[00:20] wallyworld, thumper, axw : I think I've figured out that bootstrap problem [00:20] the bootstrap command embeds a ModelCommandBase and that's never told what the name of the controller that's being bootstrapped [00:21] so it defaults to the current controller [00:21] so the correct API details are being written out to controllers.yaml [00:21] but the details for the other controller are being read out [00:21] and the other controller doesn't have any details yet [00:21] trialing a fix now [00:24] wallyworld, thumper, axw: yep, that was it [00:24] * menn0 cleans up and adds tests [00:27] wallyworld: can I pls have permission to modify the A team board? [00:44] menn0: sorry was otp, done [00:44] wallyworld: cheers [01:06] menn0: cool, thanks for fixing [01:18] wallyworld: is this a bit clearer? http://reviews.vapour.ws/r/5205/diff/4-5/ [01:18] looking [01:18] axw: awesome, thanks [01:19] wallyworld: are we still JFDIing things, or should I fix https://bugs.launchpad.net/juju-core/+bug/1603596 first? [01:19] Bug #1603596: HA often fails on azure creating virtual machine

[01:21] axw: i thinks there's more than one critical blocker? if that bug is the last one, would be good to fix that first; if not what's one more jfdi rught? [01:21] wallyworld: just that one and the one you fixed yesterday. I'll take a look at the azure one now [01:21] ok [01:21] ta [01:21] axw: if you have a bug that is also causing a CI failure and is marked crtical and tagged blocker, edit the bug to have that and use fixes-nnnn [01:22] *is not marked [01:22] mgz: okey dokey [01:22] mgz: it's not a bug related thing, we were just JFDIing things for a while because of time pressure. [01:22] but no more [01:23] menn0: oh FFS, master and model-migration conflict again [01:23] menn0: do you wanna take this one? [01:26] wallyworld: the CI run for model-migration is mostly good [01:26] there are a few failures, but I don't think they are new, and certainly not related to migration [01:26] you want to land it [01:26] so I'm thinking lets just fix the merge conflicts and merge into master [01:26] wallyworld: thoughts? [01:27] let me cast a second set of eyes on it so we can say we've done due diligence [01:27] wallyworld: http://reports.vapour.ws/releases/4148 [01:29] wallyworld: compare with the lastest curse email on master [01:29] thumper: won't the race test failures be an issue? [01:29] i thought we were gating on those now for master [01:30] races aren't [01:30] but joyent races are long known [01:30] and I should double check that the joyent deps are right [01:30] because dave fixed some of those [01:30] hmmm, i could have sworn race tests were gating [01:31] other than that, looks ok to merge [01:31] or merge regardless and if i'm wrong it will land [01:31] * thumper sets up a branch [01:34] * thumper runs the tests [01:35] anastasiamac: did we have a meeting set up for tonight? [01:36] natefinch: we did not. we said whenever u and i r available :D [01:36] anastasiamac: are you available? :) [01:36] natefinch: for u, i could be :D [01:37] natefinch: in 15mins k? [01:37] anastasiamac: sure, whenever is good for you. [01:37] natefinch: \o/ [01:54] natefinch: another issue [01:54] * redir puts it in doc [01:54] redir: good... better to find them now [02:05] thumper: yep, I'll take a lok [02:05] menn0: too late [02:05] done it [02:05] ok [02:05] and it is being merged [03:31] thumper/wallyworld/axw: fix for that concurrent bootstrap issue: http://reviews.vapour.ws/r/5262/ [03:31] easy fix, horrible test [03:31] looking [03:31] it took me 3 attempts at different approaches to arrive at this [03:31] ugh [03:31] * thumper leaves for wallyworld [03:36] menn0: jesus, you were not wrong about the test [03:38] wallyworld: yeah, pretty terrible. I almost thought about not testing the change. [03:39] you are a better man than me :-) [03:40] ha :) [03:42] thumper: is model-migration merged into master? [03:42] looks like it [03:42] menn0: yep [03:42] thumper: awesome \o/ [03:44] Bug #1604223 opened: Concurrent bootstrap fails with "no API addresses"

