[00:01] rick_h_: running a minute late [00:01] in another meeting [00:01] wallyworld: ty for the heads up [00:12] menn0: got a sec? [00:12] rick_h_: sure [00:12] can you hop in https://hangouts.google.com/hangouts/_/canonical.com/rick?authuser=1 [00:19] wallyworld, menn0: http://reviews.vapour.ws/r/5294/diff/# [00:25] thumper: we are both talking to rick, will look soon [00:26] bbl [00:27] ack [00:36] thumper: lgtm, ta [00:36] ta === frankban is now known as frankban|afk [00:57] thumper, wallyworld, do we hvae anything in the release that will break cli scripts or thta are api changes?? [00:57] if so we need to send them to the list and note them in the release notes [00:57] alexisb: not this release. when i land my queued branch after the beta is cut, that will be an api change [00:57] I didnt see anything, but it would be easy to miss something [00:57] yep wallyworld I remember that one [00:57] alexisb: but even then, there's a shim for compatibility so stuff doesn't break [00:57] as soon as htat lands we will need a note to juju-dev [00:58] yep [00:58] menn0, did you see my draft note [00:58] anything you want me to add/edit? [00:58] alexisb: sorry, I've been pulled all over the place [00:58] alexisb: looking right now! [00:59] anastasiamac, how close are we to your PR landing? === natefinch-afk is now known as natefinch [01:02] alexisb: it has not been reviewed [01:02] o yeah, natefinch ^^^ [01:02] that one is critical [01:02] alexisb: done. Two small suggestions. [01:02] if you are around [01:02] menn0, thanks [01:03] anastasiamac: link me? [01:03] natefinch: the critical PR to review is: http://reviews.vapour.ws/r/5291/ [01:03] natefinch: tyvm \o/ [01:46] later #juju-dev [01:47] bye redir [01:52] Hi all - anyone able to tell me what this means? ERROR cannot upgrade service "prometheus" to charm "prometheus": storage "metrics-filesystem" removed [01:52] ^ This is from an environment which, AFAICT, has never used juju storage [01:54] ah, looks like it's https://bugs.launchpad.net/juju-core/+bug/1599503 [01:54] Bug #1599503: Cannot upgrade charm if storage is modified, even if the service doesn't use said storage [01:55] Anyone able to suggest a workaround until such time as that bug is investigated? [01:56] natefinch, just checking that you have anastasiamac review, otherwise I will allocate to someone else [01:59] alexisb: I am looking at it. unfortunately, I'm not super familiar with how image metadata is used [02:00] natefinch, that is fine, so long as you are looking at it I am good :) [02:04] natefinch: m happy to pair up for review if it helps \o/ want to HO? [02:06] thumper, when you get back please make sure to support anastasiamac in what she needs, I have left her with the "emailing to release team" task so I can eod. [02:07] goodnight all! [02:07] alexisb: nite! [02:16] anastasiamac: HO that would be good, to help me understand what's going on, so I don't have to type out a bunch of dumb questions :) [02:18] natefinch: awesome. ill create one right now [02:19] natefinch: invite ine calendar/email [02:24] * thumper gets a cuppa before chasing up with menn0, anastasiamac, and wallyworld [02:25] thumper: flat white please [02:25] * anastasiamac does not nee to be chased. all under control for now [02:25] need* [02:52] natefinch: PR updated... unless u want to PTAL, m reading for landing \o/ [02:53] anastasiamac: I gave you a fixit then shipit. That means I trust you to fix the problems. If I didn't, it wouldn't get the shipit :) [02:53] natefinch: u r awesome. pm pressing the button then ;D [02:53] do it! :)_ [02:57] blahdeblah: raised 1599503 to critical to be addressed in next beta. but when u say "production", what version of juju are u using? 