14:00:18 #startmeeting oVirt Weekly Meeting 14:00:18 Meeting started Wed May 22 14:00:18 2013 UTC. The chair is mburns. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:18 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:00:24 #topic agenda and roll call 14:00:36 * dustins_ntap here 14:00:44 #info oVirt 3.2 update 14:00:48 #info oVirt 3.3 status 14:01:00 * danken here 14:01:00 #info conferences and workshops 14:01:04 * lvernia here 14:01:23 #info sub-project updates 14:01:26 #info other topics 14:02:14 * ofri here 14:02:23 * jb_netapp is here 14:03:30 * oschreib here 14:03:37 * mgoldboi here 14:03:39 #topic oVirt 3.2 update 14:03:55 oschreib: mgoldboi: where are we at with the 3.2 update? is it ready? 14:04:14 we uploaded RPMs two weeks ago 14:04:20 still sitting in updates-testing 14:04:20 well, we just found a nasty vdsm bug yesterday 14:04:57 danken: backport and update? 14:05:06 Unless you saw any nasty complain, I think it's safe to release. 14:05:08 I've posted a baboon-quality backport of the fix: http://gerrit.ovirt.org/#/c/14953/ 14:05:11 danken- is it a new bug? 14:05:13 but we need much more than that 14:05:34 oschreib: unfortunately, the bug is old. we only KNOW about it recently. 14:06:10 we do not change lv availability when VM is down. 14:06:20 danken: by saying much more than it- ETA? 14:06:22 hi all 14:06:47 danken: so releasing the vdsm that is in testing (4.10.3-16) won't introduce this bug 14:06:51 danken, mgoldboi: Is there a list of 3.2 bugs? 14:07:20 nope. it's not a recent regression 14:07:26 we should not block anything on this 14:07:40 ok 14:07:48 but it is important that everybody knows that a vdsm respin is pending. 14:07:58 dneary: not that i'm aware of 14:08:14 danken: then i think, if there are no other regressions known in the pending 3.2 update, the we should push it live 14:08:17 mgoldboi, We don't have a target in Bugzilla that covers all the projects? 14:08:32 in the announcement, we'll note this bug and say another vdsm will be coming soon 14:08:34 fsimonce: maybe you could help me with my backport? I do not want to foul my mouth again against myself. 14:08:45 (hi there, by the way) 14:09:16 mgoldboi: oschreib: any issues found with the 3.2 engine? 14:09:21 3.2.2 engine 14:09:44 mburns: nothing i'm aware of 14:09:49 =1 14:09:50 danken, mmm yes, I'm just worried to forget... will you ping me tomorrow too about it? (or do you need it today?) 14:09:50 +1 14:10:40 #info new vdsm issue found in 3.2, but not new in pending update 14:11:03 #info we should ship 3.2.2 and follow up with new vdsm after 14:11:09 fsimonce: no, it can wait. 14:11:17 #info engine 3.2.2 -- no issues found 14:11:34 #action mburns to push 3.2.2 packages out to 3.2.2 14:11:44 dneary: right now bugzilla is down, but i don't think releases/versioning is maintained properly on ovirt project - i'll have another look once it's up again 14:11:47 #undo 14:11:47 Removing item from minutes: 14:11:49 #action mburns to push 3.2.2 packages out to stable 14:12:36 mburns: do you have all packages update? are we missing anything 14:12:37 ? 14:12:57 mgoldboi: i have otopi, engine, ovirt-host-deploy and vdsm 14:13:08 both f18 and el6 for each 14:13:25 mburns: so cli/sdk are missing - i think michael built it 14:13:36 mburns: action item on me 14:13:38 mgoldboi: i'll hold off ovirt-node until we get the next vdsm 14:14:44 #action mgoldboi to get sdk/cli rpms to mburns for posting 14:15:06 mgoldboi: yes, i see an email from mpastern with a location internally 14:16:34 anything else for 3.2.2? 14:16:47 anything else for 3.2.x as well? 14:17:00 * mburns hoping that we can move focus completely to 3.3 after this 14:17:09 other than the vdsm and node respins 14:17:13 +100 :( 14:17:15 :) 14:17:16 mburns: when are we targeting this release 14:17:25 the 3.2.2? 14:17:29 mburns: yes 14:17:31 mgoldboi, It would really help for tracking whether we're on target for a 3.3 release, or when we need to make a 3.2 release, if we had the targets maintained 14:17:40 mgoldboi: i'll move it around later today 14:18:20 mgoldboi: sdk/cli can go later if they're not quite ready 14:18:27 or if you think they need soak time in testing 14:18:44 dneary: mgoldboi: do we have a bugzilla admin anymore? 14:18:55 mburns, Did we ever have one? 14:18:58 ykaul did most of the bz work before 14:19:08 but he's not here anymore... 14:19:10 mburns, I mean, in the oVirt project... 