13:01:49 #startmeeting 13:01:49 Meeting started Wed Dec 18 13:01:49 2013 UTC. The chair is alinefm. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:01:49 Useful Commands: #action #agreed #help #info #idea #link #topic. 13:01:56 #meetingname scrum 13:01:56 The meeting name has been set to 'scrum' 13:03:00 #info Agenda 1) 1.1 release 2) Mail list note 3) 1.2 Plan 4) Open discussion 13:03:03 anything else? 13:03:14 good for me 13:03:21 Nice. 13:04:19 #topic 1.1 release 13:04:32 https://github.com/kimchi-project/kimchi/wiki/Testing-1.1 13:04:50 royce, any update about ubuntu 13.4 tests? 13:05:36 I have nothing new, markwu reported that guests cannot be listed? 13:05:44 alinefm; autogen and make from disttarball worked 13:05:59 I'll clear the fixed bugs on the matrix 13:06:21 rotru, great! could you update the matrix? 13:06:25 royce, ok 13:06:29 alinefm; yeap 13:06:35 royce: markwu can list the guests now 13:06:36 my problem is fixed in latest code. 13:06:38 I think the most important fix we need to merge today is about logical pool 13:06:54 I think we should take a special care on Fedora 19. That is what Frobisher based on which includes Kimchi. 13:07:04 alinefm; make check is not working for me ... but I think the problem is in my kvm 13:07:05 alinefm: yes. still work with zhshzhou 13:07:09 the most important and the last one before announce the 1.1 release 13:07:30 zhshzhou: has post the patches to fix it 13:07:36 alinefm: test it on rhel. and it can works well for mpath device 13:07:55 alinefm: a problem need to look into. 13:08:11 I was testing it on ubuntu 13.10 and it doesn't list a partition 13:08:30 shaohef, zhshzhou, we need to focus on it 13:08:36 I just found the a problem. On old distributions, util-linux is 2.20. blkid does not list all recognized partitions. 13:08:40 alinefm: agree. 13:08:56 zhshzhou, any solution in mind? 13:09:00 I just check Ubuntu repo, it uses old util-linux 13:09:11 zhshzhou: $ rpm -q util-linux 13:09:11 package util-linux is not installed 13:09:25 can fdisk be used? 13:09:29 zhshzhou: problem? 13:09:39 well, we use blkid only to get partition path but as we will only work with disk/partitions, can't we assume it will be in /dev/? 13:09:40 shaohef: rpm -q $(which blkid) 13:09:42 royce: yes fdisk can used 13:09:56 Be careful of fdisk. 13:10:05 It is too low level. 13:10:21 in LVM official guide, they use fdisk output 13:10:47 zhshzhou: [user@localhost kimchi]$ blkid 13:10:48 /dev/vda1: UUID="cf4163f0-3a68-4afe-a546-35fa9ee4d9ea" TYPE="ext4" 13:10:48 /dev/vda2: UUID="VzkMUu-8HaF-bo6s-e03b-KZLF-i9YR-QX4ppF" TYPE="LVM2_member" 13:10:48 /dev/mapper/VolGroup-lv_root: UUID="aac2acac-efcb-4abc-8ad0-426b1542b373" TYPE="ext4" 13:10:48 /dev/vdb: UUID="bf68a601-64ec-4491-b91f-7840f6848a38" TYPE="ext4" 13:10:49 /dev/mapper/VolGroup-lv_swap: UUID="b675c8c1-5388-433f-9ee7-7963269198ad" TYPE="swap" 13:10:49 [user@localhost kimchi]$ rpm -q $(which blkid) 13:10:50 shaohef, royce, newer lsblk and list full path of a device, do not need blkid. 13:10:51 package /sbin/blkid is not installed 13:10:51 [user@localhost kimchi]$ which blkid 13:10:53 /sbin/blkid 13:10:56 zhoumeina; define old-linux distros 13:11:24 shaohef: my mistake, should be "rpm -qf" 13:11:30 Can we avoid too much detail? 13:12:20 zhshzhou, shaohef, can't we assume path is /dev/ ? 13:12:44 $ rpm -qf $(which blkid) 13:12:45 util-linux-ng-2.17.2-12.14.el6.x86_64 13:13:09 shaohef, too old. Let's consider alinefm's suggestion. 13:13:25 zh ok 13:14:05 continue other topic. 