[00:02] Bug #1571901 opened: "juju change-user-password --generate" is unhelpful [00:17] cmars: a really small one http://reviews.vapour.ws/r/4635/ [00:17] if you have a chance [00:17] wallyworld, sure, looking [00:18] wallyworld, awesome, thanks! [00:18] LGTM [00:18] thanks for rasing the bug [01:20] * menn0 loves the lxd provider ... so much easier to test HA [02:09] menn0: yeah, the lxd provider is awesome. I love how non-special it is. [02:27] Bug #1571914 opened: github.com/juju/juju/cmd/jujud unit tests fail if xenial is the LTS [02:27] Bug #1571916 opened: github.com/juju/juju/environs unit tests fail if xenial is the LTS [02:27] Bug #1571917 opened: github.com/juju/juju/juju unit tests fail if xenial is the LTS [02:42] wallyworld or axw: big peergrouper refactoring/cleanup http://reviews.vapour.ws/r/4636/ [02:42] menn0: does it fix the issue? [02:43] wallyworld: it might... it can't be reproduced easily (except in CI it seems) [02:44] ok [02:44] menn0: will look shortly [02:44] wallyworld: there were certainly some suspicious racy looking areas which I've elminated [02:44] code will be much better regardless [02:45] wallyworld: and these changes are very worthwhile regardless... there were some truly horrible bits [02:45] so it seemed [02:57] Bug #1567161 changed: juju2 beta3, cannot download charm, failed to download, 400

[02:57] Bug #1568176 changed: charm deployment requests invalid revision number

[02:59] why is there destroy controller, model, relation, service, unit, user.... but no destroy-machine? [03:11] wallyworld: thanks... I already fixed the typo in diff 2 [03:11] :) [03:11] menn0: changes look much nicer than what was there [03:12] wallyworld: by tomorrow we should know if they helped [03:12] indeed [03:14] natefinch: destroy is going away to be replaced by remove [03:16] for machines, relations, services, units, users [03:17] wallyworld: I'm fine with that as long as it's consistent. Given we at the code freeze, bugs-only phase.... there's not a lot of "going to be" left :) [03:17] natefinch: we will make changes as necessary till 2.0 is as it needs to be [03:17] the CLI work is still in progress [03:18] natefinch: all the terminology has been thoroughly reviewed at the highest level and agreed to [03:18] so that is what is being implemented [03:22] wallyworld: understood [03:22] :-) [03:27] Bug #1571923 opened: destroy-machine should exist in 2.0 [03:39] hi any tips on how to debug manual provider bootstrap failure? [03:39] i just get this: [03:39] 2016-04-19 03:39:14 ERROR juju.provider.manual provider.go:31 initializing ubuntu user: subprocess encountered error code 1 (Connection to 10.0.2.15 closed.) [03:41] hm, how do i get trace output? [03:44] natefinch: do you know if you left any containers running on the CI arm machine? [03:49] cherylj: was it you who had some problem with ssh terminating early and the manual provider a few months ago? [03:49] cherylj: I don't know but I can check [03:52] cherylj: I had, just cleaned them up [03:54] oh [03:54] nm [03:57] Bug #1571932 opened: github.com/juju/juju/provider/dummy unit tests fail if xenial is the LTS [03:57] Bug #1571933 opened: github.com/juju/juju/worker/provisioner unit tests fail if xenial is the LTS [04:33] Bug #1571947 opened: bootstrap --upload-tools fails with "cannot start bootstrap instance: missing tools URL" [05:41] katco: ping? [05:41] katco: probably too late for you, so I'll pick up bug 1567676 if you don't mind [05:41] Bug #1567676: windows: networker tries to update invalid device and blocks machiner from working