14:19:23 dneary: not officially 14:19:27 mburns: i think we should be ok, did we release announcement of our test repo? 14:19:45 mgoldboi: i announced the test repo 14:20:00 mburns: great - so we should be ok to release once we have vdsm 14:20:01 at least, i think i did 14:20:11 mgoldboi: vdsm isn't a blocker for 3.2.2 14:20:14 the bug is old 14:20:25 so releasing the new vdsm isn't a regression 14:20:25 mburns: the bug is curroption 14:20:40 mburns: (can lead to..) 14:20:50 mgoldboi: ok, but the current vdsm has the same bug, right? 14:21:01 mgoldboi: still,3.2.2 should go on. 3.2.0 already has it. 14:21:03 mburns: that's true - not a regression 14:21:09 mburns: ok 14:21:10 so we're not worse off 14:21:28 we should ship the one in testing now, then do async as soon as the bug is fixed 14:21:48 mburns: +1 14:21:50 dneary: i saw your mail earlier this week - what do you need happening from our side? (if we are ok to move to 3.3) 14:23:42 mburns: dneary: i'll find out whether we have bugzilla admin around (not that i'm aware of) 14:23:59 mgoldboi: ok, thanks 14:24:16 we need someone who can do admin operations on the ovirt project 14:24:24 so we can appropriately target bugs 14:24:37 and add components, etc, as needed 14:24:46 why do we need anything more than a tracker bug? 14:25:01 mburns: ah, components. 14:25:25 danken: a tracker does handle a lot of it 14:25:31 mgoldboi, ? 14:25:35 but adding target releases can be useful too 14:25:58 i think of a tracker as a list of blocking issues 14:25:58 mgoldboi, I don't understand your question. Regarding features & release planning? 14:26:15 dneary: regarding 3.3 features (should, must) do you think each feature should have it 14:26:24 whereas you could mark any bug as targeted to be fixed in a specific release 14:26:35 mburns: +1 much needed 14:26:36 mgoldboi, I'm aware of at least one feature that's listed there which will be bumped 14:26:52 ok, since you're already discussing 3.3 -- 14:26:57 dneary: me too 14:26:58 #topic oVirt 3.3 status 14:27:01 mgoldboi, So I think every feature should say what stage it's in - ready to test, in development, or in planning 14:27:22 dneary: yes, this was my action from last week, which i just didn't get to 14:27:25 And if it's in development, who's doing it, and when it's expected 14:27:39 mburns, Let's give it to mgoldboi ;-) 14:27:46 dneary: all of which should be on the feature pages 14:27:59 dneary: gladly! 14:28:05 dneary: appreciate it... i'll take care of it 14:28:22 mgoldboi, This goes back to what I said when I was in TLV: when you go to the 3.3 release planning page, you have no idea if the release date is realistic, or whether the features listed are done or in progress 14:28:26 #info need to update the release management page with feature status and release criteria for features 14:28:46 #action mgoldboi to track down vdsm/engine related component 14:28:49 Also, no idea if there are other features which would get in if someone did them, but we didn't get around to them 14:28:50 #undo 14:28:50 Removing item from minutes: 14:28:51 #action mgoldboi to track down vdsm/engine related components 14:28:59 #action mburns to handle node related components 14:28:59 Thanks guys 14:29:00 dneary: totally agreed - if you can just find me additional few hours a day.. 14:29:08 #link http://www.ovirt.org/OVirt_3.3_release-management 14:29:17 dneary: checkpoint for me early next week 14:29:24 finally, we have some flesh listed there 14:29:24 mgoldboi, Sleep is optional, you know ;-) 14:29:28 mgoldboi: i talked to dneary about the release criteria for features -- 14:29:47 i think if there a features that are critical to be in the release, we should definitely mark them as such 14:29:59 but i don't think every feature should be listed 14:30:05 danken, I would love to have, in each feature, a short test plan which people can follow on test day to check that the feature works 14:30:08 even as a *Should* 14:30:21 danken, I would put it as part of the template, and get the developer who wrote the feature to write it 14:30:55 To have a successful test day, people need a check-list of things to test 14:31:09 dneary: certainly. but the most important thing is to actually have all the features listed. which was only recently and paritally achieved 14:31:22 danken, Ah? 14:31:41 I found the page ~3-4 weeks ago, and praised it on the arch list 14:31:50 I didn't know it existed before then 14:31:51 Well, #storage still has bullets with no links 14:32:07 danken, So there are no new features in storage... correct? 