13:14:23 zhshzhou, do you have time to send a new patch version today? 13:14:28 I think it is late in China 13:14:50 I think it is a rush. 13:14:59 alinefm I'm afraid it's too hurry. 13:15:03 We need more test and review. 13:15:19 alinefm, I don't have time for such a big change. Small change can do. 13:15:20 It is better to delay the release to Friday. 13:15:54 zhshzhou, it would be a small fix, isn't it? 13:16:05 assume path is /dev and don't use blkid 13:16:19 and folks in Brazil can test it in time 13:16:27 rotru, danielhb ^ 13:16:53 There are two many devices under /dev? Not sure how do you which ones are disks? 13:17:00 alinefm, if you think this will do for the urgent release, I'm OK. 13:17:12 s/how do you/how do you/know? 13:17:22 ming, we list device names using lsblk, than we assume the full path is /dev/name 13:17:23 ming, from lsblk command 13:17:28 alinefm: let's try /dev 13:17:52 Well, my first implementation was to run "find /dev -name " but then I was asked to remove it and find another way 13:18:10 I have no problems with "find", it's more reliable than assuming than it's all on /dev/ 13:18:23 for instance, volume groups are on /dev/mapper/.. 13:18:26 sorry to interupt, any change I have to do in UI about logical pool? 13:18:29 it adds complexity 13:18:48 zhoumeina, no! I've just merged your patches about confirmation dialog 13:18:51 alinefm: too many devices on /dev. I try $ ls /dev/ just now 13:18:51 13:18:53 I really don't understand why we can't just construct the full path according to partition/disk type? 13:19:06 alinefm: $ ls /dev/ |wc -l 13:19:07 188 13:19:21 basically just assume it's /dev/sda[x] 13:19:28 markwu, yeap 13:19:50 for lvm, we can construct a '/dev/vg/lv' path for it 13:19:58 markwu, agree. 13:20:00 it's reliable enough 13:20:13 markwu. not all disk are scsi disks. 13:20:20 what about raw disk? 13:20:35 So you don't always have sda,sdb, sdc... 13:20:37 ming, I know, the device name is already included in the output 13:20:44 just an example 13:20:49 we can't assume it's all /dev/ .... again, why not run 'find . -name '? It's a performance degradation but way more reliable 13:20:51 makwu. which output? 13:21:12 danielhb, why not? 13:21:12 you can get a reasonable full path from the command line's output 13:21:29 the other option is to push kimchi support to at least Fedora 19, because then the lsblk command can provide the full device path by itself 13:21:46 well, let's move on 13:21:51 we can continue after meeting 13:22:03 the second point about 1.1 release is about translation 13:22:15 please, send out Portuguese and Chinese translations 13:22:38 I will send chinese one tomorrow morning 13:22:47 anyone would like to do that? 13:23:05 zhoumeina, it needs to be today for release 13:23:22 a volunteer for Portuguese? 13:23:49 oh, sorry, I don't have enviroment at home 13:23:51 hirata, ? 13:23:51 It seems we definitely need more time to release 1.1. 13:24:07 shaohef, could you do for Chinese? 13:24:12 alinefm: do you think it's a blocker for release 1.1 13:24:13 ? 13:24:16 alinefm: no problme 13:24:24 s/problme/problem 13:24:24 markwu, translation? 13:24:28 yep 13:24:30 shaohef, thanks 13:24:55 markwu, not a block but should be there 13:24:57 it will be great if shaohef can help 13:25:01 ok 13:25:05 shaohef, I do for Portuguese 13:25:08 ok 13:25:09 next 13:25:35 #topic Mail list note 13:25:45 alinefm; I can do portuguese 13:26:05 AdamKingIT worked hard and got our new mail list on ovirt community 13:26:12 AdamKingIT, thanks! 