Wednesday, 2017-12-13

*** openstackgerrit has quit IRC00:33
fungitc-members: commence to officing! (if you're around/conscious anyway)01:00
fungiwow, this is already a lot more people than this time last week01:00
smcginnisI'm on much later than normal. :)01:01
smcginnisAny big issues being discussed? I read through the last office hour logs but by now I've reclaimed those brain cells.01:03
fungii was marginally disappointed by the comment on the sb goal proposal which summarized as "why not just use jira"01:04
fungii would almost believe it was a troll post01:05
smcginnisI could comment on that, but would be afraid I would offend anyone from Jira.01:05
smcginnisSeems trollish to me. :)01:05
pabelangeroh, I missed that. I haven't done review for this week yet01:05
smcginnisAs someone that had to use Jira for awhile, nope.01:05
fungii responded to it in my best attempt to not offend people who make or use jira ;)01:05
pabelangerah, see it now01:07
fungii was really just disappointed someone fairly embedded in our community would make the argument that using proprietary software (even if it's "free" to use) is preferable to developing a free software alternative to that proprietary software01:08
pabelangerso, one of the confusing parts for me, is we are asking projects to move to storyboard, while not all features are finished. But, over the last few years while we have been trying to finalize the migration, we haven't grown the storyboard team.  Are we assuming, as more projects use storyboard, they will step up to write patches or features?01:09
fungithat was the idea... and even in this last round a couple people replied privately to one of my ml posts asking for details on how to get involved01:10
fungidoesn't necessarily need to be openstack people getting involved (the people doing the bulk of the work on it don't use openstack even)01:11
pabelangerright, I am assuming (hoping) the pending hosting projects spec that is being worked on, would lend nicely to maybe growing more developers in storyboard01:12
smcginnisI think that's actually been an issue. My impression earlier on was that sb was coming and some day it might be ready to use.01:12
smcginnisBut with no visual indication that it is near there, previously it was not on my radar at all as something to be concerned with.01:12
fungialso i wouldn't say that we're asking projects to move to storyboard. we're saying it's okay to choose to move to storyboard, and we're asking (via the goals proposal) if now is the time to start recommending it more heavily to the openstack community01:13
smcginnisEspecially as I think at one point we almost shut it down, right?01:13
fungiyeah, it had been really only getting development activity from hp(e), and then that team got dissolved, and some non-openstack downstream users of it stepped up to continue development01:14
pabelangerYah, I believe it was shutdown for a few months, then developer started up again (still ongoing).01:14
pabelangerI know of at least 1 project that tried to use storyboard outside of openstack for a while, but sadly migrated away from it.  I believe they did submit patch upstream too, but still opted to move away01:15
smcginnisI think making it an official goal means we are actually asking projects to move to storyboard. But I don't think most understand what, why, and how. So it actually seems a little premature to me to make it a goal.01:16
fungiwe've always believed there's an inflection point where once it gets enough requested features and enough established user base, development interest will accelerate01:16
fungiwell, teams are going to use what they're going to use. there are already plenty doing task tracking in google docs and trello01:17
smcginnisI think we need something like a regular ML post about the progress on storyboard and whether it is in a state that works for most projects before we can really make it a goal.01:17
smcginnisIt might be functionally there, but I don't think there's any visibility into that.01:17
EmilienMI've read the storyboard patch comments and fungi's comment, I have a little opinion01:18
fungibut we want to get to a point where we can stop relying on ubuntu sso, and when we stop expecting contributors to have lp/ubuntu accounts, there's a lot we can't assume about bug tracking01:18
pabelangersmcginnis: yah, reading the proposal again, it isn't clear to be that it is optional to migrate or asking projects too. Just that we are migrating to storyboard01:19
smcginnis++ I do agree getting away from lp/ubuntu accounts is a good thing. I just think we have a little more work to do before we can get there.01:19