1.25.6? [03:03] Bug #1605449 opened: juju2: command error message (usage) consistency [03:04] wallyworld: wanna catch up? [03:05] anastasiamac: all good? [03:05] sure [03:05] thumper: of course :D always, sir! [03:05] anastasiamac: do we have everything landed we need to? [03:05] thumper: landing now. will send an email once it's done.. [03:05] almost there... [03:33] Bug #1605383 changed: Tab complete causes traceback when trying to tab-complete juju commands [03:46] thumper: fyi, that unit storage bug is a legit issue when destroying machines; no user warning needed, just a code fix it appears [03:47] cool [04:19] wallyworld: how come add-credential requires the cloud/provider name on the CLI, but prompts for everything else? [04:20] natefinch: the info we were given asked that add-credential take a cloud name as a positional arg. i guess that doesn't have to be the case. i think the expectation was that the user knew what cloud they were adding a credential for since they would do it as they wanted to bootstrap that cloud [04:23] wallyworld: in theory they know all the information they're prompted for :) [04:24] huh? that's sort of irrelevant to thr question at hand. the point of add-credential is to get what they know and have juju record it in credentials.yaml. the current add-credential simpy exapcts them to nominate what cloud they are adding a credential for [04:28] wallyworld: I just mean, saying they know what cloud they want to add credential for doesn't seem like much of an argument, they should know what name they want to call the credential, too, etc. It just seems weird to require one specific positional arg, but prompt for everything else. [04:30] the use case is that they know what cloud they want to bootstrap, so instead of going through a step of prompting for that, short circuit it and just rompt for the bits that need to be interactive. but that's just one way of looking at it and was in the spec we were given [04:30] no reason why it couldn't prompt for cloud name if not supplied [04:39] I think it would be more user friendly if we did that.. it would also put it in line with how bootstrap works (i.e. if there's no arguments, go into interactive mode). [04:41] anyway, I gotta get to bed, and plan on how to move to Canada if Trump becomes President [04:44] anastasiamac: thanks, but yeah - 2.x is not any good to us; it's *production*, not "production", and it's all 1.2x :-) [04:45] anastasiamac: For now I just worked around it by reverting the charm storage code to the way it was previously. I'm about to test the upgrade now. [05:49] blahdeblah: ack === frankban|afk is now known as frankban [09:03] frobware: is there a standup today? or is that leftover from the old meetings? [09:07] babbageclunk: left over [09:07] :( [09:07] sorry [09:10] frobware: might pop my head into the Rick team standup today just for some interaction. (: [09:10] babbageclunk: sure [10:04] Bug #1544724 changed: repeatedly checks /dev/fd0 when it doesnt even exist [10:52] Bug #1605593 opened: Fix missleading values in error message: x is not a valid distro_series. [11:12] morning [11:15] Has anyone had any luck using go-guru on the juju codebase? [11:16] babbageclunk: I believe my vim-go setup uses it [11:18] perrito666: If I try to find callees using juju/juju/... as the scope one of my cores just shoots to 100% and nothing else happens. [11:19] babbageclunk: a I usually don't [11:20] but sound plausible, especially if it doesnt have any caching policy [13:17] macgreagoir: ping, morning [13:17] rick_h_: g'day [13:18] macgreagoir: howdy, as OCR today I wanted to ping and see if you can get a chance to peek at http://reviews.vapour.ws/r/5290/ before your EOD please [13:19] rick_h_: I had tried earlier, but struggled to understand it to review. I will look again. [13:19] macgreagoir: k, wrangle up someone to pair up with you then, get a hangout and go over it together [13:19] ack [13:19] macgreagoir: and run down using the review checklist which should help at least knock out chunks of the review [13:20] perrito666: Turned out I was using way too broad a scope - setting something more specific churned for a bit but then gave good results. [13:20] macgreagoir: so your pair'ing partner can limit down to things you have questions/uncertaintly on [13:20] dooferlad: can you take a few to meet up with macgreagoir to help please? ^ [13:21] macgreagoir: and in the future if you look but aren't sure make sure to leave a comment to the effect in the PR so we can see where the PR stands/etc. [13:21] rick_h_: macgreagoir sure [13:21] ty much dooferlad [13:22] macgreagoir: give me a moment to refill my water. Melting here. [13:22] nw [13:25] dooferlad: I'm in the quartz ho, when you're