13:00:57 <aglitke> #startmeeting 13:00:58 <kimchi-bot> Meeting started Wed Oct 30 13:00:57 2013 UTC. The chair is aglitke. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:58 <kimchi-bot> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:01:03 <aglitke> #meetingname scrum 13:01:03 <kimchi-bot> The meeting name has been set to 'scrum' 13:02:05 <aglitke> #info Agenda: 1) Process note 2) KVM Forum summary 3) Roundtable 13:02:22 <aglitke> Anything else? 13:02:55 <alinefm> I think this is all 13:02:58 <aglitke> The team has been very busy while I was out last week so I am going to ask for some help catching up. 13:03:03 <aglitke> great... 13:03:09 <aglitke> #topic Process note 13:03:54 <aglitke> As most of you noticed, last week alinefm was very helpful and she collected patches into a staging branch so they could be merged when I returned. 13:04:07 <aglitke> First of all, thanks to her for helping out with that... 13:04:44 <aglitke> Second: I think that worked so well that I'd like to ask her to continue helping share the responsibility of staging and merging patches. 13:05:04 <ming> aglitke, i disagree that. 13:05:10 <alinefm> aglitke, thank you for the opportunity! 13:05:11 <aglitke> It should help us to increase our velocity. 13:05:44 <ming> i think we need one from different timezone 13:06:23 <aglitke> ming: We can look into that in the future. But this is really my decision about who I want to work with. It's easier for me to coordinate with someone in my own timezone. 13:06:29 <hlwanghl> we have 9 or 10 guys in China 13:06:49 <shaohef> yes. 13:06:54 <ming> aglitke, that can not help China team. 13:07:15 <aglitke> markwu Is helping with maintainership of the kimchi-ginger branch. So I can see that responsibility evolving over time. 13:08:12 <shaohef> we should need a one familiar with kimchi code. 13:08:33 <aglitke> I am happy that shaohef has stepped up to help maintain issues. 13:08:34 <shaohef> to maintainership kimchi branch 13:08:45 <ming> aglitke, you know we need work overtime to catch a maintainer from your time zone. That is not easy. 13:08:46 <AdamKingIT> Going from 1 to 2 will give us a chance to work out the mechanics of >1. Hopefully we get so many contributors, we'll need to go 2+ real soon 13:08:57 <hlwanghl> a major part of patches are from China team, so it will help much if we have one in China 13:09:57 <alinefm> I think the main issue of getting patches merged in a good time is about reviewing 13:10:15 <aglitke> ming: Thanks for your suggestion. We will continue to work as a team to manage the difficulties with timezones. 13:10:27 <alinefm> if everyone reviews patches I can apply them sooner without having problems with timezones 13:10:30 <AdamKingIT> alinefm, we'll look forward to helping you assume this additional work! 13:10:49 <aglitke> alinefm: Agreed. The more thoroughly patches are reviewed, the easier and faster they can be merged. 13:11:08 <aglitke> Yes, thank you alinefm! 13:11:25 <pvital> sorry guys, but we are ONE team. I don't care if the commiter is in China, Europe, US or Brazil. The point is submit the patches and get them reviewed 13:11:27 <aglitke> #topic KVM Forum summary 13:11:34 <AdamKingIT> tell us about KVM forum? 13:12:03 <shaohef> maybe I can help with maintain patch. I'm familiar with git tool. And help most member of China team for their git manager and Makefile configure. 13:12:31 <ming> agtitke, not that way always. Even after full review from the others, maintain can block the integration for quite a while. 13:12:33 <aglitke> So kimchi was very well received at KVM Forum! I presented in a packed room. The audience was impressed with the clean UI and the ease of use. 13:12:51 <AdamKingIT> you have a guess how many people were packed in? 13:13:12 <aglitke> Several people asked me when/if kimchi would be ready to replace virt-manager. 