[00:08] hey perrito666, did wallyworld ping you yet about opening bugs for providers with know JES issues? (this is the destroying multiple environments with the same name issue that was brought up in Oakland) [00:09] mpontillo: you missed the fun :) most of it was over in #maas [00:09] mpontillo: it was regarding bug 1534795, if you're interested in catching up [00:09] Bug #1534795: unit loses network connectivity during bootstrap: juju 1.25.2 + maas 1.9

[00:09] cherylj: nope, in which provider? [00:10] perrito666: all of them with known issues :) [00:10] perrito666: the idea is we open bugs so we can point to them as known issues for 2.0-alpha1 [00:15] cherylj: thanks; I read through the bug. will wait to hear frobware's findings [03:03] thumper: for model migrations, I need to store the connection and auth details for the target controller [03:04] thumper: will suggested I made the API that creates the document which holds this stuff take something similar to api.Info [03:05] thumper: and I'm tempted to take a tag instead of a username, in case we end up using machine accounts at some stage [03:05] thumper: is storing a tag in the DB a no no? [03:07] hmm... [03:07] * menn0 wonders about going back to just accepting usernames [03:07] we haven't so far, but I think that this case may be fine [03:07] it does simplify some other aspects [03:07] for storing credentials [03:08] thumper: it's more flexible if any tag is allowed [03:08] agreed [03:08] i'll do that then === akhavr1 is now known as akhavr [07:54] Bug #1536025 opened: LXD provider unable to setup configuration correctly

[08:00] jam: imagemetadata worker patch is ready to go... waiting for master to unblock... this should fix the above lxd difficulty \o/ [08:03] anastasiamac: so is that bug a duplicate then? [08:03] well, not that one, the previous one [08:03] bug #1536019 [08:03] Bug #1536019: LXD provider causes imagemedata worker to bounce

[08:04] jam: i think so... for some providers, this worker should not run... this is what this patch does :D [08:08] anastasiamac: well, fwiw, i'd love it if we knew about the images that were used for LXD so we didn't have to state in "environments.yaml" that you need to run lxd-images import [08:08] anastasiamac: we could do the import ourselves, and know when we need to update the base image *and* the templated ones (if we go that far) [08:09] jam: sounds reasonable \o/ [09:06] dooferlad, ping. sync today? [09:11] frobware: sorry, yes, still able to talk? [09:11] yes [09:46] frobware: http://pactcoffee.com/spread/JAMES-CMOCKA@cp [09:46] frobware: you know, just in case [10:12] hehe, the coffee freak. morning btw [10:42] dooferlad, heh. fancy a quick HO, re dns-nameservers. :) [10:42] morning TheMue! [10:42] frobware: sure [11:28] anastasiamac: wallyworld: did we end up leaving deep debugging code in? I'm seeing a lot of "STORAGE FRONT-END" and "STORAGE BACK-END" code triggering. [11:28] but the content doesn't seem useful [11:28] as it is debugging a bunch of pointer objects [11:28] jam: not on purpose [11:29] jam: looks like it [11:29] facadeCaller: base.facadeCaller{facadeName:"Block", bestVersion: 1, caller: *pointer, closer: *pointer) [11:29] must have slipped through review [11:29] wallyworld: anastasiamac: k. prob should be fixed for at least alpha2, not sure if we've missed alpha1 [11:29] the hope is we got a bless this morning [11:29] will make alpha2 for sure [11:29] I did see a bless === ashipika1 is now known as ashipika [12:47] cherylj, beisner: am in the process of writing a conclusion to #1534795. the TL;DR is I don't believe what we've done is broken. [12:47] Bug #1534795: unit loses network connectivity during bootstrap: juju 1.25.2 + maas 1.9

