| === psivaa_ is now known as psivaa | ||
| === thumper-dogwalk is now known as thumper | ||
| === JoseeAntonioR is now known as jose | ||
| === scuttle|afk is now known as scuttlemonkey | ||
| nvalerkos | morning | 05:13 |
|---|---|---|
| nvalerkos | marcoceppi: Marco are you there? | 05:13 |
| nvalerkos | I have a problem with creating a mysql database, its not providing the password file to login as root in order to create the user/database when I add a relationship with an webapp i created | 05:15 |
| nvalerkos | I think i have my metadata.yaml wrong or something on the relationship with mysql | 05:16 |
| === med_ is now known as Guest93664 | ||
| BlackDex | Hello there. I get the following error, does some know where to look? "Service placement to machine not supported" | 10:35 |
| magicaltrout | what you doing BlackDex ? | 10:54 |
| BlackDex | magicaltrout: Deploing ;) | 11:07 |
| BlackDex | But i fond the problem | 11:07 |
| BlackDex | i pointed to a wrong machine | 11:07 |
| magicaltrout | cool | 11:08 |
| magicaltrout | that was going to be my suggestion :P | 11:08 |
| BlackDex | hehe thx ! | 11:49 |
| marcoceppi | jamespage: https://github.com/juju-solutions/layer-basic/pull/45 however, you can patch this in the charms in the future by having the following in your layer.yaml http://paste.ubuntu.com/15340870/ | 12:02 |
| marcoceppi | those will be evaluated before any pip calls are made (so if you need like buildessentials or libX-dev installed for a pip module, you can have that defined in the packages option) | 12:03 |
| jamespage | marcoceppi, oh nice - thanks! | 12:11 |
| marcoceppi | jamespage: either way, it'll land when cory_fu or someone else takes a look and +1's it | 12:11 |
| jamespage | marcoceppi, thanks for doing the fix - I was about to poke around and see if I could fix that up | 12:12 |
| jamespage | marcoceppi, btw | 12:12 |
| jamespage | http://paste.ubuntu.com/15340899/ works lovey | 12:12 |
| nvalerkos | marcoceppi: Hello Marco, I have a quick question. | 13:57 |
| nvalerkos | if you are there that is | 13:57 |
| marcoceppi | o/ nvalerkos | 13:57 |
| nvalerkos | so, I am trying to generate a new database connection with a non root user through metadata.yaml | 13:57 |
| nvalerkos | but I am not getting it right | 13:58 |
| marcoceppi | nvalerkos: sounds good, which charm are you connecting to? | 13:58 |
| marcoceppi | postgresql? mysql? | 13:58 |
| nvalerkos | I am making my own charm, I am using the mysql | 13:58 |
| marcoceppi | nvalerkos: sounds good - do you have your charm uploaded somewhere so I can follow along and take a look? | 13:59 |
| nvalerkos | marcoceppi: sent on prv message | 14:00 |
| marcoceppi | nvalerkos: okay, so your charm "requires" a MySQL interface, it's not a peer | 14:01 |
| nvalerkos | marcoceppo: I used it that, however it was connecting without the root password when using the db-relation-joined and getting a couldn't connect in log of python charm (mysql side) | 14:02 |
| marcoceppi | nvalerkos: the MySQL relations will provide database, user, password, port, hostname, you will need to check to make sure you have those values before contining since tihs is an async system | 14:05 |
| nvalerkos | oh | 14:05 |
| nvalerkos | ok got it, so the relation with requires is correct I just need to wait for complete relation credentials. | 14:06 |
| nvalerkos | Async. | 14:06 |
| nvalerkos | marcoceppi: thanks. | 14:08 |
| marcoceppi | nvalerkos: here's an example of a charm that does tomcat + MySQL and it's hook: https://api.jujucharms.com/charmstore/v5/~spagobi-charmers/trusty/spagobi-4/archive/hooks/metadatadb-relation-changed and https://jujucharms.com/u/spagobi-charmers/spagobi/ | 14:08 |
| marcoceppi | metadatadb is the name of the MySQL relation for this charm | 14:08 |
| marcoceppi | nvalerkos: the system is async, but don't block the hook waiting. If you don't have the data just exit 0 from the hook and wait for the hook to be called again | 14:09 |
| nvalerkos | marcoceppi: ok got it. I am using python | 14:10 |
| marcoceppi | nvalerkos: ah, same concept, but without all the base ;) | 14:10 |
