[00:07] Bug #1577988 opened: Revert destroy service when machine is off [01:04] redir: ericsnow: meeting? === natefinch-afk is now known as natefinch [02:28] axw: ping [02:28] menn0: pong, in a meeting atm [02:28] axw: np... let me know when you're done please [02:51] menn0: sorry, back now [02:51] either, we suck horribly [02:51] or the last CI test run had issues [02:52] * thumper gives each one about 50% [02:52] axw: np... I think I've mostly figured it out but i'll confirm with you anyway [02:52] axw: you did the debughooks work right? [02:53] menn0: yes, long ago [02:53] axw: I'm in the final throes of integrating the SSH host key stuff and noticed the debughooks tests don't do what they appear to [02:53] axw: https://github.com/juju/juju/blob/master/cmd/juju/commands/debughooks_test.go [02:53] axw: in debugHooksTests the proxy and result fields aren't used :) [02:54] axw: I guess I just wanted be sure that I wasn't missing something before fixing [02:55] menn0: :/ they used to be [02:55] menn0: don't think you're missing anything [02:56] axw: must have been later edits that broke the tests [02:56] axw: I need to rework them anyway. i'm trying to get them to pass again. [02:57] (with my changes) [03:13] Bug #1508923 changed: Support for Azure Resource Groups

[03:43] Bug #1291201 changed: Azure provider does not use storage service's affinity group

[04:29] wallyworld: I was expecting at least an "argh!" response from someone to using sync.Cond :) [04:29] axw: i can live with it :-) [04:30] seems to do what's needed nicely [04:31] wallyworld: yeah, it's appropriate IMO. people reach for channels because they're easy, but they're not always sufficient [04:31] +1 [04:31] when you have a hammer, everything looks like a nail [04:31] wallyworld: were you looking at the restore-backup bug? ISTR you were looking at something in the area [04:31] yeah [04:48] wallyworld: thanks for review, please see reply on BuildToolsTarball PR [04:48] ok [04:50] axw: yeah, i was wondering if there were more tests. it seems ok i think === frankban_ is now known as frankban [08:54] mgz: ping [08:56] mgz: unping - emailed instead [09:01] voidspace, dimitern: standup? Am I in the wrong one? You guys missed the fire alarm test! [09:01] omw [09:01] oops sorry omw [09:02] babbageclunk: eager beaver [09:02] babbageclunk: keen bean [09:45] voidspace, dimitern: porting tests to MAAS2 http://reviews.vapour.ws/r/4766/ [09:46] babbageclunk: looking [09:49] babbageclunk: LGTM [09:49] voidspace: thanks#! [09:49] -# [09:51] babbageclunk: still reviewing btw [09:51] dimitern: d'oh, sorry [09:51] babbageclunk: np :) I should've said I did [09:51] dimitern: luckily there are blockers anyway [09:54] babbageclunk: reviewed [09:59] dimitern: thanks, fixed. How do I spell jfdi for this again? [10:00] babbageclunk: $$__JFDI__$$ [10:00] dimitern: Thanks [10:01] babbageclunk: thanks for chasing all those CI issues so quickly :) [10:02] dimitern: :) [10:29] hmm - godef stopped working for stdlib packages after upgrading to xenial. Any hints for how to make it work again? [10:33] babbageclunk: what error are you getting? [10:34] dimitern: Just eg "godef: no declaration found for sort.Strings" [10:37] dimitern: still works fine for local types and cross-package ones. [10:37] babbageclunk: ah, you might have GOROOT set (or unset - can't remember what was supposed to work) [10:39] babbageclunk: I use GOROOT=/usr/lib/go-1.6 and that works with golang-1.6 from xenial/main [10:39] dimitern: thanks, I'll try that. Wonder why it stopped. [10:41] dimitern: yay, thanks! Worked great. [10:41] babbageclunk: \o/ :) [10:43] dimitern: man, felt like I had a phantom limb for a bit. [10:45] babbageclunk: re godef not working? :) [10:47] dimitern: yeah, just got so used to hitting a key to pull up the type of a thing that when it stopped I felt like I had an itch in a place that I didn't have a place anymore. [10:48] babbageclunk: I know what you mean - I use and depend on godef all the time; I only wish it worked better for stuff in export_test.go :/ [10:51] babbageclunk: I have M-. bound to godef-jump and C-M-. to godef-jump-other-window [10:52] dimitern: Man, I love godef-jump-other-window [10:54] babbageclunk: another very useful package to add is go-eldoc btw - it shows type info in the minibuffer [10:55] dimitern: ooh, just as you move around? I've been hitting C-c C-d for that, but something that was always there would be nicer. [10:56] babbageclunk: yeah, e.g. with point at 'errors.IsNotFound' it shows 'errors.IsNotFound: func(err error) bool' [10:56] (assuming errors is already imported) [10:57] babbageclunk: most of the tricks in http://dominik.honnef.co/posts/2013/03/writing_go_in_emacs/ and its follow-up are good to know [10:59] dimitern: nice, thanks - I've got go-mode set up, but not all of these. [11:55] * dimitern steps out for ~1h [12:47] * dimitern is back [12:53] dimitern: `juju help add-subnet` mentions `juju create-subnet` - which doesn't exist? [12:55] mgz: oh, yeah - it's under a feature flag "post-net-cli-mvp", but the help doc should mention it [12:55] s/should/shouldn't/ [12:57] okay, that's a new flag to me [12:57] should it work on aws/maas at present? (create-) [13:01] mgz: no, it doesn't work.. well, the flag it changes a few things if enabled, like registering extra networking commands, but the commands themselves are not fully implemented (end-to-end) [13:05] jam, fwereade ping [13:28] jam and fwereade officially owe me a beer at the sprint [13:32] voidspace, babbageclunk, dimitern, mgz, again really well done on the maas 2.0 stuff! [13:33] alexisb: indeed! it's exciting to see all coming together so quickly! [13:45] Gah, does anyone know why strace -c -f is giving me negative times? [13:48] Bug #1576509 changed: Race in macaroon-bakery