[03:46] axw: only just saw your review and it's already merging. I'll land a separate micro-PR afterwards. [03:50] menn0: it's no big deal [03:51] menn0: feel free to leave it [03:52] axw: it's ok. your suggestion is good and it's really not much effort. [04:13] wallyworld, thumper: that bootstrap fix was rejected because it's not one of the blockers. jfdi or wait? [04:14] JFDI IMO [04:14] menn0: update with new comment too :-) [04:14] wallyworld: of course :) [04:14] the change is only one line, what could possibly go wrong :-D [04:17] wallyworld, axw: in the new world of commands, I want to get the modelTag for a modelCmd [04:17] it isn't yet obvious to me [04:17] there's a CurrentModelUUID somewhere on the base command, i'd need to look it up [04:18] thumper: 1. get the client store (c.ClientStore()); 2. call store.ModelByName(c.ControllerName(), c.ModelName()) to get model details; 3. names.NewModelTag(details.ModelUUID) [04:19] thumper: see cmd/juju/model/show.go [04:19] axw: ta [04:30] ha poop [04:30] ah poop even [04:30] I have an old controller lying around [04:30] which I now can't talk do because missing creds [04:31] axw: recoverable? or just kill the machines and blow away the cache dir? [04:32] thumper: umm. I think you can get the admin password from the machine? in the agent.conf? [04:32] old password or something? [04:32] ugh [04:32] too hard [04:32] :) [04:32] thumper: not sure. there should be a command to fix it, but we don't have one [04:32] probably easiest to restart [04:33] um... [04:33] where is our config? [04:33] ~/.config? or .cache? [04:34] thumper: ~/.local/share/juju [04:35] just wondering... [04:36] but why did we choose there? [04:37] hmm [04:37] nm [04:37] thumper: IIRC, things that you need for operation to continue should go in the data dir [04:37] optional config goes in config [04:39] * thumper starts afresh [04:39] skipping windows tests review, plz :D http://reviews.vapour.ws/r/5263/ [05:12] wallyworld: what's the name you're using for the config inheritance work? [05:13] juju model tree [05:13] or juju model config tree [05:14] ok, that'll do (it's for a document i'm writing on current model migration status) [05:19] axw: no rush, when you are free, would love a review on http://reviews.vapour.ws/r/5264/ [05:20] wallyworld: ok, in a little while [05:37] yes... new command in under 350 lines http://reviews.vapour.ws/r/5265/diff/# [05:37] juju dump-model [05:37] (behind developer-mode feature flag) [05:39] * thumper out now to go make dinner [05:39] laters [05:55] wallyworld: where abouts do we generate the certs in the lxd provider? [05:55] I got lost in a twisty maze [05:56] axw: there's a juju/tools/lxdclient package [05:56] there's a UseRemoteTCP() method or something like that. it's called when config is finalised [05:56] that method generates client certs, connects to server, and gets back the server cert [05:57] the client certs are based on network interfaces and host name at the time [05:57] so it's implausible to generate ahead of time [06:01] wallyworld: I was thinking we'd only generate them at bootstrap time [06:02] wallyworld: I guess having it in DetectCredentials opens users up to accidentally storing them in credentials.yaml. would be good to auto-generate still though... [06:02] exactly [06:02] but at bootstrap time seems ok i think [06:03] i'd like to do that after the work to properly store credentials [06:03] out of the model [06:03] wallyworld: ok [06:03] for now, the PR as it exists I think adds value [06:03] model-config looks sweet now [06:04] wallyworld: model-config looks good, I'm not sure there's any point in adding the tls auth-type without the implementation though? [06:04] axw: it exists merely to define the attributes to filter out [06:04] ah, right [06:06] wallyworld: reviewed, just a few little things [06:06] axw: ta, will look after I get back from school pickup, heading out now to do that [06:35] axw: i explained the \n thing in the CLI - i could just drop the code since we don't displau authorised keys now, but figured it can't hurt to keep it? [06:37] wallyworld: I'm not strongly opposed to it, I just wondered what the reason was - wasn't mentioned in the review description. why drop the leading character? [06:38] * axw looks at comment [06:38] axw: cause i'm an idiot [06:38] typo [06:38] wallyworld: maybe just use strings.TrimSuffix(valString, "\n") then ? [06:39] yeah, that's much better [06:46] axw: wallyworld: fix of the fix :D http://reviews.vapour.ws/r/5267/ [06:46] for those of u with eagle eyes :D [06:49] axw: fast \o/ tyvm!! === frankban|afk is now known as frankban [08:47] Bug #1567708 changed: unit tests fail with mongodb 3.2 [10:18] babbageclunk: my kvm-maas scripts now work on 2.0 - please let me know otherwise [10:21] frobware: wilco, thanks. In other news it looks like the dhcp was the problem. [10:21] babbageclunk: as in lack of? [10:22] frobware: yup [10:22] frobware: just working out what I should be putting into the network info instead now. [10:40] babbageclunk: want to HO? [10:40] frobware: yeah, why not? [10:43] frobware: I'm in juju-sapphire [11:00] wallyworld: tx for the overnight merge on controller tag [11:01] perrito666: no worries [11:01] * perrito666 fears the merge with master [11:03] babbageclunk: http://pastebin.ubuntu.com/20024195/ [11:21] babbageclunk: https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1590104 [11:21] Bug #1590104: network config from datasource overrides network config from system

