/srv/irclogs.ubuntu.com/2016/06/24/#juju-dev.txt

axwwallyworld: I was thinking we'd have a separate command and API to revoke credentials00:28
wallyworldaxw: no worries, wasn't sure as both approaches can be used00:29
katcowallyworld: hey sorry that patch you reviewed is kind of a mess. it's WIP and dependent on 2 other PRs, so some of the stuff you've commented on are covered in other PRs00:39
wallyworldkatco: no worries, just wanted to give some initial feedback00:39
katcowallyworld: i'm not sure why review board doesn't do the diff between a branches parent00:40
katcowallyworld: yeah ta00:40
axwwallyworld: can you please see my reply, I'm just live testing one last time00:50
wallyworldok00:50
wallyworldaxw: lgtm00:51
axwwallyworld: ta00:51
=== thumper-afk is now known as thumper
marcoceppistill no shared filesystems in Juju?01:26
wallyworldmarcoceppi: that got bumped in lieu of other feature work01:29
marcoceppiit basically makes every web application impossible to charm still without some storage subordinate janky charm01:29
marcoceppiwordpress, mediawiki, django apps, anyhting storing uploads is just garbage01:30
wallyworldmarcoceppi: i hear you. people with higher pay grades mae those priority calls01:30
marcoceppii hear you, wish it was different01:30
wallyworldme too01:30
axwwallyworld: ok for me to land this cloud change right?01:31
thumpermarcoceppi: it is our unique funding model :)01:31
axwbefore beta01:32
axw... because they'll just pick another rev if I break things01:32
wallyworldaxw: let's just check that the previous CI run passes01:32
axwok01:32
wallyworldaxw: a minute late01:59
wallyworldaxw: sorry, thumper won't stop talking02:07
thumper:(02:07
axwheh02:07
wallyworldmenn0: you arriving sat arvo? anastasia will be there alredy and i arrive sat evening, so let's do something sunday, unless you have other plans02:23
axwwallyworld: reviewed02:31
wallyworldta02:31
menn0wallyworld: sounds good. I haven't made any plans for sunday.02:31
menn0wallyworld: I arrive sat afternoon02:32
wallyworldmenn0: ok, i arrive at airport around 20:3002:32
wallyworldwill catch up for breakfast if not before - maybe you'll be at bar drinking :-)02:33
menn0wallyworld: it has been known to happen :)02:33
wallyworldto the best of us02:34
menn0wallyworld: although I might hit some of my old haunts instead of the hotel bar02:34
menn0wallyworld: it's my old hood02:34
wallyworldnice02:34
menn0wallyworld: i'll have data so ping me on twitter or something02:34
* wallyworld doesn't do social mendia02:34
wallyworld*media02:34
menn0wallyworld: of course, I forgot :)02:35
wallyworldi'll be at breakfast first thing02:35
menn0wallyworld: email then02:35
wallyworldor hangout message02:35
menn0even better02:35
menn0and failing that I'll be at breakfast in the morning too02:35
axwwallyworld: http://reviews.vapour.ws/r/5156/03:58
wallyworldlooking03:58
axwwallyworld: I think I'll look at splitting model/controller config schemas03:59
axwnow03:59
wallyworldsgtm, hopefully won't take much03:59
mupBug # changed: 1414890, 1460685, 1491578, 150633804:25
axwwallyworld: I'm really not fond of that code I took out. I don't think it's needed, because we always generate SSH keys for the client anyway.04:35
wallyworldok, we can always change if needed04:36
=== urulama|__ is now known as urulama
ericsnowwallyworld: FYI, I have a WIP in for the "final" patch (the one that actually adds the worker)05:50
ericsnowwallyworld: http://reviews.vapour.ws/r/5157/05:50
ericsnowwallyworld: now I need to be asleep :)05:50
mupBug #1414890 opened: Network security tests fail, all ports open when using the local provider <charms> <local-provider> <juju-core:New> <https://launchpad.net/bugs/1414890>06:55
mupBug #1414890 changed: Network security tests fail, all ports open when using the local provider <charms> <local-provider> <juju-core:New> <https://launchpad.net/bugs/1414890>07:04
mupBug #1414890 opened: Network security tests fail, all ports open when using the local provider <charms> <local-provider> <juju-core:New> <https://launchpad.net/bugs/1414890>07:10
=== frankban|afk is now known as frankban
* babbageclunk sighs.08:34
babbageclunkdimitern: ping?08:35