[13:48] Bug #1576527 changed: listSuite.TestListJSON got null (showSuite too)

[13:48] Bug #1576911 changed: github.com/juju/juju/environs/bootstrap timeout (sigquit)

[13:57] * babbageclunk is out code clubbing. [14:00] Bug #1576509 opened: Race in macaroon-bakery

[14:00] Bug #1576527 opened: listSuite.TestListJSON got null (showSuite too)

[14:00] Bug #1576911 opened: github.com/juju/juju/environs/bootstrap timeout (sigquit)

[14:12] Bug #1576509 changed: Race in macaroon-bakery

[14:12] Bug #1576527 changed: listSuite.TestListJSON got null (showSuite too)

[14:12] Bug #1576911 changed: github.com/juju/juju/environs/bootstrap timeout (sigquit)

[14:19] alexisb: thanks [14:24] Bug #1578237 opened: lxd Raw response debug logging too verbose

[14:47] bogdanteleaga: got a minute? [14:49] or gsamfira? [14:50] natefinch, what's up [14:51] bogdanteleaga: I changed jujud's TLS configuration to only support TLS 1.2 last week... however, this is proving a problem for windows machines trying to download tools. Seems like the .Net WebClient doesn't always default to having TLS 1.2 enabled. [14:52] bogdanteleaga: I think it's something we can enable in the cloudconfig script we send up, but wanted to get your input on it... this code specifically: https://github.com/juju/juju/blob/master/cloudconfig/userdatacfg_win.go#L197 [14:52] yup, noticed it yesterday, got a fix, but still testing it across windows versions [14:52] bogdanteleaga: oh, awesome. Thanks! [14:54] Bug #1537153 changed: juju deploy --config option ignored when deploying a bundle <2.0-count>

[14:54] Bug #1565872 changed: Juju needs to support LXD profiles as a constraint

[14:54] Bug #1578254 opened: Race in apiserver/common and apiserver/proxyupdater

[15:02] natefinch: standup time [15:23] does anyone have the critical bugs link? [15:24] perrito666: blockers: https://bugs.launchpad.net/juju-core/+bugs?field.tag=blocker+ [15:24] tx [15:24] almost everything is in progress sweeeeet [15:24] Bug #1554819 changed: juju help create-model is misleading

