=== dreverri_ is now known as dreverri | ||
dreverri | hazmat: ping | 01:18 |
---|---|---|
hazmat | dreverri, pong | 01:21 |
dreverri | not sure if you remember but I had an issue this morning with my bootstrap node not showing up in juju status | 01:21 |
hazmat | dreverri, i remember, did you get access to the instance? | 01:22 |
dreverri | yes | 01:22 |
dreverri | nothing in /var/log/juju/all-machines.log | 01:22 |
dreverri | no other log files in the juju log folder | 01:22 |
dreverri | syslog does have this: auth: couldn't find user admin, admin.system.users | 01:23 |
hazmat | dreverri, could you paste /var/log/cloudinit-?output? | 01:23 |
hazmat | dreverri, aha i think that's the issue | 01:23 |
hazmat | dreverri, so are the juju agents running , is mongodb running? | 01:24 |
dreverri | mongodb is running | 01:24 |
hazmat | dreverri, no juju process? | 01:25 |
dreverri | no juju process | 01:25 |
hazmat | dreverri, could try manually starting it via upstart (service juju-machine-agent start), there might be more output to syslog | 01:26 |
dreverri | juju-machine-agent: unrecognized service | 01:27 |
dreverri | there are failure messages in the cloud-init-log file | 01:28 |
dreverri | I'm going to parse through that | 01:29 |
dreverri | thank you | 01:29 |
dreverri | this seems interesting: http://pastebin.com/G7x8CsUF | 01:30 |
=== gary_poster is now known as gary_poster|away | ||
hazmat | filed bug 1207590 | 02:24 |
_mup_ | Bug #1207590: ec2 provider trouble bootstrapping <juju-core:New> <https://launchpad.net/bugs/1207590> | 02:24 |
=== defunctzombie is now known as defunctzombie_zz | ||
=== defunctzombie_zz is now known as defunctzombie | ||
dreverri | hazmat: thank you for filing a bug, any thoughts on what the issue might be? | 04:11 |
hazmat | dreverri, what version of juju ? | 04:12 |
dreverri | 1.11.2-unknown-amd64 | 04:12 |
dreverri | the bootstrap node is running: 1.11.4-precise-amd64 | 04:13 |
hazmat | dreverri, where did you get that client version from? distro backports/ppa/src? | 04:13 |
hazmat | er.. where did you install juju from | 04:13 |
dreverri | https://juju.ubuntu.com/docs/ | 04:14 |
dreverri | for OS X | 04:14 |
dreverri | https://github.com/juju/juju-core/releases | 04:14 |
hazmat | ic, thanks | 04:14 |
dreverri | is that no good? | 04:15 |
hazmat | dreverri, i dunno how much active testing on mac we have. that bug doesn't look famliar. there is a newer release of juju out 1.12 @ osx via homebrew mp https://github.com/rochacon/homebrew/blob/juju-core-source-proposal/Library/Formula/juju.rb and there's another release planned for this weekend. | 04:18 |
dreverri | I will try that out | 04:19 |
dreverri | thank you | 04:19 |
thumper | marcoceppi: ping | 04:27 |
=== CyberJacob|Away is now known as CyberJacob | ||
noodles775 | adam_g: Sorry - I don't have a public one, no. All I've got at the moment is a private MP that updates a charm from the saltstack support to the ansible support. I'll try to get something public together. | 06:25 |
=== CyberJacob is now known as CyberJacob|Away | ||
rogpeppe | mornin' all | 08:18 |
=== defunctzombie is now known as defunctzombie_zz | ||
=== axw__ is now known as axw | ||
=== axw is now known as axw- | ||
m0dex | hey all, quick question (well hopefully quick) i'm using juju 1.11.4-raring-amd64 and when trying to 'juju bootstrap --constraints "instance-type=t1.micro"' it fails with "unknown constraint "instance-type"' any ideas? | 14:00 |
marcoceppi | m0dex: instance-type isn't supported yet with 1.X you'll instead need to supply something like cpu-cores=1 | 14:02 |
marcoceppi | use generic constraints cpu-cores and mem to map to the instance type | 14:03 |
m0dex | alright cool. ill give that a shot right now.. | 14:03 |
m0dex | hey any idea what version of juju i'd need to be using to get that instance-type support? | 14:14 |