ready. [13:45] gsamfira: you around? [13:49] sinzui: do you have a windows 2012 machine I could RDP into to check out its TLS support? [13:51] natefinch: developer-win-unit-tester.vapour.ws [13:55] natefinch: check your email for info about connecting [13:56] natefinch: and I promised alexisb that I will get a google doc explaining access to all CI hosts so that hyou don't need to wait for me [14:00] sinzui: awesome, thanks :) [14:02] dooferlad: ping for standup [14:14] Bug #1564662 changed: Juju binaries should be stripped [14:17] redir: ping [14:20] Bug #1605653 opened: backup-restore failed creating collection juju.blockdevices [14:23] Bug #1605653 changed: backup-restore failed creating collection juju.blockdevices [14:28] macgreagoir: want to touch base with the relations bug? [14:29] frobware: Yes, please. Standup HO? [14:29] macgreagoir: 10mins [14:29] nw [14:38] Bug #1605653 opened: backup-restore failed creating collection juju.blockdevices [14:47] balloons sinzui we havne't been putting beta builds into homebrew, this is problematic [14:48] macgreagoir: in standup HO for whenever works [14:48] marcoceppi: homebrew think juju2 is problematic [14:48] why. [14:49] marand I ahve been working with them to put juju2 into homebrew [14:49] marcoceppi: it is not compatible with juju 1. Users will get an update and break [14:49] katco:pong [14:50] we should just build a `juju load-credentials` which takes .juju/environments.yaml and converts it [14:50] marcoceppi: This the the current PR after other were abandoned https://github.com/Homebrew/homebrew-devel-only/pull/47 [14:50] redir: hey i was just looking at your race condition again trying to understand what i missed. i'm leaving a comment on the PR [14:50] redir: aside from the race detector complaining, are we seeing an actual issue from this? [14:51] sinzui, sigh.. they killed the other PR in the end eh [14:53] balloons: yes. The direction is for users to install juju or juju2 they conflict. the user chooses the version tehy need. [14:53] balloons: marcoceppi: The killer was the fact that juju's plugin detection is bong [14:54] what does bong mean? [14:56] katco: ResetStats could change (create a new one with copied values) or SetStats could remove (setting to nil) the underlying stats object during access. [14:58] katco: AFAIU Stats is only used in testing. So I think it is possible to get wrong results, but as to whether we were seeing them -- I don't know [15:02] marcoceppi: something that takes the sticky icky? [15:18] redir: i think the point i was going to make is invalid. i am worried that we have a logic race in our tests that this will fix, but at the wrong logical level [15:20] OCR: review please, http://reviews.vapour.ws/r/5295/ [15:24] mgz: ack [15:26] Bug #1605669 opened: grant-revoke User could not check status with read permission [15:31] macgreagoir: if it didn't get through before I dropped... [15:31] for simiplicty, compare with http://reviews.vapour.ws/r/5205 - it should be the inverse === JoseeAntonioR is now known as jose [16:01] mgz redir: 5295 +1 from me too. mgz, thanks for the 5205 link. [16:02] thanks guys === macgreag1ir is now known as macgreagoir [16:08] dooferlad: it's something (ip6tables, et al) between my libvirt interface and the host that causes the packets to get dropped. see nothing obvious atm [16:12] frobware: :-( [16:28] frobware: lp:1603473 updated. The work-around may not be helpful to the containers (if they just grab all active interfaces), but we'll see if it starts to help, at least. [16:31] macgreagoir: thanks [16:38] Bug # changed: 1592456, 1600404, 1603865, 1604223 === frankban is now known as frankban|afk [17:32] katco: ping [17:32] rick_h_: pong [17:32] katco: can I steal some time? [17:32] rick_h_: sure [17:32] katco: https://hangouts.google.com/hangouts/_/canonical.com/rick?authuser=1 [17:33] frobware, ping [17:33] alexisb: pong [17:33] heya frobware happy friday [17:33] on this bug: https://bugs.launchpad.net/juju-core/+bug/1604482 [17:33] Bug #1604482: MAAS bridge script should drop all 'source' stanzas from original file <2.0> [17:34] why is it critical if there is a workaround [17:36] alexisb: so... people keep running into this. and the behaviour