[15:33] Bug #1554819 opened: juju help create-model is misleading

[15:37] does anyone oppose at me taking https://bugs.launchpad.net/juju-core/+bug/1578254 ? [15:37] Bug #1578254: Race in apiserver/common and apiserver/proxyupdater

[15:42] Bug #1554819 changed: juju help create-model is misleading

[15:43] Bug #1561212 changed: register logic can lead to user lockout <2.0-count>

[15:43] Bug #1576313 changed: windows: uniter tests fail because logs get dumped to stderr

[15:45] perrito666: go for it. I was going to look at it, but hadn't gotten there yet [15:45] sorry [15:45] * perrito666 notices that all our criticals are in progres :D go team [15:45] perrito666: it's all the same to me [15:46] \o/ [15:50] sinzui: did you get in touch with the MAAS guys about the centos7 image? [15:50] natefinch: oh, sorry, they never replied. I will poke again [15:51] sinzui: do you mind if I split https://bugs.launchpad.net/juju-core/+bug/1576873 into two bugs? One for centos and one for windows? They're likely to be fixed in different ways by different people [15:51] Bug #1576873: Juju2 cannot deploy centos or windows workloads on maas 1.9

[16:00] natefinch: that sounds sensible [16:00] mgz: there's no built-in like duplicate this bug or split this bug into two or anything, is there? [16:02] mgz: nevermind, copy and paste is quick enough [16:02] natefinch: yeah, that should do [16:04] there - https://bugs.launchpad.net/juju-core/+bug/1578286 [16:04] Bug #1578286: Juju2 cannot deploy windows workloads on maas 1.9

[16:04] also edited the original to remove the windows-specific stuff [16:04] thanks nate [16:07] Bug #1578286 opened: Juju2 cannot deploy windows workloads on maas 1.9

[16:36] ericsnow: I'm here, give me a few minutes and I am ready [16:37] redir: k [16:47] Who can answer rackspace provider questions? [16:49] ericsnow: in Moonstone [16:50] ericsnow, the rackspace development account loaned to juju from cpc has a litter of monitoring entities for old juju machines from prior deployments. These might be old but I'd like you to check and ensure that the provider does clean these up correctly if it's not doing that already. [16:50] ericsnow, I'm seeing them @ https://intelligence.rackspace.com/cloud/entities [16:52] rcj: OTP, but you might want to check with mgz, abentley_, or sinzui [16:53] ^ mgz, abentley_, sinzui: Can someone from Juju track this down? I'd like to hand this off. [16:54] rcj: what exactly is a monitoring entity? [16:54] good question [16:54] rcj: my question too [16:54] something in the UI at that spot in the webapi. we see a bunch on the test account we loaned out for use in provider development. [16:55] rcj: I delete servers every few days, but not Odd_Bloke's server [16:56] sinzui: < me, earlier> I've never used RAX. < rcj> You have a server in there. < me> ... oh ... [16:56] sinzui, not Odd_Bloke's servers. this is just some entity, probably created at vm creation, that is hanging around presumably left by the juju rax provider given the entity "juju--machine-#" names. Wanted to make sure someone was aware and would look at the provider to ensure a clean user experience. [16:56] rcj: so, we're not deliberately creating those, I guess we might get them as a side effect of creating machines? [16:57] mgz, ack [16:57] Odd_Bloke: oh, it is gone now, but there was one for 2 weeks [16:57] in which case it's unclear why the matching destroy doesn't clear them [16:57] sinzui: Yeah, I know; I am not a smart man. [16:57] :p [17:00] rcj: I have no idea what to do about this [17:01] it's not even clear from the docs how you configure the behaviour [17:01] mgz: rcj: cannot find this in the UI to get a list [17:02] mgz: I think we need a bug for juju and one for ourselves. We need a cleanup script for rackspace now [17:03] ah, rackspace inteligence [17:08] https://intelligence.rackspace.com/cloud/entities [17:13] Bug #1541445 changed: empty uuid in cache.yaml after destroy-controller <2.0-count>

