=== menn0 is now known as menn0-afk [00:12] [00:31] davecheney: check now [00:32] * thumper did something [00:33] thumper: ta === menn0-afk is now known as menn0 [01:40] thumper: got it, i'll fix that stupid typo [01:55] wallyworld: just looking at the createmodel command, the new credential code is not entirely correct. there's no requirement that the attr names match between credentials and provider config. also, the provider might need to do auth-type specific things. [01:55] wallyworld: e.g. in gce, read creds from the file pointed to by the "file" attr [01:55] wallyworld: or set auth-mode in openstack [01:56] axw: that is true, there's currently no validation [01:56] wallyworld: the creds-to-config code probably needs to be split out of BootstrapConfig [01:56] wallyworld: s/validation/transformation/ [01:56] axw: the other thing is that with --config options to bootstrap, there's also no validation tehre wither [01:57] so people can pass in crap values [01:57] wallyworld: true, but not so much with credentials - they come in a standardish format in a file. we should treat them the same in bootstrap and create-model [01:57] would be good to fix all that before we ship next beta [01:57] yep [01:58] i should have left a todo in the code [04:12] https://bugs.launchpad.net/juju-core/+bug/1556630 [04:12] Bug #1556630: Timeout github.com/juju/juju/cmd/jujud/reboot

[04:12] is fixed committed [04:12] can someone kick the CI jobs to get this validated pls [04:26] Bug #1533694 changed: inconsistent juju-gui and juju status [04:32] should I be able to do a juju deploy --to lxc:1 with juju 1.25.4 and xenial? it keeps telling me "ERROR adding new machine to host unit "test/1": cannot add a new machine: machine 1 cannot host lxc containers" [04:36] https://github.com/juju/juju/pull/4708 [04:36] ^ thumper, menn0 [04:37] davecheney: looking [04:38] just a small driveby [04:42] davecheney: done [04:43] ta [04:47] wallyworld: finally got all the tests fixed. just gotta update a couple of plugins which are still using configstore, then I can remove it. [04:47] awesome [05:34] menn0: here's another little one, https://github.com/juju/juju/pull/4709 [06:07] anyone: https://github.com/juju/juju/pull/4711 << not a cleanup [06:07] an actual improvement [06:09] ^ wallyworld [06:09] test that is not hoocked up [06:09] davecheney: i think i just reviewed that one [06:09] ta [06:09] thank yo [06:09] i was scratching my head going "how does this work" [06:09] and then, oh, it doesn't... [06:20] wallyworld: https://github.com/juju/juju/pull/4712. if it's too difficult to review, I can try and break it up [06:20] axw: i'll be ok :-) [06:20] wallyworld: there's one more bit that I'm working on, which is validation of bootstrap config in jujuclient. won't change much though [06:20] ok === bogdanteleaga_ is now known as bogdanteleaga === marlinc_ is now known as marlinc [07:44] axw: lgtm with a few small nitpics. i'll be pressing to try and get this branch finally CI tested [07:45] wallyworld: thanks. just finishing up the validation changes - more test fallout, should be trivial tho [07:45] ok [07:45] so good to see that legacy stuff gone [08:37] wallyworld: when do you expect we'll be able to remove -b from restore? [08:37] axw: good question. when someone gets to do the work. i'm hoping we can get some bandwidth next week [08:37] wallyworld: shall I just say "before 2.0"? [08:38] why not :-) [09:08] dimitern: ping [09:09] dimitern: can we sync on a ppc64 bug [09:09] frobware, hey, sure [09:10] dimitern: let's use standup HO [09:10] frobware, omw [09:19] frobware, dimitern: if it's gccgo related please don't spend too much time on it... [09:24] morning [09:29] mwhudson: sure, but it's currently OK on master, broken on our maas-spaces2 branch. [10:20] mwhudson: still around? [11:08] dimitern, voidspace: as a heads up, I have another (almost complete) PR for merging latest master just in case you were also thinking the same... [11:09] frobware, sounds good +1 [11:57] what SSL/TLS protocols does the Juju controller speak? [12:24] mwhudson, dimitern: for reference, the ppc64 issue is fixed with go1.6 [12:25] frobware, sweet! but until we can use 1.6 that "fix" will solve the problem on maas-spaces2 I guess? [13:13] jam: so using `delete(m, "yup!")' means we can drop the log as that works too. :) [13:14] frobware: yay... ish/ [14:28] Bug #1554417 changed: Juju 2.0: ERROR cannot deploy bundle: machine "0" is not referred to by a placement directive (and 4 more errors)

