14:01:00 <gampel> #startmeeting Smaug 14:01:01 <openstack> Meeting started Tue Mar 8 14:01:00 2016 UTC and is due to finish in 60 minutes. The chair is gampel. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:05 <openstack> The meeting name has been set to 'smaug' 14:01:14 <gampel> hi zhonghua-lee 14:01:20 <zhonghua-lee> hi 14:01:30 <gampel> who is here ? 14:01:34 <saggi1> hi 14:02:15 <gampel> #info saggiq ,zhonghua-lee , gampel in the meeting 14:02:22 <xiangxinyong456> hi 14:02:31 <yuval> Hey 14:02:45 <gampel> #info xiangxinyong456 , yuval in the meeting 14:03:05 <gampel> Welcome Yuval, Please welcome yuval he joined the team this week 14:03:24 <zhonghua-lee> welcome 14:03:32 <gampel> Lets start or should we wait for someone 14:03:51 <gampel> #topic Smaug API v1.0 14:04:06 <gampel> I am the only one that approved it, please vote so we could merge tomorrow or fix it 14:04:22 <gampel> #link https://review.openstack.org/#/c/244756/ 14:04:55 <zhonghua-lee> I will go over it later 14:05:02 <gampel> are there any open issues in the API that someone what to talk about ? 14:05:25 <gampel> we are Missing operation log will be added in another patch 14:05:32 <xiangxinyong456> ok 14:05:40 <gampel> saggi will you take care of the next patch 14:05:48 <zhonghua-lee> looks good, no question 14:05:57 <xiangxinyong456> eran 14:06:36 <gampel> yes xiangxinyong456: 14:06:47 <xiangxinyong456> do you think we need the modify workflow in the version 1? 14:06:47 <gampel> saggi: ? 14:07:04 <saggi1> gampel: I don't understand the question 14:07:40 <gampel> #action saggi send another patch for the operation log 14:08:13 <gampel> xiangxinyong456: can you please explain i am not sure i understand 14:08:38 <xiangxinyong456> ok 14:09:07 <xiangxinyong456> about the update operation 14:09:39 <gampel> which update ? 14:10:20 <xiangxinyong456> you know we need add ,delete,select and update on the rest api 14:10:28 <gampel> yes 14:10:59 <xiangxinyong456> for example,modify the plan 14:11:13 <gampel> whats the work flow modification needed 14:12:03 <gampel> when we modify a plan it should be updated to the DB , as i discussed with @chenying in his patch 14:12:06 <xiangxinyong456> e.g. modify the checkpoint 14:12:14 <gampel> it shoudl be done only in suspended state 14:12:42 <saggi1> You can't modify the checkpoint from the api 14:12:50 <saggi1> except for changing the status for deletion 14:13:19 <xiangxinyong> ok 14:13:20 <gampel> the plan update is not relevant to the checkpoint all ready created we will serialize the plan into the checkpoint 14:13:40 <gampel> do you see other work flow modification needed ? 14:13:54 <saggi1> gampel: Checkpoint are inherently immutable 14:14:00 <xiangxinyong> modify trigger? 14:14:37 <gampel> I think if we modify a trigger we should update the operation engine 14:15:01 <gampel> and update the trigger that are registered in the system 14:15:05 <saggi1> plans can update but checkpoints are static data in the bank 14:15:46 <xiangxinyong> gampel: yeah, we need these workflow. 14:15:57 <xiangxinyong> saggi:understood 14:16:29 <gampel> xiangxinyong: ok i agree can you create a file with all the needed workflow and we could all verify them ? 14:17:38 <xiangxinyong> ok. we can gather these inforamtion and send to you 14:17:40 <gampel> #action xiangxinyong will define and verify workflow for update on all the rest API 14:17:56 <gampel> #topic REST API 14:18:14 <gampel> @chenying or @zengchen are here to update 14:18:39 <gampel> zhonghua-lee: can you update the status of the REST api 14:18:50 <gampel> #link https://review.openstack.org/#/c/286412/ 14:18:58 <gampel> #link https://review.openstack.org/#/c/286406/ 14:19:06 <gampel> #link https://review.openstack.org/#/c/287036/ 14:19:25 <gampel> I think it will be good to start with the REST see how we can merge soon 14:20:03 <gampel> zhonghua-lee: can you please update the status for @chenying and @zengchen 14:20:36 <gampel> zhonghua-lee: ? 14:21:04 <gampel> xiangxinyong: do you know the status of this patches ? 