=== CyberJacob is now known as CyberJacob|Away | ||
=== kadams54 is now known as kadams54-away | ||
=== kadams54 is now known as kadams54-away | ||
=== kadams54-away is now known as kadams54 | ||
hatch | Has anyone asked for a config field in which there are a select few options which users could choose from? Which could be represented in the GUI as a select/dropdown? | 01:50 |
---|---|---|
rick_h_ | hatch: yea, enums basically | 01:52 |
rick_h_ | hatch: it's part of the discussion around using jsonschema to represent config options | 01:52 |
=== kadams54 is now known as kadams54-away | ||
=== kadams54-away is now known as kadams54 | ||
hatch | ok - atm I am just going to give them the possible options in the description but that just feels ripe for issues | 01:52 |
hatch | typos and whatnot | 01:52 |
rick_h_ | hatch: yea | 01:55 |
hatch | it would be even better if those enum options could enable/disable subsequent config options :) | 01:55 |
hatch | when running debug-hooks on a unit the tmux session that opens is in the home directory of the unit not the charm as shown in the docs - where can I find the actual hooks? | 02:33 |
hatch | $CHARM_DIR is blank | 02:34 |
hatch | ahah I need to actually trigger the hook externally | 02:37 |
hatch | looks like we need to update these docs :) | 02:37 |
=== _thumper_ is now known as thumper | ||
=== scuttlemonkey is now known as scuttle|afk | ||
noodles775 | hatch: Yeah - it could be clearer that none of the juju env is available until you're in a hook context (ie. config-get and other juju binaries) | 04:04 |
=== erkules_ is now known as erkules | ||
hatch | yeah just a small wording change will help with that - I'll try and remember to submit a PR for that this week | 04:17 |
=== kadams54 is now known as kadams54-away | ||
=== CyberJacob|Away is now known as CyberJacob | ||
=== CyberJacob is now known as CyberJacob|Away | ||
=== CyberJacob|Away is now known as CyberJacob | ||
=== CyberJacob is now known as CyberJacob|Away | ||
=== psivaa-holiday is now known as psivaa | ||
gnuoy | Am I missing something or is charmhelpers now broken on trusty since the py3 support mp? | 13:38 |
* gnuoy goes to investigate further | 13:41 | |
marcoceppi | gnuoy: there's a bug already | 14:14 |
marcoceppi | gnuoy: https://bugs.launchpad.net/charm-helpers/+bug/1395378 | 14:14 |
gnuoy | marcoceppi, ah, ok. thanks | 14:14 |
mup | Bug #1395378: latest python3 additions seem to break with trusty python-six <Charm Helpers:In Progress by stub> <https://launchpad.net/bugs/1395378> | 14:14 |
stub | gnuoy: If you want, check out if lp:~stub/charmhelpers/py3-2 sorts things for you | 14:15 |
gnuoy | stub, thanks | 14:16 |
stub | I can't get the venvs built in a precise VM :-( | 14:18 |
mhall119 | marcoceppi: does juju work against Canonistack? I was told it did, I can bootstrap on it but `juju status` never returns | 14:22 |
marcoceppi | mhall119: it does | 14:22 |
marcoceppi | lazyPower may have some insight on this | 14:23 |
marcoceppi | mhall119: you have do to like proxies and shuttle and stuff | 14:23 |
lazyPower | mhall119: do you have your VPN setup? | 14:23 |
lazyPower | you wont get anything back from juju status util you've activated your VPN or canonistack-sshuttle - i highly recommend the VPN route | 14:23 |
mhall119 | lazyPower: no, I don't | 14:25 |
stub | gnuoy: did that work btw? I'd like to know what series you deployed too. | 14:32 |
gnuoy | stub, I haven't tried as yet tbh | 14:33 |
gnuoy | jamespage, got a sec for https://code.launchpad.net/~gnuoy/charm-helpers/haproxy-singlenode-mode/+merge/242790 ? | 15:09 |
jamespage | gnuoy, looks reasonable | 15:11 |
jamespage | is this for the 'run haproxy always' stuff? | 15:11 |
gnuoy | jamespage, it is | 15:12 |
stub | tvansteenburgh: yes, apt-get as an import side effect is disgusting. There are some others hiding in charm-helpers too... | 15:17 |
jamespage | gnuoy, +1 | 15:20 |
gnuoy | jamespage, thanks | 15:35 |
jamespage | gnuoy, https://code.launchpad.net/~james-page/charms/trusty/quantum-gateway/stable-vlan-flat-support | 15:36 |
jamespage | I'd like to get that landed asap | 15:36 |