(Ubuntu):Fix Released>

[11:26] babbageclunk: https://bugs.launchpad.net/maas/+bug/1604169 [11:26] Bug #1604169: maas login yields "ImportError: No module named 'maasserver'" [11:27] jam: ^ the maas bug I talked about this morning === rogpeppe1 is now known as rogpeppe [12:05] frobware: I'll brt, just need to grab a snack [12:05] ack [12:30] anyone know how to recompile a Go stdlib package without rebuilding the whole go install? [12:30] it used to work but no longer seems to [13:10] I believe that at some point we should accept that there is no way to be descriptive enough in a test name (in some cases) and start adding a couple of doc lines in the tests [13:26] perrito666, I heartily endorse explanatory comments all over the place [13:27] heh, my issue is with tests that usually are called something like TestThisHappensWhenX which is explanatory of what the test does, but not why, which is the important bit [13:30] bbl heating people is finally here [13:33] Bug #1604408 opened: LogForwarderSuite.TestConfigChange obtained Next, expected close

[13:36] Bug #1604408 changed: LogForwarderSuite.TestConfigChange obtained Next, expected close

[13:39] Bug #1604408 opened: LogForwarderSuite.TestConfigChange obtained Next, expected close

[14:12] frobware: Those scripts work an absolute treat! [14:14] babbageclunk: \o/ [14:14] frobware: I guess I shouldn't be surprised, but it's nice! :) [14:16] babbageclunk: what we need next is a means of describing what NICS you want connected to which subnets and/or dhcp/manual, et al [14:38] frobware: The only thing I'm having trouble with is that I don't want to allocate 16GB of disk to each node. [14:39] frobware: Even if I set VIRT_DISK_SIZE and KVM_DISK_SIZE to 10, I still get 16GB disks. [14:48] frobware: Ah, worked it out - removing the node doesn't remove the pool, so even though I'm specifying the size for add-node it's reusing the already-created image. [14:53] frobware: gah, nope - still can't see why the images are always coming up with 16G disks. [14:56] frobware: ignore, everything's fine now! PICNIC === natefinch is now known as nfinch-baby-chic === nfinch-baby-chic is now known as nfinch-babychick === nfinch-babychick is now known as nfinchbabychicks [16:30] cherylj: fyi, marking this as critical - https://bugs.launchpad.net/juju-core/+bug/1604482 [16:30] Bug #1604482: MAAS bridge script should drop all 'source' stanzas from original file

[16:31] Bug #1604474 opened: Juju 2.0-beta12 userdata execution fails on Windows if azure-provider is used

[16:31] Bug #1604482 opened: MAAS bridge script should drop all 'source' stanzas from original file

=== frankban is now known as frankban|afk === nfinchbabychicks is now known as natefinch [18:16] Bug #1604514 opened: Race in github.com/joyent/gosdc/localservices/cloudapi