13:26:21 Thanks. AdamKingIT. 13:26:27 yw. It was more persistence than working hard:-) 13:26:50 after 1.1 release announcement all patches must be send to the new list for review 13:26:54 can we use their jenkins server too? 13:26:54 I am more greedy to have gerrit setup on oVirt. :) 13:27:09 all you need to subscribe to http://lists.ovirt.org/mailman/listinfo/kimchi-devel 13:27:13 AdamKingIT: may I ask about using other facility provided by ovirt 13:27:22 like jenkins and gerrit 13:27:35 markwu, royce, ming, it is the plan 13:28:01 alinefm, great. 13:28:07 there are 2 lists: one for devel (http://lists.ovirt.org/mailman/listinfo/kimchi-devel) and other one for users feedbacks (http://lists.ovirt.org/mailman/listinfo/kimchi-users) 13:28:22 I haven't asked for other services yet, though they were pretty accommodating with the list. I think we have some local test and discussion to go before a final decision on asking don't we? 13:28:39 AdamKingIT, agree 13:29:18 right? 13:29:22 next topic 13:29:23 AdamKingIT, I think most of us agree to have a public jekin test server and gerrit. 13:29:27 anyway, we need a public server outside IBM firewal to host jenkins 13:29:27 Frobisher guys asked me about if they can pull for a daily build, so jenkins is a must 13:29:38 ming, I don't agree with gerrit yet 13:30:00 alinefm, I mean most of us. 13:30:03 agreed. gerrit more of a question than jenkins 13:30:05 so do you have any chance to talk a look at gerrit? 13:30:19 alinefm: ^^ 13:30:29 markwu, not yet but it is on my todo list 13:30:32 markwu have you set up a local gerrit we could try? 13:30:38 these days I was really focus on 1.1 release 13:31:06 AdamKingIT: http://gerrit.ovirt.org/ 13:31:18 AdamKingIT: no problem. I will try to set up an internal one. but you also have try it with ovirt or openstack 13:31:40 k 13:31:44 #topic 1.2 Plan 13:31:52 schedule: https://github.com/kimchi-project/kimchi/wiki/Planning-1.2 13:32:02 plan: https://github.com/kimchi-project/kimchi/wiki/Todo-1.2 13:32:14 I made this first plan based on backlog 13:32:26 please, take a look on it and give feedbacks 13:32:36 4 sprint? 13:32:43 Kimchi 1.2 - Sprint 3 01/29/14 -> 02/19/14 13:32:59 I think it is on China new year. 13:33:01 I keep the person assigned on backlog into plan 13:33:07 but we can change it if needed 13:33:10 I think the Spring Festival will in release 2 13:33:13 alinefm, I am sorry... I was AFK and I did not see your msg in IRC... =) 13:33:36 there is no problem with that 13:33:45 alinefm, We have many work items in backlog. 13:33:47 there are people in other location to do the work 13:34:38 ming, any one would like to add to 1.2 plan? 13:34:58 just be careful as we will have most of people out some period and we need to do all *on* time 13:35:04 alinefm, when will you release your 1.2 draft plan? 13:35:36 ming, it is already public 13:35:41 alinefm, will you have xmas holiday for several days? 13:35:52 royce, no, just 2 days 13:36:07 alinefm, I mean the task items in 1.2? 13:36:25 ok, alinefm 13:36:45 alinefm: please assign spice to me 13:36:46 ming, sorry, I didn't understand your question 13:36:59 zhoumeina, sure 13:37:20 alinefm, I mean the task items removed from backlog to 1.2 plan. Where is are the task items? 13:37:33 https://github.com/kimchi-project/kimchi/wiki/Todo-1.2 13:37:34 apporc has spice patches on the list, though some comments not yet addressed 13:38:08 shaohef, do you have any news about apporc? 