Friday, 2018-05-04

*** kumarmn has joined #openstack-tc00:14
*** kumarmn has quit IRC00:40
*** kumarmn has joined #openstack-tc01:10
*** spsurya has joined #openstack-tc01:14
*** kumarmn has quit IRC01:15
*** rosmaita has quit IRC01:35
*** mrjazzercise has quit IRC01:36
*** ricolin has joined #openstack-tc01:44
*** kumarmn has joined #openstack-tc01:58
*** kumarmn has quit IRC02:03
*** kumarmn has joined #openstack-tc02:35
*** kumarmn has quit IRC02:40
*** kumarmn has joined #openstack-tc03:12
*** kumarmn has quit IRC03:17
*** kumarmn has joined #openstack-tc03:39
*** kumarmn has quit IRC04:01
*** kumarmn has joined #openstack-tc04:31
*** kumarmn has quit IRC04:35
*** ianw has quit IRC05:06
*** kumarmn has joined #openstack-tc05:07
*** kumarmn has quit IRC05:13
*** TheJulia has quit IRC05:15
*** hogepodge has quit IRC05:15
*** TheJulia has joined #openstack-tc05:15
*** hogepodge has joined #openstack-tc05:16
*** kumarmn has joined #openstack-tc05:46
*** kumarmn has quit IRC05:51
*** ianw has joined #openstack-tc05:55
*** kumarmn has joined #openstack-tc06:19
*** kumarmn has quit IRC06:25
*** dmsimard has quit IRC06:39
*** dmsimard has joined #openstack-tc06:40
*** kumarmn has joined #openstack-tc07:21
*** kumarmn has quit IRC07:26
*** kumarmn has joined #openstack-tc07:28
*** kumarmn has quit IRC07:34
*** bauzas is now known as bauwser07:37
*** kumarmn has joined #openstack-tc07:43
*** kumarmn has quit IRC07:48
*** jpich has joined #openstack-tc07:51
*** kumarmn has joined #openstack-tc08:01
*** kumarmn has quit IRC08:05
*** kumarmn has joined #openstack-tc08:32
*** kumarmn has quit IRC08:37
*** lbragstad has quit IRC08:45
*** lbragstad has joined #openstack-tc08:45
*** kumarmn has joined #openstack-tc09:07
*** kumarmn has quit IRC09:12
*** dtantsur|afk is now known as dtantsur09:33
*** spsurya has quit IRC10:56
*** ricolin has quit IRC11:19
*** kumarmn has joined #openstack-tc11:21
*** kumarmn has quit IRC11:25
*** rosmaita has joined #openstack-tc11:53
*** kumarmn has joined #openstack-tc11:55
*** kumarmn has quit IRC12:00
*** dtantsur is now known as dtantsur|brb12:17
*** kumarmn has joined #openstack-tc12:40
*** kumarmn has quit IRC12:45
*** ricolin has joined #openstack-tc12:52
*** kumarmn has joined #openstack-tc12:52
*** kumarmn has quit IRC12:53
*** mriedem has joined #openstack-tc13:06
*** dtantsur|brb is now known as dtantsur13:09
*** kumarmn has joined #openstack-tc13:09
* ttx is back home from Copenhagen13:32
ttxdhellmann: re the bylaws change -- last time I talked to Mark radcliffe about it, he said that was the right way of saying it in legalese13:33
ttxI had to cede since legalese not english are my native languages13:34
dimsproductive trip ttx ? (copenhagen)13:37
ttxCNCF just reached Paris summit stage13:37
mugsiettx: was there a HPE party? :P13:38
ttx(they had 4300 attendees, we had 4600)13:38
ttxmugsie: there was a party in an amusement park yes13:38
ttxalthough the park was not really privatized (it'13:39
ttxs huge)13:39
mugsiettx: how did the release / growing pains / community etc chat go?13:42
smcginnismugsie: I thought they went well.13:48
smcginnisHard getting a large turnout, just like us trying to do something extra at the Summit, but there was pretty good attendance.13:50
mugsiecool. yeah, it is not a topic that a huge amount of people would go out of their way to engage with13:50
ttxWe discussed the trust scaling thing -- in particular the neat trick they have with horizontal SIGs and vertical local OWNER files...  Apparently K8s wants to break that by letting SIGs own repos13:53
ttxWhich may introduce the silo / lack of cross-coordination issues that we are encoutering13:54
ttxWe touched on various other topics, but we just did not have enough time13:55
zanebttx: I don't speak legalese, but it clearly makes no sense in English fwiw13:57
smcginnisOne difference with the SIG ownership is they have the OWNERS files that give them some finer grained control over areas of code.13:58
smcginnisAt least until they break apart the mono-repo.13:58
mugsiethey have started down that path, havent they?14:00
smcginnisStarted, but still a lot of unknowns.14:01
smcginnisThey know they need to do it, but not clear yet what the best way is to break it apart.14:01
smcginnisThey are moving out things like drivers and cloud providers, but there's still a lot of code left after that.14:01
ttxdhellmann: are you planning to post a TC status update, or should I ?14:15