[19:20] can someone approve this? It's just a markdown change -- perhaps I did it wrong. https://github.com/juju/juju/pull/5820 [19:21] balloons: LGTM [19:22] perrito666: i believe you are OCR. tal at a small change? http://reviews.vapour.ws/r/5270/ [19:22] I am, and now with heating [19:23] yay :) [19:23] sadly my office is the last portion of the wather circuit :p [19:23] balloons: you need to get the keys to reviews.vapour.ws off ericsnow then you could add yourself and your pr would be more visible :) [19:41] natefinch, do I need to tell the bot to merge then? Not sure of the formal process as I do generally hit it with the comment [19:42] balloons: I can merge it for you [19:42] balloons: done [19:42] well, except it'll fail because master is blocked [19:42] not really, just added the $$merge$$ but you will have to watch if that goes trough [19:42] ahaa [19:42] * perrito666 curses in spanish [19:43] * natefinch adds a JFDI [19:43] indeed.. I just wasn't sure if the person proposing should tell it to merge, or if the reviewer(s) do [19:43] balloons: also with that mp^ - you shouldn't be using the juju namespace for merge proposals [19:43] balloons: we had a spurious CI run because of it [19:43] mgz, right, I noticed it did a CI run [19:43] you need your git config to have different upstream and origin [19:43] I did it on github as I saw the error and thought nothing more of it [19:43] so I was led astray by gh :-) [19:44] are we still jfding changes? [19:44] you have to treat owner privs with care or I shall remove them :) [19:44] yeah, that's not really his fault. I'm kind of surprised it let him make a branch there [19:45] it is surprising the web ui would do that [19:45] well, it's made to make it easy to make minor edits to docs. [19:45] it's our fault that we automatically run CI on any branch created there [19:45] ...and that CI takes so damn long :) [19:47] working on that last part! [19:48] production deployment, beta11, AWS: 'listing instances: An internal error has occurred (InternalError)' [19:48] when trying ot provision a machine [19:48] marcoceppi: aws seems down [19:49] stokachu, us east 1 is [19:49] or was at least [19:49] stokachu balloons not according to the status anymore [19:49] Bug # opened: 1604542, 1604551, 1604559, 1604561 [19:49] balloons: yea still is, error retrieving resource count [19:49] marcoceppi: aws.amazon.com shows failure retrieving instances [19:50] stokachu: http://status.aws.amazon.com/ [19:50] marcoceppi: i see that, but im just telling you what i see [19:50] marcoceppi https://usercontent.irccloud-cdn.com/file/nH5BLoCy/ [19:50] the status is reporting it working [19:51] but the console says otherwise [19:55] stokachu: works for me http://i.imgur.com/fnlaUkl.png [19:55] so, lets pretend they fixed this, can someone help me troubleshoot that error? [19:55] rather [19:55] 'cannot run instances: Request limit exceeded. (RequestLimitExceeded)' [19:56] aren't you limited to 20 instances [19:56] or smaller? [19:56] not on this account [19:56] this is API requests [19:56] i see [19:56] http://docs.aws.amazon.com/AWSEC2/latest/APIReference/query-api-troubleshooting.html#api-request-rate [19:59] marcoceppi: what region? [19:59] us-east-1 [19:59] seems to be catching the tail end of the api outage [19:59] I got a series of other errors about root-tagging [19:59] and now it's pending again [20:00] sinzui says that's borked [20:00] try a different region [20:01] you can also see some red on our test-cloud-aws-us-east-1 health job [20:05] fcs is it really so hard to add docs to public structs? [20:06] That's why I love having golint running on save... the little squiggles drive me crazy, so I fix them ASAP. [20:11] some people might need for golint to kick them in the face to actually do something about [20:11] I would totally love a golint with that feature [20:12] haha [20:13] perrito666 is taking over Dave's role as grumpy cat, it seems. [20:13] oh, you sure are one to talk [20:13] lol [20:14] well, to be fair, maybe if we both worked at it together, we might be able to almost match dave's grumpiness. [20:15] mm, ok, it would seem that I need thumper to make sure I dont break anything in this [20:16] wat? [20:16] oh, that actually works, sweet [20:16] mm, ok, it would seem that I need one million dolars so I dont break anything in this [20:16] meh, nothing happend [20:17] heh [20:17] thumper: I am adding ControllerUser to migrations, but it would seem that migrations are per model? [20:17] perrito666: yes, migrations are only per model [20:17] controller user probably doesn't need to be migrated [20:18] thumper: ok then controller permissions are not something we will be migrating, that makes sense, since we migrate models [20:18] * perrito666 facepalms [20:18] aye [20:19] * perrito666 adds a bit of info to that pesky test that tells you that you need to migrate stuff [20:25] natefinch: your jfdi did not like windows [20:30] perrito666: juju doesn't like windows :/ [20:31] well bad luck, since we are gating landings on it [20:48] without gates there'd be no windows [20:49] *rimshot* [21:13] so if I juju set-model-config default-series=xenial && juju deploy wordpress I should expect wordpress to be deployed on xenial, yes? [21:16] Bug #1604586 opened: devel Charm Docs are out of date and examples do not work [21:23] menn0: http://reviews.vapour.ws/r/5271/diff/# [21:24] thumper: looking [21:25] menn0: ta [21:27] redir: ir fail iirc [21:28] tx perrito666 looks like it falls back if the charms doesn't support the series [21:29] but I have another error: ERROR storing charm for URL "cs:trusty/wordpress-4": delegatable macaroon cannot be obtained for public entities [21:29] think I must have something twisted. [21:29] like my understanding [21:32] thumper: done [21:52] menn0: as reviewer, here's another http://reviews.vapour.ws/r/5265/diff/# [21:57] thumper: looking [21:57] menn0: ta [22:05] katco: can you add the link to your pr to the card on the board please? [22:05] natefinch: you around to be able to give that a look over? [22:06] rick_h_: oh sorry, sure [22:06] katco: np, ty much [22:06] katco: there's a new bug added there but post the provider doc work just as a heads up, will cover in standup tomorrow [22:06] rick_h_: k [22:09] wallyworld: hey i see your comment in my PR for the config change... this change isn't in environs/config.go. should i still be writing unit tests there? [22:11] katco: it's a bit messed up. the controller config was being parsed by environs/config.go since at bootstrap we handle a single bucket of config items which are then split up. i think that's still the case [22:11] does anyone know what goes on the depends on field on rb to actually have a dependency? [22:12] katco: so as with log forward on/off, i think we can just extend the environs/config tests [22:12] perrito666: interface{} [22:12] redir: of reviewboard, sorry [22:12] missed that [22:12] perrito666: i got that, just a bad joke [22:14] menn0: and another http://reviews.vapour.ws/r/5273/ [22:16] axw: wallyworld folks arriving on the call [22:25] thumper: done and done. reboot worker one to go. [22:25] Bug #1603577 changed: backup-restore: panic: empty value for "api-port" found in configuration

