14:01:34 #startmeeting oVirt Weekly Sync 14:01:34 Meeting started Wed May 7 14:01:34 2014 UTC. The chair is doron. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:34 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:01:36 #topic Agenda and roll Call 14:01:37 #info infra update 14:01:39 #info 3.4.z updates 14:01:40 #info 3.5 status 14:01:42 #info conferences and workshops 14:01:43 #info other topics 14:02:02 let's start with infra updates 14:02:04 #topic infra update 14:02:20 dcaro: care to update us on recent updates? 14:02:30 * lvernia is here. 14:02:41 * amureini here 14:02:53 lvernia: amureini hi ;) 14:02:59 Yo :) 14:03:03 * sbonazzo here 14:03:31 most of the rpm jobs have been migrated to run per-patch, that will give us grained knowledge on when the build broke and feedbac on gerrit 14:04:09 Some slaves have been added to support the extra load generated, and we already had some issues with parallel runs, will be solved asap 14:04:41 dcaro: cool 14:05:10 the jobs have been slightly modified to accept GERRIT_BRANCH and GERRIT_REFSPEC parameters, so we can manually test a specific patch from a specific branch if needed (did that with engine 3.4.1) 14:05:18 #info most of the rpm jobs have been migrated to run per-patch. Several slaves added to support the new load. 14:05:45 * jb_netapp here 14:06:21 #info jobs are now capapable of running per branch and GERRIT_REFSPEC . 14:06:34 dcaro: thanks for the updates! 14:06:35 dcaro - any eta on the phx additional slaves? 14:06:39 Also we have started creating manual rpm creation jobs based on mock on the projects that have specific build dependencies, if you have dependency problems with manual_tarball and it's not finding some deps, send infra an email requesting a mock job for the project 14:06:52 * danken is here 14:07:28 itamar: I'm not aware of any news (no news emails), maybe eyal has more info, but he's not around 14:08:20 dcaro: thanks. 14:08:43 let's check 3.4.z status- 14:08:45 #topic 3.4.z updates 14:08:55 sbonazzo: how are we doing? 14:09:25 doron: 3.4.1 packages have been all built except for vdsm which had issues and is building right now 14:09:49 doron: we have 1 blocker left over on missing vdsm tarball which will be fixed once we have it built 14:10:04 sbonazzo: are we expecting to release it today? 14:10:26 * fabiand is here 14:10:28 doron: I think we can have it ready, but we're a bit late in updating release notes 14:10:40 doron so maybe better to postpone to tomorrow 14:11:10 sbonazzo: very well, so 3.4.1 ga tomorrows. 14:11:15 doron: ok 14:11:30 #info 3.4.1 GA expected tomorrow. 14:11:39 sbonazzo: anything else around 3.4.z? 14:12:31 doron: we've not planned it yet but I guess RC in 3 weeks, GA in 1 month 14:12:46 of 3.4.2? 14:13:34 doron: yes. bugs left from 3.4.1 will be retargeted to 3.5.0 14:13:46 doron: only critical fixes will be automatically targeted to 3.4.2 14:14:01 #info 3.4.2 planning TBD. Draft is: RC in 3 weeks, GA in 1 month. 14:14:08 k 14:14:10 let's move onm 14:14:15 sbonazzo, doron how do you identify a critical fix? should we mark bz items somehow? 14:14:50 amureini: yes- there's a blocker tracker. 14:14:54 amureini: doron: idea is to use severity/priority urgent as identification 14:15:03 #topic 3.5 status 14:15:09 amureini: and blocker of course 14:15:29 doron, sbonazzo can you provide a link to the blocker tracker? 14:15:58 amureini: not created yet, I'll send an email to devel@ovirt.org with timeline and tracker 14:16:03 doron: ^^ 14:16:10 sbonazzo, doron thanks. 14:16:27 #action: sbonazzo to end an email to devel@ovirt.org with 3.4.2 timeline and tracker 14:16:35 back to 3.5 14:16:47 who's here for gluster? 14:17:33 will get back to gluster. 14:18:02 ybronhei: here for infra? 14:19:15 ybronhei: here? 14:19:52 k, let's move to the next one... 14:20:04 sbonazzo: integration status for 3.5? 14:21:24 doron: regarding infra: report selinux policy - acked today on vdsm side 14:21:27 doron: iscsi support on hosted engine has been verified and works except for an issue we found using lvs inside vdsm groups 14:21:35 but plenty of jsonrpc changes are still pending 14:21:51 sbonazzo: inside vdsm groups? 