AskUbuntu | deploying openstack on 6 servers | http://askubuntu.com/q/327865 | 14:21 |
marcoceppi | m0dex: juju 0.7, but it's honestly not worth going all the way back to that old, soon to be if not already unsupported version of juju | 14:26 |
m0dex | alright. any idea if theres a possibility that juju doesn't want me to create such a small instance box? when i specify the EXACT specs for a t1 micro instance, its like nah dude, imma make an m1.small | 14:33 |
jcastro | m0dex: yeah that happened to me yesterday | 14:33 |
jcastro | I put in the exact numbers for a large and it makes a xlarge | 14:33 |
jcastro | I'm going to ask the core guys for instance-type support | 14:33 |
marcoceppi | m0dex: that shouldn't be the case. Try under-selling the numbers a bit. Instead of 725MB or whatever it is, do 500MB | 14:33 |
marcoceppi | it'll match that number or the next biggest unit | 14:34 |
m0dex | ok. let me give that a shot | 14:34 |
marcoceppi | jcastro: yeah, having instance-type would be awesome. | 14:34 |
m0dex | FUCK | 14:35 |
m0dex | or to elaborate, no luck. i'll undersell them even more..... | 14:35 |
jcastro | maybe it's time to file a bug | 14:37 |
jcastro | ah good news, instance-type is coming back | 14:38 |
jcastro | "2-3 weeks" is the roadmap | 14:39 |
m0dex | lol yeah i guess "juju bootstrap --constraints "mem=1M" | 14:39 |
jcastro | heh | 14:39 |
m0dex | and it still makes an m1.small. silly juju | 14:39 |
jcastro | micros do suck though, you'll have a tough time on those things | 14:39 |
m0dex | i know, but i don't think i can justify the cost of anything else yet. | 14:39 |
jcastro | marcoceppi: how did we ever get anything done without --to? | 15:34 |
jcastro | I could do this all day | 15:34 |
rick_h | jcastro: lol | 15:54 |
sidnei | hehe | 16:00 |
jcastro | rick_h: any word on those links? | 16:01 |
jcastro | the store is pretty much unusable as a browser right now. :-/ | 16:01 |
rick_h | jcastro: deploy is going out now | 16:01 |
rick_h | jcastro: IS is working on it as we speak I hope | 16:02 |
jcastro | ok | 16:02 |
jcastro | rick_h: wanna check on something for me? | 16:03 |
jcastro | rick_h: so I modified the readme in the python-django file | 16:03 |
rick_h | jcastro: k | 16:03 |
jcastro | and converted it from RST to Markdown ---> README.md | 16:03 |
jcastro | http://manage.jujucharms.com/charms/precise/python-django | 16:03 |
jcastro | warn: no README file | 16:03 |
jcastro | no idea if it's working on the actual GUI since I can't get to the README link currently | 16:04 |
rick_h | https://bazaar.launchpad.net/~charmers/charms/precise/python-django/trunk/files no readme file :/ | 16:04 |
jcastro | aha! | 16:04 |
jcastro | I'm just an idiot. Pushing ... | 16:04 |
jcastro | https://bazaar.launchpad.net/~charmers/charms/precise/python-django/trunk/view/head:/README.md | 16:05 |
jcastro | there she is! | 16:05 |
rick_h | jcastro: cool, so yea hold tight while it re-injests then | 16:05 |
marcoceppi | jcastro: stirring memories of jitsu, I feel faint and may need to lie down | 16:06 |
jcastro | rick_h: it's like what, 15 minutes? | 16:07 |
rick_h | jcastro: there abouts | 16:07 |
jcastro | Good Enough | 16:07 |
rick_h | jcastro: https://jujucharms.com/sidebar/precise/ceph-14/#bws-readme deploy done | 16:10 |
jcastro | \o/ | 16:10 |
jcastro | yeah! | 16:10 |
arosales | rick_h, is it expected for the gui to have the juju-gui md displayed with the comment at the top | 16:17 |
arosales | https://jujucharms.com/sidebar/precise/juju-gui-73/#bws-readme | 16:17 |
jcastro | I think you need a triple dash in markdown | 16:19 |
jcastro | <!--- | 16:19 |
jcastro | blah | 16:19 |
jcastro | --> | 16:19 |
jcastro | http://stackoverflow.com/questions/4823468/store-comments-in-markdown-syntax | 16:19 |