is subtly broken when it happens. I wanted to say "when juju boostraps your node, we will fix networking for you" [17:37] alexisb: I did originally take the opinion that we should rely on the workaround, but the curtin fix has not landed and it has been >4 ? weeks now [17:37] so that bug requires a curtain fix? [17:37] alexisb: happy to drop the priority [17:38] alexisb: there fix is committed, but not released. needs to get out of proposed [17:38] lets keep it on our radar but make it high [17:38] frobware, thanks for the info, I will leave you to your day now [17:38] alexisb: fine for me [17:54] * rick_h_ EOD's to pack and get ready to get out of here tonight. If anyone needs anything hit me up on my phone via email, telegram, or such [17:58] * mgz sends smoke signals to rick_h [17:58] *have* [17:58] *a* [17:58] *good* [17:58] *flight* [18:07] runRawTransaction is the one that doesnt do the model uuid mangling right? [18:11] Bug #1605710 opened: Fix and reland axw/cli-model-owner [18:11] Bug #1605714 opened: juju2 beta11: LXD containers always pending on ppc64el systems [18:16] so master is unblocked? [18:16] redir, yes sir [18:17] SO I can $$merge$$? [18:17] CRAP! quick someone block it! [18:17] yes sir [18:17] katco, it is staying unblocked tell thursday [18:17] noooo [18:17] but if you cause a regression you will get assigned to it as your top priority [18:17] i... i can write a race condition in if you want [18:17] * redir light some firecrackers and pops champagne [18:17] katco, its your own pain ;) [18:19] great someone put conflicts in my way [18:20] alright all, I am off for a quick ride, back in about an hour [18:20] happy merging [18:20] have a good ride [19:04] oh internet explorer.... thanks for wrapping this plainttext file I downloaded in 
[19:12]  ahh... windows is so crappy
=== freyes__ is now known as freyes
[19:42]  hey.  i friggin love beta12.  nice work folks!!  (resource-get slowness was killing me, i feel alive again now)
[19:46]  kwmonroe: awesome! :)
[19:53]  Bug #1605747 opened: [ juju2 beta11 ] Maas system is deployed but agent remains pending    
[20:00]  gah... I cannot for the life of me get .net to use tls 1.2
[20:01]  natefinch: the language?
[20:02]  the framework, yeah.  in this case I'm calling it from powershell... but it always requests tls 1.1 as the highest value
[20:07]  ah... beacuse it's .net 4.0
[20:09]  well shit
[20:09]  just .net get a new ver... oh, thats right, you cant
[20:10]   lol, juju depends on .net 4.5, evidently ;)
[20:10]  natefinch: since when exactly?
[20:11]  it used to work with windows 2k images that, afaik, have 4.0
[20:11]  * perrito666 admits his knowledge of the ungo-ly word of windows :p
[20:11]  but recently we changed to only supporting TLS 1.2
[20:11]  and 4.0 doesn't support TLS 1.2 as far as I can tell
[20:12]  ahh, I see
[20:12]  our userdata uses powershell, which uses .net
[20:12]  that is indeed a problem, because windows 2k is a very popular hipervisor img
[20:12]  and I dont even want to think in the consequences of upgrading .net
[20:13]  mm, isnt there a 3rd party .net implementation of tls 1.2?
[20:13]  that happens to be floss?
[20:13]  or a patch? or something?
[20:15]  kwmonroe, be sure to make use of all the extra time
[20:19]  well, I thinj the OS itself supports TLS 1.2, it's just that .net 4.0 doesn't use it, because of windows XP compatibility (at least that was suggested in an article I read, and sounds plausible)
[20:22]  should things work on mongo 3.2?
[20:22]  redir: things is a broad word mate
[20:23]  statistically speaking, stuff should not work in mongo 3.2
[20:23]  for most values of stuff
[20:23]  juju is not one of those though, it should just work (except for the tests, those wont)
[20:23]  well, I thinj the OS itself supports TLS 1.2, it's just that .net 4.0 doesn't use it, because of windows XP compatibility (at least that was suggested in an article I read, and sounds plausible)
[20:23]  Bug #1577949 changed: windows services cannot upgrade to 1.25.6       
[20:23]  Bug #1605756 opened: [ juju2 beta11 ] system show up in juju status as pending but there is no attempt to deploy in maas     
[20:24]  look at perrito666 picking up aussie slang, nice job mate!
