13:02:36 <alinefm> #startmeeting 13:02:36 <kimchi-bot> Meeting started Wed Oct 8 13:02:36 2014 UTC. The chair is alinefm. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:02:36 <kimchi-bot> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:02:36 <alinefm> #meetingname scrum 13:02:36 <kimchi-bot> The meeting name has been set to 'scrum' 13:02:56 <alinefm> #info Agenda 1) Status 2) Open discussion 13:02:56 <alinefm> anything else? 13:03:20 <royce> Good for me 13:05:10 <alinefm> #topic Status 13:05:10 <alinefm> #info Please provide your status using the #info command: #info <nickname> <status> 13:05:43 <alinefm> #info alinefm created a wiki page for the todo list for 1.4 release (https://github.com/kimchi-project/kimchi/wiki/todo-1.4) 13:06:14 <alinefm> #info info alinefm created a wiki page for 1.4 planning (https://github.com/kimchi-project/kimchi/wiki/planning-1.4) 13:06:53 <alinefm> #info alinefm sent patches to remove imported websockets code from kimchi and use the python-websockify package available in all supported distros (already merged) 13:06:59 <wenwang1> #info wenwang sent patch fix bug#414 Cannot login by clicking on the login button 13:07:20 <alinefm> #info alinefm sent patches to remove import novnc code from Kimchi and use the novnc package (on ML) 13:07:50 <alinefm> #info alinefm is working on spice code removal to depend on spice-html5 package 13:08:26 <vianac> #info vianac sent RFC for the guest cloning feature; sent final versions of patches "Display README with markdown style on Github" and "pep8: Use blacklist instead of whitelist"; reviewed patches on the mailing list; 13:08:57 <royce> #info royce has taken the whole week for holiday, no status from me this week, before holiday fix bug about change media, merged, and rebased remaining patches in ML 13:09:14 <YuXin> #info YuXin learn responsive design & investigate bootstrap grid system 13:11:03 <alinefm> royce, YuXin, wenwang1, feel free to get any unassigned task on https://github.com/kimchi-project/kimchi/wiki/todo-1.4 13:11:50 <YuXin> aline, can you send a mail with a summary about each line item 13:12:14 <alinefm> YuXin, I've already done it: http://lists.ovirt.org/pipermail/kimchi-devel/2014-September/008004.html 13:12:25 <wenwang1> alinefm: sure 13:12:45 <alinefm> there are 2 links to the ML archives in the top of the wiki page: one for the UI design and other from me about the tasks 13:13:05 <royce> I would like to take the guest disk hot plug and LDAP one 13:14:11 <alinefm> royce, sure! I have a upstream branch about LDAP - I will share it with you 13:14:43 <royce> alinefm, you have patch already? do you mind if I take that? 13:16:17 <YuXin> aline, what is the full name of SMT? 13:16:52 <alinefm> royce, I start playing with it some time ago but you can take it for sure 13:17:14 <alinefm> royce, if you want a start up: https://github.com/alinefm/kimchi/tree/alinefm/ldap =) 13:17:21 <alinefm> YuXin, hmm let me find it 13:17:29 <royce> thanks alinefm! 13:17:48 <royce> I think that task would be interesting 13:17:57 <christyp> yuxin - simultaneous multithreading 13:18:28 <YuXin> christyp, so just an option to enable/disable? 13:18:30 <christyp> but i think the LI name should be more generic, since intel calls theirs hyperthreading 13:18:55 <christyp> yuxin - there's also going to be options to chose (for power) which SMT mode you want (like, SMT2, SMT4, or SMT8) 13:19:12 <christyp> for x86, probably just enable 13:19:35 <YuXin> ok 13:20:04 <YuXin> Add support for serial console*, what is the serial console here means? 13:20:29 <alinefm> it is a text console 13:21:11 <alinefm> YuXin, I think baude has done some UI on his patches too 13:21:25 <royce> YuXin with that we do not need vnc to connect guest 13:21:50 <baude> yeah, esp for ppc, you want serial as an option 13:21:55 <baude> think latency 13:22:04 <baude> and also think about ability to copy and paste into the console 13:22:08 <baude> which spice and vnc cannot! 13:22:46 <YuXin> so just a command line tool to control vm? 13:22:46 <YuXin> ok 13:22:46 <YuXin> royce, remote command line, right? 