[00:39] menn0: right this run failed with the same message too. What's the most useful way to present the logs? Also, which logs would be useful? (I'm seeing that error in unit-mongodb-0.log.gz). [00:39] menn0: oh wait, I can more than likely see the same thing with the jenkins run, ignore me for now :-) [01:23] veebers: if the jenkins run has everything then that's perfect [01:31] menn0: can you parallelize make check in any way? [01:32] rick_h_: it is parallelised, but by package [01:32] rick_h_: so if you're running tests in just one package they get run serially [01:35] menn0: hmm, ok. I guess it might have been done but didn't give me my console back. I see xx passed, 1 skip, 4 failed. [01:36] menn0: but yea, a second run does show it running pretty parallel now that I start over. https://flic.kr/p/JQCkPg [01:37] rick_h_: yep... the way that all cores get pegged for most of the test run is also a good indication [01:37] * rick_h_ is more patient for run #2 [01:46] rick_h_: is your /tmp on tmpfs? That makes a big difference for test run time. [02:00] menn0: no, ty for the tip. i need to spend some time getting comfy with go dev and tricks like that [02:00] * rick_h_ needa an induction sprint :p [02:00] rick_h_: haha np :) [02:05] axw: big review done. phew! [02:05] axw: nice work. [02:12] menn0: TYVM [02:12] wallyworld: and you too [02:12] sure [02:43] menn0: I'm just trying to get a run with that error. I'm also trying to confirm that it's not a red-herring. I believe that there are issues as the host machine is network confined [02:48] veebers: seems unlikely that it could be related. the lease manager is a fairly internal thing. [02:50] menn0: i've had a look at a race in juju ci tests - looks like a loggo issue due to the recent changes. i can't repro, but i think i have convinced myself where the issue is https://github.com/juju/loggo/pull/18 [02:50] wallyworld: ok, give me 2 mins [02:50] sure, np [03:21] Bug #1608952 changed: Deployer: KeyError: 'uuid' connecting to environment

[03:33] wallyworld: sorry, got distracted by other work [03:33] no problemo [03:33] wallyworld: should the mutex be on the the writer itself perhaps? [03:35] menn0: the writer is an interface and my thought was that if e do that, we then require all concrete implementations to do the muxtex, whereas doing it in the context is safer? [03:35] wallyworld: yeah, fair enough [03:36] wallyworld: ship it... the PR description uses muxtex twice instead of mutex though :) [03:36] it is well know i can't type [03:37] wallyworld: I think you can blame the drugs this time [03:37] sigh, see ^^^^ [03:40] hahaha === menn0 is now known as menn0-school-run [03:43] axw: since menno is away, want a 2 line review? http://reviews.vapour.ws/r/5359/ [03:51] wallyworld: looking [03:51] wallyworld: LGTM [03:51] ta [04:38] menn0-school-run: I can't get the jenkins run of that test to fail in the same way. I'll let you know if I can get it to happen at some point === menn0-school-run is now known as menn90 === menn90 is now known as menn0 [04:38] veebers: ok bummer. I can take a look at the logs you gathered earlier if you like. [04:50] menn0: I'm not sure if they'll be interesting. It could be a red-herring due to the network confinement etc. on that machine. I'll double check and get back to you tomorrow? (i.e. I may be doing something odd there) [05:03] Is there a way to push juju tools when you manually provision a machine, or must it always pull from machine 0? [05:04] Or perhaps give it a different pull location? === frankban|afk is now known as frankban [05:46] veebers: no problems [10:28] axw: ping [11:00] Bug #1609343 opened: [aws; beta13] Juju can pick subnets from the wrong VPC with spaces constraints specified

[12:17] anyone uses vim plus ctrlp to work on juju? [12:17] * perrito666 looks at rick_h_ [12:18] perrito666: well I do use vim and ctrlp but don't consider I've "worked on juju" atm [12:19] rick_h_: ever had missing files? even tuning the max depth and limits I still miss files [12:19] perrito666: not tried it enough to say tbh [12:52] perrito666 - yeah thats weird. ctrlp has worked well for me even on large code bases.. but i admit i haven't tried it on the juju codebase [12:52] might some other bug causing it to drop files [12:56] lazyPower: very likely, in my case the files are missing even from the buffers list === redelmann_ is now known as redelmann [13:31] Bug #1609405 opened: DeploySuite.SetUpTest connection was forcibly closed