arosales | I can propose a branch, just wasn't sure how the charm browser was parsing | 16:20 |
rick_h | arosales: jcastro so we hit this with the juju-gui readme and when I checked the module we're using to parse I couldn't find it supporting that syntax directive | 16:21 |
rick_h | we've got a card/bug to try to update the module we're using | 16:21 |
jcastro | or just put the license in the code files and not in the readme like everything else? | 16:22 |
arosales | rick_h, ok thanks. | 16:22 |
rick_h | arosales: yea, checking in http://markdowntohtml.com/ two or three - doesn't matter. They don't parse it either :/ | 16:23 |
arosales | rick_h, interesting that the soruce tab --> Read.md for the gui doesn't show the comment | 16:24 |
rick_h | actually nvm | 16:24 |
rick_h | arosales: that's because in the source tab it's an html comment so it's not shown | 16:24 |
rick_h | arosales: but in the markdown parser, it's escaped? /me is checking | 16:25 |
arosales | jcastro, knows I am not a markdown expert by any means | 16:25 |
rick_h | arosales: right, the parser is turning <!-- into <!-- | 16:25 |
jcastro | rick_h: simple solution is for now to put the license at the bottom of the readme, heh | 16:26 |
rick_h | jcastro: true | 16:26 |
rick_h | jcastro: why it's not gotten fixed. Things like routing bugs > markdown parsing | 16:26 |
rick_h | jcastro: but we do know about it and would like to fix it | 16:26 |
arosales | rick_h, is the source intended to parse _all_ items in the charm directory? | 16:28 |
arosales | to jcastro's ealier point I don't see the "copyright" or "COPYING" file in the source tab listing for the juju-gui | 16:29 |
rick_h | arosales: yes, in general people like that. we've got bugs to do some basic type detection though to 'show' images and display file content as we do now | 16:30 |
rick_h | arosales: so it's only files that the back end loads. So we load everything in hooks, and other files we know charms need/etc | 16:30 |
arosales | rick_h, +1 on showing all the files they the name change to source from hooks, just wondering why those files weren't showing up | 16:31 |
rick_h | arosales: the gui can only display info on files the back end knows about. We don't load all files in a charm because there could be source packages/etc in there if the ywanted to | 16:31 |
arosales | rick_h, ah | 16:31 |
rick_h | if there are files we want to whitelist we can submit a bug on charmworld and add support for them and then they'll show up | 16:31 |
arosales | rick_h, would you like me to file a bug to whitelist copying and copyright files? | 16:31 |
rick_h | arosales: sure | 16:32 |
arosales | rick_h, thanks, bug 1207833 submitted | 16:35 |
rick_h | arosales: thanks | 16:35 |
arosales | bug 1207833 | 16:35 |
_mup_ | Bug #1207833: Juju GUI Charm Browser: Source Tab: Whitelist files such as copying, copyright, license <charmworld:New> <https://launchpad.net/bugs/1207833> | 16:35 |
marcoceppi | jcastro: we've got a few (additional) inconsistencies on the juju website | 17:26 |
jcastro | WHERE | 17:26 |
jcastro | I will fix them | 17:26 |
marcoceppi | `juju generate-config` provides links to pages like https://juju.ubuntu.com/get-started/openstack/ (amazon, openstack, hp-cloud) which I think should be removed and redirected instead to the docs page. However the openstack doc page is also pretty badly out of date (and a little light) https://juju.ubuntu.com/docs/config-openstack.html | 17:27 |
marcoceppi | So Ask Ubuntu needs to be updated and the docs example config needs to reflect actual environments.yaml layouts | 17:28 |
jcastro | I thought we were going to stacktack instead? | 17:28 |
marcoceppi | Because, lets face it, default-instance-type hasn't been a valid option for quite a while | 17:28 |
