15:01:15 <bkp> #startmeeting oVirt Weekly Sync
15:01:15 <ovirtbot> Meeting started Wed Nov 19 15:01:15 2014 UTC.  The chair is bkp. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:15 <ovirtbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:01:31 <bkp> #topic Agenda and Roll Call
15:01:38 <bkp> #info infra update
15:01:38 <bkp> #info 3.5.z updates
15:01:38 <bkp> #info 3.6.0 status
15:01:38 <bkp> #info conferences and workshops
15:01:39 <bkp> #info other topics
15:01:56 * sbonazzo here for integration
15:02:04 * mlipchuk here for storage
15:02:18 <bkp> Okay, I think we are back on track with the meeting schedule. Apologies to all for the confusion these last few weeks.
15:02:28 * lvernia is here.
15:02:44 * ofrenkel here
15:03:18 <bkp> Great, good to hear from everyone!
15:03:20 * rgolan rgolan here
15:03:21 <bkp> #topic infra update
15:03:30 * rgolan here
15:03:44 <bkp> Hey rgolan , we got you.
15:03:54 <bkp> What's the word from infra?
15:04:28 <bkp> dcaro?
15:05:29 <bkp> eedri? Anything from infra?
15:05:58 <sbonazzo> bkp: we had an outage in phx2 lab a couple of days ago
15:06:23 <sbonazzo> bkp: should be at least partially solved since I can see most of the slaves up and running
15:07:13 <sbonazzo> bkp I don't have more details on the forensic of the case
15:07:27 <danken> sbonazzo: we apparently lost our f21 slave, which we try to run functional vdsm tests on
15:07:47 <sbonazzo> danken: please send an email to infra@ovirt.org
15:07:59 <danken> sbonazzo: the state of vdsm jenkins jobs is a mess :-(
15:08:00 <ofrenkel> there are many false failures on engine patches, we really need to recover these, as jenkins becomes non reliable for developers
15:08:14 <sbonazzo> danken: it's not only vdsm....
15:08:27 <danken> sbonazzo: we communicated with dcaro, but ok.
15:08:44 <danken> sbonazzo: is opening a ticket better from your perspective?
15:09:27 <sbonazzo> danken: maybe, yes. sadly infra team is understaffed, any help from developers is welcome for maintaining jenkins jobs
15:09:38 <bkp> danken: I would say a ticket is better just from a documentation/accountability perspective.
15:09:53 <danken> sbonazzo: exactly my point - I want to own a slave, so phoracek can fix the vdsm job
15:10:16 <danken> sbonazzo: but he's waiting for 2 weeks for slave credentials
15:10:24 <sbonazzo> danken: for taking ownership of the slave I guess a formal vote on infra@ovirt.org is needed
15:11:37 <sbonazzo> danken: ok, let's push this offline on mailing list
15:12:05 <bkp> sbonazzo: Agreed. Maybe there should be a streamlining of that vote process.
15:12:55 <bkp> Anything else?
15:13:07 <sbonazzo> bkp please add action item on dcaro to review the status of danken 's request
15:13:29 <bkp> Already have it prepped.
15:14:26 <sbonazzo> bkp I guess that's all for infra for now
15:15:00 <bkp> Okay.
15:15:08 <bkp> #info Infra We had an outage in phx2 lab a couple of days ago, seems to be resolved.
15:15:11 <bkp> #info Infra We apparently lost our f21 slave, on which we try to run functional vdsm tests
15:15:14 <bkp> #info Infra The state of vdsm jenkins jobs is a mess, people are waiting a long time to get their credentials.
15:15:17 <bkp> #action Infra dcaro: Discuss on mailing list how slave ownership process can be sped up, per danken concerns in this week's meeting
15:15:20 <bkp> #info Infra The team is understafffed and looking for help from developers to maintain Jenkins jobs.
15:15:23 <bkp> Moving on...
15:15:32 <bkp> #topic 3.5.z updates
15:15:43 <bkp> sbonazzo: What's the word?
15:15:44 <danken> sbonazzo: /me just sent a formal request to own the job and its slave(s)
15:16:08 <sbonazzo> bkp: sent a status report this morning here: http://lists.ovirt.org/pipermail/users/2014-November/029282.html
15:16:29 <sbonazzo> bkp we still have 4 blockers open and 5 regression to be reviewed
15:16:36 <sbonazzo> bkp RC is scheduled for next week
15:16:48 <sbonazzo> bkp so we need an ETA on the blockers
15:17:26 <sbonazzo> bkp also bug count is still high, remaining bugs must be scrubed and postponed to later releases if not blocking / urgent
15:18:56 <bkp> sbonazzo: That RC is for 3.5.1, correct?
