02:01:24 <zhenguo> #startmeeting mogan
02:01:25 <openstack> Meeting started Thu Apr 20 02:01:24 2017 UTC and is due to finish in 60 minutes.  The chair is zhenguo. Information about MeetBot at http://wiki.debian.org/MeetBot.
02:01:26 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
02:01:28 <openstack> The meeting name has been set to 'mogan'
02:01:59 <liusheng> o/
02:02:45 <zhenguo> liusheng, Xinran_, zhangyang: o/
02:02:58 <zhangyang> o/
02:03:21 <Xinran_> o/
02:03:25 <zhenguo> seems others will be late
02:03:46 <zhenguo> as always, the agenda:
02:03:46 <zhenguo> #link https://wiki.openstack.org/wiki/Meetings/Mogan#Agenda_for_next_meeting
02:04:02 <zhenguo> let's jump in
02:04:09 <zhenguo> #topic announcements and reminders
02:04:22 <zhenguo> The tempest gate is still broken :(
02:04:42 <zhenguo> sorry for the inconvenience
02:05:07 <liusheng> any ironic guy tring to fix that ?
02:05:55 <zhenguo> liusheng: seems some gates turns to green, but still some fails
02:06:11 <liusheng> zhenguo: :(
02:06:59 <zhenguo> liusheng: I will try to look into it today
02:07:30 <zhenguo> another thing is we have replaced /types with /flavors in our URI.
02:07:41 <zhenguo> maybe also need to change /instances to /servers.
02:08:01 <liusheng> zhenguo: we can try to skip the tempest test if we are urgent to merge something
02:08:08 <liusheng> zhenguo: that is cool
02:08:39 <zhenguo> liusheng: hah, let's try to fix it instead make it non-voting :D
02:08:52 <liusheng> zhenguo: lol
02:09:23 <zhenguo> Mogan UI is comming, I will try to add the instances panel by this week
02:09:44 <zhenguo> maybe we need a separated panel group like Compute
02:09:52 <zhenguo> I would like to call it Bare Metal Compute
02:10:19 <liusheng> zhenguo: awesome!
02:10:21 <zhenguo> with some panels like Compute, overview, instances, images, keypairs
02:10:44 <liusheng> zhenguo: you are horizon expert!
02:10:51 <zhenguo> liusheng: ol
02:10:52 <zhenguo> lol
02:11:19 <zhenguo> but not sure if it's a good idea to totally separated from Compute
02:11:30 <zhenguo> let's try to review the dashboard when it's ready
02:12:44 <zhenguo> as we don't have any moganclient released, I added the python-moganclient repo to mogan-ui test requirements
02:13:11 <zhenguo> and a seperated devstack plugin like other horizon plugins
02:13:57 <liusheng> zhenguo: maybe we can release moganclient in pypi
02:14:19 <zhenguo> liusheng: yes, but maybe it's better to wait for the new flavor, wdyt?
02:14:28 <liusheng> zhenguo: yes, sure
02:14:51 <shaohe_feng> zhenguo: liusheng: I'm here
02:14:58 <shaohe_feng> morning all
02:14:59 <zhenguo> liusheng: as there's a way to add non released client to requirements, it's not that urgent :D
02:15:09 <liusheng> zhenguo: yes, I know that
02:15:10 <zhenguo> shaohe_feng: o/
02:17:01 <zhenguo> shaohe_feng: hi, do you have anything want to talk in this topic :D
02:17:39 <shaohe_feng> zhenguo: Let me go through the msg
02:18:28 <shaohe_feng> zhenguo: what's change,  about client?
02:19:08 <zhenguo> shaohe_feng: not client, we talk about whether we need to release the client
02:19:54 <zhenguo> shaohe_feng: does the OSIC affect you Intel guys :P
02:20:24 <shaohe_feng> zhenguo: waiting for msg :)
02:20:42 <zhenguo> shaohe_feng: hah
02:21:02 <shaohe_feng> zhenguo: we defer the client release just for the new flavor?
02:21:19 <zhenguo> shaohe_feng: not only
02:21:28 <zhenguo> as we still have many things to refactor maybe
02:21:32 <zhenguo> it's not stable
02:21:44 <zhenguo> and we don't have a strong requirement to release the client
02:22:00 <shaohe_feng> zhenguo: how many works for the client?