ttxNot much merged last week...14:15
*** melwitt is now known as jgwentworth14:15
ttxAlso passed the required time for approval. I can W+1 it if you want14:16
ttx is pretty close too14:18
jrolllbragstad and I are having trouble tracking down approval process for the openstack-specs repo, does anyone remember or have a pointer to docs on that?14:23
*** hrybacki has joined #openstack-tc14:27
dhellmannttx: if you have one last one ready to go that would be good; otherwise I can do it later today14:33
dhellmannttx: I need to get those scripts you have for identifying things that are ready to be approved so I don't have to do all of that date math by hand14:34
dhellmannjroll : there used to be a group led by thingee that managed that, but it's mostly defunct now. we might need to revive it if you think there's something that needs to go through a global specs process.14:34
ttxI have nothing ready to go :)14:34
dhellmannttx: ok, I'll pull something together then14:35
ttxYou could also totally decide that Friday is not the day you'll post that :)14:35
dhellmannthat's true, maybe that should be a monday thing :-)14:36
jrolldhellmann: lbragstad had this spec, would be good to have buy-in, I'm not sure how formal we want it to be:
lbragstadwe contemplated the idea of making it a community goal, but it's a lot of work and not likely something a single project would get done in a release14:39
dhellmannI hate to derail you while we figure out what some sort of approval process might look like14:39
lbragstadthe other option we have it tags14:39
dhellmannI'm coming to this cold (I saw that you were working on it, but haven't read it). Can you describe why you think it's a "global" spec instead of a keystone spec?14:40
* jroll has to run for a bit but will read back14:41
lbragstadthanks for the assist jroll :)14:41
jrollno prob :)14:41
lbragstadwe've had a bunch of efforts recently to move project policies into code and provide documentation14:41
lbragstadwhich let's projects leverage all the work we've been doing on oslo.policy14:42
lbragstadlike being able to deprecate policies in favor of other ones, rename policy rules, and so one without breaking users or deployers14:42
lbragstad(pretty much the same process that we use for configuration values)14:42
*** hongbin has joined #openstack-tc14:43
lbragstadnow that nearly all of the infrastructure is in place, we have the tools to introduce a set of default roles available out-of-the-box14:43
dhellmannthat all sounds good14:43
lbragstadthat should all behave consistently across services14:44
lbragstadwhich should hopefully lead to easier use for operators because they don't have to reinvent wheels just to get a read-only role =/14:44
lbragstadand, ideally, inter-op would benefit, too14:44
lbragstadit also might make it easier for deployments with strict security requirements to adopt openstack because it offers things like this by default, and so on14:45
dhellmannall of that sounds good and like something we should be done. can the keystone team drive the changes needed?14:46
lbragstadso - the big drive in Dublin was to try and come up with a set of roles that we should offer by default14:46
lbragstadwell - we've introduce a bunch of the tooling to oslo.policy and done most of the bits in keystone14:47
lbragstad(e.g. system scope and system role assignments)14:47
lbragstadnow we're at the point where we need to start helping other projects consume it14:47
lbragstad(i have a few patches up to nova to get things started)14:47
lbragstadbut - with that, i assume we'll need consensus on the overall approach and buy-in.14:48
lbragstadwhich is where the spec comes in14:48
dhellmannwho do you imagine would be in the group who would approve that?14:48
lbragstadany project looking to offer those roles by default?14:48
dhellmannok, I'm trying to figure out a way to do this that doesn't rely on us setting up a group that will have trouble meeting and agreeing (as the global specs thing did in the past)14:49
lbragstadfeel free to poke holes here14:49
dhellmannso I wonder if the keystone team just continues to own it, but goes to each of the other major teams looking for input, how well might that work?14:49
lbragstadthat might work14:49
lbragstadthe other approach that i thought was interesting was project tags14:50
dhellmannthis would have been ideal as a forum topic14:50