15:19:01 <sbonazzo> bkp yes
15:19:40 <bkp> sbonazzo: Also, should there be a status on the addition of el7 support to the 3.5.x cycle? Issues with the test builds?
15:19:55 * tal here
15:19:59 <sbonazzo> bkp: as far as I know el7 build is completed
15:20:07 <sbonazzo> bkp it should be fully testable on el7
15:20:36 <sbonazzo> bkp: help testing nightly snapshot before RC is welcome
15:21:10 <sbonazzo> bkp http://www.ovirt.org/OVirt_3.5.z_Release_Management#Nightly_Builds
15:21:17 <bkp> Coolness. Anything else?
15:21:50 <sbonazzo> bkp that's all for 3.5 on integration / releng side
15:22:13 <dcaro> bkp: sbonazzo I suppose I'm a little late :/
15:22:42 <bkp> dcaro: A bit. :) Did you have anything to add to the infra discussion?
15:23:25 <bkp> sbonazzo: Acknowledged on the 3.5 news.
15:23:57 <bkp> #info 3.5.z updates Full report at http://lists.ovirt.org/pipermail/users/2014-November/029282.html
15:24:00 <bkp> #info 3.5.z updates Highlights from report: 4 blockers open, 5 regressions to be reviewed
15:24:03 <bkp> #info 3.5.z updates RC for 3.5.1 is scheduled for next week, so ETA needed on blockers
15:24:06 <bkp> #info 3.5.z updates Bug count is still high (150), review and either scrub or postpone to next release if not blocking or urgent.
15:24:09 <bkp> #info 3.5.z updates .el7 support for engine is completely built and fully testable on el7. Help testing nightly snapshot prior to RC next week is welcome.
15:24:44 <dcaro> bkp: we did an engine upgrade and jenkins upgrade last week, the engine upgrade is not finalized yet (had some issues with vdsm, looking into it) but the system is stable now, have to plan with caution to finish the upgrade
15:25:04 <danken> speaking of 3.5 - has anybody noticed a problem installing on el6.6 hosts?
15:25:05 <bkp> #topic infra update part 2
15:25:07 <dcaro> (engine is 3.5, there's on ehost with vdsm 3.5, all the others are 3.4)
15:25:39 <sbonazzo> danken: installed yesterday for log-collector testing, no issue on engine side
15:26:20 <danken> there's a confusing report about libvirtd interfering with vdsm
15:26:22 <danken> Bug 1162539 -        libvirtd cannot be started using upstart if started manually using sysvinit
15:26:31 <danken> I wonder if anybody noticed something similar
15:26:34 <bkp> dcaro: ETA on that plan?
15:26:54 <mlipchuk> dcaro: We currently investigating an issue which could be related when creating a volume on RHEL 6.6
15:27:25 <mlipchuk> dcaro: but that is still under investigation
15:27:54 <dcaro> mlipchuk: good to know, can you add me to the list if any? So I'll be aware?
15:28:11 <dcaro> bkp: not really yet, but hopefully next week
15:28:17 <mlipchuk> dcaro: though the version is 3.4
15:28:30 <bkp> #info Infra An engine upgrade and jenkins upgrade were performed last week. The engine upgrade is not finalized yet (had some issues with vdsm, looking into it) but the system is stable now, have to plan with caution to finish the upgrade. (Hoipefully next week)
15:28:32 <mlipchuk> dcaro: this is the bug https://bugzilla.redhat.com/show_bug.cgi?id=1160620
15:28:34 <bkp> #info Infra We currently investigating an issue which could be related when creating a volume on RHEL 6.6, but that is still under investigation
15:29:31 <dcaro> mlipchuk: oh, no the issue I had was that vdsm removed all the network configuration so the host was unreachable, strangely after one day the host came back online (I rebooted it like 50 times in the meantime, trying to get ILO console to work)
15:31:31 <bkp> #topic 3.5.z updates part 2
15:31:42 <bkp> #info 3.5.z updates There may be a problem installing on el6.6 hosts. Bug 1162539 libvirtd cannot be started using upstart if started manually using sysvinit. Let's check this one out, please.
15:31:51 <bkp> Okay, let's keep going...