EmilienMI actually understand both point of views, someone contributing to a OpenStack project, who just need a tool that work and seeing a new tool a risk to have less features than before and loose some time on something that is maybe not worth it01:19
smcginnispabelanger: Yeah, and it would seem odd to me to have an optional goal really.01:19
fungismcginnis: the sb devs were trying to do a blog for a while highlighting new features and use cases, but i haven't kept up with it recently01:19
EmilienMon the other hand, building some tooling that fits with how we build OpenStack, with the risk of re-implementing some tools that already exist out there (jira, trello, etc) but not really open-source01:20
EmilienMmy opinion is a bit balanced tbh, and I haven't commented yet in the review01:21
smcginnisSide note, but there are some open source Trello alternatives. I think it would be nice to make that available for teams at some point.01:21
fungismcginnis: i guess it's in need of some new articles01:21
EmilienMsmcginnis: I remember Summit/PTG sessions about all these tools01:21
EmilienMI remember ttx / mordred leading the discussions, it was a few years ago I think01:22
fungismcginnis: well, the idea with sb is that it combines kanban and worklist features with bug tracking01:22
pabelangerI recently was pointed to by SoftwareFactory team, but haven't really look to much into it01:22
EmilienMin my opinion, we shouldn't make it a goal for Queens, but rather take a few (big) projects and migrate them01:23
fungiso we could also host a second trello-like kanban (e.g. wekan?) i suppose if people think there's a lot they want out of a kanban that they're not getting from sb's implementation01:23
EmilienMso during queens we could get more feedback on things went and improve the process for the rest of the projects01:23
pabelangerEmilienM: I believe it is proposed for Rocky currently01:24
EmilienMsorry not queens01:24
smcginnisGetting a few more projects using SB could definitely help.01:24
EmilienMI think my opinion is not much about the tool itself01:24
EmilienMit's about the changes we make01:24
fungiEmilienM: well, it's being proposed as a goal for rocky anyway01:24
smcginnisBut then there is a big group of more interelated ones (Nova, Cinder, etc) that would kind of all need to make the jump at the same time.01:24
EmilienMZuul v3 had been quite disruptive to be honest01:24
pabelangeri think it was even suggested getting the current projects to help answer questions or advocate SB to projects that haven't migrated01:25
EmilienMi'm afraid some community members just get tired of these big changes every cycles01:25
fungibut i also think it's a bit large of a goal for a 6-month cycle unless the majority of the projects who are considering using it have already migrated01:25
smcginnisEmilienM: zuul v3 is a good point. I am worried at some point we're going to lose some of the less dedicated contributors because they get frustrated.01:25
EmilienMI personaly don't launchpad01:26
fungiif we get serious about longer release cycles, i think more involved goals like that become a bit more tenable01:26
EmilienMI personaly don't like launchpad01:26
smcginnisI guess kind along the lines of contributor happiness that cdent has been championing.01:26
EmilienMbut I don't want our contributors running away because we keep creating tools and making big changes at every cycle01:26
EmilienMthey'll probably troll "ok are we going to re-write gerrit?"01:26
EmilienMI'm half joking here01:26
pabelangerwell, gerrit did rewrite their web interface01:27
EmilienMif I would be an occasional contributor who has to maintain some code for my company, I would have an hard time to follow everything that happens in our community01:27
EmilienMI guess for us, it's easy because we're full time on this thing01:27
fungiit seems like the larger we get as a community, the more diverse the infrastructure needs for various projects become and the harder it is to get them all to use something consistent anyway01:27
EmilienMbut I've talked with customers who try to contribute and they keep saying things change/move super fast01:27
EmilienMI guess, my opinion is this one: I like to think we can go away from Launchpad - let's just make it step by steps and not target one cycle/one goal.01:28
fungimaybe the foundation's idea to start having other non-openstack communities will help with that, if we can also start splitting off things that aren't openstack and say that it's okay for them to be different, do different things, use different tools, et cetera01:28