[20:24]  I might be spending too much time with down unders
[20:24]  perrito666: juju
[20:25]  I see so tests won't pass
[20:25]  so no juju doesn't work with 3.2 yet
[20:25]  if you increase timeout to something near 2 days they will actually
[20:25]  lol
[20:26]  for 10 points, guess what is the part of juju that does not work well with mongo 3.2?
[20:26]  uh, the part that uses mongo?
[20:26]  lol
[20:26]  mmm it doesn't even find it's path
[20:27]  no one will say juju-conn suite?
[20:27]  redir: ah, that is new
[20:27]  heh.... we said tests
[20:27]  redir: you have to tell the tests where is mongo
[20:27]  the tests are what, like 10x slower on 3.2?
[20:27]  natefinch: I think it was 100x
[20:27]  OMG
[20:28]  we deviced a way so they would run at the same speed, it required some hacking of the mongo options plus a ramdisk plus I cant recall what other thing
[20:28]  its all the fault of the new storage for mongo, it is fast for anything but what we do in the tests
[20:28]  if you make the tests run with the legacy storage they pass fast
[20:29]  the thing is, that storage engine is EOL
[20:29]  * perrito666 knows more about this subject than he would like
[20:30]  perrito666: finds it if I set JUJU_MONGOD
[20:30]  yes, that is the one
[20:31]  and unintstall all the old mongos
[20:31]  the mongo using tests are not all that smart
[20:31]  which I did first
[20:31]  I have another adjectives for it
[20:31]  I have a few myself
[20:31]  lol
[20:31]  but thought I'd try out the tests while I shepherd some merges
[20:31]  this deserves t-shirts
[20:31]  with the new mongo pkg
[20:32]  redir: unless you intend to be merging for ~30min*100 better try something else
[20:32]  10gen printed plenty of shirts
[20:32]  which might be part of the problem
[20:32]  perrito666: 3 merges...
[20:33]  ~ 27 minutes each
[20:33]  last one starting now
[20:39]  oh, so its 20 days, you might want to find something else to do
[20:42]  Bug #1575794 changed: Agent config format version should be changed for 2.0     
[20:42]  Bug #1584565 changed: Documentation regarding user management is out of date    
[20:42]  Bug #1590947 changed: TestCertificateUpdateWorkerUpdatesCertificate failures on windows     
[20:45]  Bug #1575794 opened: Agent config format version should be changed for 2.0     
[20:45]  Bug #1584565 opened: Documentation regarding user management is out of date    
[20:45]  Bug #1590947 opened: TestCertificateUpdateWorkerUpdatesCertificate failures on windows     
[20:48]  weird, it says we have .net framework 4.5 installed, but powershell is still using 4.0
[20:52]  nope, I'm wrong... .net 4.5 is clr version 4.0.30319.17001 - 18400 ... makes perfect sense
[20:54]  Bug #1575794 changed: Agent config format version should be changed for 2.0     
[20:54]  Bug #1584565 changed: Documentation regarding user management is out of date    
[20:54]  Bug #1590947 changed: TestCertificateUpdateWorkerUpdatesCertificate failures on windows     
[20:54]  Bug #1605767 opened: MachineSuite.TearDownTest no reachable servers      
=== natefinch is now known as natefinch-afk
[21:18]  Bug #1605769 opened: LXD Bootstrap Failure: Cannot add new machine     
[21:18]  Bug #1605770 opened: firewallerSuite.TearDownTest inst.Dial() failed       
[21:18]  Bug #1605776 opened: annot dial mongo to initiate replicaset: no reachable servers      
[21:42]  Bug #1605777 opened: munna/juju2 is too slow to deploy windows images       
[21:55]  alrighty all, spending some time with family after a long week and before  I travel
[21:55]  have a great weekend everyone
[21:58]  alexlist: tc
[21:58]  oops mt
[22:18]  Bug #1605790 opened: Unable to initialize agent   
[23:07]  perrito666: the full suite on 3.2 only takes about 2 minutes longer than 2.6 for me.
[23:10]  redir: nice, most likely using the old storage engine
[23:10]  if not, wow