dhellmannttx, do we have space to add ^^ to the forum?14:50
dhellmannah, good14:50
lbragstad but it doesn't appear to be working14:50
dhellmannso then we need to make sure as many teams as possible end up represented in that room14:51
lbragstadi was also wondering if it would make sense to have a tag called asserts:basic-rbac14:51
dhellmannso I think a keystone-spec, with lots of solicited input, followed by implementation in at least one project and then a community goal is the way to push this through14:51
dhellmanna tag is a good intermediate step, sure14:51
dhellmannbetween getting 1 project done and having the goal14:52
lbragstadi suppose if we end up making it a goal, then that defeats the purpose of the tag14:52
hrybackilbragstad: can we submit a spec past freeze in this instance?14:52
hrybackidhellmann: an initial goal is to do the work in Keystone and Barbican and use them as examples14:53
lbragstadhrybacki: i wouldn't have a problem with it because it's technically not a "new" specification in terms of the keystone team or buy-in from the keystone team14:53
* hrybacki nods14:53
dhellmann lbragstad : right, we would use the tag at first and then drop it when we have the goal done14:53
lbragstadbut i'm inclined to open it up on the mailing like to make sure people are OK with it, and we can formally document the exception14:53
zaneblbragstad: would getting it prototyped in some projects for Rocky and then community goal for Stein be achievable do you think14:53
dhellmannalthough going to the trouble of setting up the tag might be a waste if we're only going to drop it later14:54
lbragstaddhellmann: ++ the more you say that the more i agree14:54
lbragstadzaneb: i think a community goal is going to be an insane amount of work and it's going to require a lot of coordination14:54
lbragstadbut... i wouldn't be opposed to trying it14:54
lbragstadi just fear that it would stretch multiple releases14:55
zanebworse than mox? ;)14:55
hrybackilbragstad: so with regard to testing -- would we want to bring Felipe's spec back into keystone as well?14:55
dhellmannI think we want to hold back a bit on planning big goals for stein other than finishing the python 3 work. that deadline is looming large.14:55
zanebso maybe we publicise that this is likely to be a goal for T and try to get people working on it sooner if they can?14:56
hrybackizaneb: I think that is a sane approach. It will give us time to work out proper testing w/ Patrole as well14:57
dhellmannyeah, we would want to settle on the roles and have a PoC before making it a goal, so that seems good14:57
lbragstadi'm basing most of my estimates on the work we did for the policy and docs in code goal15:04
lbragstadand i have a feeling changing default policies to be more sensible is going to be less trivia15:04
lbragstad1.) because it involves changing things 2.) requires additional testing15:04
lbragstadwhere as the policy in code goal was really just porting things from a file to code15:05
dhellmannlbragstad : yeah, that seems like a reasonable assumption15:05
ttxdhellmann: we have a full room available all day on Tue/Wed/Thu for additional topics / continuation discussions15:08
dhellmannttx: I think lbragstad pointed to a session already on the schedule, so we're set15:08
ttxso yes we can add, the trick being to find a time where people can15:08
mnaseri think what lbragstad is bringing up is quite useful, just to throw in my operator hat, rbac right now is a bit of pain and has really weird issues15:23
mnaserespecially when it affects things like this - - which renders the whole idea of domain/projects/sub-projects pretty much un-usable15:24
openstackLaunchpad bug 968696 in OpenStack Identity (keystone) ""admin"-ness not properly scoped" [High,In progress] - Assigned to Adam Young (ayoung)15:24
mnaser(fun bug to go through that's there since 2012 if anyone wants to have a look)15:24
lbragstadmnaser: :)15:29
lbragstadyeah - i'm hoping the adoption of scope_types fixes that15:29
mnaserbut the only way it can be really fixed is with by all projects adopting it, if a single project doesn't adopt it, then it makes this whole thing a bit useless unfortunately.15:30
lbragstadit's a massive amount of work15:30
lbragstadwhich is where the tooling helps15:31
mnaserhow do we not break existing deployments is the hard thing as well15:32
mnaseri don't know how far you've gotten about nailing that down15:32