babbageclunkdimitern: I hacked up the postgres charm last night - it's pretty neat!08:43
dimiternbabbageclunk: hey08:47
dimiternbabbageclunk: sorry, still reading the headlines...08:47
babbageclunkdimitern: man, tell me about it.08:47
dimiternbabbageclunk: so what about the charm?08:47
dimiternbabbageclunk: you added bindings?08:48
babbageclunkdimitern: No, just in reference to your email about charming - I was hacking workload version into the postgresql charm.08:48
dimiternbabbageclunk: ah :) I see08:49
babbageclunkdimitern: but it was fun, and easier than I'd anticipated08:49
babbageclunkdimitern: hmm - about bindings - are we (or maybe charmers) going to need to come up with standard names for spaces so that bindings match up between different charms?08:50
dimiternbabbageclunk: it is neat, getting into charming for sure is - I was surprised, and as they say here 'appetite grows along with eating more of it' :)08:50
babbageclunkdimitern: ha08:50
babbageclunkdimitern: ...or is there a way to say "this binding corresponds to this space" when deploying?08:51
dimiternbabbageclunk: yeah - `juju deploy .. --bind '<ep1>=<sp1> <ep2>=<sp2> ..'`08:52
babbageclunkdimitern: Ah, hadn't seen that. Cool.08:53
dimiternbabbageclunk: or just --bind '<sp>' to bind all to the same (can also mix those 2)08:53
babbageclunkdimitern: Would you be able to look at this review? It's pretty straightforward. http://reviews.vapour.ws/r/5152/08:53
dimiternbabbageclunk: sure08:54
dimiternbabbageclunk: LGTM09:01
voidspacedimitern: babbageclunk: be at standup in 509:01
dimiternok09:04
dimiternfrobware: standup?09:04
aim__dimitern: yep, trying...09:14
frobwaredimitern: did you see the latest missing 'search' comments in /etc/resolv.conf. https://bugs.launchpad.net/juju-core/+bug/157594009:35
mupBug #1575940: LXC containers under MAAS get no "search <domain>" entry in resolv.conf when deployed with juju2 <cdo-qa-blocker> <landscape> <juju-core:Triaged> <https://launchpad.net/bugs/1575940>09:35
dimiternfrobware: looking09:36
dimiternfrobware: that's beta9 though..09:37
frobwaredimitern: right, the changes you landed will only be available in 10 or a daily ppa09:37
dimiternfrobware: yeah, but I'll try to repro those last steps with b10 + 1.9.3 and 2.0b8 maas09:38
frobwaredimitern: great, thanks09:40
* dimitern thinks .. at least brexit didn't happen just before a release.. that would've been bad.. oh wait a minute! :/10:21
andrey-mpjamespage: Am I right that next release of OpenStack charms will be 07/16? If it so - when last changes can be made for this release? Is it possible to merge my changes for nova-compute https://review.openstack.org/#/c/331249/ ?11:22
jamespageandrey-mp, hey!11:23
jamespageandrey-mp, https://launchpad.net/charms/+milestone/16.07 due 07/2811:23
andrey-mphi!11:23
jamespagewe freeze two weeks in advance of that date - so I think that's the 14th11:23
jamespagefor features to land; I have your review on my todo list11:24
andrey-mpgood, thanks )11:24
jamespageandrey-mp, do you have an example ephemeral backend charm as well?11:24
andrey-mpwill wait for comments11:24
andrey-mpyeah - https://jujucharms.com/u/cloudscaling/scaleio-openstack11:24
andrey-mpthis is charm for connecting ScaleIO storage as ephemeral to nova-compute11:25
andrey-mpthis charm also connects ScaleIO storage to cinder via storage-backend11:25
dimiternfrobware: ping11:42
frobwaredimitern: pong11:46
dimiternfrobware: re bug 157594011:46
mupBug #1575940: LXC containers under MAAS get no "search <domain>" entry in resolv.conf when deployed with juju2 <cdo-qa-blocker> <landscape> <juju-core:In Progress by dimitern> <juju-core 1.25:Triaged> <https://launchpad.net/bugs/1575940>11:46
dimiternfrobware: I've reproduced it, working on a fix11:46
frobwaredimitern: what's the issue?11:47
dimiternfrobware: good we caught this before beta10 is out, and I strongly suspect it affects 1.25 too11:47
frobwareooh11:47
dimiternfrobware: adding a comment to the bug now - basically the issue is with `if !haveDNSConfig { // parse resolv.conf }` (in worker/provisioner/broker.go11:48
dimiternfrobware: i.e. it happens if we only have dns servers (which is the only thing we can get from MAAS 1.9 API), but not search11:49
