[00:00] redir: did you want to talk? [00:01] sure [00:01] where? [00:04] redir: standup ho is fine with me [00:04] OK [00:04] brt [00:06] Bug #1333162 changed: cannot bootstrap on openstack provider: Multiple security_group matches found for name 'XYZ', use an ID to be more specific. [00:06]

=== stokachu_ is now known as stokachu [01:41] wallyworld, axw: can we schedule some time next week to talk about the impact of model config tree on migrations? [01:42] menn0: sure. any day after 9am is fine for me. some days I can do earlier, some I can't [01:44] axw: 9am your time right? [01:44] menn0: yup [01:44] cool [01:44] menn0: gotta do school dropoff [01:44] axw: let's wait until wallyworld responds and we'll make a time [01:45] anytime suits me [01:45] 11am works [01:46] * menn0 opens time zone planner [01:47] ok, that works for me [01:47] i'll put it in the calendar [01:49] wallyworld, axw: done [01:51] menn0: not shown up for me yet - you made it next monday? [01:51] checking [01:52] you made it last monday [01:52] :-) [01:52] wallyworld: I'm a moran [01:52] fixed [01:52] we've all been there :-D [01:52] so it i [01:52] s [01:53] https://cdn1.vox-cdn.com/thumbor/cdRv1O8Rg-_aVC7N81Z6ROrV6_A=/0x0:524x349/1310x873/cdn0.vox-cdn.com/uploads/chorus_image/image/25006307/moran.0.jpg [01:53] lol [01:54] classic [02:09] Bug #1428439 changed: retry-provisioning launches instances for containers; cannot retry containers at all

[02:09] Bug #1612478 opened: remove retry-provisiosing command

[02:31] am I doing something wrong? When I marshal from json to a generic structure in Go, I get map[string]interface{}, when I do it with gopkg.in/yaml.v2, I get map[interface{}]interface{} ... what's up with that, anyone know? [02:48] I would guess that yaml is wrong (but no, yaml does that iirc) [02:49] Menn0 I would be grateful if you publish the instructions (even if crude) for the proxy in juju, the creation of a new lxd instance takes forever [02:50] perrito666: ok, I'll try and do that today [02:50] Just a hint would be enough [02:50] I am at near 25 per bootstrap [02:51] Or more [02:55] wallyworld: can you make a new repo in github.com/juju called cloudschema? [02:55] sure [02:56] cloudschema? did we want ot more general? [02:56] it [02:56] not sure [02:57] is this for the jsonschema stuff? [02:57] right now it's not cloud config specific, but it is "juju config that we'll want to use for interactive prompts" specific [02:58] wallyworld: current API: https://godoc.org/github.com/natefinch/cloudschema [02:58] right, so i can't ser why "cloud" is relevant, it's just a schema lib [02:59] wallyworld: because there's metadata that isn't validation schema, like "secret" and metadata for generating CLI prompts [02:59] that's not cloud either [02:59] passwords are a generic concept [02:59] yes, true [03:00] clischema? [03:00] gojsonschema is taken it looks like [03:00] is it reall just CLI; the same thing could be used got GUI [03:00] true [03:00] uxschema? [03:00] what's in gojsonschema? [03:00] that's the old jsonschema we forked from someone else [03:01] it's what this will replace [03:01] that and environschema [03:01] hmmm, so we really want this new one call gojsonschema. why not gojsonschema.v2 [03:01] no I really don't want to call it gojsonschema [03:01] that's what it is [03:02] isn't it? [03:02] it's jsonschema... you don't need "go" [03:02] if you want, we can call it jsonschema [03:02] the convention with go projects is goblah right? [03:02] but jsonschema works [03:03] the problem is that it's not really just jsonschema. it's jsonschema with extra junk [03:05] hmmm [03:05] i hate naming stuff [03:07] natefinch: because it's off juju, maybe jsonschema is ok, since it's a jsonschema lib for use in juju? if it does jsonschema plus extras, i think that's ok? [03:08] it's not not jsonschema [03:08] that's fine [03:11] natefinch: done [03:12] wallyworld: sweet, thank you [03:12] wallyworld: and thanks for the input on the name [03:12] sure :-) [03:12] bikeshed :-) [03:12] names matter more than we think :) [03:14] lol [03:14] $ ls [03:14] LICENCE LICENSE README.md schema.go schema_test.go [03:14] crazy brits [03:17] natefinch: it's only you guys in the US that spell differently to the rest of the whole world [03:17] we know the difference between nouns and verbs [03:17] licence is a noun. license is a verb [03:18] advice is a noun. advise is a verb [03:18] etc [03:18] hmm, good example. [03:18] we also understand the difference between least to most significant. YYYY-MM-DD or DD-MMM-YYYY [03:18] not YYYY-DD-MM [03:19] no no, we do MM-DD-YYYY [03:19] same diff :-) [03:19] still out of order [03:19] but yes, still dumb [03:52] code review anyone? [03:52] https://github.com/juju/jsonschema/pull/1 [05:43] Bug #1612500 opened: Can't migrate back to a controller due to user permissions [06:01] wallyworld: would you please take a look: http://reviews.vapour.ws/r/5423/ [06:01] axw: sure, just got to go to school pickup, will lookin a bit [06:01] wallyworld: np, thanks [06:27] axw: so did you discover why rate limiting wasn't kicking in? or was that a red herring? [06:27] wallyworld: red herring. it is, I verified it works - I think their subscription is just so busy it never stopped getting rate limited [06:28] wondered as much [06:28] i think storage accouny key used to be a pointer rather than a slice before the sdk change right? [06:29] wallyworld: it was a pointer to a struct that had multiple keys in it [06:29] wallyworld: also it never used to have this permission thing [06:30] ah, right ok. lgtm. code looks ok, i'll take your word thst it works [06:33] wallyworld: thanks [06:33] np === akhavr1 is now known as akhavr [09:34] Bug #1603910 changed: model-level log forwarding not supported

