17:00:21 <jroll> #startmeeting ironic
17:00:21 <openstack> Meeting started Mon Apr 18 17:00:21 2016 UTC and is due to finish in 60 minutes.  The chair is jroll. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:24 <openstack> The meeting name has been set to 'ironic'
17:00:25 <jroll> hey everyone
17:00:26 <TheJulia> o/
17:00:26 <NobodyCam> o/
17:00:26 <lucasagomes> o/
17:00:26 <rpioso> \o
17:00:27 <devananda> o/
17:00:31 <jroll> as always, we have an agenda:
17:00:33 <jroll> #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting
17:00:36 <jroll> light today :)
17:00:39 <NobodyCam> :)
17:00:41 <mat128> o/
17:00:43 <alineb> o/
17:00:44 <rama_y> <rama_y> o/
17:00:44 <jroll> #chair NobodyCam devananda
17:00:45 <openstack> Current chairs: NobodyCam devananda jroll
17:00:49 <NobodyCam> week before summit
17:00:55 <rama_y> o/
17:00:59 <jlvillal> o/
17:01:00 <jroll> #topic announcements and reminders
17:01:01 <krotscheck> o/
17:01:04 <krtaylor> o/
17:01:05 <jroll> so
17:01:08 <jroll> one week til summit
17:01:09 <rloo> o/
17:01:21 <jroll> I expect everyone will be busy preparing
17:01:22 <godp1301> \o
17:01:29 <vdrok> o/
17:01:36 * NobodyCam can't wait to see everyone who is going in actual physical form
17:01:39 <jroll> our design summit schedule is here:
17:01:40 <dtantsur> o/
17:01:41 <jroll> #link https://www.openstack.org/summit/austin-2016/summit-schedule/global-search?t=ironic%3A
17:01:48 <sambetts> o/
17:01:51 <jroll> please do read through things and prepare for the sessions
17:01:59 <jroll> I'll be fleshing out etherpads a bit more this week
17:02:24 <jroll> in other news, our gate is pretty hosed right now, a couple people are digging on it a bit
17:02:26 * jroll finds bug link
17:02:27 <rloo> would be good if the etherpads included links to specs that we could look at before next week
17:02:40 <jroll> #link https://bugs.launchpad.net/diskimage-builder/+bug/1571492
17:02:42 <openstack> Launchpad bug 1571492 in Ironic "gate-tempest-dsvm-ironic-pxe_ssh-dib failing : test_baremetal_server_ops ssh timeout" [Critical,Confirmed]
17:02:53 <jroll> rloo++ I will be adding those
17:02:58 <krtaylor> we could make a list in the planning etherpad
17:03:06 <krtaylor> #link https://etherpad.openstack.org/p/ironic-newton-summit
17:03:10 <rloo> thx jroll. esp if the specs don't exist yet ;)
17:03:23 <jroll> krtaylor: well, I moved that content into the individual etherpads
17:03:28 <jroll> so I'd prefer the specs there
17:03:42 <krtaylor> jroll, sure, understood
17:03:44 <jroll> and I'll blow away the planning pad as well with links to individual etherpads and other info
17:03:57 <rloo> ++ for specs in the individual specs. that way all the info will be there about a particular session
17:04:33 <rloo> jroll: i think it is good to keep the planning pad around, for historical reasons?
17:04:45 <jroll> rloo: maybe? I guess I can push it down
17:04:51 <jroll> etherpad does have history, too
17:05:17 <rloo> jroll: yeah, as long as the etherpad isn't deleted, right?
17:05:31 <jroll> rloo: yeah, I'd only delete the content, not the etherpad itself
17:05:53 <rloo> jroll: oh, sorry, ok if you're only deleting the content :)
17:06:00 <jroll> cool
17:06:26 <jroll> one last thing here, I'm considering releasing a new ironic version this week, any opposition?
17:06:34 <jroll> I want to get us into the habit of more frequent releases
17:06:52 <jroll> and there's some good stuff in there I'd like to get out to users
17:07:13 <jlvillal> jroll: Should we set a goal for more frequent releases?  Every X weeks?
17:07:21 <rloo> i'm fine with releasing as long as it before Wed. In case releasing later and something breaks and 'no one' is around...