[13:31] Bug #1609407 opened: remoteFunctionalSuite.TestUsingTCP no such network interface

[13:31] balloons: Are you there? [13:31] katco, babbageclunk: quite a tiny review fixing bug 1609343, when you have a moment please! [13:31] Bug #1609343: [aws; beta13] Juju can pick subnets from the wrong VPC with spaces constraints specified

[13:35] mgz: I am having trouble installing juju today, juju-core gives me 1.25.6 and the juju package does not seem to include the juju binary. [13:36] fwereade_: do you have time to help dimitern by reviewing http://reviews.vapour.ws/r/5363/ please so it might make b14? [13:36] mbruzek: well, that's interesting [13:36] rick_h_, looking [13:36] mbruzek: using what ppa? [13:36] fwereade_: ty [13:36] mgz: but tab completion works like a charm! [13:37] mbruzek: I have reports to the opposite with other setup [13:37] so... things in general are not quite right [13:37] rick_h_, fwereade_: thanks! :) [13:38] mgz: I have both devel and stable in my source.list.d directory: [13:38] deb http://ppa.launchpad.net/juju/stable/ubuntu xenial main [13:38] deb http://ppa.launchpad.net/juju/devel/ubuntu xenial main [13:38] mbruzek: so, the fix is likely you just need to install `juju-1.25` but I'm a little concerned the dep chain is wrong for the packages you hae installed [13:38] dimitern: care to trade for fwereade_'s branch http://reviews.vapour.ws/r/5362/ please? [13:38] dimitern: before jumping into the next pool [13:39] rick_h_: sure - I'll have a look before digging in around bug 1608105 [13:39] Bug #1608105: LXD no longer activates all interfaces on initial deploy when using MAAS2rc3 and JUJU Beta13 [13:39] ugh, the ppa packaging is just all a mess still >_< [13:39] mgz: what if I wanted 2.0-beta13? [13:39] that's what just 'juju' should get you if you have the devel package installed [13:40] but I'm not sure that mixing devel and stable ppas actually works how we want atm [13:40] mgz: It was necessary to have both devel and stable at one point to get juju installed. [13:40] mgz: I did not do this to anger you, it was needed at one point [13:41] mbruzek: yeah, and it certainly should work [13:41] mgz: I could remove stable and try again if you think that will work [13:41] we just have too many different sets of packaging and they don't all play nicely with each other [13:42] mbruzek: pastebin `dpkg -l "juju*"` first [13:42] fwereade_: you've got a review ;) [13:43] mgz: http://paste.ubuntu.com/22028864/ [13:43] dimitern, cheers :) [13:44] * dimitern whew.. that's got to be the longest bug and PR descriptions for a fix with diff size less than 1/3 of the length of both those descs.. [13:45] dimitern: :P [13:45] mbruzek: thanks, go ahead and futz now and see if you can get a working set? [13:46] dimitern, and so do you :) [13:46] fwereade_: tyvm! [13:47] rick_h_: so how's the process now - with +1 I need to ask e.g. mgz (hey there ;) for a second +1 and test the QA steps? [13:48] mgz: I removed the stable ppa, updated and got this message when trying to install juju-core:juju-core is already the newest version (1.25.6-0ubuntu1~16.04.1~juju1). [13:48] dimitern: no, you can just check the QA steps locally using the doc for any shortcuts if it's something that's pretty QA'able [13:48] dimitern: nothing there involves the QA team, it's all checking each other [13:49] rick_h_: I see - but do I still need 2x LGTM or not? [13:49] it would be nice to actually have functional test coverage for subnet stuff on aws [13:49] dimitern: no, we're not doing 2x +1 atm [13:49] dimitern: more +1 and QA OK [13:50] rick_h_: ok then, thanks [13:50] mgz: there are a couple in featuretests/ [13:51] mgz: yes, we need to get some more of that for the AWS stuff since it does't require too much custom setup/etc like MAAS [13:51] mgz: can you give http://reviews.vapour.ws/r/5363/ a QA OK using the steps there please? [13:52] dimitern: it's on our own team to check QA. Maybe natefinch for katco or fwereade_ [13:52] dimitern: unless mgz is just curious :) [13:54] mgz: This is me futzing: http://paste.ubuntu.com/22029893/ [13:54] mgz: Shall I open a bug or is this just my system messed up? [13:54] * dimitern leaves it for standup [14:00] mbruzek: wanting to install juju-1.25 seems right [14:01] dimitern: fwereade_ natefinch ping for standup [14:01] omw [14:02] rick_h_: omw [14:11] mgz: right but how do I get 2.0 ? [14:12] mbruzek: a bug is probably a good idea regardless [14:12] we need something to track sorting out the ppas [14:13] mgz: juju-core or is there a packaging project? [14:14] just juju-core [14:23] mgz: https://bugs.launchpad.net/juju-core/+bug/1609437 [14:23] Bug #1609437: Unable to install juju 2.0 using apt [14:23] ...that's a scarier bug title than it should be :) [14:24] mgz: sorry what would you like me to change it to? [14:25] mbruzek: don't worry about it, I cna edit [14:25] do ash you wish. I am not feeling creative today [14:25] rick_h_: so apparently my bug is no longer a blocker; same gameplan? [14:27] katco: /me looks [14:27] rick_h_: you responded to that thread last night it looks like [14:27] balloons: oh, was #1603221 the one that axw noted was not implemented in lxd? [14:27] Bug #1603221: Charms utilizing storage fail on LXD