=== akhavr1 is now known as akhavr === akhavr1 is now known as akhavr [10:47] anyone up for a review of this please? it just adds a (failing) test for some currently broken functionality. fix in subsequent PR. https://github.com/juju/juju/pull/5984 [10:54] rogpeppe: LGTM [10:54] dimitern: thanks! [11:05] dimitern: i just realised that to save CI test time, it's better if I submit the fix as well in the same PR, so I've proposed this: https://github.com/juju/juju/pull/5985/files [11:05] dimitern: that has the PR you just reviewed as the first commit [11:05] dimitern: and the fix as the second [11:06] rogpeppe: on it [11:06] dimitern: feel free to comment directly on the second commit: https://github.com/juju/juju/pull/5985/commits/9d3478263f1d5fd4be0adb038e0811c25c9641a0 [11:07] rogpeppe: ok, but where are your QA steps? :) [11:07] dimitern: check that juju still works, i guess :) [11:08] rogpeppe: that's preferable, but it should be possible to trigger the issue you're fixing? [11:08] dimitern: no, the issue only occurs in tests that I've seen [11:08] rogpeppe: or it's one of those races that's hard to trigger? [11:08] dimitern: because generally apiserver instances aren't torn down [11:09] rogpeppe: ah, so it's not observable otherwise? [11:09] dimitern: yes [11:09] rogpeppe: ok, housekeeping then ;) [11:09] dimitern: you could look at it like that, i guess [11:10] dimitern: but it would impact in any real situation where we try to tear down workers synchronously when there are outstanding blocked API calls [11:11] rogpeppe: right [11:11] * dimitern adds rogpeppe's fork as remote to run the tests [11:12] rogpeppe: wow, you're like voidspace - hardly ever deleting merged branches [11:13] dimitern: i never delete merged branches [11:13] dimitern: why bother? :) [11:13] rogpeppe: for the sake of anybody pulling regularly from your fork I guess ;) [11:13] dimitern: it'll only make a difference the first time they pull... [11:13] rogpeppe: and seeing a couple of screenfuls each time heh [11:16] Bug #1612624 opened: Bootstrap fail on MAAS if ipv6 is disabled [11:34] rogpeppe: LGTM [11:34] dimitern: ta! [11:41] mhilton: i just realised i pushed $$merge$$ having assumed you'd LGTM'd the branch [11:41] mhilton: and you haven't, sorry! [11:41] mhilton: any chance of a retrospective... ? [11:43] rogpeppe: eheh [11:43] mgz: :) [11:54] morning, happy friday [12:06] dimitern, katco: could you take another look at http://reviews.vapour.ws/r/5366/diff please? [12:07] babbageclunk: sure, in a moment [12:07] dimitern: Thanks! [12:46] here's the branch that exposed the bug fixed by my earlier PR. it splits the controller and model APIs. reviews much appreciated! http://reviews.vapour.ws/r/5426/) [12:47] dimitern: ?^ [12:50] rogpeppe: will do after I finish babbageclunk's review [12:53] Bug #1612645 opened: New EC2 region in India (ap-south-1 - Mumbai)