15:32:00 <bkp> #topic 3.6 status
15:32:14 <bkp> Any burning issues for 3.6 this week?
15:32:23 <sbonazzo> bkp sent a status report this morning here: http://lists.ovirt.org/pipermail/users/2014-November/029283.html
15:32:36 <sbonazzo> bkp no blockers as expected so early in the release cycle
15:33:13 <sbonazzo> bkp discussion about release process / release criteria should have been closed last week, but no progress since its opening
15:34:00 <sbonazzo> bkp added  External Project Schedules to the report for having a look at what will be out with ovirt
15:34:11 <lvernia> sbonazzo: If there's no input, things should probably stay as proposed, shouldn't they?
15:34:45 <sbonazzo> lvernia: 2 proposal
15:34:47 <sbonazzo> 1) keeping the same release criteria we had for 3.5 [2] 2) review the proposed changes [3] and prepare new release criteria for 3.6
15:34:57 <sbonazzo> [2] http://www.ovirt.org/OVirt_3.5_release-management#Release_Criteria [3] http://lists.ovirt.org/pipermail/devel/2014-September/008695.html
15:35:12 <bkp> lvernia sbonazzo I would say so. If you don't chime in, then that's implicit approval.
15:35:13 <sbonazzo> lvernia: poll open :-)
15:36:17 <sbonazzo> bkp so both option acknowledged or indifferent? just pick one and go with that?
15:36:52 <lvernia> sbonazzo: You suggested [3] due to deficiencies you noticed in [2]?
15:37:01 <sbonazzo> lvernia: yep
15:37:06 <bkp> sbonazzo: I would hope not.
15:37:09 <lvernia> So there's one vote for [3] over [2] :)
15:37:21 <sbonazzo> lvernia: thanks :-)
15:37:28 <bkp> But lvernia has a point.
15:38:07 <sbonazzo> bkp: I agree
15:38:28 <bkp> I will make on more entreaty here to get people to chime in.
15:38:36 <bkp> Anything else?
15:38:48 * lvernia Googles "entreaty".
15:39:18 <sbonazzo> bkp scheduling / timeline is not yet defined for 3.6 hope to get a cleaner view on what's going to be in for next week
15:39:21 <bkp> lvernia: See: harrass
15:39:42 <lvernia> Well, I see both [2] and [3] suggest 3.6 be released 6 months after 3.5.
15:40:08 <lvernia> Was there any input on why it wouldn't be 6 months?
15:40:45 <sbonazzo> lvernia: http://lists.ovirt.org/pipermail/users/2014-November/028875.html
15:41:28 <lvernia> Ahh yes, I remember reading those :)
15:41:29 <lvernia> Thanks.
15:42:00 <lvernia> And you're saying the deadline for deciding was a week ago? :)
15:42:32 <sbonazzo> lvernia: well...  yes according to existing process
15:42:41 <lvernia> Which is what we're voting on :p
15:42:48 <sbonazzo> lvernia: yep
15:43:17 <bkp> Let's move that offline, and try to get more input on all of these.
15:43:24 <sbonazzo> bkp:  +1
15:43:29 <lvernia> Yes, same as two and three weeks ago...
15:43:48 * rgolan have no clue on what this is about
15:43:57 <sbonazzo> lvernia: well, at least we have your vote :-)
15:44:09 <lvernia> rgolan: Deciding on a time frame for 3.6.
15:44:17 <bkp> Well, if no one cares, then we'll leave it to team leads.
15:44:17 <lvernia> rgolan: Nobody seems to care so no decision is made :p
15:44:37 <rgolan> lvernia: I figured that out from the mail archive... :P
15:45:02 <bkp> #info 3.6 status Full report at #info 3.5.z updates Full report at http://lists.ovirt.org/pipermail/users/2014-November/029282.html
15:45:05 <bkp> #info 3.6 status No blockers! Well, that's expected.
15:45:07 <bkp> #info 3.6 status External Project Schedules to the report for having a look at what will be out with oVirt
15:45:10 <bkp> #info 3.6 status Discussion about release process / release criteria should have been closed last week, but no progress since its opening
15:45:13 <bkp> #action 3.6 status All: If you want to have a say in the release process, get in on the discussion:  http://www.ovirt.org/OVirt_3.5_release-management#Release_Criteria http://lists.ovirt.org/pipermail/devel/2014-September/008695.html
15:45:17 <bkp> #info 3.6 status scheduling / timeline is not yet defined for 3.6 hope to get a cleaner view on what's going to be in for next week.