13:22:52 <royce> we use this text console just like you ssh to a guest, but this does not need network 13:24:15 <royce> YuXin , you can understand that way: 13:26:01 <royce> hejie and zhengsheng seems have patches in vdsm before about this, baude if you would like, you can take a look for some reference:) 13:26:09 <YuXin> ok 13:26:09 <YuXin> UI to allow user adds volumes from different pools to a Template 13:26:09 <YuXin> template should be configured against storage pool, not volume per my understanding 13:26:09 <YuXin> so what does this line item mean 13:26:49 <alinefm> today the user can only specify one volume size to be created in a pool 13:27:19 <alinefm> we should support user adds multiple disks to a template, ie, multiple disk sizes to different pools 13:27:42 <alinefm> example: I want a template with 2 disks, one 10G on default pool and other 5G on NFS pool 13:27:44 <baude> royce, i have submitted working patches several times to the list 13:28:03 <alinefm> baude, you still need to resend it according to review comments 13:28:25 <baude> i sent a fixed patch to you alinefm asking questions ... which I am waiting on answers for 13:28:26 <royce> baude, that's great, haven't opened my mailbox these days due to holiday 13:28:35 <baude> royce, this was months ago 13:28:51 <royce> o_O 13:29:11 <royce> sorry I haven't noticed, I will take a look on that 13:30:28 <alinefm> well, let's move on to open discussion session 13:30:33 <alinefm> #topic Open Discussion 13:30:34 <baude> i can just redo the patch and resend if you guys prefer 13:30:40 <alinefm> what are the topics for today? 13:30:50 <royce> alinefm, will it be good to assign disk when creating vm and just add one root disk to template? Because if we want to have 3 disks, one from pool A, the other two from pool B, things will become complicated 13:31:35 <YuXin1> why can not have 2 diskes from the same pool? 13:31:36 <alinefm> royce, why? 13:31:55 <alinefm> YuXin, you could have too 13:33:04 <alinefm> royce, I think in {disks: {size: ..., pool: ..., }, {size:..., pool:...}} 13:34:24 <royce> Specify one by one seems a burden for me I think... 13:35:25 <christyp> some customers might want to have a lesser accessed disk on a slower/cheaper type of storage 13:35:30 <YuXin> so in template, it should be 13:35:31 <YuXin> disk1 10G Pool-A 13:35:31 <YuXin> disk2 10G Pool-A 13:35:31 <YuXin> disk3 10G Pool-B 13:35:31 <YuXin> disk4 10G Pool-B 13:36:33 <YuXin> always configured against a pool with a disk size, not a volume 13:36:43 <alinefm> royce, but user will do that just once as it is on template 13:36:54 <royce> christyp, got u, but I think that is from a single guest level, not from template level 13:37:14 <christyp> royce: i guess i could see it going both ways 13:37:14 <alinefm> YuXin, for iSCSI and SCSI instead of the size we will have the volume name as those pools are read only 13:37:28 <alinefm> and in those cases the disk should be sharaeble 13:37:31 <royce> because template is for most people, and a single guest is to customize this template of his own needs 13:38:02 <YuXin> iSCSI and SCSI is read-only for create volume, but volumes in those pool are not read-only 13:38:10 <alinefm> YuXin, correct 13:38:39 <YuXin> so a template is created with iscsi volume, then multiple vms can be created from this template 13:38:45 <alinefm> royce, let's say I want to all my VMs have access to a NFS pool as it contains some data needed by them 13:38:48 <YuXin> then multiple vms share the same volume 13:38:52 <alinefm> should I add it vm by vm? 13:39:09 <YuXin> I think so 13:39:21 <YuXin> when iscsi, only specify a pool without size 13:39:35 <YuXin> when create vm from that template, list volumes for selection 13:40:34 <alinefm> YuXin, it is also an idea 13:40:43 <alinefm> but we should do it prior to any other change on template 13:40:48 <alinefm> as it is independent 13:41:00 <alinefm> YuXin, if you want I can add a task for it 13:41:11 <YuXin> ok 13:41:36 <royce> alinefm, your case makes sense , for iscsi and scsi pool I think add disk is better 13:41:51 <alinefm> royce, do you mean what YuXin proposed? 13:42:08 <royce> yes, I mean we need both cases 13:42:17 <YuXin> royce, do you mean add disk to template or vm? 13:42:36 <royce> the scsi and iscsi volumes are disk level so need to add to vm 13:42:44 <alinefm> royce, YuXin, I don't want to insert a lot of information while creating the VM 13:42:51 <alinefm> it must be as simple as possible 13:42:53 <royce> and size and pool seems a template level thing, right? 