02:22:08 <liusheng> the releasing is also easy
02:22:27 <zhenguo> shaohe_feng: not much works, but we need to increase the client version number
02:23:02 <zhenguo> liusheng: sure
02:23:25 <shaohe_feng> zhenguo: OK. So it it not a emergency thing.
02:23:34 <zhenguo> yes, I think so
02:24:03 <zhenguo> ok, if there's no other things, let's move on
02:24:08 <shaohe_feng> zhenguo: OK. maybe we can focus on the high priority
02:24:17 <zhenguo> yes
02:24:26 <zhenguo> #topic task tracking
02:24:38 <zhenguo> #link https://etherpad.openstack.org/p/mogan-task
02:24:53 * zhenguo gives folks a few minutes to review
02:26:52 <zhenguo> for the new flavors, I would like to  add a POC patch first
02:28:07 <zhenguo> and I find we have a instance type projects table, but no access control exposed to users, really weired, hah
02:29:33 <shaohe_feng> zhenguo: who will improve it?
02:29:39 <liusheng> zhenguo: I found that long ago
02:29:54 <liusheng> zhenguo: and seems I have asked some guy
02:30:07 <liusheng> zhenguo: lol
02:30:19 <zhenguo> hah, we can add  that support first
02:31:13 <zhenguo> if there's no one want to take it, I will :D
02:31:54 <liusheng> zhenguo: do we need a separated instance type project table ?
02:31:55 <zhenguo> the access control URI may look like /flavors/flavor_uuid/access, wdyt?
02:32:06 <liusheng> zhenguo: or just add a column in instance_type ?
02:32:27 <zhenguo> liusheng: maybe it's better to be separated, for indexing
02:33:03 <liusheng> zhenguo: will check it more when reviewing :D
02:33:12 <zhenguo> liusheng: and I'm thinking to make cpus, memory, disks, nics all inseperated tables from flavors
02:34:58 <zhenguo> the keypairs patch is blocked by the gate
02:35:02 <liusheng> zhenguo: that may cause many tables association
02:35:04 <liusheng> zhenguo: yes
02:35:32 <zhenguo> liusheng: yes, but not sure the advantages and disadvantages
02:35:45 <liusheng> zhenguo: but you can download the mogan patch and client patch to test the keypars commands :D
02:35:47 <zhenguo> liusheng: do you know why nova make all things separated from instances
02:36:05 <zhenguo> liusheng: sure, thanks for the big patch!
02:36:24 <liusheng> zhenguo: np
02:36:51 <liusheng> zhenguo: not all things in Nova are separated from instances I think, may need to dig more.
02:37:13 <zhenguo> liusheng: yes, let's look into it
02:37:37 <zhenguo> after the keypair patch landed, I will add it to configdrive
02:38:42 <shaohe_feng> zhenguo: Good. Keypair is too important to users.
02:38:51 <liusheng> zhenguo: that will be convenient for functionalities testing
02:39:24 <shaohe_feng> liusheng: Good job.
02:40:14 <zhenguo> but users may need to import nova keypairs to mogan, it's somewhat convoluted
02:41:01 <zhenguo> but we can't use nova keypairs directly :(
02:41:52 <liusheng> hah, the keypair-create api provide two functionalities, 1. create a new keypair; 2. import an existing keypair
02:42:04 <liusheng> zhenguo: may that can help
02:42:07 <shaohe_feng> zhenguo: do you means user will deploy both nova and mogan?
02:42:39 <shaohe_feng> liusheng: import the existing keypair from nova?
02:42:43 <zhenguo> liusheng: yes, they can import the keypairs for vm to mogan
02:43:11 <zhenguo> shaohe_feng: yes, nova for vms and mogan for bms, in a hybrid env
02:43:14 <liusheng> shaohe_feng: yes, the mogan db or Nova db only record the public key, and can be query
02:43:17 <shaohe_feng> then I think maybe keypairs put in nova DB is a not a good idea.
02:43:42 <shaohe_feng> myabe another project should take care keypairs.