gnuoy | jamespage, err, I thought I already had. sorry about that | 15:37 |
jamespage | gnuoy, you acked and I landed next | 15:37 |
jamespage | thats for stable | 15:37 |
gnuoy | jamespage, +1 | 15:37 |
jamespage | gnuoy, thanks | 15:38 |
=== scuttle|afk is now known as scuttlemonkey | ||
=== scuttlemonkey is now known as scuttle|afk | ||
=== scuttle|afk is now known as scuttlemonkey | ||
=== kadams54 is now known as kadams54-away | ||
sayon | hey there, i am wondering if i can upgrade from juju 0.7 to the newer juju-core, without having to re-bootstrap the whole environment again? | 16:42 |
marcoceppi | sayon: that's a great question | 16:45 |
marcoceppi | sayon: as you may be aware a lot has changed in the latest releases of juju, let me find out if there is a migration path | 16:47 |
sayon | marcoceppi: i know, right :) | 16:47 |
lazyPower | oh wow 0.7 to 1.20.x? woo thats a huge update. | 16:47 |
sayon | lazyPower: yea, i tried to upgrade once juju-gui was introduced but i could not figure out how to do it without rebootstrapping | 16:48 |
marcoceppi | sayon: so we've since introduced a "juju upgrade-juju" command but I don't think that will work on an environment that's as old as 0.7 | 16:49 |
hazmat | sayon, there's isn't anything supported from pyjuju/zookeeper to juju-core/mongodb | 16:49 |
sayon | marcoceppi: it won't i already tried, there is no upgrade-juju command :( | 16:50 |
sayon | hazmat: i thought so | 16:50 |
hazmat | sayon, capturing the topology from juju status > and capturing the data in env and then rebootstrapping and restoring by hand .. its a pretty manual process.. juju-core does support in place upgrades. | 16:51 |
=== kadams54-away is now known as kadams54 | ||
=== kadams54 is now known as kadams54-away | ||
sayon | i am running in alot of trouble since i upgraded my maas installation on ubuntu 12.04 | 16:51 |
hazmat | sayon, doh.. | 16:52 |
hazmat | sayon, was that recently? | 16:52 |
sayon | hazmat: it has been some month i guess, everything worked just fine till i wanted to add another machine | 16:53 |
hazmat | sayon, there was a security fix that got pushed for maas in the last few weeks that broke pyjuju .. pyjuju has been eol'd for a few years. | 16:53 |
sayon | hazmat i guess thats what i am facing right now | 16:53 |
hazmat | sayon, if you want to keep the env, i'd recommending apt-pinning maas to the older version | 16:53 |
hazmat | the security fix was around the object storage capabilities of maas, and is only relevant if maas was exposed directly to untrusted users which is not common. | 16:54 |
sayon | hazmat, tried that already but i am running into problem since the database structure seems to have changed alot | 16:54 |
hazmat | ugh | 16:55 |
sayon | when downgrading it fails to convert from the newer to the older database version | 16:55 |
* hazmat tries to dig up details | 16:56 | |
hazmat | sayon, do you have the error when adding a new machine to the environment? | 17:00 |
hazmat | ie from log | 17:00 |
hazmat | it looks like the change was specifically around how charm urls are handed to units | 17:00 |
hazmat | sayon, also what version of maas is the newer one? | 17:01 |
sayon | well when i add a new machine it gets commisioned in maas and is the ready | 17:01 |
sayon | but "juju status" won't list the new machine and may not acuire it for charm installations | 17:01 |
sayon | hazmat: its maas 1.2+bzr1373+dfsg-0ubuntu1~12.04.5 | 17:02 |
hazmat | sayon, interesting.. thats a different issue.. the commit i was referencing hasn't landed in 1.2 series yet... do you know what the other version of maas was.. apt-cache policy maas should have it | 17:06 |
sayon | hazmat: wow, seems like it was 0.1+bzr482+dfsg-0ubuntu1 | 17:07 |
hazmat | sayon, hmmm.. so nutshell i'd recommend a bootstrapping that env again using the juju stable ppa on precise https://launchpad.net/~juju/+archive/ubuntu/stable?field.series_filter=precise | 17:11 |
hazmat | juju-core can self-update itself.. | 17:11 |
hazmat | i'm asking on #maas re best version/ppa for precise | 17:12 |