[06:45] Bug #1571982 opened: Centos7 machines fail to run cloud-init on Azure === rogpeppe3 is now known as rogpeppe [08:02] anyone here know much about how the lxd provider works? [08:04] dimitern: ping, 1:1? [08:04] frobware: oops, omw - sorry [08:18] Bug #1572022 opened: status randomly shows one of the peer relations, not both [08:34] Bug #1572022 changed: status randomly shows one of the peer relations, not both [08:44] perrito666, please ping me when you have a bit of time to chat about restores, I want to validate my assumptions [08:44] mgz: awake again already? [08:46] Bug #1572022 opened: status randomly shows one of the peer relations, not both [09:01] babbageclunk: dimitern: frobware: easy one to start the day http://reviews.vapour.ws/r/4638/ [10:19] frobware, voidspace: I can confirm MAAS 2.0 imposes the same restriction (for both devices and machines) on physical interfaces (can only be linked to untagged vlans) [10:19] here's the python script converted to py3 and finally working: http://paste.ubuntu.com/15927539/ [10:20] and here's a paste of the output showing the issue: http://paste.ubuntu.com/15927550/ [10:21] morning btw [10:44] updated the script to make it easier to switch maas versions: http://paste.ubuntu.com/15927766/ [10:50] frobware, voidspace: fyi, filed bug 1572070 [10:50] Bug #1572070: MAAS 2.0 cannot link physical device interfaces to tagged vlans, breaking juju 2.0 multi-NIC containers [11:22] wallyworld: https://bugs.launchpad.net/juju-core/+bug/1571832 -- what do you think about changing the tools storage to combine controller+model catalogies (model overlaying), so any tools added to the controller are available in all models? [11:22] Bug #1571832: Respect the full tools list on InstanceConfig when building userdata config script.

[11:22] catalogues* [11:23] axw: i think that makes sense, since model tools are sort of related to those of the host controller anyway [11:23] in that there's a version dependency [11:24] wallyworld: ok, I'll see what I can do later on then [11:25] azure is buggered on master atm, fixing that first [11:25] ok [11:29] frobware, voidspace: http://reviews.vapour.ws/r/4639/ fixes bug 1567676, please have a look when you have a moment [11:29] Bug #1567676: windows: networker tries to update invalid device and blocks machiner from working

[11:30] dimitern: looking [11:31] frobware: ta! [11:38] dimitern: done [11:39] frobware: thanks! [11:49] Bug #1571082 changed: autopkgtest lxd provider tests fail for 2.0

[12:01] frobware: updated http://reviews.vapour.ws/r/4639/ btw [12:01] dimitern: looking [12:02] dimitern: looking [12:03] dimitern: what do we consider invalid now? [12:04] frobware: what's considered invalid hasn't changed, just how we handle it [12:04] dimitern: I guess my concern is should we simply drop the validation and the respective tests? [12:05] dimitern: it seems a little counter intuitive to validate and then ignore [12:05] frobware: we could, but I'd prefer to keep it at least for a while - easier to catch issues once we start creating bridges from juju [12:06] frobware: or anything else on the machine that needs to use those device names (ip commands, etc.) [12:16] dimitern: if it is invalid shouldn't we encode it in a way that makes it safe to go into the db? [12:16] dimitern: if invalid names are not invalid it does seem weird - especially logging at Warning level [12:16] dimitern: assuming we're sure that's the right thing to do LGTM [12:16] dimitern: but like frobware, if there's no such thing as an invalid name then why even check? [12:17] dimitern: I leave it in your hands though [12:17] voidspace: cheers [12:18] dimitern: when will your branch removing AddressAllocation for MAAS land? [12:18] dimitern: as it impacts my devices work [12:19] voidspace, frobware: let me give you an example - on a windows machine, "Local Network Connection #2" is a valid name, but trying to save this into state goes via the API server which runs on Ubuntu, and has different criteria for valid names (much more restrictive) [12:19] voidspace: it landed yesterday [12:19] dimitern: I understand why we need to allow invalid names [12:19] Bug #1572102 opened: Juju could indicate if the LXD image is out-of-sync with upstream

[12:19] dimitern: what I don't understand is why we still need to check if they're invalid [12:20] dimitern: ah, so it has landed! [12:20] frobware: it's still safe to store it in mongo, no encoding needed [12:20] dimitern: I saw that the AddressAllocation flag was still used - but didn't see that it was just to emit a warning [12:20] cool [12:20] frobware, voidspace: ok, I think you convinced me to drop the tests and just keep the rest [12:20] dimitern: so let's drop the tests - they don't test anything we rely on [12:24] ashipika: picking on you since you're the only OCR that's online. would you mind taking a look at https://github.com/juju/juju/pull/5215 if you have time? [12:25] axw: my pleasure [12:26] frobware: done, do you want a last look before I hit the button? :) [12:26] dimitern: looking [12:27] dimitern: go for it [12:27] sinzui: in case you see the azure issue in CI before my fix lands, I'm pretty sure it's covered by https://bugs.launchpad.net/juju-core/+bug/1571947, and fixed by the branch linked [12:27] Bug #1571947: bootstrap --upload-tools fails with "cannot start bootstrap instance: missing tools URL" [12:28] * dimitern hits the button [12:31] axw: is it worth adding a note describing why the URL checking was reverted? [12:31] ashipika: yeah I guess. I'll create a tech-debt bug and link it [12:32] axw: ack [12:33] axw: great job! tyvm [12:33] now with that fix I can actually do some work with juju :) [12:33] dimitern: :) [12:36] axw: ship it [12:36] ashipika: thanks [12:43] Bug #1572116 opened: cloudconfig/instancecfg: SetTools should ensure URLs are set [13:24] thank you axw [13:46] dimitern: pong, hey [13:49] katco: hey, np - just fixed bug 1567676 and as you were assigned to it wanted to check before I started [13:49] Bug #1567676: windows: networker tries to update invalid device and blocks machiner from working