13:13:59 <pvital> aglitke, did your talk recorded? 13:14:04 <aglitke> AdamKingIT: We were in a smaller salon but at least 60 I would assume. aliguori was there and several other maintainers from libvirt, etc. 13:14:11 <AdamKingIT> we have a little ways to go, but cool that are thinking that way 13:14:20 <alinefm> oh very good! 13:14:26 <hlwanghl> a good beginning 13:14:28 <aglitke> pvital: Yes it was. I will let everyone know once the videos for the Forum have been released. 13:14:30 <AdamKingIT> 60 is a pretty good crowd 13:15:02 <aglitke> I was able to demonstrate ISO streaming as a preview feature and people were very excited about it. 13:15:29 <aglitke> One comment that I think we should take seriously is the need to have kimchi packaged in distros. 13:15:48 <aglitke> Most people will try a new project by first installing the distro package. 13:16:03 <aglitke> Only then will they download the source and make improvements. 13:16:15 <aglitke> We have offers of help from the Fedora community already. 13:16:24 <pvital> aglitke, nice! we can use twitter, facebook and google+ to share the video and improve the marketing of kimchi 13:16:24 <AdamKingIT> great 13:16:29 <aglitke> ANd we should really work hard to add Ubuntu and openSUSE to the list. 13:17:00 <aglitke> pvital: Yes, very good suggestion. Anyone should feel free to market kimchi in their own social circles. 13:17:21 <alinefm> aglitke, what is the process to that? 13:17:25 <alinefm> if there is one 13:17:33 <aglitke> Kimchi is seen to directly address a gap we've had in the KVM tools for a long time. 13:17:50 <aglitke> alinefm: Each distro has its own process. 13:18:26 <aglitke> It can be a bit involved, but we have skills in this area already. 13:18:33 <pvital> aglitke, I have some experience with RPM and can help with it 13:18:35 <aglitke> I have done a fedora package for MOM 13:18:47 <aglitke> pvital: Excellent. You could look at openSUSE 13:19:02 <pvital> sure! 13:19:04 <aglitke> we already can build rpms today. We just need to get our package included in the official repos 13:19:33 <shaohef> but we need more features, then we can get our package included in the official repos 13:19:40 <aglitke> So... I guess that's all I have to report from KVM Forum. 13:20:00 <aglitke> shaohef: I think 1.0 is really good enough to start packaging. 13:20:08 <aglitke> It can already do some cool things. 13:20:18 <aglitke> You never want to wait until your finished to start. 13:20:49 <royce> In official repos will push us to evolve 13:21:07 <aglitke> So in summary, great work guys! People like what they see. 13:21:20 <aglitke> #topic Roundtable 13:22:11 <shaohef> aglitke: https://github.com/kimchi-project/kimchi/issues/217 13:22:21 <royce> I'd like to to ask guys about deep scanning and vm edit~ 13:22:34 <aglitke> Since I am a bit behind in the patch queue, if you have a series that is really ready to be merged, past the subject line of the patch series here. I will prioritize review for these sets. 13:22:59 <shaohef> aglitke: AdamKingIT: can you help to Make Kimchi page at Wikipedia in English? 13:23:20 <royce> [project-kimchi][PATCHv3 0/6] Using vm uuid inside model 13:23:22 <shaohef> aglitke: AdamKingIT: we will Make Kimchi page at Wikipedia in Chinese. 13:23:23 <alinefm> aglitke, I guess shaohef patches about distros are ready 13:23:33 <AdamKingIT> Yes, I'll look into it 13:23:38 <aglitke> royce: that is merged 13:23:44 <alinefm> [PATCH V7 0/7] support distros 13:23:58 <royce> uhhh, sorry, aglitke 13:24:07 <aglitke> royce: np :) 13:24:13 <aglitke> royce: let's talk deep scanning 13:25:01 <ming> I think debug report V11 is quite ready now. 13:25:16 <shaohef> aglitke: host monitor? 13:25:20 <aglitke> ming: ok... I will take a review. 13:25:30 <royce> that's good, aglitke, [project-kimchi][PATCHv2 4/4] deep scan: add model support 13:25:46 <shaohef> aglitke: for CPU usage? we should show every cpu usage? 