marcoceppiagunturu: you can juju action status <UUID> or juju action fetch <UUID> to get more details00:25
metsukedoes anyone use juju with a cinder backend and find it to be robust?00:31
=== Guest75342 is now known as CyberJacob
=== urulama__ is now known as urulama
Razvahey! is there any know bug regarding network interfaces not named eth?09:39
axinohi charmers, can I get your attention on https://code.launchpad.net/~chris-gondolin/charms/trusty/nrpe/fix-sub-postfix/+merge/285692 ? cc gnuoy` thedac10:21
=== med_ is now known as Guest5506
jamespag`gnuoy, working through beisner's 16.02 testing sweepup merges11:31
gnuoyjamespag`, ah, ok. I said I'd do that if you want to leave that to me11:32
jamespag`gnuoy, i have cycles inbetween things11:32
gnuoykk, thank you11:32
Razvahey! is there any know bug regarding network interfaces not named eth?11:43
magicaltroutthats like the most cryptic question ever11:43
jamespag`beisner, gnuoy: hey can we switch amulet test cases to use /usr/bin/env python instead of /usr/bin/python?11:44
jamespag`I'd like to be able to execute them in a venv (think under tox) but right now that's not possible11:44
gnuoyjamespag`, I have no objection11:44
Razvamagicaltrout I'm using Ubuntu 14 and trying to install OS using Juju. It seems that as long as I'm using ethX...the install will work. If I use the "native" naming (enoX, emX etc) it will basically...fail.11:45
RazvaOS = OpenStack11:45
magicaltroutjamespag` and co are your people for that ;)11:45
jamespag`Razva, can we see a log?11:46
jamespag`will help with triage as to where you problem lies...11:46
=== jamespag` is now known as jamespage
jamespageI was developing a tick11:48
jamespagebeisner, your landme's are landed11:51
Razvadarn netsplits.11:51
marcoceppimagicaltrout: I shot over a pull req https://github.com/OSBI/juju/pull/111:52
magicaltroutits like the 90's all over again11:52
marcoceppiI really want m4 instance types in Juju11:52
magicaltroutRazva: if you didn't see james asked for a pastebin log11:52
magicaltroutoh thats mine marcoceppi thanks for that, I just ignored it as I assumed it was against the pull i submitted :)11:53
marcoceppimagicaltrout: our builder will fail pull req that haven't been formated according to golangs `go fmt`11:53
magicaltroutsorry about that11:54
marcoceppiwallyworld: this has been updated https://github.com/juju/juju/pull/4426 any chance we can get a JFDI?11:54
wallyworldmarcoceppi: np, doing it now11:55
wallyworldwill be in beta111:55
marcoceppiwallyworld: \o/11:55
jamespagegnuoy, which bits of https://pastebin.canonical.com/148351/ still need review?12:24
jamespagebeisner, yah - https://code.launchpad.net/~james-page/charms/trusty/neutron-gateway/xenial/+merge/28586212:26
jamespagehas some more bits there as well12:26
beisnerjamespage, ah good deal12:26
=== med_ is now known as Guest37570
wallyworldmarcoceppi: merged \o/12:51
marcoceppiwoop woop!12:51
marcoceppithanks for the fix magicaltrout12:51
magicaltrouthopefully I'll find some time to actually learn some go and do something more exciting than copy and paste12:52
magicaltroutright, I have a quick one about publishing and stuff that I don't quite get12:53
magicaltroutso I have a charm in github12:53
magicaltroutand supposedly I can publish that to the charm store12:53
magicaltroutso in my local repo I can run: juju publish --from=. cs:~f-tom-n/trusty/gitlab cs:~f-tom-n/trusty/gitlab12:55
magicaltroutor something like that?12:55
magicaltrouteven though I've not created anything in launchpad or bzr?12:55
=== perrito667 is now known as perrito666
marcoceppimagicaltrout: theoretically, yes, it'll look something like this:12:58
magicaltroutnice caveating12:58
marcoceppicharm push .12:58
marcoceppisince the charm command knows who you are, and the name of the charm it'll just upload to cs:~f-tom-n/trusty/gitlab12:59
marcoceppithe new charm 2.0 stuff is private beta atm, but I can see about adding you if you're interested12:59