[14:28] rick_h_: yeah [14:28] katco: ok, then let's play normal OCR please [14:28] fwereade_: dimitern and if it's EOD please work with katco to help do QA with the shared info before you head out [14:29] rick_h_: sounds good [14:30] fwereade_: mind if I go ahead and land your ErrMissing branch for trunk? we want a new change to test. [14:30] review for anyone interested, fairly simple code: http://reviews.vapour.ws/r/5364/ [14:30] mgz, be my guest :) [14:31] Bug #1609437 opened: Unable to install juju 2.0 using apt [14:53] poor axw... did we ever figure out what was going on with his review? http://reviews.vapour.ws/r/5355/ [14:55] rick_h_: sure, will check with katco [15:00] babbageclunk: ping [15:00] dimitern: hey [15:00] babbageclunk: hey :) I've seen your comment re not seeing the expected subnet [15:00] babbageclunk: can you give me more details about it? [15:01] dimitern: Sure - I got an error running `juju add-space test subnet-096b3f72 subnet-0c9eaf65` [15:02] dimitern: error: invalid arguments specified: "subnet-096b3f72" is not a valid CIDR [15:03] babbageclunk: hmm .. odd - I'm trying now to see if I'd made a typo [15:05] dimitern: If I look in the AWS console under network interfaces (which might not be the right place) I can see subnet-0c9eaf6 but not 096b3f72 [15:05] babbageclunk: have a look at the dropdown menu above on the left (Services v), pick VPC from it and then Subnets [15:06] dimitern: Oh, ok - I can see it in there. [15:07] dimitern: So I don't know why I was getting that error then. [15:08] babbageclunk: you're right - that error wasn't exactly expected :/ [15:09] babbageclunk: it will work in 3 steps though (with update the QA steps shortly) - i.e. add-space test; add-subnet subnet-xxx test; add-subnet subnet-yyy test; [15:10] dimitern: ok cool - happy to try doing it again in a bit. [15:15] babbageclunk: if you can - great! ;) [15:28] btw the shared aws account has these additional users, which never logged in: natefinch, alexisb; while wallyworld, and mfoord haven't logged since 2015 - please ping/mail me if you want account setup info (or password reset, etc. if you forgot) [15:37] Bug #1609437 changed: Unable to install juju 2.0 using apt

[15:37] Bug #1609463 opened: upgrade-charm command is inconsistent with deploy when using local charms === frankban is now known as frankban|afk [17:03] * rick_h_ goes for lunchables [17:07] Bug #1609494 opened: grant-revoke: disabled user still in admin read-write

[17:16] Bug #1609494 changed: grant-revoke: disabled user still in admin read-write

[17:19] Bug #1609494 opened: grant-revoke: disabled user still in admin read-write

[17:38] bbl [18:28] katco: default is a keyword. It unfortunately is a very useful variable name, and there aren't many good alternatives. Suggestions more than welcome. I agree dflt is terribad. [18:57] katco (or anyone else) could use a review on this (sorry no reviewboard link, not sure what happened) https://github.com/juju/schema/pull/13 [19:00] I wish hexchat had a hotkey to change font size [19:13] natefinch: lol oops forgot about that [19:13] natefinch: d3fault [19:14] natefinch: dfault [19:14] * katco shrugs [19:15] dfault is pretty good [19:15] sorta sounds like a DJ name [19:16] lol [19:27] gah, reflection is so horrible... once you start using it, you can't ever take it out, because you never know who is relying on it :/ [19:30] I'm pretty sure we're always passing a map[string]interface{} to coerce, but because we're using reflection, I can't know for sure... in theory it can take map[Stringer]interface{} too, but I'm pretty sure no one does that. [19:33] * rick_h_ goes to get the boy from camp [19:41] back [19:42] katco natefinch: byOmission sort of applies to default in a very pompous way [19:43] Bug #1609041 changed: Race in github.com/juju/juju/cmd/modelcmd