13:25:46 <ming> I am working with Hong liang to deliver the UI part later. Anyway, the backend can be integrated first. 13:25:49 <aglitke> royce: You feel it is ready for merge? 13:26:05 <royce> deep scanning's lacking test because we don't have the psedo iso gen 13:26:19 <royce> I think it is ready, aglitke 13:26:28 <aglitke> ok... I will review. 13:26:51 <aglitke> also... everyone else can help to review these high priority patch sets 13:26:55 <royce> If you like my way wrapping task id in storagepool resource 13:26:57 <apporc> aglitke: I think [project-kimchi] [PATCH V3 0/5] Add spice support in KIMCHI is ready to be reviewed too. 13:27:13 <aglitke> ok 13:27:14 <apporc> aglitke: shaohef had tested that . 13:27:33 <shaohef> aglitke: yes. seem it is good. 13:27:35 <aglitke> alinefm: Where are we at with iso streaming? I know you've made some progress narrowing down the issues. 13:27:42 <ming> royce, I think we can enhance the task implementation a bit. 13:27:57 <alinefm> aglitke, I sent last week the a patchset for review 13:28:02 <alinefm> waiting for review =) 13:28:22 <alinefm> [PATCH 0/7 V2] ISO Streaming 13:28:27 <AdamKingIT> SOME people must have been away at a conference or something :-) 13:28:28 <aglitke> alinefm: So the feature tests can identify if an out of band DNS resolution is required? 13:28:43 <aglitke> AdamKingIT: :) 13:29:07 <hlwanghl> host tab: processors, memory dynamic statistics showing a line chart, should we include this? 13:29:14 <alinefm> aglitke, I used socket.gethostbyname() to get the ip associated to the hostname 13:29:26 <aglitke> alinefm: ok 13:29:27 <alinefm> not sure it is the best approach 13:29:41 <aglitke> hlwanghl: I am not sure about that history chart... 13:29:45 <ming> alinefm, Can you give a full description about how to test ISO streaming? 13:29:55 <aglitke> It doesn't fit our presentation style very well. 13:29:58 <AdamKingIT> hlwanghl you have a "host" patch ready for review? 13:30:01 <ming> I am really eager to test that feature. 13:30:06 <shaohef> aglitke:https://github.com/kimchi-project/kimchi/issues/198 more want to support spice. 13:30:37 <royce> ming, with alinefm's workaround it can be used I guess 13:30:46 <zhoumeina> we have just sent patchset to support spice 13:30:55 <alinefm> ming, it is on the patch cover letter but basically you just need to applu my patches and specify an URL to creata a template 13:31:05 <ming> royce, I would like a all one on on FC18 or FC19. 13:31:16 <hlwanghl> AdamKingIT most parts are ready, i'll integrate the back end api and thensend out 13:31:19 <ming> a all on --->all in one 13:31:25 <aglitke> Cool. I am eager to re-enable the part of the UI where you can create a template from a known remote iso (select from a list). 13:31:28 <pvital> AdamKingIT, hlwanghl, aglitke: I can review the host tab patches once it covers part of kimchi-ginger work!! 13:31:31 <shaohef> alinefm: why do you need to get the hostname? we can get the cherrpy.url to get the ip or hostname 13:31:37 <royce> I think that can do, because it works around libvirt and qemu problem 13:32:17 <royce> shaohef, qemu-io got stucked because the mis-use of curl when resolving the hostname 13:32:18 <aglitke> pvital: Thanks! 13:32:21 <markwu> aglitke, I can't say the plugin support patch is ready to merge, but I would like to know if you like the direction. 13:32:28 <hlwanghl> pvital great 13:32:43 <aglitke> markwu: Yes... I owe you a review on it. Do you cover the UI piece as well? 13:32:45 <alinefm> shaohef, it is about the URL for remote ISO 13:33:10 <alinefm> shaohef, on FC19 qemu gets stuck because it could not resolve the hostname, but if we use ip it works fine 13:33:53 <markwu> aglitke, I didn't cover the UI part. 13:33:55 <shaohef> alinefm: the ip is same with the cherrpy server? 