14:21:55 danken: thanks for updating, trying to get a hold on an infra rep. 14:21:56 doron jmoskovc is working with vdsm team on that 14:22:06 itamar: vgs 14:22:34 itamar: there's a thread on devel mailing list about that 14:22:52 sbonazzo: ok, thanks. 14:22:56 sbonazzo: I'm aware of it and it's not trivial. saw danken's response. 14:23:17 sbonazzo: "ovirt-guest-tools-iso similar to rhev-guest-tools-iso and uploaded to ISO domain by default" - status/eta (currently its "design") ? 14:23:54 itamar: yes, still in design, lev is still working on it 14:24:10 doron: here 14:24:12 sbonazzo: ok, well, low risk to anything else i guess, so can keep at it. 14:24:21 DWH running on separate host? 14:24:21 reports running on separate host? 14:24:22 doron: went for few minutes :/ sorry 14:24:25 eta? 14:24:36 ybronhei: will ping you later on, please stay around. 14:24:55 itamar: both of them and also websocket proxy on separate host have patches under verification 14:25:18 does "under verficiation" means gerrit verify, or merged and being tested? 14:25:34 itamar: we've suggestions for doing certificatoin enrollment in different ways than we have in those patches 14:26:14 and "guest agents repository (debian, suse, ubuntu, etc.)"? 14:26:20 sbonazzo: this is only for websocket iiuc, not for the others. 14:26:29 itamar: means "verified by developer" and near to be merged but we have comments that need decision to be taken on how to do some details 14:26:51 i see 33 and 37 in "green", which means "done" - as in "merged" ? 14:26:58 doron: certification enrollment may be needed also on reports 14:27:00 are they both merged? 14:27:08 sorry, 33 and 36 14:28:03 * sbonazzo looking at google docs 14:28:59 itamar: not merged yet but have verification flag and +1 flags 14:29:12 sbonazzo: then shouldn't be green... 14:29:15 itamar: should be merged before 3.5.0 alpha this week 14:29:24 itamar: yellow? 14:29:48 nothing... 14:30:17 itamar: ok 14:31:10 these two may get removed from tracking i guess: 14:31:10 ovirt-guest-tools-iso similar to rhev-guest-tools-iso and uploaded to ISO domain by default 14:31:13 guest agents repository (debian, suse, ubuntu, etc.) 14:31:18 let's see at end of call. 14:31:31 itamar: ok 14:33:17 itamar: want to move to next team? 14:33:30 doron: that should be all for integration on 3.5. We're going to release 3.5.0 alpha on friday 14:33:43 doron: yes, lets go back to infra. 14:33:51 sbonazzo: thanks. 14:34:01 sbonazzo: so for 3.5. alpha there is no integration item which is ready to be in that build? 14:34:06 ybronhei: here? 14:34:16 yes 14:34:39 :) 14:34:39 itamar: we're trying to get iscsi support and websocket on separate host in 14:34:51 ybronhei: infra 3.5 updates? 14:35:11 ok, the table is updated , some features on progress, some already in 14:35:51 ybronhei: features missed FF, 14:35:55 is there an ETA? 14:36:02 currently we plan to have all that stated for 3.5 14:36:19 about ETA I don't know much, will be cleared more next week 14:36:46 ybronhei: we are already a week after the original date. so we need to start focusing. 14:37:04 for example, validate selinux policy at cluster level is 'postponed', yet danken just mentioned there is agreemnt on it? 14:37:27 well, I just merged it on vdsm side 14:37:30 yes i know, the features are in good shape. not merged already, but i assume most of them will be during this week 14:37:44 I did not notice it was declared as postponed. 14:37:51 can we remove all the postponed items for tracking or not? 14:37:55 it didn't afaik 14:38:09 wait, ill just verify it 14:39:00 any of the non green items planned for the alpha build, planned for friday? 14:39:04 yes, you can 14:39:25 no.. 14:39:53 yes or no? 14:40:10 we can remove the postponed , won't be ready this week 14:40:27 ok. 14:40:58 ETA for AAA till the end of may, the rest during the next 2 weeks 14:42:02 thanks. 14:43:00 let's move on. 14:43:18 network is next. lvernia your updates? 14:43:23 doron: Yes. 14:43:28 doron: So, things aren't looking great. 14:43:47 bridge_opts isn't in yet because of the dependency on the init_scripts bug. 14:44:21 lvernia: is there a reasonable eta for it, or should we push it forward? 