02:45:04 <zhenguo> yes, but we can't rely on nova keypairs API, as Mogan env is not with Nova always
02:45:12 <liusheng> shaohe_feng, zhenguo yes, we should consider the standalone scenario of mogan deployment
02:45:28 <zhenguo> yes
02:46:25 <zhenguo> ok, instance fault patch seems ready
02:46:40 <shaohe_feng> zhenguo: yes.
02:46:59 <zhenguo> shaohe_feng: do you know why luyao implement the hasattr built-in function ?
02:47:30 <shaohe_feng> zhenguo: yes.
02:47:47 <shaohe_feng> zhenguo: the build_in method hasattr for python3 has different action with python2
02:47:59 <shaohe_feng> zhenguo: it has side-effect
02:48:22 <zhenguo> shaohe_feng: oh,
02:48:43 <zhenguo> do other projects implement it again?
02:49:01 <shaohe_feng> zhenguo: it try to load the "fault" attr, if can not load, py3 throw a exception.
02:49:05 <zhenguo> sorry, I don't get time to look through the whole patch
02:49:21 <shaohe_feng> zhenguo: but python2 just pass it, and return false.
02:49:31 <zhenguo> shaohe_feng: you mean hasattr will raise exception?
02:49:41 <shaohe_feng> zhenguo: yes. just for py3
02:49:50 <zhenguo> shaohe_feng: oh, tha't odd
02:50:00 <shaohe_feng> zhenguo: we can check other project.
02:50:14 <zhenguo> shaohe_feng: ok, thanks
02:50:27 <shaohe_feng> zhenguo: oh, maybe we can improve the version object lib.
02:50:40 <shaohe_feng> zhenguo: luyao's way can work around.
02:50:52 <liusheng> shaohe_feng: really, I just tested in py3, it return a False
02:51:38 <zhenguo> liusheng: you mean no exception?
02:51:44 <liusheng> zhenguo: yes
02:51:52 <liusheng> Python 3.5.2 (default, Nov 17 2016, 17:05:23)
02:51:53 <shaohe_feng> liusheng: zhenguo: we can check the py3 doc for hasattr
02:51:53 <liusheng> [GCC 5.4.0 20160609] on linux
02:51:53 <liusheng> Type "help", "copyright", "credits" or "license" for more information.
02:51:53 <liusheng> >>> class test(object): pass
02:51:53 <liusheng> ...
02:51:53 <liusheng> >>> a=test()
02:51:56 <liusheng> >>> hasattr(a, 'aa')
02:52:07 <shaohe_feng> liusheng: yes, this example can work.
02:52:18 <zhenguo> ok, we can check after the meeting
02:52:49 <zhenguo> seems no much time left
02:53:02 <zhenguo> zhangyang: hi, any update for cloudboot driver?
02:53:23 <zhenguo> Xinran_: how's the attach/detach interfaces work?
02:53:24 <shaohe_feng> liusheng: but fault field it use also python reflect to add it. some different with your test.
02:54:11 <liusheng> shaohe_feng: oh, got it.
02:54:20 <zhangyang> zhenguo: going to have a meeting with cloudboot guys today, and start coding next week. really slow :(
02:54:39 <zhenguo> zhangyang: hah,
02:55:02 <zhenguo> zhangyang: we can have a deep discussion during the bug smash.
02:55:32 <zhangyang> zhenguo: yeah, really expect that.
02:55:52 <zhenguo> zhangyang: :D
02:56:46 <zhenguo> liusheng: maybe you can update the eptherpad to move console task to Done :D
02:56:47 <Xinran_> I plan to apply these two fonctions in the class InstanceNetworksController. Is it possible ?
02:57:06 <zhenguo> XinRan_: sure
02:57:10 <liusheng> zhenguo: oh, ok
02:57:23 <zhenguo> Xinran_: maybe networks/interfaces ?
02:59:13 <zhenguo> ok, we can discuss this more after the meeting
02:59:19 <zhenguo> anything else here?
02:59:36 <shaohe_feng> no
02:59:42 <zhenguo> ok
03:00:05 <zhenguo> #topic open discussion
03:00:14 <zhenguo> who's got a thing:)
03:00:54 <zhenguo> ok, It's almost time, thanks everyone for joining the meeting!
03:01:00 <zhenguo> #endmeeting