magicaltrouthmm someone lied then and told me I could do it! :P13:00
marcoceppiyou could do it, if you have early access to the tool - I imagine we're pretty close to public beta though esp since we're submitting them to Xenial13:00
magicaltrouthello there people who know more than I do14:28
magicaltroutsubmitting a layered charm14:29
magicaltrouthow on earth do you do it?14:29
magicaltroutor do you submit the built thing?14:29
rick_h_magicaltrout: you submit the built thing14:29
rick_h_magicaltrout: and submit any layers you've added to http://interfaces.juju.solutions/14:29
rick_h_if they're useful to others out there14:29
=== Ursinha_ is now known as Ursinha
lazyPoweror  include in your finally assembled charms readme a breadcrumb trail back to your upstream layers source control system for bug reporting purposes, and the ~charmers will appreciate your attention to detail.14:39
magicaltroutah yeah good point14:40
iceyany support yet for controlling a remote lxd controller / juju env from an OSX host? The lxd machine would be running Xenial and be accesible over a local network14:50
beisnericey, that would just rock.  better yet, control multiple remote lxd hosts and voila, a rockin local cloud.14:51
iceybeisner: yes, that is why I'm excited about lxd!14:52
iceyI'm trying to set it up from the release notes now :)14:52
rick_h_icey: no, we'll look at adding more awesome lxd support, but it'll happen in the next cycle14:52
iceyawww rick_h_ :(14:52
iceylooking at the docs I'm not sure why I couldn't use the remote lxd host?14:54
rick_h_icey: it might be something you can hack up. I just think the juju provider assumes it's a local lxd14:55
iceyrick_h_: it looks like there's config for remote-url so I'm hopeful :)14:55
rick_h_icey: let me know how it goes. I've been wrong before, but thought that didn't make the current cut for feature freeze14:55
iceyif I get it working, I'm blogging about it because it'll be like LOCAL PROVIDER ON A MAC!14:56
iceylazyPower: ^^14:56
rick_h_even better :)14:56
lazyPowericey: you said the magic word!14:56
lazyPowerman its 2016 and i have cut exactly 0 blog posts :/ i feel like i'm slackin14:57
rick_h_lazyPower: I'm comin for you!14:57
rick_h_lazyPower: for blog posts :P14:57
rick_h_lazyPower: need to get you doing one on resources maybe. Surely you've got some charm that can use it :P14:57
lazyPowertempt me with a good time whydontchya :D14:57
lazyPowerI have a few ideas14:58
lazyPowerand a weekend to make it happen14:58
rick_h_lazyPower: good stuff, just giving you a hard time :)14:59
iceyso far so good, rick_h_:15:01
icey./juju bootstrap -m lxd15:01
iceyBootstrapping model "lxd"15:01
iceyStarting new instance for initial controller15:01
iceyLaunching instance15:01
iceyfrom my mac15:01
rick_h_icey: interesting15:01
icey$ lxc list15:02
icey|                        NAME                         |  STATE  |       IPV4       | IPV6 | EPHEMERAL | SNAPSHOTS |15:02
icey| juju-12f338d5-04f6-4628-88cb-0d18179c8d74-machine-0 | RUNNING | (eth0) |      | NO        |         0 |15:02
iceyon the remote host :)15:02
suchvenuHi Kevin15:02
jaceknkjackal: cory_fu: hey. Thank for the review, I updated my collectd layer so if any of you has time I would love re-review: https://bugs.launchpad.net/charms/+bug/153857315:03
mupBug #1538573: New collectd subordinate charm <Juju Charms Collection:Triaged> <https://launchpad.net/bugs/1538573>15:03
josequestion, let's say I have a config value that I want to be different for each instance, is there a set it with juju set?15:06
* icey sadface15:06
iceylazyPower rick_h_: 2016-02-17 15:05:19 ERROR cmd supercommand.go:448 no registered provider for "lxd"15:06
iceyERROR failed to bootstrap model: subprocess encountered error code 115:06
iceythat's after we downloaded the tools15:07
iceythat's also in the cloud-init-output.log on the lxc container started to be the controller15:08
iceybut I can bootstrap this lxd environment from on that host15:09