14:44:25 The Neutron appliance has some unforeseen complications, due to usage of RabbitMQ and not necessarily QPID. 14:44:38 doron: lvernia: the initscripts bug is on QA 14:44:46 and there is a build for it in brew 14:45:00 apuimedo: Sounds good, does it imply a time frame for its release? 14:45:13 lvernia: I've been putting pressure on qa 14:45:41 I have set my eyes on next week, but I can just pressure :P 14:45:53 lvernia: line 43 - what's 'mostly done' 14:45:53 I know what "mostly dead" is ;) 14:45:57 doron: As for ethtool_opts, I'm hoping we can get the REST support in as well by tomorrow and color it green. 14:46:11 itamar: One GUI patch missing, which is nice to have but definitely not detrimental. 14:46:21 itamar: I think it can be merged as a bug fix. 14:46:56 so bridging_opts is pending on an init_script version to be released? 14:47:01 for fedora? rhel? 14:47:16 itamar: el6 14:47:39 it is a workaround for presumably a kernel shortcomming that is not present on el7 and fedora 14:47:44 apuimedo: and vdsm code is merged or can't be merged until its fixed? 14:48:08 itamar: the patch includes a vdsm specfile requirement to the brew package 14:48:22 and we are delaying merging it so that it does not break building 14:48:48 apuimedo: for .el6, but you could merge it and release note not to use it on .el6 yet? (without the dependnecy) 14:49:20 itamar: we certainly could do that if danken agrees to it 14:49:27 itamar: we could, but we'd need to disable the unit test 14:49:38 danken: the functional test 14:49:44 danken: on .el6. do we have it running on fedora and .el7? 14:49:51 (hmmm, did we setup an .el7 host yet?) 14:50:33 itamar: I am not aware of an el7 host running functional tests 14:50:46 but this one would pass there 14:51:58 doron: the mtg is scheduled to end in 7 minutes. are we going to go through all the or will we stop there? 14:52:04 danken: well, do we have an ETA for the initscripts (weeks/months?) 14:52:29 mskrivanek: I'd like to try and finish as much as possible. 14:54:00 itamar: we're nagging chrispy for this 14:54:09 * fabiand can give a short update on Node features for 3.5: Look good so far, still uncertainty about generic registration, virtual appliance is a bit behind schedule 14:54:12 danken: ok, i guess we can revisit next week. 14:54:13 doron, ^^ 14:54:16 shouldn't be more than 1 week or two 14:54:35 fabiand: generic registery is noted as 'postponed'? 14:54:37 fabiand: got it, is there an eta for appliance? 14:54:50 fabiand: ETA for hosted engine? 14:55:13 itamar, hoste dengine depends a bit on the appliance for testing. Hopefully we can start testing next week 14:55:24 doron, Hopefully we cna get the first appliance this week 14:55:49 itamar, on the registration - IIRC we did not want to postpone it 14:56:05 so why does it say 'postponed' ? 14:56:38 * fabiand checks 14:58:08 derez, thanks for those bugs :) 14:58:21 * mskrivanek can't stay another hour, will reconnect in ~20 minutes, I doubt we'll be finished by then;) 14:58:41 mskrivanek: I'll ping you last. 14:59:01 doron, sorry, I have to drop (additional meetings) - will attempt to reconnect later, but can't promise anything 14:59:17 amureini_away: you;re up next... 15:01:34 until we hear from fabiand, I'll provide the sla status; 15:02:05 we have 2 features postponed. there's 1 which will probably slip. 15:02:07 doron, for numa I'd like someone to review http://gerrit.ovirt.org/#/q/status:open+project:vdsm+branch:master+topic:numa-support,n,z 15:02:32 danken: it's there for review. 15:02:43 doron: fabiand already posted the status above? 15:02:46 itamar, doron - I don't know why it was set as postponed. rbarray is working on it and we target it for 3.5. 15:02:49 oh, on our question. 15:02:50 rbarry, 15:03:04 fabiand: eta? 15:03:23 doron: yes, but someone should join me in reviewing it 15:03:31 msivak: fromani: could you help there? 15:03:34 I guess at least 2 weeks, itamar 15:03:40 danken: reading 15:03:48 doron - no SLA item in the alpha build? 