dimiterneasy to fix at least - just need to check whether we have both fields set11:50
frobwaredimitern: bleh11:51
frobwaredimitern: that's horrible11:51
dimiternfrobware: yeah :/11:52
mupBug #1595937 opened: there is no way to configure the availability zone for neutron's dhcp agents <juju-core:New> <https://launchpad.net/bugs/1595937>12:02
frobwaredimitern: do you have some time to talk about the fan?12:15
dimiternfrobware: yeah, but can it be a bit later?12:16
dimiternfrobware: I need to go to the bank now (again)12:16
frobwaredimitern: yep sure - np.  Selling sterling? :-D12:16
dimitern:D12:16
dimiternfrobware: I wish12:16
frobwaredimitern: I'm just going to invent myown12:17
dimiternfrobware: there's the fix for dns search missing: http://reviews.vapour.ws/r/5159/12:17
dimiternvoidspace, babbageclunk: ^^ please take a look12:17
* dimitern steps out, back in ~30m hopefully12:17
voidspacedimitern: LGTM12:18
dimiternvoidspace, babbageclunk, frobware: thanks guys!13:18
dimiternfrobware: hey, you wanted to sync re the fan?13:50
=== redelmann is now known as rudi|wfh
mupBug #1588575 changed: allwatcher_internal_test has intermittent failure <intermittent-failure> <juju-core:Fix Released by dave-cheney> <https://launchpad.net/bugs/1588575>14:08
=== alexisb-afk is now known as alexisb
babbageclunkdimitern, frobware, voidspace: review please? http://reviews.vapour.ws/r/5161/14:17
dimiternbabbageclunk: looking14:18
babbageclunkThanks!14:18
dimiternsinzui: hey, did we cut 2.0 beta10 yet?14:25
dimiternsinzui: FWIW there's a fix landing as we speak, which will be good to be in beta1014:25
sinzuidimitern: yes we have14:26
dimiternsinzui: ok, then the fix will be in beta11 I guess14:26
sinzuidimitern: yep. We will release next week14:27
dimiternok14:27
alexisbkatco, ping14:36
katcoalexisb: pong14:36
alexisbperrito666, ping, you still around today??14:37
cheryljdimitern: what fix?14:38
mupBug # changed: 1467331, 1486712, 1567951, 1567963, 1568090, 1571792, 1576003, 1593299, 1593566, 1593761, 1593838, 1594232, 1594335, 1594580, 1594855, 1594875, 159527814:38
dimiterncherylj: for bug 157594014:38
mupBug #1575940: LXC containers under MAAS get no "search <domain>" entry in resolv.conf when deployed with juju2 <cdo-qa-blocker> <landscape> <juju-core:Fix Committed by dimitern> <https://launchpad.net/bugs/1575940>14:38
natefinchwelp.... our meals will be cheaper at the sprint at least.14:38
cheryljdimitern: oh, you figured it out?!  awesome!14:39
dimiterncherylj: yeah :)14:39
perrito666alexisb: I am intermittently14:41
cheryljdimitern:  I bet ahasenack is relieved that he wasn't going crazy with that bug.14:43
cheryljdimitern: thanks for tracking it down14:43
dimiterncherylj: :) yeah - if I looked more carefully earlier I would've been able to repro and fix it earlier..14:44
ahasenackcherylj: :)14:45
natefinchsinzui: has your MAAS come back up?  Wondering how the HA CI test fared14:46
dimiternahasenack: sorry ;)14:46
sinzuinatefinch: yes, let me see if I can find you two results. master consistently fails, but model-migrations passed14:46
ahasenackI was the last standing witness, since cgregan wasn't seeing it anymore14:46
katcodoes anyone know if we can still accept patches from people in the UK?14:47
dimiternI bet he didn't go and set dns servers on subnets directly though, and it can't be repro'd with a default maas installation14:47
natefinchkatco: lol14:47
dimiternkatco: don't be nasty to the poor guys :D14:48
sinzuinatefinch: on master, HA is never reached: http://reports.vapour.ws/releases/4099/job/functional-ha-recovery-maas-1-9/attempt/814:48
natefinchsinzui: ok.  Does the same test pass on other substrates?14:48
sinzuinatefinch: the model-migrations branch passes http://reports.vapour.ws/releases/4097/job/functional-ha-recovery-maas-1-9/attempt/614:49
katcodimitern: lol wasn't meant to be a jab. commentary on isolationism, ramifications of that vote14:49
sinzuinatefinch: Yes, the test passes on aws, azure, and openstack14:49
dimiternkatco: :) jk14:49
katcodimitern: how are you btw14:49
natefinchsinzui: ok, good. That jives with my manual testing.14:49
dimiternkatco: tired.. I'm looking forward to getting back from the sprints hehe14:50
dimiternbut also for some beers while at them14:50