[12:59] rogpeppe: I'll take a look at yours while he's doing mine! :) [13:15] babbageclunk, dimitern: ta! [13:16] rogpeppe: Would it be hard to split the Api -> API change into a separate diff? It'd be a lot easier to follow the controller API / model API change. [13:16] babbageclunk: i hoped you wouldn't say that :) [13:17] babbageclunk: it would be quite hard now, unfortunately [13:17] rogpeppe: oh no - squished commits? [13:17] babbageclunk: yeah [13:18] oh well [13:18] babbageclunk: perhaps you can just slide your eyes over the spelling changes :) [13:19] rogpeppe: I guess I'll have to! :) [13:20] babbageclunk: thanks [13:20] babbageclunk: i should've thought at the time [13:23] Bug #1612653 opened: bootstrap fails in beta15 when installing via snap [13:52] rogpeppe: reviewed - looks good, just one bit of confusion on my part. [13:52] babbageclunk: tyvm [13:53] Bug #1612658 opened: LXD no activate interface is bonding is used in MAAS2rc4 [13:53] babbageclunk: OpenControllerAPI landed in a PR just recently [13:54] rogpeppe: Ah, right - thought it could be something like that. [13:58] rick_h_: I'm going to miss standup. My wife is sick in bed, so I'm with the kids. Might get some work done this afternoon, but not sure. I'll send a summary email of where I am. [13:59] and if anyone wants to do a review: https://github.com/juju/jsonschema/pull/1 [13:59] natefinch: k [13:59] natefinch: please make sure the kanban board is up to date with links/etc [14:00] natefinch: of particular interest is where the windows bu g is [14:00] katco: dimitern standup time [14:00] owm [14:01] rick_h_: I didn't do much with the windows bug, unfortunately... maybe that was a prioritization failure on my part, if so, I am sorry. I was finally getting somewhere with the jsonschema stuff, so I was trying to push through with it. [14:02] natefinch: I've promised that bug by next beta. We need to get it in. it's carried over and block OIL for multiple weeks atm [14:02] natefinch: so that bug is priority #1 please [14:02] rick_h_: right. My apologies. I'll do my best to get it done today. [14:03] babbageclunk: sorry it took me so long, but you have an initial review [14:03] babbageclunk: I'm happy to chat / clarify anything after standup [14:18] rogpeppe: reviewed yours as well, finally [14:18] dimitern: thanks! === natefinch is now known as natefinch-afk [14:28] katco: If you have five minutes now, stand-up HO? [14:28] macgreagoir: still there :) [14:28] Cool, cheers :-) [14:44] macgreagoir: well for 1, it looks like there's more recent precedent for utilizing testing.RunCommand without the goroutine [14:44] macgreagoir: so i would say at the very least, the way you started to do things is more correct [14:45] macgreagoir: e.g. take a look at cmd/juju/controller/listmodels_test.go [14:45] axw: has some very nice tests [14:47] macgreagoir: and it is done by mocking the api client, which is what i remembered doing, so i would definitely hold this up as the way to properly write tests against commands [14:47] macgreagoir: does that sound ok? [14:53] dimitern: cool, thanks [14:59] dimitern: The connected model checking was something Will asked for. Is he around today? Can't see any holiday on the calendar. [15:00] katco: Cheers [15:01] macgreagoir: hope that helps :) [15:01] (my irssi bell seems quite, sorry) [15:01] macgreagoir: lol no worries [15:01] *quiet [15:04] katco: That might be where unregister tests ideas were scraped from :-) [15:04] macgreagoir: :) the cargo culting is real. just gotta make sure you're copying the right stuff! ;p [15:14] babbageclunk: sorry, not around today [15:14] rick_h_: Stink - thanks! [15:27] babbageclunk: he may have forgotten about GetAuthTag [15:28] babbageclunk: sorry, I guess you got conflicting suggestions :/ [15:28] babbageclunk: let's have a chat with Will next week I guess.. [15:30] dimitern: I don't think GetAuthTag does what he wanted, but I can't really explain why he wanted the model check - I do feel like it's a bit redundant given the check in the ctor. [15:30] babbageclunk: ok [15:30] babbageclunk: I'll have a look at his earlier comments [15:31] babbageclunk: and possibly update mine, but a bit later (need to go out ~1h) [16:07] anyone here use juju at azure? [16:11] Bug #1612717 opened: Pinger facade not implemented on controller websocket connection [16:14] Bug #1612717 changed: Pinger facade not implemented on controller websocket connection [16:29] Bug #1612717 opened: Pinger facade not implemented on controller websocket connection [16:29] Bug #1612722 opened: Juju cannot destroy models/machine, all testing substrates are exhausted