13:38:21 AdamKingIT: yes, I will continue to do it in 1.2 13:38:22 there is sometime he is out 13:38:25 alinefm: no 13:38:41 shaohef, ok 13:38:46 zhoumeina, I will assign it to you 13:38:51 alinefm: have no news for a long time. 13:38:55 alinefm: ok 13:39:10 so we need contact him to check if he can continue. If not, we can take over his work if he doesn't mind 13:39:10 I noticed someone didn't have tasks on some sprint, we can assign them to follow the review 13:39:28 alinefm: will you add ming's backlogs to 1.2? 13:40:00 zhoumeina, not sure about is "ming's backlog" 13:40:12 but kimchi backlog is on 1.2 plan 13:40:22 s/about/what 13:40:31 some thing like exception.... 13:40:56 code refactoring 13:41:16 zhoumeina, (alinefm) Improve error feedback on sprint 2 13:41:27 markwu, (alinefm) Restructure backend model (framework) on sprint 1 13:41:31 zhoumeina, parts of them. I think we can give a better sizing of all the items to see if we can add more into 1.2 relase. 13:41:41 oh, I see 13:41:51 ok 13:42:16 If we think we can do more, then lets get the high priority items done early, leaving time for the more 13:42:39 ming, we also have 60 issues to close 13:42:47 alinefm, when you select the items from the backlog, do you have a basic rule? 13:43:26 ming, yes 13:43:49 alinefm, what is that? 13:44:13 some people already demonstrate interest in do, patches depend on mail list, tasks needed by kimchi 13:44:20 ming, any problem with that? 13:45:37 alinefm: what about the storage attach and detach? 13:45:41 I think we should not be too optimistic about velocity, if we merge too soon without thourough test and investigation, we will have piles of bugs to fix 13:46:04 rotru, agree 13:46:13 alinefm, I think we should have a broad discussion on what are the basic rules. 13:46:17 agreed, though I also think we are capable of a higher velocity 13:46:18 royce: agree. 13:46:32 +1 13:46:32 zhoumeina, do you want to do it for 1.2? 13:46:50 Broad discussion on how do we prioritize the task items? 13:47:28 alinefm: I 'm not sure I can hold the back-end, but certainly Ok for UI 13:47:51 zhoumeina, maybe we can add it to sprint 4 13:47:54 zhoumeina: we can give you some help 13:47:58 yes 13:48:00 shaohef: thanks 13:48:24 during sprint most of china folks will be out for holidays so we should avoid add a lot of tasks there 13:48:33 alinefm: put it sprint 4 please 13:48:38 zhoumeina, sure 13:50:08 and waiting widget 13:50:13 zhoumeina, do I assign it to you? or leave it in blank? 13:50:26 alinefm: you can assign it to me 13:50:43 done 13:50:48 If I have some problem I will ask back-end guys for help 13:50:57 royce, you were assigned to backend stuffs on backlog 13:51:16 keep it on 1.2 plan? 13:51:32 zhoumeina; you can count on me 13:51:39 ok, alinefm, for cdrom in sprint 3 you can keep my name there 13:51:46 rotru: thanks 13:52:19 done 13:52:20 I think we should have an estimate on the sizing of the task after this meeting for the task assigned and re-visit the task if there is any issue. 13:52:23 zhoumeina, a lot of helpers! =) 13:52:36 :-D 13:52:42 great 13:53:06 any more updates? 13:53:16 What do you all think about holding a planning poker session? 13:53:49 texas hold'em ? 13:53:51 alinefm; I am going to have a lot o things of ginger to do .. but you can assign one or two tasks for me in the sprints ... 