[14:28] Bug #1556961 opened: i/o timeout from mongodb

[14:58] Bug #1556630 changed: Timeout github.com/juju/juju/cmd/jujud/reboot

[15:37] Bug #1557020 opened: Spurious output with kill|destroy-controller [15:37] katco: ping [15:37] rick_h__: she's out sick this morning [15:38] natefinch: ah ok, thanks for the info [15:38] rick_h__: welcome [15:38] * rick_h__ stops waiting in call [15:38] doh [15:38] * rick_h__ sends chicken noodle soup katco's way [15:41] rick_h__: oh you can wait, she will eventually be back [15:41] *lol* [15:42] perrito666: most helpful person around :) [15:42] you know me [15:46] what SSL/TLS protocols does the Juju controller speak? [15:46] Bug #1557020 changed: Spurious output with kill|destroy-controller [16:05] sparkiegeek: tls 1.2 [16:05] dimitern: frobware: I guess networking meeting is not on? [16:06] dimitern: frobware: given that no-one else is here... [16:07] voidspace, yeah, sorry about that [16:07] dimitern: jay didn't come either [16:12] right, well too much to do anyway :) [16:13] first meeting about kindergarten for our oldest child is today. Info packet was set in comic sans and poorly written.... [16:14] natefinch: thanks [16:16] sparkiegeek: welcome [16:28] voidspace: can we sync on MAAS 2.0; standup HO [16:28] rogpeppe: you around? [16:28] natefinch: i am [16:29] rogpeppe: I'm adding a --resource flag to charm publish, so you can specify resources to publish with the charm... however, the code doesn't actually exist on the charmstore yet to support that. The tests for publish currently all run a full-on charmstore... so I'm uncertain how to proceed with testing the added functionality. Any suggestions? [16:30] natefinch: in my view, you can't test it. [16:31] natefinch: you could implement a mock charmstore endpoint, but what confidence would that given you that the code would eventually work when run against a real charmstore [16:31] ? [16:32] rogpeppe: it would ensure I am parsing the CLI correctly and passing that data onward and upward. It can't ensure that the real charmstore actually has that endpoint or deals with the information correctly, that's true. [16:33] rogpeppe: I would note that I am required by a very long management chain to deliver this functionality (presumably with tests) in the next week-ish.. and I am sure we cannot also implement the serverside work in that time. [16:33] natefinch: if that's all you need to do, then i'd implement a fake charmstore handler that implements or fakes the required endpoints [16:33] rogpeppe: I think that's the best I can do, given the constraints of my current reality :) [16:34] natefinch: it probably is [16:46] dimitern, voidspace: http://reviews.vapour.ws/r/4157/ [16:48] frankban, looking [16:48] dimitern: ty [16:50] frankban, oops sorry :) tab completion failed [16:50] frobware, LGTM [16:51] dimitern: you can get to the bug from the commit [16:51] dimitern: I have a quick one as well ;-) http://reviews.vapour.ws/r/4154/ [16:51] frankban, sure :) looking [16:52] ty [16:52] Bug #1557052 opened: Can't bootstrap LXD on juju2.0-beta2 (xenial): no registered provider for "lxd"