lbragstad^ that's the one reason why i'm grateful we have policy in config15:32
lbragstaddespite defining policy rules in json being a total pain and not very ergonomic15:32
lbragstadit gives operators a way to say15:33
lbragstad"nope, i don't want the new thing, i want my custom policy that i've developed over the last X years"15:33
lbragstador whatever15:33
lbragstadso - they keep all of that in their policy files and they never turn on
mnaseri guess the blocker is no one has put up the time to push these fixes down to projects?15:34
lbragstadkinda - ayoung did a bunch of work on this, but it's such a massive change that it's kind of a bear to do by yourself15:35
lbragstadso i'm hoping a PoC and clear documentation helps us make it easier for people to grasp15:36
lbragstadand enable them to work on it15:36
lbragstadwe've also got a bunch of things available to make processing this stuff for consuming project easier (the hand off of token scope between keystonemiddleware and oslo.context)15:43
*** dansmith is now known as superdan15:45
*** EmilienM is now known as EvilienM16:09
*** jpich has quit IRC16:31
*** dtantsur is now known as dtantsur|afk16:32
fungias someone popped up in the infra channel a little while ago asking, i ran some numbers to get an updated core reviewer count17:00
fungiat present count we have 760 distinct people with permission to approve changes to one or more of the 801 deliverable repositories for official openstack project teams17:00
fungii don't have numbers on how many have actually exercised it over some recent period, that's just determined by analyzing permission via acls17:01
dhellmannfungi : do you have a script somewhere to pull those numbers?17:10
dhellmannI was considering adding info to my stats tool to indicate whether a person was a core reviewer on any project17:10
fungidhellmann: it's but needs to work at the moment17:10
dhellmannalthough my current analysis doesn't really need to take that into account17:10
funginote that the term "core reviewer" is sort of a nebulous, social term we've been using which doesn't directly map to gerrit's configuration. the closet approximation i can analyze is accounts included (directly or transitively) in groups with the ability to approve changes on some branch of a deliverable repo for an official team17:12
fungier, s/closet/closest/17:13
dhellmannyeah, that's what I would have done, too17:13
dhellmannI'm trying to figure out how to count +1 votes in a useful way17:13
dhellmannI thought if the person giving +1 had +2 rights on *any* repo, then they have shown that they understand that spamming +1s doesn't gain trust17:14
fungiwith that script and maybe some minor tweaks you could indeed map a specific vote to whether or not the reviewer has approval rights on the repo associated with it17:14
dhellmannI also thought if the person had previously given a -1 and followed up with a +1 that would indicate they had actually reviewed the change17:14
fungiyeah, that's a useful heuristic17:14
dhellmannalthough I haven't figured out if I can get the historical -1s out of the gerrit data, yet17:15
fungiperhaps only if the -1 and +1 don't appear on the same patchset17:15
dhellmannI need to experiement with that more17:15
dhellmannoh, yeah, that's a good check17:15
fungiand you can get historical patchset data out of the changes methods in the api, you just need to pass some extended options i think17:16
fungialso, that will probably slow down your queries by orders of magnitude17:16
dhellmannright now the tool works by running a query and storing copies of the results locally, then to process the results it loads everything from the local cache17:19
dhellmannthe idea was that we might have a list of arbitrary review IDs, not linked by common topics or matching a query that was easy to express17:20
dhellmannso there are 2 steps in the process "download a bunch of data" and "process the data"17:20
*** ricolin has quit IRC17:49
*** rosmaita has quit IRC18:22
*** kumarmn has quit IRC19:00
*** rosmaita has joined #openstack-tc19:00
*** kumarmn has joined #openstack-tc19:01
*** kumarmn has quit IRC19:11
*** kumarmn has joined #openstack-tc19:12
*** kumarmn has quit IRC19:17
*** kumarmn has joined #openstack-tc19:17
*** kumarmn has quit IRC19:33
*** kumarmn has joined #openstack-tc19:34
mnaserforgive me, still a little new for all of this.  my travel plans shifted and i'm able to attend the leadership meeting, however, there's no listed location.  has that not been decided yet or is it somewhere that I can't find -- ?19:38