kjackalThank you for your effort jacekn , I just changed the status of the bug ticket to new so that it comes back up on the review queue (I think it is needed)15:11
kjackalI haven't seen anything like this jose15:11
iceyjose: the closest you will probably get is to deploy the charm multiple times to give them different configuration values as all units share the same configuration15:11
joseright, right. I'll experiment a bit with actions and otherwise juju run. thanks!15:12
=== ericsnow_ is now known as ericsnow
marcoceppiicey: use --upload-tools15:23
iceymarcoceppi: ERROR failed to bootstrap model: no matching tools available15:23
marcoceppiicey: that was /with/ upload-tools?15:23
iceythat may be the part that rick_h_ says is missing right now15:23
iceyyes marcoceppi, it gets a LOT farther without upload tools15:23
iceygoes through to ERROR cmd supercommand.go:448 no registered provider for "lxd"15:23
iceymarcoceppi: machine I'm trying to bootstrap from is a Mac15:24
=== Ursinha_ is now known as Ursinha
iceyif that helps15:24
marcoceppiicey: you've got alpha2?15:25
jamespagewolsen, re https://bugs.launchpad.net/charms/+source/cinder/+bug/146830615:26
mupBug #1468306: Missing os_region_name for mult-region cinder with nova <cpec> <hitlist> <openstack> <sts> <cinder (Juju Charms Collection):Fix Committed by billy-olsen> <nova-cloud-controller (Juju Charms Collection):In Progress by billy-olsen> <nova-compute (Juju Charms Collection):In Progress by15:26
mupbilly-olsen> <https://launchpad.net/bugs/1468306>15:26
jamespagecinder landed - nova-cc and nova-compute need a little change - but looking good15:26
marcoceppiicey: interesting15:26
jamespagemainly cause I jumped you with mitaka templates...15:26
rick_h_icey: hmm, yea I think it's something that we *should* be able to do without much more work, but with freeze and such it'll be something next cycle unfortunately15:27
rick_h_icey: 2.0 has folks heads down atm15:28
iceyyeah rick_h_; ideally bootstrap lxd from mac will be there for 2.0 ;-)15:28
thedacgnuoy: Can I ask you to merge this please. Already approved https://code.launchpad.net/~ajkavanagh/charm-helpers/add-service-checks-lp1524388/+merge/28560415:49
gnuoythedac, sure15:51
=== kwmonroe_ is now known as kwmonroe
magicaltroutwhoop its only taken 2 days to write a very basic charm... its like belgium take 2 \o/16:08
lazyPowermagicaltrout - it gets easier every time you do it :)16:12
lazyPower<3 me some layers and no more boilerplate16:13
lazyPowerand broken english apparently. but i digress, its quite pleasant :D16:13
magicaltroutindeed, the chaming bit wasn't what took time, its the parsing of the configs and stuff for configuration changes that took some figuring out16:13
lazyPowermagicaltrout - we riffed on the bus right?16:13
magicaltroutwe did indeed16:14
lazyPowernice :)16:14
* lazyPower adds you to the list "my people"16:14
magicaltrouter right16:20
magicaltroutnext random one16:20
magicaltroutI running trunk.... as you do16:20
magicaltrouti tried running bundletester but it seems to be using the non trunk install I have as its attempting to run bootstrap16:21
magicaltrouteven though my env is bootstrapped16:21
magicaltroutsound plausible? and can i get it to use my trunk stuff16:21
magicaltroutI have my PATH set to use the trunk binary16:21
marcoceppimagicaltrout: IIRC, bundletester resets your path for you ;)16:22
marcoceppilet me take a look16:22
wolsenjamespage, thanks for the review, I'll take a look at the comments on nova-cc and nova-compute16:26
gnuoythedac, landed16:29
magicaltroutother question16:38
magicaltroutactually, scrap it16:38
magicaltrouti'm an adult I can work it out ;)16:38
=== skay_ is now known as skay
=== skay is now known as Guest63313
=== Guest63313 is now known as skay
marcoceppimagicaltrout: feel free to ask anyways :)16:51
magicaltroutnope I'm cool, apart from the charm updating appearing sluggish ;(16:59
magicaltrouthttps://jujucharms.com/u/f-tom-n/gitlab/trusty/ up to date gitlab for the masses, I shall write a few more tests and get it into the store17:02