17:07:32 <lucasagomes> release early, release often :-)
17:07:32 <NobodyCam> rloo: ++++
17:07:47 <jroll> rloo: yep
17:07:51 <devananda> jroll: minor or point release?
17:07:53 <jroll> jlvillal: I'll think on it :)
17:08:03 <jroll> devananda: it'll be major, some deprecated things were removed
17:08:17 <NobodyCam> jroll: is there a complete list?
17:08:23 <mat128> NobodyCam: release notes :)
17:08:33 <NobodyCam> :p
17:08:35 <rloo> so removing deprecated things always == major release?
17:08:40 <jroll> NobodyCam: yep, http://docs.openstack.org/releasenotes/ironic/current-series.html
17:08:42 <devananda> jroll: wait, 6.0 ?
17:08:50 <jroll> detailed list: https://github.com/openstack/ironic/commits/master
17:08:52 <jroll> devananda: yes
17:09:17 * devananda takes a look at 5.1..current
17:09:18 <jroll> rloo: removing a config is typically a major release afaik
17:09:52 <jroll> as it will break if folks have not updated per the deprecation notice
17:10:39 <jroll> devananda: do you have concerns with releasing 6.0.0?
17:10:45 <rloo> jroll: oh. yuck. in that case, for future, we should get in the habit of removing whatever in one release instead of spreading across several releases in the same dev cycle
17:10:55 <NobodyCam> jroll: just the two [agent] items removed it I read that correctly
17:10:58 <jroll> rloo: why?
17:11:00 <jroll> NobodyCam: yes
17:11:03 <NobodyCam> :)
17:11:15 <rloo> jroll: i dunno. cuz a major release seems major.
17:11:35 <jroll> rloo: it shouldn't, though, the whole idea is that releasing is a normal thing
17:11:39 <rloo> jroll: and 'just' to have a major release cuz we removed 2 configs.
17:11:59 <mat128> rloo: there's no point in waiting for most stuff (and more risk) between "major" releases
17:12:07 <JayF> I mean, flipping the major version number should mean exactly 1 thing: "operator, read some release notes before deploying this"
17:12:11 <rloo> jroll: seems easier for the operators, to do the legwork/checks for one release with 'all' removed things.
17:12:21 <JayF> I don't think there's any point in bundling more into a release or avoiding major releases
17:12:24 <jroll> rloo: then they can follow stable channels
17:12:27 <JayF> ++
17:12:30 <jroll> an operator should always read the release notes
17:12:50 <rloo> i'm not an operator, it was just a thought :)
17:13:36 <sambetts> I assume we wouldn't be maintaining stable for every major release?, so we'd have essentially a Ironic LTS release?
17:13:38 <devananda> I do not see anything jumping out of the output of "git log HEAD ^5.1.0 --no-merges" which would require a significant upgrade -- aside from those two config options
17:13:50 <jroll> sambetts: correct
17:13:55 <JayF> sambetts: aiui we only have stable branches along the 6m releases
17:13:55 <jroll> devananda: correct
17:14:03 <jroll> right, what jay said
17:14:03 <lucasagomes> sambetts, right, stable ones follow the openstack rules
17:14:07 <rloo> I'm not against releasing often, i'm just thinking that if a major release is needed any time we remove something that has been deprecated, it would be nice to remove whatever we want to remove, in the same major release (of a development cycle). that's all.
17:14:28 <devananda> I agree with rloo, fwiw
17:14:36 <NobodyCam> just a question
17:14:50 <NobodyCam> are we going to be removing anything else later this cycle?
17:14:56 <jroll> with my ops hat on, I disagree. I'd rather have one or two things to check per deploy, than a laundry list every 3-6 months
17:15:04 <jroll> NobodyCam: probably
17:15:18 <mat128> ++
17:15:40 * JayF agree with jroll
17:15:45 <jroll> proposal: release 6.0.0 now, and pose the question for the future on the ML?
17:15:51 * NobodyCam just may want ot avoid jump several major rev # in a single cycle ...
17:15:56 <jroll> s/now/when our gate is unbusted/
17:16:00 <jroll> NobodyCam: why? it's just a number.