[12:54] frobware: awesome, thank you! [13:01] fwereade, jam, voidspace meeting [13:03] alexisb: omw [13:39] perrito666: answered your question about that drink called Ayran [13:40] TheMue: you drank that? [13:42] perrito666: yes, very refreshing. yogurt and milk with a little bit salt and then beat until frothy [13:42] I should try [13:42] perrito666: it's a national drink in Turkey [13:42] perrito666: they had no beer there, so I asked for something authentic [13:49] dooferlad: tell me more about this ethernet thing you talk of... [13:49] voidspace: It is like WiFi, but with wires. [13:49] dooferlad: my wifi never works either... :-) [13:50] voidspace: :-) [13:50] not entirely true... [13:50] * voidspace lurches to lunch === beisner- is now known as beisner === StoneTable is now known as aisrael [14:46] Bug #1533750 changed: 2.0-alpha1 stabilization [14:49] Bug #1533750 opened: 2.0-alpha1 stabilization [14:52] Bug #1533750 changed: 2.0-alpha1 stabilization [14:55] Bug #1533750 opened: 2.0-alpha1 stabilization [14:58] Bug #1533750 changed: 2.0-alpha1 stabilization [14:59] Can I get a quick review? http://reviews.vapour.ws/r/3578/ [15:12] cherylj, beisner: my observations are in https://bugs.launchpad.net/juju-core/+bug/1534795/comments/31 [15:12] Bug #1534795: unit loses network connectivity during bootstrap: juju 1.25.2 + maas 1.9

[15:25] Bug #1536215 opened: 1.25.0: deployment times out - system is deployed successfully by maas 1.9 but juju state never transitions from pending

[15:25] Bug #1536230 opened: backups create help text is wrong about downloading

[15:58] beisner, ping [16:01] Bug # changed: 1497312, 1517611, 1525280, 1527349, 1533469, 1533751, 1533792, 1533849, 1536215 [16:03] abentley sinzui - have we tested a 1.20 => 1.25 upgrade? [16:04] lazypower: yes it does work. [16:04] awesome, i recall 1.22 had some problem child days, so this is great news [16:05] lazypower: Yes, we test an upgrade from 1.20 for everything: http://reports.vapour.ws/releases/3522/job/aws-upgrade-20-trusty-amd64/attempt/508 [16:05] thanks for the confirmation sinzui [16:06] awesome, thanks abentley sinzui - you just calmed a dusty old charmers fears :D [16:09] lazypower: Though the script exits early if the target is 2.0+ because upgrades from 1.20 -> 2.0 are not supported-- you have to go through 1.25 [16:09] abentley makes total sense. this is for a user upgrading from 1.20 => 1.25 [16:26] frobware, pong. in a mtg atm. [16:26] beisner, ack. ping me when you have some time; would like to discuss your concerns. [16:37] Bug # changed: 1390585, 1486553, 1527681, 1528217, 1531064, 1534353, 1534795 [16:37] Bug #1536215 opened: 1.25.0: deployment times out - system is deployed successfully by maas 1.9 but juju state never transitions from pending

[16:42] frobware, so i think the lowest common denominator in the topic is this: bug 1536262 ...and that is why we see what we see in 1.25.3. thoughts? [16:42] Bug #1536262: dns-* e/n/i are misplaced [16:55] hi cherylj, frobware, if we can get 1.25.3 into ppa:juju/proposed, that will unblock our openstack CI metal lab. is that something that is planned in the next day or so? === tasdomas` is now known as tasdomas [16:57] beisner: yes, we're aiming for today. [16:57] beisner: we just wanted all the dust to settle on the questions around DNS and MAAS [16:57] cherylj, frobware - sweet. yeah so there is still dust but given a sudden realization that indentation in the e/n/i file is meaningless, it's old dust. [17:48] natefinch: is this what you meant? http://reviews.vapour.ws/r/3579/ [18:08] ericsnow: yep :) [18:13] natefinch: thanks for pointing that out :) [18:14] ericsnow: no problem. I've done the exact same thing a half a dozen times... write a whole bunch of code and then realize half of it is covered in the stdlib. === marlinc_ is now known as marlinc === _stowa_ is now known as _stowa [18:55] Bug #1536324 opened: juju backups restore times out trying to connect to API server