EmilienMthe first reaction when I said to my group we'll have another bug tracker they said "again a new tool?"01:29
smcginnisThat's what I've seen from the k8s community.01:29
fungiEmilienM: honestly, the infra spec simply says that at some point we'll approach the tc about saying it's okay for most projects to migrate to sb. since there's no formal mechanism to do that, the goals process was chosen as the closest option01:30
EmilienMI didn't say it, they did. We need to consider this feedback.01:30
pabelangerpersonally, it is the same issues with slow moving projects too. I myself have had problems contributing on my owntime, I've always found best to learn more about the contributors of the community and keep asking to get things merged01:30
fungii agree the goals process is not a great fit for this01:30
EmilienMinstead of "ok folks you have 6 months to change your bug tracker", it would be "how can we help to make this happen, in your pace and with the features that you need"01:31
fungiyeah. so far it's involved interviewing heavy lp users from various teams to try and work out feature gaps in sb and prioritize implementation for them01:32
pabelangerI think cmurphy did a great post to ML last week (or week before) about keystone. Saying they can't migrate yet due to private bugs. So, that would be a good example to help with what EmilienM just said01:32
EmilienMI spent ~1h / day on Launchpad doing bug triage - I can tell that switching to a new tool will take time. Second thing is we have a lot of bugs in Launchpad. ~500 new bugs at every cycle on average.01:32
fungiyes, i estimate simply importing nova's bugs from lp will take the better part of a week based on some preiminary testing01:33
EmilienMso all that said, before making it a goal we should have a pilot, and not a small project. Something like Nova, Neutron or TripleO.01:34
EmilienMa project willing to make the change, and willing to feel some pain during the migration, if some things go wrong01:34
pabelangerEmilienM: I thought there was talk about tripleo being okay with the migration to SB already?01:34
EmilienMonce the pilot migrated, we can say "ok this huge project migrated, we fixed a bunch of sb issues, we can now think at others"01:35
EmilienMpabelanger: i don't think tripleo is blocker, we're willing to migrate01:35
EmilienMtripleo is always happy to follow and contribute how we can01:35
pabelangerEmilienM: okay, that's how i remember it. Will to migrate01:35
EmilienMwe need more than one champion for this goal because troubles we'll have a lot for sure01:36
pabelangerYah, I mentioned to dhellmann in passing that I'd be interested in help diablo_rojo with storyboard. But I admit, i don't use storyboard much, so I cannot comment much on how the usage will affect a project01:37
EmilienMI'll comment what I said today on IRC in the gerrit patch01:38
fungiwe've had increasingly large teams (aside from infra early on) moving to sb. octavia and monasca a few months ago01:38
fungikolla-kubernetes migrated just last week, and the other kolla repos will probably be migrated soon01:38
EmilienMagain, I'm very positive about the idea - but I'm (too?) realistic about it being a goal.01:38
EmilienMfungi: cool, good to know01:38
EmilienMhow much bugs?01:38
pabelangerI know we in infra have add some issue, and people have proposed fixes. Nothing that I remember as blocker01:40
fungi5k bugs total in sb so far according to `select count(*) from stories;`01:40
pabelangeroh, how many migrated :)01:40
fungigood question... my sql is not to great though. need to constrain to values under a certain integer in the id column01:41
fungier, not so great01:41
fungiwe use an offset for non-migrated stories to avoid future collisions with lp bug numbers01:42
fungiso in theory we can tell by counting the rows where the story id is less than the offset01:43
smcginnisselect count(*) from stories where id < xxx ?01:43
fungimigrated so fa01:44
fungiand yeah, that's the syntax i ended up guessing at01:44
fungistories under 2000000 are migrated and correspond to lp bug numbers so that we can easily change urls in systems like gerrit later01:45
fungiany with ids 2000000 and up were created new in sb01:46
fungibut yeah, development on our blockers _is_ progressing01:49
fungifor example, i'm very excited about
fungithat's the biggest missing piece now for private vulnerability reporting, and also should solve one of the major concerns raised for docs reporting (so we can have embedded urls which prepopulate story descriptions with information about the page/file where the issue was found)01:50