| marcoceppi | bash | 14:10 |
| nvalerkos | ah | 14:10 |
| nvalerkos | ok | 14:10 |
| nvalerkos | will do. thanks again. | 14:10 |
| nvalerkos | I saw that when you first gave me that charm link yesterday and did not gave any attention to it. | 14:11 |
| jamespage | thedac, coreycb, beisner: got time for a quick one? | 14:11 |
| jamespage | https://review.openstack.org/#/c/291146/ | 14:11 |
| beisner | jamespage, sure. 2-for-1? :-) https://review.openstack.org/#/c/290943/ https://review.openstack.org/#/c/288483/ | 14:12 |
| jamespage | beisner, sure - looking | 14:13 |
| beisner | jamespage, ps, we've got the charm-single config file repo live now. https://github.com/openstack-charmers/bot-control/tree/master/config/charm-single | 14:13 |
| beisner | so ceph single is good, and the odl ones get the right vars, but odl seems to be trusty-only | 14:13 |
| jamespage | beisner, I just put up a fix for odl - it was missing systemd support for wily | 14:13 |
| beisner | jamespage, sweet, thanks | 14:14 |
| jamespage | I had that in an old bzr branch that I had some other stuff in | 14:14 |
| jamespage | picked it out - hopefully should work OK now | 14:14 |
| beisner | jamespage, did you see the c-h fork that one of the odl charms is using? | 14:14 |
| jamespage | beisner, yes | 14:14 |
| jamespage | beisner, openvswitch-odl is a bit of a mess tbh - we need to move it to layers properly | 14:15 |
| jamespage | beisner, it now passes amulet smoke at least | 14:15 |
| jamespage | https://review.openstack.org/#/c/287153/ | 14:15 |
| jamespage | beisner, just kicked off a full run of that | 14:16 |
| jamespage | beisner, another fairly trivial one - https://review.openstack.org/#/c/289249/ | 14:18 |
| beisner | jamespage, oooo. idea. we should have more magic words to kick of relevant mojo specs. in this case, it'd be the ssl-everywhere specs. it would take rejigging the specs a bit to use the change branch, but doable and probably what we want. | 14:19 |
| jamespage | sounds awesome | 14:19 |
| jamespage | tinwood, not sure whether you spotted but https://review.openstack.org/#/c/288733/ needs a rebase | 14:25 |
| tinwood | jamespage, Thanks for spotting this. I was waiting on gnuoy's keystone_v3 to get merged before I did the rebase. I'm guessing that's gone through, so I'll rebase after I push this fix to charm-helpers. | 14:27 |
| jamespage | tinwood, it has | 14:27 |
| tinwood | thanks, will do. | 14:28 |
| jamespage | thedac, narindergupta: hey - I'm due to push the nuage charms live today - but I think they are all in a private team I can't see | 14:31 |
| jamespage | https://bugs.launchpad.net/charms/+bug/1510672 | 14:32 |
| mup | Bug #1510672: Nuage Canonical Charm: vrsg <Juju Charms Collection:Fix Committed> <https://launchpad.net/bugs/1510672> | 14:32 |
| narindergupta | jamespage: let me give you access for same. | 14:33 |
| jamespage | narindergupta, ta | 14:33 |
| narindergupta | jamespage: you are part of team now and will subscraibe you that bug | 14:34 |
| narindergupta | jamespage: done | 14:35 |
| narindergupta | jamespage: please let me know once live so that we can comnunicate the same to nuage team as well | 14:37 |
| jamespage | marcoceppi, ok I'm going to put these nuage charms into the charm store using charm upload/publish etc... | 14:40 |
| marcoceppi | jamespage: gl | 14:40 |
| jamespage | marcoceppi, if I place them under 'charmers' does that make them the official charm? | 14:40 |
| marcoceppi | jamespage: no, that's not really how it works | 14:41 |
| jamespage | marcoceppi, ok I'm missing a step then | 14:41 |
| marcoceppi | jamespage: you upload then publish them /somewhere/ | 14:41 |
| jamespage | I could see how to go from development to published | 14:41 |
| marcoceppi | ie ~nuage/trusty/nuage-charm | 14:41 |
| marcoceppi | or whever | 14:42 |
| jamespage | ok | 14:42 |
| marcoceppi | then, you just "charm promulgate" | 14:42 |
| jamespage | oh - ok | 14:42 |
| jamespage | lemme try that | 14:42 |
| jamespage | and that makes that revision the actual official one then... | 14:42 |
| marcoceppi | jamespage: pretty much | 14:44 |