sayon | hazmat: that means reinstallation of all machines used with juju charms, right? | 17:12 |
hazmat | sayon, sadly yes | 17:13 |
hazmat | that version of maas was ancient, and pyjuju has been eol'd for a while | 17:14 |
sayon | hazmat: ok, i think i will then do a complete reinstallation with the new LTS version of ubuntu | 17:15 |
sayon | and not stick with precise for any longer | 17:15 |
sayon | but thanks alot for your help and support hazmat and marcoceppi! | 17:16 |
hazmat | np | 17:18 |
=== scuttlemonkey is now known as scuttle|afk | ||
=== scuttle|afk is now known as scuttlemonkey | ||
=== scuttlemonkey is now known as scuttle|afk | ||
=== scuttle|afk is now known as scuttlemonkey | ||
skay | avoine: I'd like to be able to run collectstatic, any objection to the idea of having it happen during install and upgrade? | 19:03 |
skay | avoine: also, there is the ansible branch you have, and I'm wondering if most all of hte python code in hooks.py is going away and it will turn in to a lightweight shim around ansible calls? | 19:04 |
avoine | skay: The idea was to use a subordinate for that: http://bazaar.launchpad.net/~patrick-hetu/+junk/django-contrib-staticfiles/files | 19:05 |
avoine | skay: also Simon proposed to use dj-static: https://code.launchpad.net/~bloodearnest/charms/precise/python-django/trunk/+merge/235430 | 19:07 |
avoine | skay: yes, migrating to Ansible is the long term plan | 19:07 |
skay | for that dj-static branch -- I would leave whether or not someone uses dj-static up to the project | 19:09 |
skay | but the dollectstatic command still needs to be called, and if it should be handled as a subordinate, maybe document it in the readme? | 19:09 |
skay | I happen to be using dj-static right now, and I handle the dependencies in my project's requirements file | 19:10 |
skay | so python-django installs it for me accordingly | 19:10 |
skay | on another note, when find_django_admin_cmd fails to find the admin, would you want sys.exit(1) there? | 19:11 |
skay | I'm not clear on when things should fail silently or not, but that one seems like maybe it shouldn't fail silently | 19:12 |
avoine | skay: can you give me the line number? | 19:14 |
avoine | skay: Simon's patch would be better since it is not recommended to use django.contrib.staticfiles in production | 19:15 |
skay | avoine: http://bazaar.launchpad.net/~patrick-hetu/charms/precise/python-django/pure-python/view/head:/hooks/hooks.py#L206 | 19:18 |
=== rcj is now known as Guest93149 | ||
avoine | I think I did that to catch the cases where django is not installed | 19:22 |
avoine | skay: do you have a case where this should not append? | 19:22 |
skay | avoine: I'll try to dig up what was happening to me. somehow the method was returning None and then the charm was attemping to call None syncdb etc. | 19:23 |
avoine | ouch | 19:24 |
skay | avoine: I'm not sure how much you want to dig in to it. I was working on a branch of this to allow for pip extra args and was doing --no-index --file-files and had hte wrong version of something, if memory serves | 19:24 |
skay | avoine: so it didn't install django properly? I thought perhaps that should put the unit in to an error state earlier than it did | 19:24 |
skay | it ended up looking like an error with the pgsql relation rather than earlier | 19:24 |
skay | so, the method allows for the django-admin command not to be found, and doesn't fail. perhaps it should | 19:25 |
avoine | skay: no because you could have installed django from a debian package so I'm not checking if it exist earlier | 19:25 |
skay | avoine: oh that makes sense. | 19:26 |
skay | I figured there was a reason | 19:26 |
avoine | skay: yeah the pgsql will also fail if django is not installed | 19:26 |
avoine | *relation | 19:26 |
=== roadmr is now known as roadmr_afk | ||
=== rcj is now known as Guest31019 | ||
=== roadmr_afk is now known as roadmr | ||
=== scuttlemonkey is now known as scuttle|afk | ||
=== fuzzy_ is now known as Fuzai | ||
=== Guest31019 is now known as rcj | ||
=== rcj is now known as Guest68049 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!