marcoceppimagicaltrout: nice!17:02
magicaltroutyeah, I'm just migrating a client17:08
magicaltroutso the stuff that isn't yet charmed I need to charm17:08
magicaltroutso I'll do my best to get it to a decent standard for general consumption17:08
magicaltroutI need to finish off the PDI charm me and kwmonroe were working on tomorrow17:09
magicaltroutthats the missing piece of the puzzle for this migration, but all that is missing is some random actions, the majority of the charm now works17:09
kwmonroeand what a fine charm that was magicaltrout!  the memories bring much joy ;)17:09
kwmonroemy favorite part was when cory_fu looked at our python in disgust.17:10
kwmonroeexit 0 for life!17:10
magicaltrout"pfft you could have done that shit in bash"17:10
magicaltroutI stuck with it though, I did gitlab in bash as well for no real reason other than, I could ;)17:10
kwmonroeheh, awesome.17:11
josebdx: ping17:16
lazyPowerinteresting, has anyone seen a charm actively dismantle security features such as  enabling password auth and root login via ssh?17:20
cory_fulazyPower: No.  Why would a charm want to do that?17:32
lazyPowercory_fu - good question, and its been noted in the review17:33
iceymarcoceppi: I managed to get it working by bootstrapping from the lxd machine and copying over Juju's files to my mac18:04
iceynow I can work with lxd from my Mac, I just can't bootstrap18:04
iceylooks like it may be having some other issue now though, juju shows a new machine as pending for a while...it seems to be trying to get tools from the host machine rather than the bootstrap node -_-18:09
marcoceppiicey: this is because the bootstrap state-server is probably configured as to the wrong address18:23
iceyI bootstrapped from the served lxd is running on, then tried depoloying from a remote machine; deploy worked, container came up, but then it looks like yeah, the state-server was configured wrong; the bootstrap container had an ip on both the lxcbr and the normal network18:24
iceylazyPower: your consul charm seems to be broken now19:05
lazyPowericey - maintainer is in metadata and the readme :D19:07
lazyPowerdas not meeeeee19:07
iceyThe Consul charm maintainers are: - Charles Butler ( @chuckbutler ) ...19:09
* lazyPower touches his nose19:10
lazyPowerwhats goin on with it icey ?19:10
iceyhashicorp pulled everything off of bintray19:10
lazyPowerthats a config.yaml update19:10
iceyDL links have to change to releases.hashicorp.com19:10
iceynot a version thing...19:10
lazyPowercan you update config.yaml, give a quick test and PR that back upstream?19:11
beisnerjamespage, all clear on:  https://code.launchpad.net/~james-page/charms/trusty/neutron-gateway/xenial/+merge/28586219:11
iceythere are literally no longer any versions on bintray19:11
lazyPowerthats fine19:11
beisnerthedac, ^ as a prev. reviewer on that, can you have another peek?19:11
lazyPoweryou should be able to change the source_url19:11
thedacbeisner: sure19:11
lazyPowerto whatever, just grab the sha1sum of the bin and stuff that in config19:11
iceyno source_url...19:11
beisnerthedac, tyvm19:11
iceylazyPower: i'm going to have to muck around in the hooks/consul.py to fix it19:12
lazyPowericey rewrite the charm, give me a PR19:12
* icey seems to be doing that a lot lately19:12
lazyPowerwe like your work, what can I say?19:12
lazyPowericey matt and i can update that with a source_url/hash config option19:13
lazyPowerdid they drop deb delivery all together?19:13
iceyit's now thgouth https://releases.hashicorp.com/consul/0.5.2/consul_0.5.2_linux_amd64.zip19:14
thedacbeisner: jamespage: merged \o/19:14
iceywith shasums at https://releases.hashicorp.com/consul/0.5.2/consul_0.5.2_SHA256SUMS19:14
beisnerthedac, right on!19:14
iceylazyPower: almost ready to test this change and then I'll send a PR :)19:14
lazyPowerwoo \o/19:14
lazyPowericey - install_remote() will handle that natively19:15
iceybasically a lot of {0}_{1} => {0}/{0}_{1}19:15