[13:53] dimitern: feel free to take it, i was spinning getting vmaas set up [13:53] dimitern: but please keep me updated on it. it looked strange to me, like an impossible code path. what's your idea? [13:53] katco: it's fixed now and landed [13:54] dimitern: what was it? [13:55] katco: the seemingly impossible path actually turned to be quite valid - a windows machine calling the api on ubuntu, which in turn tries to validate that api args as linux interface names :) [13:55] Bug #1572145 opened: kvmProvisionerSuite.TestContainerStartedAndStopped no event arrived

[13:58] dimitern: ahhhh! [13:59] :) [13:59] dimitern: that makes me feel better haha [13:59] dimitern: i was wondering if perhaps GOOS was being locked during compiled-time or something [14:00] katco: fortunately not :) thanks for looking into it though, I know the code is somewhat gnarly and hard to follow at places, but it's improving already === natefinch is now known as natefinch-afk [14:11] bbl ~1h [14:14] Bug #1572159 opened: juju deploy: cannot specify resource revisions [14:20] I keep having the urge to all-caps JUJU when I mention it in text. [14:20] (Also when I say it aloud, although to a lesser extent.) [14:21] That's normal, right? [14:21] :) depends on the locale [14:21] heh [14:23] Bug #1572159 changed: juju deploy: cannot specify resource revisions [14:26] Bug #1572159 opened: juju deploy: cannot specify resource revisions [14:38] frobware, voidspace, babbageclunk: one seriously complicated review guys: http://reviews.vapour.ws/r/4641/ :) [14:38] dimitern: SOUNDS LIKE A CHALLENGE! [14:40] :) [14:59] Bug #1569898 changed: cmd/pprof: sporadic test failure

[14:59] Bug #1571737 changed: Race in maas provider storage

[15:02] Bug #1569898 opened: cmd/pprof: sporadic test failure

[15:02] Bug #1571737 opened: Race in maas provider storage

[15:08] Bug #1569898 changed: cmd/pprof: sporadic test failure

[15:08] Bug #1571737 changed: Race in maas provider storage

[15:11] rogpeppe: FYI, I've commented on bug #1572159 [15:11] Bug #1572159: juju deploy: cannot specify resource revisions [15:13] ericsnow: thanks, i just responded [15:13] ericsnow: and it worked for me [15:13] rogpeppe: great! :) [15:13] ericsnow: out of interest, where in the code are the revnos detected? [15:14] rogpeppe: resource/resourceadapters/deploy.go [15:14] rogpeppe: via handleResources() in cmd/juju/service/deploy.go [15:29] Bug #1572159 changed: juju deploy: cannot specify resource revisions [15:32] Bug #1572159 opened: juju deploy: cannot specify resource revisions [15:38] Bug #1572159 changed: juju deploy: cannot specify resource revisions [15:52] dimitern, voidspace, frobware: Something for review - http://reviews.vapour.ws/r/4642/ [15:56] babbageclunk: almost done [15:59] babbageclunk: reviewed === redir_afk is now known as redir === natefinch-afk is now known as natefnich === natefnich is now known as natefinch [16:27] dimitern, frobware: Thanks! [16:28] voidspace: don't forget to update the document! :) [16:29] babbageclunk: hah, true enough :-) [16:29] I won't [16:30] voidspace: I'm going to replace the locks on fakeController with channels, then should I look at devices? [16:39] babbageclunk: I've started on devices, but mostly just understanding (or not) what's there [16:39] babbageclunk: we can share the work [16:40] babbageclunk: I'm doing AllocateContainerAddresses [16:40] voidspace: ok cool. I'll do this and then grab another one [16:41] cool [16:56] Bug #1572237 opened: juju rc1 loses agents during a lxd deploy [17:02] well juju tests with 1.6 in windows are broken with a big B [17:05] Bug #1572237 changed: juju rc1 loses agents during a lxd deploy [17:08] Bug #1572237 opened: juju rc1 loses agents during a lxd deploy [18:06] Well, there's a dumb bug [18:07] upgrade-charm barfs if you include the trailing slash for the path for a local charm [18:16] huzzah, fixed my bug === redir is now known as redir_lunch [19:44] Bug #1570759 changed: apt-get install juju does not install /usr/bin/juju