14:32:11 ;-) 14:32:33 dneary: fwiw, we've had the release management pages for a few releases, now 14:32:43 and they're always linked from the meeting minutes... 14:32:51 dneary: there are new features 14:33:00 they are just not linked AFAIK 14:33:15 dneary, danken the (storage) features without the pages should indeed be updated , although at this point can't gaurentee they'll be ready for the 3.3 timeline 14:33:42 amureini: this is why we have the must and should 14:34:24 mgoldboi, indeed. mburns, can you add an AI on me to make sure all the storage features are marked correctly? 14:34:45 mburns: ^^please do 14:35:09 well... actuallly the timeline is an issue on the SLA side as well.... 14:35:16 #action amureini to ensure that all storage features have feature pages and are updated to reflect status (and risk of missing 3.3) 14:35:24 mburns, 10x. 14:35:25 another important AI is to add a HOW TO TEST section to each feature page. 14:35:52 I know that *my* pages lack them :-( 14:36:53 danken: it's a good input 14:37:04 mgoldboi: how do you want to handle adding "How To Test" 14:37:12 danken: but i'm afraid we missed this release on this item 14:37:15 want me to action you to track all non-node features? 14:37:24 and i'll add node? 14:37:49 mburns: we can try it with checkpoint a week before testday? 14:38:17 mgoldboi: sure 14:38:19 mburns: go for it - action items on me today 14:38:21 and most importantly, all feature owners should know about this. 14:38:40 #action mgoldboi to get people to add "How to Test" section to feature pages for non-node features 14:38:50 danken: believe me, they will know 14:38:56 #action mburns to add How To Test section to feature pages for Node features 14:39:10 danken: yep, that's the action items... 14:39:15 #chair dneary mgoldboi 14:39:15 Current chairs: dneary mburns mgoldboi 14:39:30 ofri: SLA features are at risk? 14:39:48 things are not going as fast as expected 14:40:10 it seems we wont make it to feature freez on the end on May 14:40:15 how critical are these for the 3.3 release? 14:40:32 and how much of a slip are we talking about? 14:40:49 we are talking about two weeks....l 14:40:52 amureini: actually, same questions for the storage features you think might miss^^ 14:40:58 I am afraid that sla is not alone there. 14:41:29 ok, let's back up a second then 14:41:32 we have serious concern about network quantum integration 14:41:43 we'll go through each area... 14:42:04 #info sla group is at risk for features being ready for dev freeze 14:42:06 mburns: i feel we are shooting in the dark here, as dneary stated - we need to know what we must deliver and estimate than 14:42:23 #info estimated delay to get them in ~2 weeks 14:42:37 mgoldboi: yes, we need that info for making the decisions 14:42:59 but getting ballpark estimates for how much delay is needed can be done now 14:43:23 mburns: ok, let's go with it 14:43:36 danken: what about other networking features? 14:43:48 and how much delay on quantum integration? 14:44:11 mgoldboi, Thanks! 14:44:49 mburns: they have troubles, too. but they are less obligatory for the release 14:45:48 danken: any concept of how much they're going to slip? 14:45:49 no code is written for http://www.ovirt.org/Features/Network_Labels - only hacks 14:46:10 mburns: and we just found out that http://www.ovirt.org/Features/Normalized_ovirtmgmt_Initialization has an unfortunate collision with ovirt-node 14:46:14 danken: Labels feature definitely won't be ready for 3.3. 14:46:31 Requires a solid design first. 14:46:44 I mean, for the feature freeze. 14:46:46 * mburns wonders if he should ask what *will* be ready 14:46:57 migration net is in 14:46:57 mburns: good question... 14:47:16 i get the feeling that this discussion is too big for this meeting... 14:47:58 mgoldboi: how about, in the process of getting the feature pages updated, we also get status on each feature 14:48:10 and an idea of how critical it is for the release 14:48:27 and how long it would need to get in 14:48:29 mburns: +1 14:48:38 mgoldboi: and have that for next week's meeting 14:48:56 sounds good 14:48:57 mburns: optimistic - but let's try 14:49:00 mburns: I would LOVE to have http://www.ovirt.org/Features/Normalized_ovirtmgmt_Initialization playing nice on ovirt-node for the release 14:49:25 mgoldboi: opesimistic? 