iceyhandle downloading the zip, download a separete shasum, confirm and drop in place?19:15
iceylazyPower: ^19:15
lazyPowershould, the sum i think is expected to be handed to install_remote19:16
lazyPowervs fetching the shasum file19:16
lazyPowerbut it may just work, i haven't tried handing it a url19:16
iceylazyPower: this way we can get the new shasum by just updating the version in config though :)19:16
iceyit shouldn't break again given that it's /their/ own CDN to handle downloads19:16
lazyPowerwe can hope19:16
marcoceppihey magicaltrout, where do you keep your gitlab layer? I have a few things I'd like to contribute19:29
magicaltroutfunny, I don't recall submitting my charm to the review queue ;)19:49
marcoceppimagicaltrout: I'm really keen on gitlab, I tried to write a charm for it a few years ago, I run it myself manually20:13
magicaltroutmarcoceppi: the scale out stuff with omnibus isn't great, plus i'm not a fan of the complete package stuff either20:15
magicaltroutbut https://gitlab.com/gitlab-org/gitlab-ce/blob/master/doc/install/installation.md20:15
magicaltroutI don't mind doing that20:15
magicaltroutbut that involves manually compiling Ruby20:15
marcoceppimagicaltrout: well upstream doesn't always know what's best ;)20:15
marcoceppithe omnibus is fine, we can just modify the configuration afterwards20:16
magicaltroutokay, so install via omnibus, and an exploded shell script20:18
magicaltroutand then if someone connects an alternative postgres then the hook updates the config and shunts the DB over to the separate DB?20:18
magicaltroutyesterday i was thinking about doing similar with Nginx even though it doesn't support proper clustering and scale out yet20:19
marcoceppimagicaltrout: thta's what I'm thinking, do an export and import of the database when related20:19
marcoceppion relation removal, remote dump the db locally then allow the removal20:19
marcoceppibut that's a bit more involved20:20
marcoceppijust starting out I like the idea of going single -> scaleout20:20
* marcoceppi will submit a few pull req20:20
magicaltroutokay marcoceppi, if you fancy tackling the DB as and when, I'll fiddle with optionally decoupling Ngnix and the crappy script20:22
marcoceppimagicaltrout: sure, sounds fun!20:23
magicaltroutif hacking DB connection hooks is your idea of fun :P20:23
balloonsso marcoceppi, I get to bug you about GSOC again. Have you gotten your ideas on paper yet? I can fill them in the wiki even if needed. We need to have them done by Friday to submit the app20:24
marcoceppiballoons: yeah, I had some questions20:24
balloonsI know jose is ok with doing some charm tasks.. But we could just assign the rest to him too ;-)20:24
marcoceppiballoons: is this targeted like Google Code In?20:24
balloonssure, ask away20:24
joseme wat?20:24
balloonsmarcoceppi, no it's not the same. This is 1 on 1 mentoring, and over the summer. so May - August20:25
marcoceppiballoons: then I have a few ideas of charms we'd want to write20:25
joseand for university students who are 18 or older, and currently enrolled in a university program20:25
marcoceppithat I can help get on paper20:25
marcoceppigive me like, 40 mins, otp20:25
balloonsright, so presumably they have more experience than the hs students20:25
balloonsmarcoceppi, brillant. jose, did you have charm ideas, or are you simply open to mentoring whatever marcoceppi comes up with?20:26
joseballoons: I can probably be the backup for Marco's ideas20:26
balloonstag-teaming them is a good idea20:26
lazyPowercharmschool all the things20:28
=== urulama is now known as urulama__
joselazyPower: you will mentor as well? fantastic!20:34
iceyis it possible to make a charm that will optionally be a subordinate?21:05
rick_h_icey: not currently. It's declared in metadata to be a subordinate or not21:05
iceyrick_h_: that's what I thought; thanks!21:05
iceyis it possible to relate two charms through the peer relation?21:07
marcoceppiicey: no, but you can use the same interface21:10