[17:42] ericsnow: same thing different branch http://reviews.vapour.ws/r/4767/ [17:42] sinzui, mgz I'd might need a fix on the upstart service (1.25); What's the target branch for doing a merge proposal? [17:43] 1.25 [17:43] ^ niedbalski [17:43] Bug #1578327 opened: juju 1.25.5: Juju Login with invalid user tag should return CodeUnauthorized

[17:43] niedbalski: https://github.com/juju/juju/tree/1.25 [17:44] sinzui, yep, but afaik this is part of packaging and it seems that the debian sources are not on the same tree, right? [17:44] niedbalski: That is the branch we will make 1.25.6 from. [17:46] niedbalski: 1.25 does have different packaging rules, but to be clear NO juju has upstart or systemd rules in packaging. [17:47] niedbalski: jujud generates its system files on installation. So jujud supports both systemd and upstart. It chooses which files to generate based on the host OS. this is also true for windows and centos [17:48] ericsnow: tx [17:49] sinzui, cool. It seems that the service definition template lives in github.com/juju/juju/service/upstart/upstart.go, thanks!. [17:49] fab [17:50] balloons: r=me [17:53] sinzui, ack. I'll remove the tweaks to the jobs to call my code and enable the flow [17:54] :) [18:28] Bug #1578337 opened: no command to remove controllers [18:31] Bug #1578337 changed: no command to remove controllers [18:36] I was running `juju upgrade-juju --upload-tools' and ran into "ERROR POST https://192.168.19.150:17070/model/63ffd949-a9c4-4319-8b70-c33d505cc7ef/tools?binaryVersion=2.0-beta7.2-xenial-amd64&series=trusty,quantal,saucy,yakkety,wily,precise,vivid,xenial,raring,utopic Forbidden Port" - is this a known issue? [18:43] Bug #1578337 opened: no command to remove controllers [19:11] the true test of a tiling wm... my mouse cursor disappears =| [19:29] katco: i3? [19:29] yep [19:29] redir: yep [19:29] katco: mine doesn't [19:29] redir: doesn't what? [19:29] disappear [19:30] redir: oh, no, i wouldn't expect it to. i've been using i3 for a few years, i've just triggered some kind of bug [19:30] oic [19:31] redir: and i expect it's not a bug in i3 [19:31] my cursor disappears about once a month, but I use unity [19:31] natefinch: redir: it seems to happen when i mouse over the ubuntu terminal [19:31] natefinch: redir: not all the time, but when the bug happens that seems to be what i was doing [19:31] running the 1.x tests wipes out all open terminals or spams them with systemd messages [19:31] katco: easy fix, don't do that ;) [19:31] o.0 [19:32] natefinch: you're absolute right. i should be using the emacs terminal ;p [19:32] katco: ahhh, I have terminator set as sensible-terminal [19:32] redir: yeah, the systemd thing with the 1.x tests is super annoying [19:33] did you guys open a bug for that? [19:35] mgz: you mean for the mouse thing? [19:35] I mean for running the tests doing weird things [19:36] mgz: I think there was one open for it.... don't remember exactly [19:48] * redir thought it was a feature [19:49] ericsnow: another same thing different branch review http://reviews.vapour.ws/r/4768/ -- no major rush as the target is blocked. [19:50] brb reboot [20:35] Bug #1578373 opened: unblock panic: invalid memory address or nil pointer dereference

[20:50] moved a true unit test to a test suite that doesn't do setup or teardown, shortened test length by 2/3rds :/ [20:50] anyone could check this very brief yet critical patch? http://reviews.vapour.ws/r/4769/ [20:51] natefinch: \o/ keep doing stuff like that. we'll widdle away at this problem yet! [20:51] whittle? [20:51] i don't know how to spell that. [20:51] still sounds foreign to me :p [20:53] Bug #1578376 opened: Cannot add MAAS credentials through juju add-credential

[20:53] Bug #1578383 opened: incorrect extension on resource upload [20:53] katco: it's just one of many reasons why I don't like testsuites... it's too easy to fall into the trap of just adding another test to the existing suite, even if it doesn't need any of the setup or teardown. [20:54] natefinch: yes. it would be much better to share functionality by making shared code functions, not methods. [20:56] katco: yep [20:58] katco: I'm pretty sure 'widdle' is as in, to pee on [20:58] mgz: o.o [20:59]