[19:46] Bug #1609041 opened: Race in github.com/juju/juju/cmd/modelcmd

[19:47] perrito666: hmmm... context for that comment? [19:47] katco: you where discussing a replacement for the word default [19:48] perrito666: ah [19:48] and I used my super spanish powers to find an old and unused one :p [19:49] haha [19:51] haha [19:55] Bug #1609041 changed: Race in github.com/juju/juju/cmd/modelcmd

[20:53] thumper: sorry! Here now if you want to do a quick chat? [20:54] babbageclunk: ok [21:22] wallyworld, ping [21:22] yo [21:22] yo dawg [21:22] zup [21:23] wallyworld, did you impliment the show-model command [21:23] no [21:23] was multi model work i think [21:23] thumper, was it you? [21:23] wallyworld: how are you feeling? [21:23] much better today thanks [21:24] had a raging headache and sore throat yesterday [21:24] wallyworld: didn't realize you were under the weather :( glad you're feeling better [21:24] wallyworld: cool, then I wont feel guilty saying I addressing your comments and you should take a second look :p [21:24] lol [21:24] katco: tis all good :-) [21:25] i don't know how, but even a remote globally distributed team still seems to demonstrate properties of infection clusters [21:28] katco: visruses spread using computers you know [21:29] wallyworld: i bet i transferred it from my compooper when i got my drink out of its cup-holder [21:29] seems plausible [21:32] well we are clever enought to get together once every semester, when its winter for half of the team [21:32] we are not that smart [21:33] anastasiamac, running late [21:44] alexisb: not sure if you're there, but we lost you [21:49] katco: that answers the question [21:52] perrito666: apparently :) [21:56] Bug #1558703 changed: PatchValue unsafe for SetUpSuite

[22:01] redir: that bug that's assigned to you about grant/revoke: I have a fix for it already. it's bundled up with some other changes, not sure if we want to gate on them [22:02] katco: nfi what's up with reviewboard, but my PR was reviewed on github :) [22:02] axw: ah ok :) good to hear [22:02] axw: point me to the redir issue (redir is on holidays iirc) [22:03] axw: your code was so good RB knew it didn't need to be reviewed [22:03] axw: I am giving some love to grant/revoke [22:03] perrito666: https://bugs.launchpad.net/juju-core/+bug/1609494 [22:03] Bug #1609494: grant-revoke: reenabled users missing from list-users

[22:03] katco: heh, probably more like it was +1500/-1600 lines and it fell over :p [22:03] haha [22:05] I think i would gladly give time of my life to change reviewboard for something else that we can think is better for a few months before starting hating that too [22:06] perrito666: i think all diffs and comments should be pasted here, in irc. [22:06] katco: I think emacs got to your nervous system and its not you talking [22:06] wallyworld: is the leads call no more? [22:06] axw: yeah, seems so [22:07] i thought we were still going to have it but seems not [22:07] cool, got up at 5:45 for no reason -_- [22:07] axw: i only found out by chance when i saw the meeting invite come through yesterday and it was cancelled [22:13] wallyworld, ping [22:13] sorry lost network, back now [22:13] ok, see you in 1:1 [22:24] thumper: I've just done a big card cleanup. All the cards on the onyx board that were still relevant have been moved to the A team board [22:25] menn0: cool [22:25] menn0: can we chat about the i/o timeout bug? [22:25] thumper: sure [22:26] 1:! [22:26] thumper: i'm there already [23:38] Bug #1600257 changed: Broken bash completion with old ppa packages present

[23:40] veebers: FYI this is the branch for juju-ci-tools: https://code.launchpad.net/~axwalk/juju-ci-tools/cli-model-owner [23:40] veebers: ignore the one linked from the PR [23:44] Bug #1600257 opened: Broken bash completion with old ppa packages present

[23:44] axw: axk [23:44] :-\ ack [23:47] Bug #1600257 changed: Broken bash completion with old ppa packages present