15:45:20 <bkp> #info 3.6 status Still looking at 6 months out in both release process proposals, but danken has a proposal to extend the cycle. http://lists.ovirt.org/pipermail/users/2014-November/028875.html
15:45:24 <bkp> #action 3.6 status All: let's examine and vote on proposal to extend the cycle. http://lists.ovirt.org/pipermail/users/2014-November/028875.html
15:45:27 <bkp> Okay, any other 3.6 issues?
15:45:39 <lvernia> bkp: danken was the one suggesting shortening the cycle.
15:45:49 <lvernia> bkp: If I understand correctly.
15:46:01 <bkp> Ah, misread the thread.
15:46:18 <danken> lvernia: right. I'd like to have more small cycles
15:46:18 <rgolan> bkp: lvernia which I think is kind make sense for big features
15:46:18 <lvernia> bkp: I would say there's disagreement on the ideal length of a minor version :)
15:46:30 <danken> which means more test days
15:46:41 <danken> and less horrible slips in scheduule.
15:46:56 <bkp> #info 3.6 status CORRECTION: Still looking at 6 months out in both release process proposals, but danken has a proposal to change the cycle. http://lists.ovirt.org/pipermail/users/2014-November/028875.html
15:47:00 <bkp> #action 3.6 status CORRECTION: All: let's examine and vote on proposal to chage the cycle. http://lists.ovirt.org/pipermail/users/2014-November/028875.html
15:47:35 <bkp> Unfortunately, we need to discuss offline, have to wrap this up.
15:47:42 <sbonazzo> bkp ok
15:47:46 <bkp> #topic conferences and workshops
15:47:47 <rgolan> danken:  this would probably mean the feature evolves differently than putting all in into a single version no?
15:48:01 <bkp> #info LISA14 went pretty well. Booth traffic was solid, gshereme and bkp answered a fair amount of general questions about oVirt, and talked to a lot of sysadmins with specific queries on oVirt.
15:48:05 <bkp> #info FOSDEM is the next big event. OSAS will be very much on the scene, and oVirt has applied for a booth at the event. There is still time to get your talks into the IaaS and Virt devrooms.
15:48:09 <bkp> #info Now that travel is wrapped up, I am working on getting meetup and workshop materials ready for anyone who wants to present on oVirt at a local event.
15:48:24 <bkp> Any other conference news?
15:49:21 <bkp> Okay, moving on.
15:49:24 <rgolan> bkp: do we know what people at FOSDEM would like to here on oVirt or we just come up with something (not very interesting probably...)
15:49:29 <sbonazzo> bkp 11th Intl. Conf. on Open Source Systems CFP Open Through Jan. 10 call for papers http://www.oss2015.org/call-for-papers/
15:49:38 <rgolan> bkp: s/here/hear
15:50:40 <bkp> rgolan: JUst submit something you think is interesting.
15:50:45 <bkp> #info 11th Intl. Conf. on Open Source Systems CFP Open Through Jan. 10 call for papers http://www.oss2015.org/call-for-papers/
15:51:11 <bkp> #topic other topics
15:51:23 <bkp> #info A big welcome to our new OPW interns! Be ready to assist with any of their questions as they lurk about the channel and the lists!
15:51:26 <bkp> #info It was brought to my attention this morning that ovirt.org/stats has not been working for a while. I have asked misc to look into it.
15:51:29 <bkp> #info Speaking of stats, downloads for oVirt engine in Oct. totaled 1460 from resources.ovirt.org. This is down from 2194 downloads in Sept. Likely now that 3.5 is rolling out, this trend will likely change.
15:51:33 <bkp> #info One of the big goals we should have for 2015 is figuring out a way to determine actual installs of oVirt. If anyone has ideas on how we can pull such information together, let's get a discussion going on the users mailing list.
15:51:37 <bkp> #info Finally, there is an informal survey out on the users mailing list now about extending oVirt support to Debian/Ubuntu. Got thoughts on this? Here's the thread: http://lists.ovirt.org/pipermail/users/2014-November/029209.html
15:52:00 <danken> rgolan: yes. it means that big feature should have smaller milestones. which I think is a Good Thing
15:53:02 <bkp> Does anyone want to raise an issue? discuss something?
15:54:04 <bkp> Going once?
15:54:49 <bkp> Okay, thanks, see you all next week.
15:55:01 <bkp> #endmeeting