14:21:31 <gampel> I review them all i think that they are almost there 14:21:39 <zhonghua-lee> gampel: ok 14:21:57 <zhonghua-lee> chenying, adn chenzeng's work are in good progress 14:21:57 <gampel> please everyone review today or tomorrow so we can merge them 14:22:09 <xiangxinyong> gampel: we will 14:22:33 <gampel> I think we should create priority of all the patches so we could start merging 14:22:42 <gampel> I think the fisrt to go in should be the REST 14:22:42 <zhonghua-lee> now, we have much work on UI 14:22:55 <zhonghua-lee> gampel: totally greee 14:23:10 <gampel> Ok next topic 14:23:28 <gampel> #Protectables status & issues 14:23:50 <gampel> #topic Protectables status 14:24:13 <zhonghua-lee> I think we should think about how to get the child resource 14:24:45 <zhonghua-lee> right now , we get all volume from volume plug-in 14:25:02 <zhonghua-lee> it's not easy to understand 14:25:10 <zhonghua-lee> I think 14:25:16 <gampel> yes each Protectables retruns its type only on the get_child_r.. 14:25:48 <saggi1> zhonghua-lee: We know, it was something we thought about originally when designing the protectables. 14:25:57 <gampel> we had two option to do it top down or down to top 14:26:07 <saggi1> There are two ways. Have the VM Protectable in charge of getting the dependencies of all types 14:26:21 <saggi1> or have each type be reponsible for itself. 14:26:23 <zhonghua-lee> yes, i know why do that but, I think we should change the method name 14:26:40 <gampel> if this name is confusing no problem 14:26:44 <saggi1> zhonghua-lee: I don't mind, do you have a suggestion? 14:26:46 <gampel> whats your suggestion 14:27:29 <zhonghua-lee> saggi1: how about support a method wiht filter function 14:27:55 <gampel> I am not sure i understand ? 14:28:44 <zhonghua-lee> I mean we can use the list_resouce method, and add a filter parameter 14:29:03 <zhonghua-lee> so that this method can filter resource by parent resource 14:29:41 <saggi1> so changing it to filter_by_parent_resource()? 14:30:01 <saggi1> or list_by_parent_resource()? 14:30:05 <zhonghua-lee> saggi1:something like that 14:30:27 <gampel> we must provide the resource 14:30:31 <gampel> as a param 14:30:34 <zhonghua-lee> list_by_parent_resource sounds betterr 14:30:34 <saggi1> of course 14:30:38 <saggi1> the interface is the same 14:30:41 <saggi1> it's just the name 14:30:52 <saggi1> list_by_parent_resource(parent_resource) 14:30:59 <zhonghua-lee> there are some difference 14:31:06 <saggi1> ? 14:31:33 <zhonghua-lee> I think "child resource" means the resource under the parent resource 14:31:34 <gampel> zhonghua-lee: : I think that what you are proposing is not possible 14:31:50 <zhonghua-lee> e.g. volume is the child resource of VM 14:31:59 <saggi1> dependent 14:32:19 <zhonghua-lee> gampel: why? 14:32:33 <gampel> if it is changing the name no problem 14:32:44 <zhonghua-lee> gampel:it just change the name 14:32:57 <gampel> ok so no problem 14:33:53 <zhonghua-lee> gampel: it's just my suggestion. 14:34:12 <gampel> so get_dependent_ .. 14:34:31 <gampel> I am not sure i follow whats the proposed new name 14:34:49 <xiangxinyong> get_dependent_by_resource? 14:35:06 <zhonghua-lee> list_resource_by_parent? 14:35:27 <yuval> get_dependent_resources() ? 14:35:27 <gampel> Ok fine with me 14:35:37 <zhonghua-lee> some like that, whatever do not contain the "child" 14:36:01 <gampel> Ok saggi lets vote 14:36:06 <zhonghua-lee> s/some/something 14:36:44 <gampel> saggi: whats you view I am fine with both 14:36:54 <zengyingzhe_> what's the options? 14:37:06 <gampel> Hi welcome :) 14:37:18 <gampel> get_dependent_resources( parent) 14:37:22 <zengyingzhe_> sorry i'm late. 14:37:25 <gampel> and list_resource_by_parent 14:37:29 <gampel> no problem 14:37:51 <gampel> zengyingzhe_: we need an update of the protactable status 14:38:52 <gampel> Ok lets vote get_dependent_resources( parent) or list_resource_by_parent 14:39:22 <saggi1> I'm partial to get_dependent_resources 14:39:43 <zengyingzhe_> get_dependent_resources +1 14:39:47 <xiangxinyong> get_dependent_resources 14:39:49 <yuval> get_dependent_resources 14:39:57 <gampel> ok we got a winer 14:40:24 <gampel> #action rename fetch_child to get_dependent_resources 14:40:37 <zengyingzhe_> OK, i'll change this method name tomorrow. 14:40:46 <gampel> zengyingzhe_: can you please update of the protactable status 14:41:26 <zengyingzhe_> you mean at etherpad? 14:41:52 <gampel> no i mean the patches 14:42:03 <gampel> #link o ProtectableRegistry https://review.openstack.org/#/c/281783/ 14:42:17 <gampel> #link oCinder https://review.openstack.org/#/c/285611/3 14:42:30 <gampel> #link Nova https://review.openstack.org/#/c/286542/3 14:42:53 <gampel> #link protectable RPC handlers https://review.openstack.org/#/c/285921/ 14:43:10 <gampel> whats the status and what are we still missing 14:44:17 <gampel> zengyingzhe_: ? 