13:34:03 <aglitke> markwu: ok. 13:34:19 <aglitke> shaohef: No. this is the IP of the server that hosts the ISO image 13:34:23 <hlwanghl> markwu we can work together on UI part 13:34:24 <aglitke> it can be anywhere 13:34:28 <ming> markwu, I think plugin might not a good word. I think it is just another application, not a plugin. 13:35:05 <aglitke> I prefer plugin. Application hints that it runs as a separate process. 13:35:07 <AdamKingIT> I remember a patch from meina to cover the UI extension. Did I dream that? 13:35:13 <aglitke> The plugin is extending the rest api 13:35:41 <markwu> and it's 'plug gable' indeed :) 13:35:56 <shaohef> aglitke: [PATCH 00/13] network related patches and the UI is ready. UI can create/delete/start/stop/list network. 13:35:59 <zhoumeina> AdamKingIT: UI extension? 13:36:04 <AdamKingIT> y 13:36:15 <aglitke> shaohef: Does the net create use macvtap , etc? 13:36:19 <zhoumeina> yes 13:36:36 <shaohef> aglitke: I will integrate all patches into one, and let every one try 13:36:36 <alinefm> AdamKingIT, I saw that too 13:36:39 <shaohef> aglitke: yes 13:36:47 <zhoumeina> I wan to dicuss that UI extension with you too 13:36:53 <aglitke> shaohef: That is excellent. 13:37:02 <aglitke> I am anxious to try it. 13:37:07 <AdamKingIT> I'll look for it and review. We may want to start to incorporate a notion of how auth will work for extensions. I saw an interesting talk by the paypal api arch last week that made me think... 13:37:14 <aglitke> It seems we have patches out for nearly all S1 features. 13:37:30 <royce> sorry to interrupt one more ask, guys do you like my way using sub-resources to deal with vm disks and nics, snapshots and so on? when referring to vm edit we will not touch its subresources~ 13:37:50 <aglitke> royce: Yes, I think that is a good idea. 13:37:53 <ming> aglitke, so it is a extension implemented by a seperate application. Extension may be better. 13:37:55 <shaohef> aglitke: if an ether nic exists create macvtap, or a linux bridge exists , then create a linux bridge network. 13:38:02 <aglitke> It will allow us to add/remove disks and nics. 13:38:08 <zhoumeina> AdamKingIT: I wonder what is the orginal purpose of autogenerate tabs by add tab htmls? 13:38:38 <royce> thanks aglitke, then we can move on on those edit features~ 13:38:45 <shaohef> royce: . 13:38:46 <pvital> aglitke, I think we need a review from UI side to [PATCH] StoragePool: Add NFS pool from pradeed 13:38:54 <AdamKingIT> zhoumeina: sorry I didn't follow the question. maybe we can chat afterwards, or tomorrow 13:39:13 <royce> shaohef? 13:39:14 <aglitke> zhoumeina: hlwanghl: Can you guys review the UI for [PATCH] StoragePool: Add NFS pool 13:39:16 <shaohef> royce: yes. I write my snapshot use subresources as VM already. 13:39:27 <royce> yeah, I know, shaohef 13:39:45 <zhoumeina> We want to make UI customize or just more convient to add tabs? 13:39:53 <shaohef> royce: it is convenience to edit VM 13:40:46 <royce> for the [PATCH] StoragePool: Add NFS pool, this patch melt all changes in a big one seems... 13:40:49 <zhoumeina> AdamKingIT: ok 13:40:51 <AdamKingIT> zhoumeina: main goal is to support functional extensions, so new capabilities manifest in the UI 13:41:11 <aglitke> zhoumeina: The plan is that extensions will just add their own tabs but will not modify existing tabs. 13:41:12 <shaohef> aglitke: I will integrate all patches include UI into one, and let every to try. and then split it for review. 13:41:23 <shaohef> aglitke: about the network. 13:41:47 <aglitke> shaohef: Is that necessary? I think it is easy enough to apply a series (as long as it does not need a rebase). 13:42:23 <zhoumeina> AdamKingIT: that is what I think 13:42:43 <shaohef> aglitke: someone think it is not convenience to save as so many patches and apply them. 13:42:45 <ming> shaohef, I always like a slef-containe package. But I am a developer, I can apply you patch into my workspace and install them correctly. 