs.openstack.org/kilo/config-reference/content/section_compute-scheduler.html
08:19:09 <joehuang> #info flavor and aggregate "Configure scheduler to support host aggregates" section in
08:19:12 <joehuang> http://docs.openstack.org/kilo/config-reference/content/section_compute-scheduler.html
08:19:18 <joehuang> #link "Configure scheduler to support host aggregates" section in
08:19:22 <joehuang> http://docs.openstack.org/kilo/config-reference/content/section_compute-scheduler.html
08:19:34 <joehuang> #link http://docs.openstack.org/kilo/config-reference/content/section_compute-scheduler.html
08:19:41 <joehuang> sorry, paste many times
08:19:57 <joehuang> good
08:20:29 <SAshish> its okay, yeah will try this
08:20:31 <joehuang> for client, I saw the patche
08:20:53 <joehuang> good
08:21:07 <joehuang> any topic on client?
08:21:11 <goutham_> yea we need to do some modifications so
08:21:13 <goutham_> gave -1
08:21:40 <joehuang> ok. so review after your update
08:22:02 <goutham_> nothin as of now working on tests cases once it is done i will update it .i have to discuss about the micro version in
08:22:07 <goutham_> the keypair part
08:22:08 <goutham_> :)
08:22:21 <joehuang> ok
08:22:51 <goutham_> i mean to get the enabled users
08:23:31 <goutham_> so i have gone through the document as per my understanding we may have to create nova object twice
08:23:38 <joehuang> I think we may have to use fixed version as for now
08:24:40 <joehuang> versions discovery is a little complex
08:24:47 <goutham_> yes because i have tried with this as well  "from novaclient import api_versions"
08:25:18 <joehuang> Ashish, how about your suggestion?
08:25:29 <goutham_> with this api-version we can create only the major version and in mitaka devstack the major version is 2.0
08:26:32 <SAshish> Goutham: have you checked how it is done in other projects?
08:26:34 <goutham_> but with 2.0 u cant get the list of keys for a given user
08:26:36 <SAshish> senlin, mistral
08:26:39 <SAshish> etcc.
08:27:03 <joehuang> The latest version of Nova should be 2.37 or so
08:27:22 <goutham_> <SAshish> :no i didnt go through the other project
08:27:47 <SAshish> it is always better if we have dynamic versioning so that it will be easy to maintain down the line
08:27:54 <joehuang> the major version is 2, and minor version
08:27:59 <joehuang> sure
08:28:11 <joehuang> if we can use dynamic version, it'll be much better
08:28:23 <SAshish> provided all latest version have backward compatibility
08:28:27 <goutham_> >>> latest = api_versions.discover_version(nova,api_versions.APIVersion("2.latest")) >>> latest <APIVersion: 2.25>
08:28:28 <joehuang> Ashish's suggestion is good, refer to Mistral or senlin
08:28:29 <SAshish> which it will have
08:28:57 <goutham_> this is what i got in local devstack (mitaka) have to check with newton
08:29:01 <joehuang> use latest is also one way
08:29:15 <goutham_> sure i will look into mistral and senlin
08:29:35 <goutham_> yes to have the latest one i am afraid we may have create object twice
08:29:35 <joehuang> you need to update mitaka devstack
08:29:58 <SAshish> so Joe, Functest will have all newton clients right?
08:30:12 <joehuang> one second, I'll check
08:30:40 <joehuang> #link https://wiki.opnfv.org/display/functest/OpenStack+python+clients
08:31:52 <joehuang> FUNCTEST will use fixed version
08:32:04 <joehuang> that's the conflict happened
08:33:42 <joehuang> we are using older client version, Jose made a mistake
08:34:25 <SAshish> shall we need to upgrade our client versions
08:35:10 <joehuang> We use version: 0.2.2 in Functest, the client requirements for this tag is older
08:35:19 <joehuang> https://github.com/openstack/kingbird/blob/0.2.2/requirements.txt#L41
08:35:51 <joehuang> Functest install python-openstackclient==2.3.0
08:36:07 <joehuang> Kingbird install python-openstackclient>=2.1.0 # Apache-2.0
08:36:35 <joehuang> I'll try the Functest container tomorrow
08:36:49 <SAshish> okay
08:37:11 <joehuang> BTW, Tricircle has just been accepted as OpenStack official project
08:37:16 <SAshish> yeah, congrats
08:37:19 <SAshish> nice :)
08:37:29 <joehuang> so I am not so crazy busy in the application defense
08:37:48 <goutham___> congrats joe
08:37:59 <joehuang> thanks to all of you for the consideration in multisite
08:38:04 <SAshish> when there can be such news for Kignbrid
08:38:39 <joehuang> Needs to discuss with you and Dimitri, shall we make kingbird more publicly and being one of openstack official project?
08:39:02 <SAshish> yeah, we have to
08:39:18 <joehuang> We have developed kingbird in Multisite, OPNFV, this is not visible in OpenStack community
08:39:31 <SAshish> correct,
08:39:44 <joehuang> We may have conclusion in next weekly meeting how to do that
08:39:55 <SAshish> +1
08:40:00 <joehuang> Good.
08:40:08 <joehuang> other topics?
08:40:16 <SAshish> intel labs
08:40:20 <SAshish> job failure
08:40:22 <joehuang> please
08:40:34 <SAshish> can we restart the jumphost
08:40:35 <joehuang> It's because josh has removed kingbird job
08:40:41 <SAshish> its damn slow
08:41:14 <joehuang> could you send a mail in opnfv mail-list, if nobody object, just restart it
08:41:27 <SAshish> yeah
08:41:47 <joehuang> I'll check the client version issue in Functest container
08:42:07 <joehuang> #action fix the client issue in Functest
08:42:38 <joehuang> #action consulting whether it's ok to jumphost restart in OPNFV mail-list
08:43:08 <joehuang> #info discuss whether to make kingbird as one OpenStack official project in next weekly meeting
08:43:26 <joehuang> more topics?
08:43:38 <SAshish> @action Goutham: check mistral and senlin for nova client versions
08:43:38 <collabot`> SAshish: Error: "action" is not a valid command.
08:43:48 <joehuang> use #
08:43:48 <SAshish> #action Goutham: check mistral and senlin for nova client versions
08:43:56 <joehuang> gerat
08:43:59 <joehuang> great
08:44:02 <SAshish> no other topics right now
08:44:05 <SAshish> :)
08:44:20 <joehuang> ok, let's conclude the meeting, thank you, have a nice day
08:44:28 <SAshish> thanks, Good bye
08:44:35 <joehuang> bye
08:44:45 <joehuang> #endmeeting