14:44:30 <zengyingzhe_> OK, i got it. 14:44:55 <saggi> Hi, got disconnected 14:45:07 <gampel> welcome back 14:45:07 <zhonghua-lee> right now , nove,cinder plug-in are ready 14:45:22 <saggi> good jobs! 14:45:25 <saggi> job 14:45:33 <zhonghua-lee> how about network,image plug-in? 14:45:38 <gampel> very good job please ask everyone top review them so we could merge tomorrow 14:45:51 <gampel> yuval do you want to take glace image 14:45:59 <yuval> Yes 14:46:02 <gampel> ok 14:46:28 <gampel> networking should be easy as we look at it as one entity 14:46:46 <zengyingzhe_> I'll start to work on rest protectable plugins tomorrow. 14:47:22 <xiangxinyong> :) 14:47:23 <gampel> I think we all ready merged it 14:47:33 <gampel> @chenying did it 14:47:47 <zhonghua-lee> gampel:yeah, but after review :) 14:48:08 <gampel> Lets skip the ยท Protection Plugin & Service 14:48:27 <gampel> I think that the developers of that part are not here 14:48:36 <gampel> #topic UI dashboard status 14:49:00 <gampel> xiangxinyong or zhonghua-lee do you what to update on the status 14:49:16 <xiangxinyong> i have submit two patches, please review them 14:49:32 <saggi> xiangxinyong: it's in my queue 14:49:40 <gampel> yes in mine as well 14:49:51 <gampel> please ask all tyhe team to review it as well 14:50:02 <gampel> are there any issues ? 14:50:18 <gampel> open issues ? 14:50:35 <xiangxinyong> not yet 14:50:38 <zengyingzhe_> Has the protectable resource paging issue discussed yet? 14:51:20 <gampel> let discuss this @chenying include it in his REST patch 14:51:42 <gampel> but we can push the RPC and Plugin part to second phase 14:52:22 <gampel> xiangxinyong: :) very good job keep us up to date and will try to review 14:52:33 <gampel> zengyingzhe_: what do you think 14:53:10 <zengyingzhe_> I don't see why we need paging for protectable resources. 14:53:11 <xiangxinyong> gampel: np 14:53:32 <zengyingzhe_> it's represented as a graph 14:53:43 <zengyingzhe_> right? 14:54:16 <gampel> yes but the user can access the API of only a Volume lets say and a user could have 10000 14:54:31 <saggi> We might need it if the user has a lot of VMS 14:54:39 <saggi> So in the UI we'll need to page it 14:54:50 <zhonghua-lee> saggi:+1 14:55:07 <zhonghua-lee> especially the method "list_resource" 14:55:25 <saggi> For dependencies we might not need ti 14:55:34 <saggi> since we need to hold them all in memory anyway 14:55:49 <saggi> but listing could be a problem 14:56:08 <xiangxinyong> but how to paging for protectable resources in a tree structure? 14:56:14 <gampel> i agree it is not argent but from user experience and performance it will be required in second phase 14:56:38 <xiangxinyong> because the resource is listed by a tree 14:57:04 <zengyingzhe_> this means we must record all resources info in some place, such as DB or memory 14:57:21 <zhonghua-lee> xiangxinyong: you mean UI? 14:57:23 <zengyingzhe_> right now, there's no such implementation. 14:57:26 <gampel> Yes if the user as 10000 VMs you will have to show all the VMs and it will probebly take a very long time 14:57:38 <xiangxinyong> zhonghua-lee: yeah. 14:57:38 <saggi> Most openstack listing APIs support paging so we just forward it. 14:58:06 <gampel> I think that we will have to support this but it can be done in second phase 14:58:10 <zhonghua-lee> xiangxinyong:we can use async loading 14:58:15 <xiangxinyong> zhonghua-lee: the resources are listed in a tree. 14:58:30 <gampel> zengyingzhe_: do you agree 14:58:43 <xiangxinyong> zhonghua-lee: it is not related with async and sync. 14:58:57 <zhonghua-lee> xiangxinyong: what's th problem 14:59:03 <dneary> Joining a few minutes late today 14:59:10 <zengyingzhe_> gampel, yes, i'll think through it to find how to implement. 14:59:19 <gampel> we could open a bug about it so we will not forget and currently be focused on the end to end integration 14:59:35 <zengyingzhe_> sure 14:59:37 <gampel> so we could merge all teh protactabole patchs 14:59:38 <zhonghua-lee> gampel:+1 14:59:57 <dneary> Sorry - wrong channel 15:00:01 <xiangxinyong> zhonghua-lee: the resource is listed by parent, i page the whole tree 15:00:07 <gampel> #action zengyingzhe_ open a bug about the pagination support in the protactable 15:00:14 <gampel> sorry thank you every one 15:00:21 <gampel> our time is done 15:00:33 <gampel> by end thank you for all the good work 15:00:36 <xiangxinyong> 3ks 15:00:37 <zhonghua-lee> xiangxinyong:let's disucss later 15:01:01 <zengyingzhe_> Thank you all, bye 15:01:02 <gampel> #endmeeting 15:01:02 <xiangxinyong> zhonghua-lee: ok 15:01:09 <gampel> #endmeeting