13:43:09 <shaohef> ming: thank you. 13:43:32 <aglitke> ok... So we all have a long list of patches to review. 13:43:36 <shaohef> aglitke: a good way to download too many patches and apply them. 13:43:52 <hlwanghl> aglitke: sure 13:43:52 <aglitke> Thanks in advance to everyone for helping us clear the patch backlog by reviewing code. 13:43:55 <shaohef> aglitke: do you use mutt? 13:44:02 <aglitke> shaohef: Yes! 13:44:12 <royce> aglitke, do we have a tool to measure our review? 13:44:15 <zhoumeina> aglitke: shaohef:we have los of issues in github that have not add tags and send out 13:44:15 <aglitke> it makes managing patches much easuer 13:44:47 <aglitke> royce: No not really. There are tools that can scan Reviewed-by tags on the mailing list. 13:44:50 <shaohef> zhoumeina: I will tags them with aglitke 13:44:56 <zhoumeina> some of those issues created long ago 13:45:08 <aglitke> But I have found that when you track that stat it encourages people to do more superfluous reviews. 13:45:10 <royce> sometimes our boss ask us our review activities, I will look for some~ 13:45:29 <aglitke> I'd rather people spend more time doing a smaller number of deep. substantive reviews. 13:45:30 <royce> :)))) 13:45:45 <royce> that's true 13:45:55 <shaohef> aglitke: got it. most of us use thunderbird. :) thunderbird it not convenience to manage so many patches. 13:45:59 <aglitke> So it's a hard stat to track. 13:46:11 <aglitke> shaohef: Yeah, that is why I do not use it. 13:46:18 <aglitke> Maybe Evolution is better? 13:46:28 <ming> Anyway, gerrit may be better to tack review stats. 13:46:33 <AdamKingIT> I think it comes under the heading of "be careful what you measure, its what you are going to get" 13:46:40 <shaohef> aglitke: Evolution better than mutt? 13:46:50 <ming> And gerrit can easily apply a set of patches 13:46:57 <aglitke> shaohef: Not for me, but maybe for you. 13:47:06 <aglitke> ming: We will not use gerrit. 13:47:12 <AdamKingIT> I didn't have much luck w/ evolution, though I have forgotten what happened 13:47:21 <aglitke> heh 13:47:23 <shaohef> aglitke: I will have a try Evolution. 13:47:27 <shaohef> aglitke: thank you. 13:47:34 <royce> If we are going to being in repo, a jenkins server may help? 13:47:55 <pvital> I heard that gerrit is one of the best tools to review patches 13:49:12 <shaohef> aglitke: substantive reviews is more important than measure review 13:49:34 <pvital> royce, Jenkins is a powerful CI tool (have worked a lot with Jenkins), but IMHO we need improve a little bit our unit tests to use it 13:49:34 <aglitke> pvital: Some people like it. This team has experience with it from oVirt. In my opinion, it encourages bad behavior. 13:49:43 <shaohef> royce: yes. apporc can help set up jenkins server 13:49:47 <aglitke> It is not as inclusive as a mailing list. 13:50:04 <aglitke> Because discussion is hidden from plain sight. 13:50:20 <pvital> aglitke, understand 13:50:34 <royce> right, pvital 13:51:37 <aglitke> Any other topics today? 13:51:54 <pvital> aglitke, yes 13:52:05 <shaohef> ping 13:52:30 <zhoumeina> aglitke: we have not get a conclution if we want to support IE8 from now on. 13:52:41 <pvital> aglitke, I have engineers to work with backend of the following sprint 2 and 3 tasks that covers kimchi-ginger use cases: Storage: New Storage Pool based on local disk (storage); Add new yum repository (from ISO and from URL) (host); Software update (host); and host: shutdown/reboot (host) from Sprint 3 13:52:41 <aglitke> zhoumeina: I think we should drop it. 13:52:42 <shaohef> anyone can response me? 13:52:52 <aglitke> shaohef: I see you 13:52:58 <zhoumeina> I agree 13:53:12 <aglitke> pvital: ok 13:53:19 <pvital> aglitke, Can I assign the engineers to work on them? Someone desagree? 