14:49:33 danken: we are discussing some network stuff with mark in #vdsm-networking at freenode 14:49:35 and when I say that - I mean to have ovirt-node not depend on bridges when configuring networking... 14:50:04 danken: very unlikely unless we slip the entire ovirt release by a month, and even then -- i couldn't promise 14:50:33 mgoldboi: i think we need as much of this as we can for next week's meeting 14:50:44 so we can make changes to the schedule 14:51:59 ok, so we have actions to gather the information needed for next week 14:52:08 anything else we should cover today for 3.3? 14:52:15 that isn't going to depend on that information 14:52:26 mburns: can you estimate the amount of effort for allowing bridgeless network config in ovirt-3.3? 14:53:00 danken: not really, at the moment 14:53:54 danken: having a bridge on the network is something that was an implicit assumption 14:54:05 i have no idea how wide-ranging the side-effects would be 14:54:16 if we removed that assumption 14:54:41 and my team is pretty stretched at the moment 14:55:57 mburns: it would help us a lot. because the currently suggested workaround is: add brth0 in ovirt-node. remove breth bridge and add ovirtmgmt bridge ovirt-host-deploy. remove ovitrmgmt from Engine (for cluster with non-VM management network) 14:56:19 that's too insane, even for me. 14:56:47 danken: i'll try to get someone to look at it 14:57:10 but i can't make any promises until we do some investigation 14:57:23 but we can discuss this outside the meeting 14:57:29 #topic Conferences/Workshops 14:57:38 dneary: any updates that you know of here? 14:57:46 * mburns notes theron is not around 14:58:01 mburns, Theron wrote a post workshop report after China 14:58:26 dneary: did that go to a list somewhere? 14:58:29 It was going over & back with OVA and Intel, but I think it's close to final. Sounds like the event went well - over 100 people at the workshop 14:58:43 excellent 14:58:53 Theron's on a plane right this minute (I spoke to him at the gate an hour or so ago) 14:59:06 #info post workshop report for Shanghai workshop is in the works and close to final 14:59:18 dneary: i'm guessing it will be made available in the near future? 14:59:30 There are no workshops coming up that I'm aware of, but we are planning a big oVirt developer meeting at the KVM Forum in Edinburgh. 14:59:37 mburns, Yes, it will 14:59:52 #info report to be made available soon 15:00:02 #action Theron to send event report from Shanghai to various mailing lists, and get it on the website 15:00:09 dneary: do you know anything about the cloud/virt conference in SF in August? 15:00:12 i just rebooted my server and my host wont' restart 15:00:21 i'm getting this: StoragePoolMasterNotFound: Cannot find master domain: 'spUUID=13615644-7a31-489c-ac1f-e025c77ee1ba, msdUUID=a6cb74ba-f9dc-4fa0-a354-78e99891a0c4' 15:00:29 nothing has changed =\ 15:00:46 mburns, We'll have more news next week about that 15:00:50 #info planning a big ovirt developer meeting at KVM Forum in Edinburgh in November 15:00:59 #info update on cloud/virt conf coming next week 15:01:40 dneary: anything else you know about that should be captured here? 15:01:58 Summit presentations related to oVirt? 15:02:17 dneary: are you asking if there are some? 15:02:24 There are a number of oVirt/RHEV related sessions at the Red Hat Summit next month. 15:02:25 or if people want to give them? 15:02:36 If that's something worth talking about 15:02:43 probably just a bullet 15:03:02 dneary: any idea if all Shanghai preso's are ready? 15:03:09 ie- uploaded 15:03:30 doron, No idea, I'm afraid 15:03:33 #info there are a number of ovirt/RHEV related talks scheduled for Red Hat Summit (schedule: http://www.redhat.com/summit/agenda/glance.html) 15:03:50 dneary: worth following this, as I got a few mails about it. 15:04:02 ewoud: anything from the infra team? 15:04:13 dneary: mburns maybe something else which should be menioned here 15:04:28 There are at least 6 or 7 - including a rpesentation by Yours Truly on the value of community 15:04:39 while in Shanghai, we discovered ovirt.org is unofficially blocked in China. 