[19:50] morning folks [19:50] * thumper trawls through emails [19:50] sinzui: what's the status of a maas 2 ci test? [19:50] thumper: morning [19:50] thumper: we have a machine ready for us to install it. Installation will start tomorrow [19:51] sinzui: ok. will be good to see this get some further testing [20:01] alexisb: morning === redir_lunch is now known as redir [20:55] cherylj - ran into that same issue i bumped my head on last week and got enough to give you a proper bug this time around https://bugs.launchpad.net/juju-core/+bug/1572312 [20:55] Bug #1572312: Juju fails to deploy allocate node due to bad http response: 400 Bad Request [20:56] lazyPower: good news! menn0 has already fixed that issue :) [20:56] oh yeah? [20:57] lazyPower: yeah, see bug 1569054 [20:57] Bug #1569054: GridFS namespace breaks charm and tools deduping across models [20:57] Nice! That makes so much sense now why it was intermittent === natefinch is now known as natefinch-afk [21:53] Allah is doing [21:53] sun is not doing Allah is doing [21:53] moon is not doing Allah is doing [21:54] stars are not doing Allah is doing [21:54] planets are not doing Allah is doing [21:54] galaxies are not doing Allah is doing [21:54] oceans are not doing Allah is doing [21:55] mountains are not doing Allah is doing [21:55] trees are not doing Allah is doing [21:55] mom is not doing Allah is doing [21:55] dad is not doing Allah is doing [21:55] boss is not doing Allah is doing [21:55] job is not doing Allah is doing [21:56] dollar is not doing Allah is doing [21:56] degree is not doing Allah is doing [21:56] medicine is not doing Allah is doing [21:56] customers are not doing Allah is doing [21:56] you can not get a job without the permission of allah [21:57] you can not get married without the permission of allah [21:57] cmars: do you have 5 min? [21:57] nobody can get angry at you without the permission of allah [21:57] wallyworld, sure [21:57] light is not doing Allah is doing [21:57] cmars: https://plus.google.com/hangouts/_/canonical.com/tanzanite-stand [21:58] fan is not doing Allah is doing [21:58] businessess are not doing Allah is doing [21:58] the question is, what is op doing? [21:58] america is not doing Allah is doing [21:58] fire can not burn without the permission of allah [21:59] knife can not cut without the permission of allah [21:59] rulers are not doing Allah is doing [21:59] governments are not doing Allah is doing [21:59] sleep is not doing Allah is doing [21:59] hunger is not doing Allah is doing [22:00] food does not take away the hunger Allah takes away the hunger [22:00] water does not take away the thirst Allah takes away the thirst [22:00] seeing is not doing Allah is doing [22:00] hearing is not doing Allah is doing [22:00] wtf [22:01] seasons are not doing Allah is doing [22:01] weather is not doing Allah is doing [22:01] humans are not doing Allah is doing [22:01] animals are not doing Allah is doing [22:01] the best amongst you are those who learn and teach quran [22:02] one letter read from book of Allah amounts to one good deed and Allah multiplies one good deed ten times [22:03] hearts get rusted as does iron with water to remove rust from heart recitation of Quran and rememberance of death [22:03] heart is likened to a mirror [22:03] when a person commits one sin a black dot sustains the heart [22:06] to accept Islam say that I bear witness that there is no deity worthy of worship except Allah and Muhammad peace be upon him is his slave and messenger [22:06] read book www.fazaileamaal.com [22:06] read book www.muntakhabahadith.com [22:07] need spiritual teacher visit www.alhaadi.org.za [22:07] allah created the sky without any pillars [22:07] allah makes the sun rise from the east and sets it in the west [22:07] allah makes the day into the night and the night into the day [22:08] allah gives life and Allah gives death [22:08] all creation are useless,worthless,hopeless [22:08] can not do [22:08] can not benefit [22:08] can not harm [22:08] allah is the doer of each and everything [22:09] when Allah wants us to stand we stand [22:09] Bug #1511537 changed: Failed to load cookies EOF

[22:09]