fungifrom the vmt side the idea is that we can have a recommended reporting url documented in prominent places which creates default-private stories01:52
fungifor use when reporting suspected vulnerabilities01:52
fungiso the reporter doesn't need to know to set the "create as private" checkbox when initially filing one01:53
EmilienMI write my thoughts on
EmilienMI hope it reflects well my opinion and what I wrote here01:55
pabelangerI'll read up in the morning, will be calling it EOD in another minute or so01:56
fungiEmilienM: thanks, well-stated in the review02:05
EmilienMthanks for the feedback02:05
*** lbragstad has quit IRC02:25
*** liujiong has joined #openstack-tc03:25
*** mriedem has quit IRC03:28
*** rosmaita has quit IRC04:29
*** diablo_rojo has quit IRC04:49
*** alex_xu has quit IRC06:26
*** alex_xu has joined #openstack-tc06:27
*** liujiong has quit IRC07:52
*** liujiong has joined #openstack-tc07:53
*** kumarmn has joined #openstack-tc09:03
*** gcb has joined #openstack-tc09:06
*** kumarmn has quit IRC09:07
*** jpich has joined #openstack-tc09:11
*** dklyle has quit IRC09:20
*** liujiong has quit IRC10:09
*** dtantsur|afk is now known as dtantsur10:48
*** gcb has quit IRC11:47
*** gcb has joined #openstack-tc11:48
*** cdent has joined #openstack-tc11:50
*** purplerbot has joined #openstack-tc12:22
*** purplerbot has quit IRC12:32
*** purplerbot has joined #openstack-tc12:33
*** rosmaita has joined #openstack-tc13:19
*** mriedem has joined #openstack-tc13:56
*** openstackstatus has quit IRC14:36
*** lbragstad has joined #openstack-tc14:37
*** openstack has quit IRC14:39
*** openstack has joined #openstack-tc14:41
*** ChanServ sets mode: +o openstack14:41
*** openstack has quit IRC14:43
*** openstack has joined #openstack-tc14:46
*** ChanServ sets mode: +o openstack14:46
*** kumarmn has joined #openstack-tc14:52
*** marst has joined #openstack-tc14:56
*** ianychoi has quit IRC15:04
*** cdent has quit IRC15:58
*** ianychoi has joined #openstack-tc16:06
*** dklyle has joined #openstack-tc16:21
fungittx: i can't wait to see the variety of feedback we get on moving to year-long cycles. i have no doubt people are going to raise all sorts of benefits and drawbacks i'm not imagining16:27
smcginnisFor sure.16:28
pabelangerexciting times16:30
*** diablo_rojo has joined #openstack-tc16:30
ttxso far the vendors/distros have been pretty supportive16:34
ttxmost of them have switched to ignoring one every two releases already16:35
ttxand have to do n->n+2 support by themselves16:35
ttxso they would prefer us to cut a branch every year and support yearly upgrade paths instead16:35
smcginnisI've gotten some concerns from developers on changes to cadence, but vendor conversations have all been positive about the change.16:39
persiasmcginnis: from curiosity, how many of those are social vs. focused on freezes, etc.16:51
smcginnispersia: How many of the concerns from developers you mean?16:53
smcginnispersia: My feel is it's mostly social.16:53
persiaThanks for the confirmation of my expectations :)16:54
*** jpich has quit IRC17:15
dhellmannleave it to ttx to stir the hornets' nest the day before my PTO starts ;-)17:16
smcginnisdhellmann: Maybe good timing? :)17:17
dhellmannheh, maybe.17:18
*** diablo_rojo has quit IRC17:26
*** diablo_rojo has joined #openstack-tc17:31
*** dtantsur is now known as dtantsur|afk17:39
dimsLOL dhellmann18:49
dimsdhellmann : ttx had a co-conspirator in smcginnis18:49
smcginnisttx, dhellmann, and I all had some good discussions around this. :)18:54
*** openstackstatus has quit IRC19:27
fungithere were quite a few people who read the draft of that too, making sure it didn't sound _too_ frightening19:35
*** openstackgerrit has joined #openstack-tc19:36
openstackgerritSean McGinnis proposed openstack/governance master: Mark Cinder policy in code as done
dhellmanndims : yeah, I'm just giving ttx a hard time; we talked about this a bit last week20:05
*** openstackstatus has joined #openstack-tc20:28
*** ChanServ sets mode: +v openstackstatus20:28
*** thingee has quit IRC20:43
*** johnthetubaguy has quit IRC20:43
*** thingee has joined #openstack-tc20:44
*** johnthetubaguy has joined #openstack-tc20:45
*** marst has quit IRC22:24
*** harlowja has quit IRC22:54
*** diablo_rojo has quit IRC22:58
*** diablo_rojo has joined #openstack-tc22:59
*** kumarmn has quit IRC23:25
*** harlowja has joined #openstack-tc23:42
*** kumarmn has joined #openstack-tc23:51
*** kumarmn has quit IRC23:55

Generated by 2.15.3 by Marius Gedminas - find it at!