17:16:06 <NobodyCam> true
17:16:08 <rloo> oh, yeah, i'm good with releasing 6.0.0 as long as it is early this week.
17:16:14 <sambetts> Deprecation period for stuff would still be a whole cycle right? or would it then be 6 months from when the deprecation merged?
17:16:38 <jroll> same deprecation rules as openstack official policy: over a cycle boundary AND at least 3 months
17:16:46 <lucasagomes> sambetts, IMHO cycle, to follow OpenStack
17:16:50 <jroll> we'll likely always follow that
17:17:03 <devananda> I'm fine with dropping 6.0.0 soon, with the understanding that Newton-final might end up being based on not-6.x (it could be 10.x, who knows)
17:17:09 <jroll> https://github.com/openstack/ironic/commits/master
17:17:28 <jroll> devananda: yep, that's my assumption
17:17:29 <devananda> however, I also don't think that signalling to operators 5.1.x -> 6.0.0 as a major upgrade is the right message. there are no "major" changes in here
17:17:29 <NobodyCam> ya, I am good with dropping to.
17:17:45 <devananda> it's too late -- we dropped two CONF options, so it is a major upgrade
17:18:00 <devananda> but in the future, I'd like to batch up the deletion of planned removals a bit better
17:18:03 <jroll> devananda: yep, dropping configs is always a major upgrade
17:18:04 <rloo> unless we want to roll back the CONF options removal?
17:18:05 <devananda> but anyhow, that is a topic for ML or summit
17:18:33 <jroll> let's not revert things
17:18:38 <jroll> for the sake of a number
17:18:40 <NobodyCam> devananda: +1 on batching the removal's :)
17:18:51 <NobodyCam> oh ya let NOT revert
17:18:56 <godp1301> let's not revert things++
17:19:23 <rloo> jroll; that's fine. no revert. will the release notes indicate that there are no major changes, but major release cuz of config options. in the 'prelude' part I guess?
17:19:41 <jroll> rloo: I don't know yet
17:20:02 <[1]cdearborn> o/
17:20:02 <jroll> major versions designate "we broke a thing", not "we shipped a major feature"
17:20:03 <lucasagomes> devananda, yeah batch such changes up seems fine, we can perhaps use the first release of the cycle as the release to remove all the deprecated configs from the previous cycles
17:20:23 <devananda> lucasagomes: I think that would be good
17:20:39 <jroll> I still disagree that batching deprecation drops is 'better', but I'll save that for email
17:20:50 <NobodyCam> lucasagomes ya with a single clear note saying this is what being removed ++
17:21:00 <NobodyCam> hehehe
17:21:20 <jroll> okay, shall we move on?
17:21:28 <lucasagomes> +1 to move on
17:21:31 <godp1301> \o/
17:21:31 <NobodyCam> +1
17:21:46 <jroll> any other announcements/reminders?
17:22:34 <jroll> moving on
17:22:41 <jroll> #topic subteam status reports
17:23:11 <jroll> #link https://etherpad.openstack.org/p/IronicWhiteBoard
17:23:16 <jroll> starts at line 80
17:23:22 <jroll> sorry, lost that tab for a moment :P
17:23:28 <jroll> I'll give folks a few minutes to review
17:24:26 <rloo> dtantsur: is the critical bug the gate breakage?
17:24:36 <dtantsur> yep
17:24:37 <jroll> yes
17:25:13 <jroll> by the way, huge thanks to dtantsur for doing bug triage, I don't thank you enough for that
17:25:29 <rloo> ++++ definitely thx dtantsur
17:25:31 <jroll> and also big thanks to rloo for sending the subteam reports over email weekly and cleaning the agenda
17:25:40 <dtantsur> =^_^=
17:26:04 <rloo> :)
17:26:24 <JayF> We do have a new patch up re: Nova stuff, to fix a longstanding bug in the nova driver
17:26:28 <rloo> jroll: wrt network isolation. Not sure I'll have time this week but if I do -- do you think the code is ready to go in, just needs reviews?
17:26:33 <JayF> it's not linked in the report so going to find and link here
17:26:59 <jroll> rloo: I didn't get to review it last week, but I believe so (and the CI is green!)