15:04:56 #info presentations from Shanghai will be posted soon (some issues with wiki posting, but working that out) 15:05:09 doron: oh, that's nice... 15:05:19 mburns: yes 15:05:23 mburns: I started discussing it with Theron 15:05:32 ewoud: ok, one sec... 15:05:38 as ovirt currently has zero visibility in China 15:05:51 doron, http://www.ovirt.org/Special:ListFiles <- This shows a number of the Shanghai presentations have been uploaded. 15:06:02 dneary: checking 15:06:18 I don't know if they've been linked in from the agenda 15:06:30 How does one get a website unblocked in China? 15:06:36 I have no idea how to do that 15:06:45 * mburns doesn't either 15:06:49 dneary: long story, but 15:07:18 the final result is we need a local site in china, do not ask why, as the answer has political aspects. 15:07:27 basically 15:07:46 how does an open source project have a "site" in china? 15:07:49 we need ovirt.cn which will replicate the contents of ovirt.org and also 15:07:58 oh, website 15:08:05 * mburns thought you meant physical site 15:08:07 allow to replicate ovirt.cn things into ovirt.org. 15:08:44 sounds like an infra team mission! 15:08:51 don't you need a Chinese company to have a Chinese website? 15:09:04 it is so. also, we may want to support sibling projects 15:09:14 doron: does it have to be a separate site that replicates or can we alias it to our openshift site? 15:09:16 such as gluster. 15:09:48 this seems like a big topic, can we take it to the mailing list? 15:09:52 ewoud: I suggest we take it to a different conversation. I just wanted to make sure this comes up here, as we are loosing ovir tin China now. 15:10:09 Noted. 15:10:10 mburns: worth following up next week with Theron as well. 15:10:11 doron: perhaps infra@ and board@ 15:10:19 mburns: sure. 15:10:21 maybe with direct cc for theron 15:10:23 doron: sounds like a better idea 15:10:29 anyone got any ideas oon how to help me troubleshoot? 15:10:34 i've got people asking why the vm's are down 15:10:41 ok, moving on 15:10:41 mMessage Storage domain does not exist: ('8780c881-4623-4083-871f-cf6ffcc71745',) 15:10:49 #topic infra report 15:10:53 ewoud: ok, all yours 15:10:58 blindrage, Sorry - you walked into the middle of a meeting. What's the issue you're having? 15:11:04 mburns: cool 15:11:07 s /rhev/data-center/mnt/insight-slave.startechnology.com\:_media_storage_ovirt_data/8780c881-4623-4083-871f-cf6ffcc71745/ 15:11:07 dom_md images 15:11:15 my main host won't activate 15:11:18 keeps going to failed 15:11:20 infra started installing VMs on alterway machines 15:11:22 was working, system reboot 15:11:37 #info infra installing virtual machines on alterway hosts 15:11:39 i'm getting that message about 8780c…. not existing, but it's mounted 15:11:40 blindrage: we're in the middle of a weekly meeting. any chance you can wait a few minutes? 15:11:45 that's fine 15:11:46 last week foreman.ovirt.org was set up, obasan is now working on setting up monitoring.ovirt.org 15:12:02 #info foreman instance is up, monitoring.ovirt.org is underway 15:13:12 ewoud: anything else? 15:13:43 mburns: sorry, colleague had a question 15:13:46 np 15:14:31 we're looking into setting up puppet on the infra 15:14:45 doesn't puppet go hand-in-hand with foreman? 15:14:49 it does 15:15:09 so don't we already have it b/c we have foreman? 15:15:16 or can you setup foreman without puppet? 15:15:24 but there are many ways to set it up and I've been slacking on proposing how to set it up 15:15:29 which puppet modules etc 15:15:34 ok 15:15:46 #info puppet setup is being looked at as well 15:15:54 we currently have an empty git repo, but puppet is running on a few hosts 15:16:14 also, puppet 2.7 (in fedora) can't connect to puppet 2.6 (in epel) 15:16:19 so work in progress there 15:16:42 regarding rackspace hosts, there's still problems with the installer freezing 15:17:06 #info some puppet incompatibilities with fedora (2.7) vs epel (2.6) versions 15:17:19 #info rackspace hosts having issues with installers freezing 15:17:26 ewoud: is someone working on that issue? 15:17:31 there's a support case I think, but since we have a limited number of VPN connections (1), it's harder to have multiple people working on it 15:18:07 I think dcaro is now looking into it together with eedri 15:18:29 #info dcaro and eedri are working on the issue with rackspace 15:18:56 that's it for now 15:19:05 ok, thanks 15:19:12 #topic other topics 15:19:23 anything else before i close the meeting? 15:20:00 going once... 15:20:21 twice.. 15:20:44 gone. 15:20:46 #info nothing else 15:20:49 thanks all! 15:20:51 #endmeeting