[19:29] Bug #1536324 changed: juju backups restore times out trying to connect to API server

[19:29] Bug # opened: 1536333, 1536336, 1536337, 1536340 [19:59] Bug #1536345 opened: backups create and backups restore defaults clash

=== mbarnett` is now known as mbarnett [21:01] was the LXD provider available in 1.25 behind a feature flag? [21:02] or was it just 1.26/2.0 ? [21:02] katco: ^^? [21:02] thumper: just 1.26/2.0 [21:02] ta [21:02] thumper: np. specifically 1.26-alpha2 [21:02] k, [21:16] thumper: I very much doubt that machine-dep-engine is going to merge this week === menn0_ is now known as menn0 [21:17] thumper: there are unblessed changes in master which conflict [21:18] thumper: so I have to wait for them to get blessed and then merge and fix the conflicts and then get machine-dep-engine blessed [21:18] thumper: there aren't enough days left in the week for that [21:18] ugh [21:21] thumper: I might merge machine-dep-engine right up to latest master in the hope that it'll get blessed. It might save someone less familiar with the changes in the feature branch from having to deal with the conflicts. They look non-trivial. [21:21] (next week) [21:25] * thumper needs moar coffee [21:29] OT is any of you in the go slack channel, I filled the form to get the invite but didnt :( [21:32] perrito666: I can ping the maintainers about it [21:39] hey davecheney, are you still looking for some bugs to take? [21:39] perrito666: https://gophersinvite.herokuapp.com/ [21:44] * thumper stops procrastinating and dives into the code [21:56] natefinch: oh, that changed [21:57] perrito666: evidently === natefinch is now known as natefinch-afk [22:08] natefinch-afk: tx btw, I am in [22:11] Bug #1536378 opened: resolver loop error in maas-1_8-OS-deployer test

[22:14] Bug #1536378 changed: resolver loop error in maas-1_8-OS-deployer test

[22:23] Bug #1536378 opened: resolver loop error in maas-1_8-OS-deployer test

[22:28] WTF!!!! [22:28] I thought that we stopped panicing on unknown series? [22:29] I'm using version.MustParseBinary with "3.4.5-zesty-amd64" [22:29] and it panics! [22:30] mm, I thought it too [22:37] that's pretty disappointing [22:42] thumper: i can work on that if you like [22:42] but if you use the Must form, it will panic [22:42] that's why the Must form exists [22:42] if you don't want a panic, don't call mustParseBinary [22:43] I didn't expext ParseBinary to return an error on unknown series [22:43] that's fine, just gone with "trusty" [22:43] thumper: what should it do if the series is unknown ? [22:43] davecheney: wrt work, cherylj had a particular bug I think [22:43] davecheney: not sure... [22:44] it would beunwise to return something like "2.0.5-unknown-unknown" [22:44] I thought it would have just created a binary with a series it didn't know about [22:44] like "zesty" [22:44] 'cos lots of caller expect that any version value is accurate [22:44] and not find tools etc [22:44] hmm, that makes senes [22:44] that would fit with the notion of Parse [22:44] however that isn't important just now [22:44] I'm working around it [22:44] it just parses the form provided [22:45] if you want something else, ParseValidSeries ? [22:45] ^ guess [22:55] davecheney, thumper, katco: I notice I'm invited to this core leads call in 5 minutes.... is that on purpose? [22:55] natefinch-afk: did you go to the earlier one today? [22:55] same question [22:56] natefinch-afk: it's for if you wish to ask questions about the sprint [22:56] katco: oh [22:56] alexis combined normal weekly team call with team lead call [22:56] for debrief [22:57] natefinch-afk: thanks for the reminder [23:03] natefinch-afk: coming? [23:04] katco: I can come for a bit