17:27:14 * lucasagomes waits for the nova link
17:27:22 <rloo> jroll: good to know. I'll try to find some time.
17:27:33 <jroll> thanks
17:27:34 <JayF> https://review.openstack.org/#/c/306670/ is a fix for Bug 1503453, longstanding in the nova driver
17:27:35 <openstack> bug 1503453 in OpenStack Compute (nova) "unavailable ironic nodes being scheduled to" [Medium,In progress] https://launchpad.net/bugs/1503453 - Assigned to Jesse J. Cook (jesse-j-cook)
17:27:36 <JayF> reviews appreciated
17:28:21 <vdrok> one more bug, one more nova change - https://review.openstack.org/#/c/306010/ :)
17:28:24 <dtantsur> oh, is it the problem we've experienced in tripleo?
17:28:34 <dtantsur> I mean, the double instance assignment
17:31:06 <jroll> JayF: ^^
17:31:31 <JayF> dtantsur: I'm not sure; I think this is for another case
17:32:03 <JayF> dtantsur: this is where a node is in some state, like deploy failed or maintenance, where it'd usually show up as 0 resources available, but because of $failures, the instnace_uuid can get left on and readded
17:32:12 <jroll> should I go ahead and open discussion?
17:32:13 <JayF> dtantsur: I reccomend reading the bug as it is complex and a pita :)
17:32:20 <dtantsur> sure, thanks
17:32:30 <jroll> #topic open discussion
17:32:33 <jroll> anyone got anything?
17:33:03 <JayF> If anyone is going to the summit and wants to be a part of informal meetups for dinner/drinks/etc, if you put your info here: https://etherpad.openstack.org/p/ironic-summit-arrivals it'll make organizing a lot easier :)
17:33:25 <JayF> I'll put all names there into a group hangout chat before the summit so we have a place to coordinate
17:33:35 <rloo> thx JayF. I thought krtaylor had offered to help organize something cuz he is local.
17:33:47 <jroll> ^ a note on that: group hangout gets spammy as heck, join at your own risk (I promise it's fun though)
17:33:50 <JayF> rloo: I'm not organizing anything except getting all the hangouts usernames + arrival times in oneplace
17:34:07 <JayF> jroll: well also, with hangouts, you can mute a single group conversation :) part of why it's a good medium
17:34:16 <rloo> JayF: would it be useful to add a column for evening availability for a dinner or something?
17:34:21 <thiagop> maybe krotscheck can help choose the barbecue place
17:34:29 <JayF> rloo: *shrug*. It's an etherpad, you're welcome to modify if you think it's helpful :)
17:34:35 <JayF> I just wanted ot make sure nobody got left out of the group hangout
17:34:40 <krotscheck> Dependso n whether you want to wait in line or not.
17:34:41 <rloo> or maybe we'll play by ear. i am not volunteering.
17:34:44 <JayF> just b/c they were new, or from a different timezone, or whatever
17:34:44 <krotscheck> SEc, lemme find the good one.
17:34:44 <thiagop> krotscheck: sorry, krtaylor
17:35:01 <jroll> the one with the line is franklin's :P
17:35:02 <NobodyCam> krotscheck: I hear you can play folks to wait in line for you
17:35:40 <lucasagomes> jroll, oh I heard about this franklin's. That's the bbq place that people starting queue at 8am right?
17:35:44 <jroll> so if there's no meeting topics here, let's talk about food back in the main channel so I can get up and get a drink :P
17:35:45 * NobodyCam luv's allthings BBQ
17:35:49 <jroll> lucasagomes: probably closer to 6am :P
17:35:54 <lucasagomes> gosh heh
17:35:57 <krtaylor> I was looking at a coupleof places with good food/beer
17:36:03 <lucasagomes> we need a placeholder there then :-D
17:36:07 * TheJulia has this feeling lots of heartburn will be scheduled for next week
17:36:11 <TheJulia> due to food
17:36:12 <jroll> I'd skip the keynotes and wait it out monday, but they're closed mondays :/
17:36:29 <jroll> anyway, thanks for a good meeting y'all
17:36:31 <jroll> #endmeeting