[22:25] coffee first [22:32] menn0: the reboot one is just a backport [22:32] doesn't need its own review IMO [22:35] * perrito666 would be a bit happier if we stoped passing State around as if it where a hot potato and started passing a reduced interface of state [22:46] thumper: ok cool. I thought it sounded familiar :) [22:46] hmm... [22:46] merge bot is rejecting a 1.25 merge, but juju.fail doesn't show any blockers [22:47] ah... a private bug [22:47] hence juju.fail doesn't see it [22:51] wallyworld: any idea on the status of this: http://reviews.vapour.ws/r/5173/ [22:51] ugh... none of the core description stuff has any json serialization tags [22:51] perhaps we should add some [22:53] looking [22:53] menn0: it can be discarded for now, i'll do that [22:54] wallyworld: cheers [22:58] wallyworld, thumper do you guys have time to meet before the standup? [22:58] yeah [22:58] I will jump on teh hangout [22:58] alexisb: can do, i have a 1:1 with redir, hopefully he will be ok if we do that later [22:59] wallyworld, it is not super urgent [23:00] wfm [23:08] thumper: I just went through the review checklists with http://reviews.vapour.ws/r/5265/. It wasn't too painful and I found a bunch more stuff. [23:09] menn0: how r we maintaing review checklist? is it part of our wiki? or juju/juju/docs? [23:09] maintaining*? [23:10]