katcodimitern: :)14:50
natefinchsinzui: interesting that it passes on the model migration branch... thanks for that info, that could be very useful14:50
dimiternsomebody didn't use the pre-push hook recently..14:51
dimiternchecking: go vet ...14:51
dimiternprovider/openstack/firewaller.go:234: arg c.environ.Config().UUID() for printf verb %s of wrong type: (string, bool)14:51
dimiternprovider/openstack/firewaller.go:354: arg c.environ.Config().UUID() for printf verb %s of wrong type: (string, bool)14:51
dimiternprovider/openstack/firewaller.go:368: arg c.environ.Config().UUID() for printf verb %s of wrong type: (string, bool)14:51
dimitern 14:51
natefinchdimitern: I never use the pre-push hook.  If we don't enforce it on the bot, it's not a real rule.  But my editor runs go vet on save, so....14:52
dimiternnatefinch: good point, it should be part of what 'make check' does14:53
natefinchdimitern: I also never use the makefile :)14:53
dimiternnatefinch: but the merge bot does ;)14:53
natefinchdimitern: indeed.14:53
babbageclunkforgot - I need to go home now while my wife goes to the doctor, but I'll be back online from there.14:56
katcoredir: standup time15:03
alexisball I will be out for a bit this morning, call my cell if somehting is urgent15:09
=== alexisb is now known as alexisb-afk
perrito666bbl mail me if you need me urgnently ill be on and off all day15:11
=== frankban is now known as frankban|afk
=== alexisb-afk is now known as alexisb
perrito666is there a way to ask state if a model is controller model?16:55
mupBug #1594865 changed: 2.0 beta8: deployment with sphere as provider - bootstrap node - no space left on /  <oil> <juju-core:Invalid> <https://launchpad.net/bugs/1594865>17:05
voidspacefrobware: simple-ish one if you're still around http://reviews.vapour.ws/r/5162/17:09
frobwarevoidspace: remind me again... what's the deal with omitempty and time?17:11
voidspacefrobware: it just doesn't work...17:12
voidspacefrobware: it gets omitted all the time17:12
voidspacefrobware: even when it's there - so you have to use a pointer17:12
frobwarevoidspace: asked for an explanation in the review.17:12
frobwarevoidspace: as in we add something explainging what doesn't work17:13
voidspaceok17:13
frobwarevoidspace: I'm rushing this; will look on monday as I need to go out17:16
perrito666K going to London bbl17:16
katcoperrito666: safe travels17:16
voidspacefrobware: perrito666: bye o/17:16
perrito666Man this referendum really broke my 1us 1eu chain :p17:18
mupBug #1596038 opened: typo in cloud definition yields "ERROR cloud <foo> does not require credentials" <landscape> <usability> <juju-core:New> <https://launchpad.net/bugs/1596038>17:29
mupBug #1596039 opened: juju add-cloud should tell you what it did <landscape> <usability> <juju-core:New> <https://launchpad.net/bugs/1596039>17:29
mupBug #1596045 opened: Juju says windows mongo: invalid version <blocker> <ci> <mongodb> <regression> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1596045>17:54
perrito666oi, I am partially back, now from the airport18:19
alexisbperrito666, either that was a very short flight to london or something happened wtih your travels18:23
perrito666alexisb: Cordoba airport18:23
perrito666I was here incredibly early18:23
perrito666exiting my lovely country requires bureaucracy so I decided that I wanted room just in case18:24
natefinchyeah, I prefer to be super early, too.  Much rather mess around in the airport for an hour or so, than be running and scared I'm going to miss my flight.18:25
perrito666ouch my screen is starting to show some serious screen burn I hope it will ive until laptop refresh18:25
katcoperrito666: what's the period on laptop refreshes?18:45
perrito666katco: 3ys I believe18:47
katcoperrito666: cool18:47
natefinchyeah, 3 years18:55
perrito666that would be march for me18:55
katcoericsnow: ping18:55
natefinchJuly for me :)18:56
ericsnowkatco: pong18:56
katcoericsnow: hey, trying to use the stub framework, but when i do .CheckCall i get this: "... mismatch at .FuncName: unequal; obtained "github.com/juju/juju/apiserver/observer/fakeobserver.(*Instance).Join"; expected "Join""18:56
katcoericsnow: is there a way to make it not care about the namespace + type? what am i doing wrong?18:57