13:54:01 AdamKingIT. That is a great idea. That looks more like a community way. 13:54:08 rotru, there are some ones with your name 13:54:13 rotru, could you take a look? 13:54:41 alinefm; I just see 1 for sprint 1 13:54:42 I'll look into what we might be able to use to hold the session... I'll be out for the holiday's, but we can get started w/ sprint 1 in the meantime 13:56:18 sprint 1 starts today/tomorrow so we need to have agreement in these items 13:58:25 alinefm; I am ok with my item for sprint 1 13:58:33 well, seems everyone is comfortable with sprint 1 13:58:56 rotru: Do you need me to do the iSCSI and FC UI? 13:59:24 alinefm; I'd like to take 'Manage guest CDROM device', if no one has already taken it 13:59:44 rotru, royce asked for it 13:59:51 alinefm; thats ok 13:59:56 zhoumeina; I think so 14:00:09 rotru:ok 14:00:09 sure, you can take it rotru 14:00:34 1.2 plan for me is great, I'm little worried about our patches cannot receive enough review, we really need to say more NOs to others patches:) 14:00:41 zhshzhou: do you need a help for the iscsi support? 14:01:12 markwu, thanks. Firstly I need suggestions on the API refactoring. 14:01:18 yes 14:01:19 royce patch: yes, other patch: no ;-) 14:01:31 markwu, after the API is steady. I need help from front-end 14:01:34 zhshzhou: I mean the ui part 14:01:34 rotru,alinefm: we do need to also track assignments on the ginger pieces... 14:01:59 markwu, I read code implementing the logical pool UI. Not trivial. 14:02:13 heheheh, royce patch: Neither! AdamKingIT 14:02:20 zhoumeina: zhshzhou need your help on the iscsi storage pool 14:02:30 ok 14:02:35 no problem 14:02:38 rotru; we have done this already 14:02:43 zhoumeina: thanks 14:03:20 fnovak; we have done this already 14:03:36 royce; thanks! 14:03:37 markwu, shaohef, I just found a steady method to locate device node full path. 14:03:44 fnovak: it's already synced in alinefm's plan 14:04:03 zhshzhou: that's great 14:04:04 https://github.com/kimchi-project/kimchi/wiki/Storage-volume-attachment-and-detachment 14:04:21 Alinefm, I would like to add this to 1.2 plan. 14:04:25 firstly lsblk -P, it prints major and minor device number. 14:04:25 then "cat /sys/dev/block/major\:minor/uevent" 14:04:28 ming, it is there 14:04:29 sprint 4 14:04:45 Storage volume for VM is basic virtualization requirement. 14:04:45 zhshzhou: ok. you can update your patch with it or just share your idea with alinefm and ask her help to update it. 14:05:03 markwu, zhshzhou sure 14:05:08 shaohef markwu, In uevent there is "DEVNAME=XXX", then it should be /dev/XXX 14:05:12 you don't need to work overnight 14:05:30 rotru, markwu : ok, i saw parts,, just was worried when rotru said he had lots of ginger pieces, but coudl take some others.. That it meant there was another list somewhere! 14:05:52 Alinefm, thanks. But it is still in backlog. 14:06:09 fnovak: yes. it should be post to another code repo 14:06:12 zhshzhou: I will do Chinese translation this night. Any help you can ask me. 14:06:12 fnovak, specific ginger tasks aren't in kimchi plan 14:06:18 ming, really? 14:06:32 ming, please, remove it from backlog 14:06:48 fnovak; =] sorry, I should have not said "LOTS" 14:06:50 shaohef: thanks! 14:07:33 our time is over 14:07:40 but we can continue after meeting 14:07:54 ming, any other task you would like to add to 1.2 plan? 14:08:03 or anyone else ^? 14:08:11 alinefm, I am looking into the plan and will come back to you later after I have full sizing 14:08:25 it is an interactive work so we can continue talking about it 14:08:32 ming, great 14:08:57 unfortunately we don't have time to do open discussion today 14:09:05 #endmeeting