marcoceppiin the provides/requires section21:10
iceymarcoceppi: so if I use the same interface, they will relate?21:14
marcoceppipeers is only for inter-unit relations21:14
* icey is sad21:14
iceyI'll have to use the same hook for both a peer and vspecific relation then21:14
iceyalright, thanks marcoceppi!21:15
ChrisHolcombecmars, do you know if the az zone gets set for ec2 instances ?21:42
balloonsmarcoceppi, how's it coming along?21:57
marcoceppiballoons: any chance I can get back to you tomorrow?22:02
marcoceppiI have to finish these FFe and I'm OTP atm22:02
balloonsmarcoceppi, yep. Tomorrow is fine. Since they are in your head, it should just be as simple as finding some time to get them written down.22:03
=== natefinch is now known as natefinch-afk
apuimedoGood God gracious, the lxd deployer in juju 2.0 is so much faster than the juju-deployer in returning control22:35
apuimedogood job guys22:35
apuimedothe juju status is also much better22:35
marcoceppiapuimedo: glad you like! it's been awesome to see what the juju-core people have been doing to make juju much nicer to use22:42
apuimedomarcoceppi: the only thing I don't know is why the "overrides" section in my bundle doesn't work. Has it changed in v4?22:43
apuimedomarcoceppi: agreed, the UX is miles better22:43
marcoceppiapuimedo: overrides aren't officially part of the bundle spec, sadly22:45
marcoceppiapuimedo: so they won't work in native juju deploy22:45
marcoceppiapuimedo: we're working on defining how overrides should/will work in the next version of juju22:45
apuimedothat's a bit inconvenient22:45
apuimedolooking forward to that22:46
marcoceppiapuimedo: it's not osmething we've forgotten, we just want to make sure the experience is good22:46
blahdeblahAnyone know why juju mangles /root/.ssh/authorized_keys, and what I can do to fit in with that, whilst still enabling my charm to write a working entry in there?22:46
marcoceppiapuimedo: Mark made some mention of "flavors" of a bundle onthe mailing list recently https://lists.ubuntu.com/archives/juju/2016-February/006545.html22:46
marcoceppiblahdeblah: I wasn't aware juju mangles that - what provider are you using?22:47
blahdeblahmarcoceppi: openstack (in Canonistack)22:47
* apuimedo is ashamed. I've not followed the list recently22:47
* blahdeblah grabs sample22:47
marcoceppiapuimedo: I admit I even have a hard time keeping up with all the mailing lists I watch22:48
marcoceppiblahdeblah: interesting. I suppose it's so the operator can juju ssh <machine-#> which would be a root login22:48
apuimedoyeah... With the openstack ones I have more than enough22:48
blahdeblahmarcoceppi: Except that the main point of it is to break that: http://pastebin.ubuntu.com/15103537/22:50
marcoceppiblahdeblah: wat.22:50
blahdeblahmarcoceppi: read the fine pastebin :-)22:51
marcoceppiblahdeblah: I did22:51
marcoceppithat was my reaction22:51
marcoceppiblahdeblah: I don't see that in my 2.0-alpha2 deployed service22:51
blahdeblahI asked in #juju-dev (twice) why they do that, and I got crickets.22:51
marcoceppiblahdeblah: is that a problem with the image you're using?22:51
blahdeblahmarcoceppi: this is stable22:51
marcoceppipfft, stable22:52
marcoceppiwho uses that! ;)22:52
blahdeblahand the keys are clearly marked as being put there by juju - look at the last field22:52
marcoceppiblahdeblah: sure, but the first parts of the file22:52
blahdeblahmarcoceppi: don't make me come over there! :-)22:52
marcoceppiI'm not sure that's juju22:52
marcoceppilet me boot up my 1.25.3 vm22:52
beisnerthedac, up for landing 2 test updates?  takes n-c-c and dashboard to the maximum passable test level, and marks the end this pass of test updates from me.22:54
thedacsure, I'll take a look22:55
beisnerthedac, appreciate it22:56
thedacbeisner: just want to confirm that lines 215 - 220 are correct on the nova-cc. >= trusty_mitaka use legacy_ratelimit? Seems like that should be the "future" not legacy23:04

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