ericsnowkatco: it uses whatever you passed to Stub.AddCall()18:58
katcoericsnow: ah ok. see if i can clean that up a bit, thanks18:58
ericsnowkatco: I'd recommend using a string literal in that call18:58
katcoericsnow: i'm using reflection; seems more robust18:59
alexisbanyone on the channel used dave's pprof command?19:09
mupBug #1596066 opened: Juju never achieves HA on maas 1.9 with trusty <blocker> <ci> <maas-provider> <regression> <trusty> <juju-core:Triaged> <https://launchpad.net/bugs/1596066>19:09
natefinchalexisb: https://medium.com/@tjholowaychuk/profiling-golang-851db2d9ae24#.vlmpkj2tg19:23
mupBug #1596066 opened: In some configurations, juju never achieves HA on maas 1.9 with trusty <blocker> <ci> <maas-provider> <regression> <trusty> <juju-core:Triaged> <https://launchpad.net/bugs/1596066>19:24
mupBug #1596071 opened: api documentation should be written from a user perspective <landscape> <juju-core:New> <https://launchpad.net/bugs/1596071>19:24
natefinchericsnow: it looks like we don't ever call Priority.decode?19:42
ericsnownatefinch: hmm, I thought there was a test that roundtripped it19:43
ericsnownatefinch: thanks for that review :)20:28
marcoceppirick_h_ alexisb what's the over under on the api changing agian for beta11?20:37
rick_h_marcoceppi: I'd not take any odds on that20:37
marcoceppirick_h_: I really /really/ want to update bundletester/amulet/deployer/python-jujuclient but if beta11 is likely to have more breakage I don't want to have developers constantly patching to keep up20:38
rick_h_marcoceppi: I'd really hoped to have things frozen but it's not worked out that way and I've learned my lesson that I just don't think I can say anything with a really firm commitment there.20:41
rick_h_marcoceppi: the big rename branch of Tim's missed some areas of the code that we're debating if it needs to be done or not now20:42
rick_h_marcoceppi: so I know if several chunks that aren't determined as "go" or "no go" yet that will be breaking api changes20:42
marcoceppirick_h_: cool, I wasn't looking for a promise, just if there were still murmors20:42
rick_h_marcoceppi: lots of murmors20:42
marcoceppirick_h_: cool, we'll evaluate the level of work, and keep an eye out for the next few days next week if there's any announcement of changes20:42
marcoceppi+1 to the mails that have gone out so far announcing changes before they land20:42
marcoceppiit's really helped us from getting soured when betas drop20:43
rick_h_marcoceppi: yes, we'll be continuing that for sure. It's a case of "the least we can do" :)20:43
marcoceppi\o/20:44
babbageclunkericsnow, natefinch: would either of you be able to review a bugfix for me? http://reviews.vapour.ws/r/5163/20:49
ericsnowbabbageclunk: sure20:50
babbageclunkericsnow: thanks!20:51
ericsnowbabbageclunk: LGTM20:52
babbageclunkericsnow: awesome.20:52
babbageclunkericsnow: See you on Monday!20:53
katcooh dear god. emacs truly can do anything: https://github.com/Fuco1/clippy.el20:54
babbageclunkkatco: ha ha, I should install that.21:04
katcobabbageclunk: it looks like you're trying to install clippy.el21:04
babbageclunkericsnow: mind reviewing the dependency change for that too? http://reviews.vapour.ws/r/5164/21:04
babbageclunkor katco, would you have a moment to review this? http://reviews.vapour.ws/r/5164/21:14
katcobabbageclunk: ship it21:14
babbageclunkkatco: sweet, thanks21:15
mupBug #1596104 opened: Can't kill a controller if the substrate is gone <v-pil> <juju-core:New> <https://launchpad.net/bugs/1596104>22:24
katcoanyone still around?22:24
redirkatco: :)22:25
katcoredir: yay! i have a small review for you22:25
redirk22:25
katcoredir: http://reviews.vapour.ws/r/5165/22:25
katcoredir: it doesn't require any juju knowledge, so if you're comfortable, i think you can +1 it. it's gating some of my other work22:25
katcoredir: and i am just now seeing i should have implemented Increment and Decrement in terms of Add(n)22:27
redirshould I look later?22:28
katcoredir: no, new version up22:28
mupBug #1596104 changed: Can't kill a controller if the substrate is gone <v-pil> <juju-core:Invalid> <https://launchpad.net/bugs/1596104>23:34

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!