marcoceppi | jcastro: I wasn't sure what should point where | 17:28 |
marcoceppi | but everything we have for openstack is out of date | 17:28 |
marcoceppi | So, I figured you'd know best what to sort and where | 17:29 |
marcoceppi | openstack_s3 was a hack added to the python juju | 17:29 |
marcoceppi | "hack" | 17:29 |
jcastro | ok so step one would be to update the docs? | 17:30 |
marcoceppi | docs and ask ubuntu, are probably sufficient for now | 17:30 |
jcastro | man, why yuo bring this up when nick is already off for the weekend. :) | 17:30 |
marcoceppi | haha ;D | 17:30 |
jcastro | ok so I don't know what is correct wrt. an openstack config | 17:30 |
marcoceppi | jcastro: http://paste.ubuntu.com/5940862/ | 17:31 |
=== defunctzombie_zz is now known as defunctzombie | ||
jcastro | fixed in both places | 17:34 |
marcoceppi | jcastro: ta! | 17:37 |
jcastro | I'll talk to nick about what the plan for AU/docs integration or non-integration will be | 17:37 |
jcastro | in the meantime updating in both places | 17:38 |
marcoceppi | jcastro: we've basically postponed stacktack integration, but if it's a blocker for content we can elevate it again | 17:38 |
jcastro | yeah the thing is is we can't remove the AU question, all the juju config questions have nice google juice | 17:38 |
marcoceppi | stacktack integration is technically "done", but we've had problems getting it to work consistently | 17:38 |
jcastro | They link to the docs though, so we should be fine | 17:39 |
=== CyberJacob|Away is now known as CyberJacob | ||
AskUbuntu | What's the difference between the python-django and django-rainbow charms? | http://askubuntu.com/q/327938 | 18:11 |
jcastro | marcoceppi: did we put deployer in a new place yet? I built a stack locally and I'm ready to test it on HP | 18:28 |
marcoceppi | jcastro: not that I'm aware of | 18:28 |
jcastro | arosales: I have a ~charmers proposal I'd like to run by you | 19:21 |
arosales | jcastro, for sure do you want to cover that today? | 19:21 |
jcastro | so I was thinking | 19:22 |
jcastro | if we should do autolanding charm updates with tarmac | 19:22 |
jcastro | so like let's say I do a merge proposal | 19:22 |
jcastro | and then marco +1's it | 19:22 |
jcastro | and mims +1's it | 19:22 |
jcastro | tarmac auto lands it without us having to manually merge | 19:22 |
jcastro | this would make it trivial for us to like review doc changes | 19:22 |
jcastro | you could just read the diff, and then approve/disapprove | 19:23 |
jcastro | without having to manually merge the submission | 19:23 |
jcastro | it's one less step for the reviewers | 19:23 |
AskUbuntu | Sharing SSH keys during a Juju Charm deployment | http://askubuntu.com/q/327967 | 19:25 |
marcoceppi | jcastro: that's the hope with the charmtester stuff | 19:28 |
marcoceppi | is to gate on tests actually passing before a merge is even looked at | 19:28 |
arosales | marcoceppi, Charm tester also came to mind when jcastro mentioned tarmac | 19:29 |
arosales | jcastro, I don't think there are any issues with using something like tarmac. I think the work would be to integrate that into charm tester | 19:29 |
marcoceppi | tarmac is fine, that's just a mechanism and test runner for landing changes. I think we definitely want to use the idea of gating merges on tests though, for sure | 19:30 |
* marcoceppi cross fingers for this cycle | 19:30 | |
jcastro | arosales: yeah so I totally didn't think about tests | 19:31 |
arosales | I think once we have a good test story the next step, as you mention, is automating more of merge requests | 19:32 |
arosales | but the core philosophy should be centered around quality | 19:32 |
arosales | but yes a generally +1 to CI | 19:32 |
=== CyberJacob is now known as CyberJacob|Away |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!