15:03:51 danken: msivak is away and will be bak next week where he will be glad to help. 15:04:03 itamar: true. 15:04:17 I'll be happy to help as well 15:04:18 itamar, If that's to late - and that could be imo because it's the vein between - Node and Engine - we should probably postpone it 15:04:52 itamar: Backend for registration is about 50% done. I'd expect to have it posted by next Tues/Wed, but then I'm off for a couple of days, and testing may be difficult 15:04:53 fabiand: we can keep it for now. so many others are late by other folks... 15:05:00 Ack 15:05:35 doron: isn't next week too late? 15:05:45 danken: for what? 15:06:04 doron: numa is a 3.5 feature, isn't it? 15:06:10 or is it posponed, too? 15:06:20 danken: it is a 3.5 feature. 15:07:06 danken: and msivak reviewed most in april 15:08:00 doron: still, there are pending patches with API changes 15:08:23 danken: next week isn't too late at this point. we are pre-alpha. we are allowing features to converge this time till end of the month. 15:09:00 so just to conclude the sla 3.5, everything not postponed / slipped should be in during next week or the week after. 15:09:37 since storage is away, 15:09:45 is ecohen here? 15:10:05 she is, but awels should represent UX. 15:10:17 awels: 3.5 updates? 15:10:46 doron: Oops, sorry about that. yes, we merged the updated look and feel patch on Monday 15:10:56 awels: cool ;) 15:11:02 anything else relevant for3.5? 15:11:02 awels: one (and only) ux item is in :) 15:11:08 a big one... 15:11:20 good timing to get feedback on it from alpha build. 15:11:22 doron: yes really big one, that was a pain 15:11:40 good job ux team. 15:11:54 doron: REST api is still out like before. 15:12:10 awels: yes, it's not tracked any more for 3.5. thanks. 15:12:27 ydary: care to update on dwh for 3.5? 15:13:14 fromani: I'd love to see your review of http://gerrit.ovirt.org/#/c/26876/4/vdsm/virt/sampling.py 15:13:27 danken: sure thing, added to my TODO 15:13:29 I guess not. 15:14:01 itamar: since virt storage and others are missing, I guess we should ask for offline update with eta to the devel list? 15:14:52 I'll take it as a yes.... 15:15:00 doron: actually, 15:15:09 ? 15:15:18 doron: I think worth an interactive discussion - will try to ping them later for more data. 15:15:45 very well. 15:16:29 in the meantime let's move on; 15:16:31 #topic conferences and workshops 15:16:58 Offline updates from bkp; 15:17:20 #info Rene Koch will be presenting another workshop in Vienna this weekend. bkp gave a talk at the CentOS dojo on Monday night. Light attendance, but lots of VMware people who were eager to hear more. 15:18:03 #info bkp is giving three talks this week at FISL 15 in Porto Alegre, Brazil. 15:18:18 doron: sec about 3.5 15:18:26 and finally, If the local point person for OpenStack Israel wants to coordinate with bkp for the show, reach out to me by email. 15:18:29 what about live merge? 15:18:29 danken: ? 15:18:41 it's not there yet 15:18:54 danken: live merge (delete snapshot) coding 15:19:02 in any meaningful definition of feature freeze 15:19:03 danken: waiting for the storage and virt updates. 15:19:08 danken: we did not review storage today. 15:19:26 untill then, 15:19:28 #topic other topics 15:19:33 itamar: doron: I'm trying to understand our state of afairs 15:19:47 it's not clear to me when is real feature freeze. 15:19:48 danken: in which area? 15:20:10 doron: When should I reject changes to API? 15:20:10 danken: branching day. 15:20:25 which is when? 15:20:30 danken: end of month. 15:20:30 danken: sec 15:20:34 iirc 15:20:57 itamar: true http://www.ovirt.org/OVirt_3.5_release-management 15:21:06 Branching: 2014-05-30 15:21:13 danken: but i'm concerned with how many of the items are not ready at alpha. need to think a bit about this. 15:21:29 itamar: need to think on alpha. 15:21:50 ie- should we push alpha a week later? 15:21:58 " Feature freeze: 2014-05-01" is meaningless, at least to me. 15:22:14 doron: there are enough items 'in' for alpha i think. (even just the gui merits it...). but not sure a broad test day. 15:22:25 danken: yes, we need to revisit the terms/dates. 15:22:29 danken: I agree, we should enforce it next time. 15:22:44 doron: or postpone... 15:22:52 itamar: postpone what? 15:22:59 doron: dates of the milestones. 15:23:14 itamar: hence push alpha. 15:23:25 doron: but we do need the milestone of "before feature freeze" in which we start to converge. 