13:43:33 <YuXin> aline, we can try to simplify it, but current design has problem 13:44:26 <alinefm> royce, yes - I think we 3 are saying the same 13:44:56 <YuXin> for example, if a iscsi pool, when create vm, can we automatically pick up a volume? 13:45:31 <royce> currently not, I suppose we'd better to change to this way 13:45:33 <alinefm> based on each criteria? not in use yet? 13:45:42 <YuXin> yes 13:45:49 <YuXin> a volume not in use 13:46:49 <alinefm> what about the disk size? 13:46:59 <alinefm> maybe we could continue asking the disk size for the user 13:47:28 <YuXin> each lun in an iscsi pool has a fixed size per my understanding, can that be changed? 13:47:38 <royce> nope 13:48:15 <alinefm> my point is, if we pick a volume automatically we can pick a 1G volume which is not enough to create a VM 13:48:22 <alinefm> ie, an easy point of failure 13:48:53 <royce> alinefm, previously markwu has proposed to form all volumes in scsi/iscsi to a VG then use it, the volume size can be dynamic, do we want to consider this usage? 13:49:36 <alinefm> royce, create a LVM with iSCSI/SCSI volumes without the user be aware? 13:49:45 <alinefm> if those volumes are used for other proposals too? 13:51:07 <royce> There are cases we have original data in the volumes...But we loose elasticity 13:51:25 <alinefm> royce, YuXin, if you think we need to move the volume selection to the VM, I am OK if we list the non-used volumes to user to he/she picks one 13:51:40 <YuXin> ok 13:52:09 <YuXin> add a task for tat 13:52:12 <alinefm> because automatically select one is very dangerous from my view 13:52:21 <YuXin> agree 13:52:28 <royce> OK 13:53:44 <YuXin> go back to "UI to allow user adds volumes from different pools to a Template" 13:53:59 <YuXin> so this line item also need to be changed 13:54:04 <alinefm> sounds good: "Move iSCSI/SCSI volume selection to VM creation instead of Template view" 13:54:05 <alinefm> ? 13:54:21 <royce> I think that still need to remain, YuXin 13:54:27 <YuXin> why 13:54:40 <YuXin> add volume to template, does not make sense as we discussed 13:55:02 <alinefm> YuXin, maybe we need to change the statement 13:55:15 <royce> :) disks from different pools 13:55:18 <alinefm> but it is to add multiple "disk size/pool" 13:55:18 <YuXin> so it just means to specify a size, right? 13:55:27 <YuXin> ok, got it 13:55:56 <alinefm> YuXin, if I change "volumes" to "disks" is better to understand? 13:56:05 <YuXin> ok 13:56:29 <alinefm> YuXin, royce, look: https://github.com/kimchi-project/kimchi/wiki/Todo-1.4 13:56:51 <alinefm> I added the iSCSI/SCSI task and changed 'volumes' to 'disks' on "UI to allow user adds volumes from different pools to a Template" 13:57:43 <royce> Thanks for the patient explanation :)))alinefm 13:57:54 <YuXin> I will take a further look tomorrow and take over some line items 13:58:08 <alinefm> royce, it is my pleasure 13:58:24 <alinefm> I missed you all on last scrum meeting 13:58:29 <alinefm> YuXin, thanks 13:58:39 <alinefm> royce, YuXin, wenwang1, and don't forget the RFC's 13:58:42 <alinefm> or mockups 13:58:48 <YuXin> sure 13:58:51 <royce> Yes! We too! 13:59:00 <alinefm> just to we get agreement before implementation 13:59:25 <wenwang1> sure alinefm 13:59:42 <alinefm> any other topic for today? 14:00:06 <alinefm> oh I forgot to mention to you about the planning 14:01:04 <alinefm> instead of having 2 devel sprints and 2 stabilization sprints, we will have one single devel phase until 11/19 14:01:20 <royce> ok 14:01:23 <alinefm> from 11/19 to 12/10 it is code freeze - only bug fixes will be merged 14:01:57 <alinefm> and the final week (from 12/10 to 12/17) it is hard code freeze which means only blocks bug fixes will be merged 14:02:06 <alinefm> and the announcement on 12/19 14:02:13 <alinefm> makes sense for you? 14:02:19 <YuXin> ok 14:02:23 <wenwang1> ok 14:02:38 <royce> Good plan! we do not want to submit patch on dead lines 14:02:56 <alinefm> royce, yeap! me too =) 14:03:02 <royce> quality can't gaurentee 14:04:07 <alinefm> anything else? 14:05:16 <alinefm> any comments? doubts? suggestions? 14:06:11 <royce> not any more:) 14:07:42 <alinefm> well, any time you can ping me here or send email do ML 14:08:02 <alinefm> thanks everyone for joining 14:08:32 <alinefm> #endmeeting