*** kumarmn has quit IRC19:38
clarkbmnaser: it is typically in or near the conference venue19:38
mnaserclarkb: i see, will an email be sent out sometime to announce more details or is that listed somewhere that i can't find?19:39
clarkbmnaser: ya usually Alan Clark sends email iirc. I'll ask around about it19:39
mnaserclarkb: thank you19:39
dhellmannmnaser : yeah, we have a location, let me find that info and share it with everyone19:42
dhellmannoh, wait, maybe that was something else19:42
*** kumarmn has joined #openstack-tc19:42
clarkbI've pinged foundation side too and asked if we can update the wiki if it is known19:42
dhellmannyeah, sorry, I was thinking of the diversity happy hour19:43
*** kumarmn has quit IRC19:44
*** kumarmn has joined #openstack-tc19:45
*** kumarmn has quit IRC19:47
*** kumarmn has joined #openstack-tc19:48
*** kumarmn has quit IRC19:48
*** kumarmn has joined #openstack-tc19:48
*** kumarmn has quit IRC19:57
*** kumarmn has joined #openstack-tc19:58
*** kumarmn has quit IRC19:58
*** kumarmn has joined #openstack-tc19:58
*** kumarmn has quit IRC20:10
*** kumarmn has joined #openstack-tc20:17
*** zaneb has quit IRC20:17
mugsieIs there info on the BoD/TC/UC dinner yet?20:18
*** kumarmn has quit IRC20:18
*** kumarmn has joined #openstack-tc20:18
dimsmugsie : haven't seen anything yet20:43
dhellmannmugsie : ttx thought it was going to be saturday evening due to a conflict with a diversity event on sunday, but I don't think we have times or locations yet20:43
dhellmannI'm holding off on rebooking my flight until I have confirmation20:44
EvilienMI think ttx was expected me to help on the location20:44
dhellmannas it is I expect to land at 6:00 Saturday20:44
EvilienMand I was about to ask what kind of place we aim for20:44
EvilienMbecause I'm a bit of a foody here and I can really give different options20:44
dhellmannI would have thought the foundation would be organizing that20:45
dhellmannEvilienM : do you mean what sort of food?20:45
* dhellmann would prefer a place where seafood was not the only option20:46
EvilienMif we're looking for a fancy place, I know where the food (french) is amazing20:46
EvilienMand no loud music, really nice to speak20:46
dhellmannis that going to be good for a group of 50+?20:46
dhellmann24 board members; 13 tc; ? UC; ? foundation staff20:47
dhellmannof course this late we don't know how many of those will be coming, do we20:47
EvilienMon my list there is also which is one of the best restaurants in town IMHO20:47
EvilienMfor 50+ ?? wow. bacchus not sure then20:47
*** kumarmn_ has joined #openstack-tc20:49
dhellmannyou should get a head-count from ttx20:49
dhellmannI just know these have been big in the past20:49
dhellmannEvilienM : what's good for breakfast?20:49
EvilienMdhellmann: ahah I have what you need20:50
EvilienMboth are my favorites20:51
jrollyou can't post these links when I'm hungry already :|20:52
EvilienM is also really good, we loved it20:52
*** kumarmn has quit IRC20:52
EvilienMjroll: really I love the city for its food diversity20:52
* dhellmann thinks he should have planned to stay further into the city so these places were on the way to the convention center20:53
EvilienMdhellmann: you're fan of scotch iirc?20:55
dhellmannsome, yes20:55
EvilienMthere is a scotch place, let me find it, I went there one time it was really cool20:55
dhellmannit looks like the bar where we had the tc dinner last time is closed now; that's too bad20:57
dhellmannthey were really cozy20:57
EvilienMthis one, I really like the place too20:57
EvilienMthey have all sorts of scotch afik20:57
dhellmannI'll just plan to follow you around all week :-)20:58
EvilienMwe'll finish at my place opening bottle of french wine that I brought from home ;-)21:02
dhellmannsounds perfect21:02
dhellmann /me has to go figure out dinner for tonight now21:03
*** mriedem is now known as mriedem_yard21:13
*** kumarmn_ has quit IRC22:26
* mugsie is landing at 17:30 on Sat - this should be fun :)22:39
*** rosmaita has quit IRC22:41
*** hongbin has quit IRC23:01
*** kumarmn has joined #openstack-tc23:02
*** kumarmn has quit IRC23:03
*** kumarmn has joined #openstack-tc23:03
*** mfedosin has quit IRC23:25
*** kumarmn has quit IRC23:58

Generated by 2.15.3 by Marius Gedminas - find it at!