[16:52] frankban, feel free to drop the issue, but I think it's good to add easy references to bugs where appropriate [16:52] argh! [16:52] frobware, ^^ [16:52] dimitern: aimed at me? :) [16:53] frobware, :) [16:53] lol [16:54] dimitern: cool part is that dimitern comment was appropriate to my branch too [16:54] frankban, yours LGTM btw :) [16:54] dimitern: ty [17:03] dimitern, voidspace: one more (merge master): http://reviews.vapour.ws/r/4158/ [17:05] frobware, cheers, will look shortly [17:06] dimitern: thanks. would be good to try and get this ^^ in today as we'll have a CI run to look at tomorrow [17:07] frobware: great [17:07] voidspace: can we sync; standup HO [17:08] frobware: sure [17:08] rogpeppe: do you have a link to info on how to set up my machine to run the rest of the tests? For example, I don't have elastic search running on my machine, which seems like it's necessary [17:09] frobware: I'm there [17:09] rogpeppe: actually... just looked at the readme... is makesysdeps all I need? [17:09] natefinch: i think so [17:09] frobware, LGTM [17:09] rogpeppe: I guess I can try it and see ;) [17:09] natefinch: you can run the tests with ES disabled too [17:10] rogpeppe: oh, that would be nice :) [17:10] rogpeppe: especially since make sysdeps just failed :) [17:10] natefinch: but it's not too bad - if you've installed it, you can just run: sudo service elasticsearch start [17:10] natefinch: and the tests will use it [17:12] natefinch: you can export JUJU_TEST_ELASTICSEARCH=none [17:12] natefinch: but that does disable quite a few tests [17:13] rogpeppe: heh, it disables all the publish commands. Perhaps that's not the way to go about it. [17:15] natefinch: i like the charmstore approach to testing against ES more than juju's approach to testing against mongo - it uses the same server for all tests rather than starting a new instance for each tested package. [17:17] natefinch: FYI, someone suggested that I get ES directly from the site and not install via the make file [17:17] natefinch: it worked for me [17:19] ericsnow: do we need a specific version? [17:21] natefinch: I'm running 1.7.5 (downloaded, unpacked to dir under $HOME, and run straight from unpacked dir/bin/elasticsearch without -d) [17:22] natefinch: I left it running in a separate terminal so I could deal with it more directly [17:30] lol, installing elasticsearch 1.7.4: The package is of bad quality. The installation of a package which violates the quality standards isn't allowed. This could cause serious problems on your computer. Please contact the person or organisation who provided this package file and include the details beneath. [17:30] er .5 that is === natefinch is now known as natefinch-lunch [17:31] how exactly can dpkg know that? [18:28] Bug #1557102 opened: Cannot unregister from a controller

=== natefinch-lunch is now known as natefinch-afk [18:34] off to battle the school system's bureaucracy, wish me luck [18:34] luck [18:40] Bug #1557102 changed: Cannot unregister from a controller

[18:52] Bug #1557102 opened: Cannot unregister from a controller

[19:04] Bug #1557124 opened: Controller is orphaned for remotely registered user

=== natefinch-afk is now known as natefinch [19:59] Bug #1557143 opened: help text for juju list-users needs improving

[19:59] Bug #1557146 opened: help text for juju register needs improving

[19:59] Bug #1557148 opened: help text for juju add-user needs improving

[20:16] frobware: yay for 1.6 fixing things [20:18] ericsnow: are we skipping standup today? [20:19] natefinch: wasn't planning on it [20:34] mwhudson: this was my fix for gccgo. http://reviews.vapour.ws/r/4157/ [20:34] mwhudson: only on maas-spaces2 atm, but if you have or are aware of a better fix... [20:34] frobware: the better fix is https://bugs.launchpad.net/ubuntu/+source/golang/+bug/1536882 [20:34] Bug #1536882: upload golang1.6 package for trusty [20:35] :) [20:35] which reminds me [20:35] voidspace, frobware you guys still around? [20:35] are you guys still using gccgo on ppc64el for xenial? [20:35] because if you are, please stop [20:35] (gccgo is less buggy in xenial, but still) [20:36] alexisb: I can be [20:36] mwhudson: not sure about xenial [20:37] mwhudson, no === natefinch is now known as natefinch-afk [23:02] davecheney: thanks for the unused cleanup. I've been meaning to do that, didn't know which tool to use tho [23:03] axw: wallyworld do you people also have DSTd this weekend? [23:04] not in queensland or western australia [23:04] perrito666: what he said. you? [23:05] davecheney: oh, now I see why I didn't know which tool to use. it was announced yesterday ;p [23:06] axw: what tool are you talking about? [23:06] wallyworld: https://godoc.org/honnef.co/go/unused -- tool davecheney used for the PR that removed all the dead code [23:06] axw: huh, my ide has done that for over 12 months :-) [23:07] don't need to wait for go to catch up [23:07] wallyworld: across package boundaries? [23:07] yup [23:07] mkay. I'd like to see the code for that [23:07] it's open source. there's a who inspections framework [23:07] whole [23:07] all sorts of inspections are written [23:08] wallyworld: but it looks so ugly :p [23:08] pycharm with golang plugin [23:08] perrito666: have you even seen it? [23:09] yes I did, I once even did an editor that "borrowed extensively" from it :p [23:10] i use the gtk theme, looks native [23:11] wasnt pycharm a closed source thing? [23:12] no, never [23:12] oh I see there is a paid version and a free version