13:53:42 <aglitke> Are the tasks free in https://github.com/kimchi-project/kimchi/wiki/Todo-1.1 ? 13:54:00 <pvital> aglitke, the back-end tasks, yes 13:54:14 <aglitke> ok. Then yes, feel free to add names by the tasks there 13:54:21 <pvital> I guess hlwanghl and mnz are working in the UI side 13:54:35 <zhoumeina> yuxin and dingxin too 13:54:42 <pvital> aglitke, ok thanks! 13:54:50 <AdamKingIT> pvital: I can work w. you on the items 13:55:01 <pvital> AdamKingIT, sure! 13:55:03 <aglitke> AdamKingIT: Thanks! 13:55:21 <zhoumeina> AdamKingIT: you can assgin me some work at sprint1 13:55:41 <markwu> so all of the tasks listed by pvital can be accepted in kimchi's master branch? 13:55:45 <AdamKingIT> I won't forget you :-) 13:55:49 <royce> Storage: New Storage Pool based on local disk (storage)-->can I ask what it this one? pvital 13:56:02 <pvital> aglitke, question? How can I proceed with new features requests? 13:56:19 <ming> markw: where are the tasks you are taking about? 13:56:25 <pvital> aglitke, for example: I requested support to raw type of VM image disks 13:56:37 <markwu> Storage: New Storage Pool based on local disk (storage); Add new yum repository (from ISO and from URL) (host); Software update (host); and host: shutdown/reboot (host) from Sprint 3 13:57:02 <shaohef1> pingh 13:57:04 <shaohef1> ping 13:57:10 <shaohef1> hello all 13:57:17 <pvital> shaohef, you're alive! 13:57:19 <AdamKingIT> shaohef1: pong 13:57:21 <shaohef1> do I miss some message? 13:57:21 <pvital> :-S 13:57:34 <aglitke> markwu: Yes, they all should be written in a cross platform, distro-neutral fashion 13:57:50 <pvital> ming, royce: https://github.com/kimchi-project/kimchi/wiki/Todo-1.1 13:58:42 <markwu> so the host management is also the coverage of kimchi, right? that's my understanding. but fnovak told me it belongs to kimchi-ginger. I am very confused about the line between kimchi and kimchi-ginger 13:58:56 <ming> pvital, thanks. 13:59:00 <AdamKingIT> that line bears some discussion... 13:59:16 <ming> markwu, we may setup a call with Frank about that. 13:59:31 <AdamKingIT> kimchi doesn't aim to be primarily in the host management business, but some features are useful for the KVM admin 13:59:45 <AdamKingIT> some or more host unique, and belong in ginger 13:59:48 <pvital> markwu, I understood that we have both goals: host and KoP support :-D 13:59:51 <royce> thanks I see, pvital, it is about the LVM pool 13:59:53 <ming> markwu, IMO, ginger is for power platform and most of the tasks are host management. 14:00:08 <aglitke> markwu: It is confusing. We need to decide which things belong in the main host tab and which ones will go into the extra tab 14:00:22 <markwu> aglitke, exactly! 14:00:23 <ming> pvital: what is KoP? 14:00:29 <pvital> royce, yeah! we need support to add new disks in a LVM pool 14:00:30 <aglitke> One example is the feature for creating a storage pool from a physical disl 14:00:47 <pvital> ming, KVM on Power 14:00:49 <aglitke> that is really beyond the scope of kimchi core 14:00:59 <ming> pvital, are you work on KVM on Power? 14:01:14 <aglitke> So the API to enumerate host disks and convert them to storage pools will probably belong on the ginger tab 14:01:30 <aglitke> markwu: We will need to have a review on a per-feature basis. 14:01:47 <markwu> but it's definitely helpful for virutalization management 14:01:50 <pvital> ming, http://www.youtube.com/watch?v=cLQI20LI6EQ&feature=share 14:01:52 <markwu> ok 14:01:53 <aglitke> Basic host mgmt: reboot, update, debugreport,. etc is ginger core and belongs on the kimchi host tab 14:02:13 <royce> pvital, we don't have access of youtube:( 14:02:15 <aglitke> ok guys. top of the hour. Thanks for joining. 14:02:16 <pvital> ming, I'm working on kimchi-ginger, that supports KoP and host side operations :-D 14:02:23 <aglitke> #endmeeting