[16:44] Bug #1612687 opened: Juju not showing real states of deployed nodes via MAAS. [16:58] katco: can you please review https://github.com/juju/juju/pull/5987 [17:10] here's some apiserver cleanup; reviews appreciated: https://github.com/juju/juju/pull/5988/files [17:25] rick_h_: sorry, was eating lunch. tal [17:27] katco: all good, decided it was a clean revert so let it rip [17:27] rick_h_: ok. redir's +1 is probably sufficient as well. verifying a revert requires no knowledge of juju :) [17:27] katco: :) [17:29] man, we really need to stop writing doc lines that are tautological [17:33] katco: can you peek at https://github.com/juju/juju/pull/5988/files then which is more interesting? [17:35] rick_h_: ok [17:35] katco: ty [17:38] Bug #1612744 opened: MachinerSuite.TestMachinerMachineAssignedUnits agent should be terminated

[17:38] Bug #1612745 opened: rackerSuite.TestGainLeadership number of ClaimLeadership is wrong

[17:38] Bug #1612747 opened: DebugHooksServerSuite.TestRunHook signal: killed

[17:46] * rick_h_ gets some lunch with that done [18:09] ejat: i use juju/azure [18:10] kwmonroe: okie .. its working now .. thanks [19:11] bbiab [19:19] * rick_h_ runs to get my truck and the boy from summer camp. have a good weekend folks [19:30] Bug #1612775 opened: resource get fails, resource already staged [20:02] it would be extremely nice of us to name the same thing in the same way across facades [20:02] perrito666: novel idea [20:03] authorizer and state are called something different across all facades [20:04] and just found one without authorizer [20:13] sinzui: are you still having problems with that windows deploy CI test? I seem to be able to deploy a windows charm just fine on azure [20:13] natefinch: I think not. I am still wating for one to complete to say all is well [20:13] natefinch: once juju broke the reate limit. azure locked our for many hours [20:13] sinzui: great :) [20:14] oops [20:19] natefinch: I see some azure successes. most failed though.I suspect the failure were at the tail end of the lockout. I will retest. [20:27] sinzui: hey does qa have any plans to release the qa tools as a snap? [20:27] katco: we expect to do dailies [20:27] sinzui: daily releases of the qa tools? [20:28] katco: oh, no. no plans to snape jujuc-ci-tools or juju-release-tools [20:28] sinzui: sorry, jujuc-ci-tools [20:28] sinzui: having just gone through the setup process, it would be awesome to just download a snap [20:29] maybe balloons will want try it. Since snap doesn't work with windows, centos, or osx, we are dedicated to working with a common solution. [20:30] sinzui: ah good point [20:30] Bug #1612793 opened: bootstrap azure/southcentralus is failing in 2.0-betaX [20:39] ... obtained int = 2 [20:39] ... expected int = 2 [20:39] evidently [20:44] yay my recator compiles for the 1st time [20:44] refactor [21:39] Bug #1612335 changed: Azure rate limit leads to catastrophic failure for subscription

[21:39] Bug #1612722 changed: Juju cannot destroy models/machine, all testing substrates are exhausted

[21:48] Bug #1612836 opened: Cannot bring up hosted model machines in azure

[22:14] rick_h_: ping, don't suppose you're on [22:34] mmm, ppas dont work nicely with my deb proxy