15:23:47 itamar: yes, but it should be something else which is not FF. 15:23:57 ie- FF should be meaningful. 15:23:59 doron: i didn't get a feeling so many more items would go in with the extra week, but i'm not against pushing it out a week. 15:25:09 for me, "feature freeze" means "APIs are defined and committed, though implementaiton may be buggy". this is not the case today. 15:25:29 anything else is a misnomer. 15:26:08 danken: I agree, but not for time, as it was delivered in this way to the tems. 15:26:15 (teams) 15:26:45 danken: I'd make FF a hard limit with minimal exceptions (final review, etc). 15:26:50 * mskrivanek back 15:27:01 mskrivanek: care to share virt status? 15:27:04 danken: that would be api freeze, not feature freeze (which should be feature complete in my view...) 15:27:10 the sheet is updated now 15:27:31 things worth waiting for are edit running vms, 15:27:35 mskrivanek: line 62 - green but not done? 15:27:37 virtio-rng 15:27:51 it is done 15:27:56 mskrivanek: maybe partially(vdsm only)? 15:27:56 no comment means "done" too:) 15:28:07 itamar: about naming, you can take a look here: https://fedoraproject.org/wiki/Releases/21/Schedule 15:28:10 mskrivanek: please edit to done... 15:28:28 yes, as Dan said, API-wise I still hope it can make it by next week, but the rest won't make it. Still it would be a big step 15:28:29 ie- if the feature is in, it's in. or not. 15:28:38 itamar: we're not in any way "feature complete" either... 15:28:46 doron: it's never as simple as that:) 15:29:03 mskrivanek: why define it asa feature if it's not in? 15:29:07 sometimes we implement a slightly different thing... 15:29:08 mskrivanek: so why is it green? 15:29:16 ie- it could be infra work we do not track. 15:29:44 itamar: which one? 15:29:54 mskrivanek: ok, i think we were talking two different lines... 15:30:16 mskrivanek: quesiton on line 63 - v2v work - any feature page / something about this one to read? 15:30:54 itamar: still in proof of concept stage. Arik made some progress and is putting up a plan how to interact with it 15:31:12 we also need to consider the future v2v replacement 15:31:25 mskrivanek: with which one? the old perl virt-v2v? the guestconv, or the new C-based virt-v2v? 15:32:00 doron: I do not know about "teams" but I can surely attest about myself: the requirements regarding calendar were not clear to me. not even now. 15:32:38 danken: api will shut down on branching day. 15:32:48 itamar: I *think* it's the perl one for now, but i may be wrong. 15:33:07 danken: as well as any new additions. It'll be mostly final bits being ported from master. 15:33:17 mskrivanek: ok. i assume line 65 i can remove? 15:33:53 itamar: yes, basically all the white lines 15:34:20 how is 65 different than 69? 15:34:56 except 66 which even in its partiality is a foundation for line 79 15:35:17 65 is engine-side only, 69 is the data sent between vdsm and engine 15:35:59 line 85 has been repeatedly asked for - how far is it? 15:36:42 mskrivanek: --^ 15:36:51 vdsm patches are posted for "forced" reboot and for engine-controlled reboot. Which should partially take care of the feature. however even that is at risk for 3.5. 15:37:54 the exact request in 85 may not be technicaly feasible so we want to get at least something there and see if it is enough 15:38:23 there are other enhancements which may address the actual need - like boot menu support 15:38:46 mskrivanek: ok. I'm leaving only this white line for now to discuss it some more. 15:39:01 itamar: and 66 please 15:39:18 ok 15:39:49 mskrivanek: though i think 71 is even smaller... 15:40:04 also as for 74, the core instance types functionality is finished, but there are things to still work on (create instance permission, enable/disable instance type, etc) 15:40:13 ok. thanks. 15:40:16 ydary? 15:40:18 71 is a small nit, just didn't get to it and it's too late to start since we're in May 15:40:25 amureini_away? 15:43:07 itamar: I suggest amureini_away will update offline. 15:43:34 doron: I'll try to ping him later... 15:43:48 k, so let's close this session. 15:44:06 thanks everyone for showing up. We'll probably update the schedule. 15:44:16 #endmeeting