| jamespage | marcoceppi, bleh - charm-store does not know I'm a member of nuage-charmers... | 14:47 |
| jamespage | I suspect regular injestion from lp? | 14:47 |
| marcoceppi | jamespage: what's the error you're getting? have you `charm login` yet? | 14:47 |
| jamespage | "unauthorized: access denied for user "james-page" | 14:48 |
| jamespage | " | 14:48 |
| jamespage | marcoceppi, I've been publishing aodh all morning so I know I'm normally ok | 14:48 |
| jamespage | https://jujucharms.com/u/james-page/aodh/9 | 14:48 |
| jamespage | I guess I can put them under charmers for now and move them later | 14:49 |
| marcoceppi | jamespage: huh, looks like you're a member as of 2 mins ago ;) https://launchpad.net/~nuage-charmers | 14:49 |
| jamespage | marcoceppi, yeah - I just created that team :-) | 14:49 |
| marcoceppi | jamespage: I wouldn't moving later sounds painful | 14:49 |
| marcoceppi | jamespage: try logging in again? | 14:49 |
| marcoceppi | jrwren: help ^? | 14:50 |
| marcoceppi | err, frankban ^? | 14:50 |
| rick_h__ | jamespage: will have to log out/in to catch the new team? | 14:51 |
| jamespage | rick_h__, how do I logout? | 14:51 |
| jamespage | being a dumbass | 14:51 |
| rick_h__ | jamespage: oh, from cli? | 14:51 |
| rick_h__ | jamespage: rm -rf .go-cookies | 14:51 |
| jrwren | rm ~/.go-cookies, IIRC | 14:51 |
| rick_h__ | jamespage: think it's still there | 14:51 |
| jamespage | ta that did the trick | 14:52 |
| marcoceppi | \o/ | 14:52 |
| * marcoceppi pines for the new charm command | 14:53 | |
| jamespage | working nicely | 14:55 |
| jamespage | narindergupta, https://jujucharms.com/u/nuage-charmers/ | 14:55 |
| jamespage | rick_h__, marcoceppi: the series in metadata support is great bt | 14:56 |
| narindergupta | jamespage: thanks | 14:56 |
| jamespage | w | 14:56 |
| rick_h__ | jamespage: <3 | 14:56 |
| marcoceppi | jamespage: yeah, I can't take credit for that, but it makes charming so much easier | 14:56 |
| narindergupta | jamespage: vrsg is not listed yet. Are we releasing vrsg as well? | 14:56 |
| jamespage | narindergupta, I have slow upstream internet :-) | 14:57 |
| marcoceppi | jamespage: what a concept, you're the bottleneck to the store and not some ingestion process ;) | 14:57 |
| jamespage | marcoceppi, lol | 14:57 |
| narindergupta | jamespage: thanks | 14:58 |
| jamespage | marcoceppi, I was local deving this morning and deploying on serverstack - boucing via the charm-store was neat | 14:58 |
| jamespage | narindergupta, all done | 14:58 |
| narindergupta | jamespage: thanks I can see it now | 14:58 |
| marcoceppi | :D jamespage with the new channel stuff, it'll be even more awesome to rev around in the store | 14:58 |
| jamespage | narindergupta, a bundle would be a nice next step :-) | 14:58 |
| narindergupta | jamespage: yeah thay have bundle and i will ask team to work on with latest upstream charm and submit for review | 14:59 |
| jamespage | narindergupta, sure | 14:59 |
| jamespage | thedac, all promulgated - did it new style | 15:02 |
| jamespage | thedac, I'll explain later | 15:02 |
| marcoceppi | jamespage: looks great https://jujucharms.com/nuage-vrs/ | 15:04 |
| jamespage | marcoceppi, indeed | 15:04 |
| jamespage | something is odd with interface-rabbitmq - always takes forever to clone | 15:05 |
| lazyPower | marcoceppi - just one thing, and i'm guilty of this too. but the card says "by: undefined" | 15:11 |
| lazyPower | we're missing a data point on the charms to populate that field in the card | 15:11 |
| marcoceppi | lazyPower: that's something for the charmstore | 15:12 |
| magicaltrout | also it would be cool (if you can't already) to embed a paragraph of the readme into the card | 15:13 |
| beisner | jamespage, shall we land openvswitch-odl? full passes too. https://review.openstack.org/#/c/287153/ | 15:15 |
| jamespage | beisner, go for it | 15:15 |
| jacekn | could one of the charms tell me if hold up on https://bugs.launchpad.net/charms/+bug/1538573 is expected? AFAICT it just needs to be merged | 15:16 |
| mup | Bug #1538573: New collectd subordinate charm <Juju Charms Collection:Fix Committed> <https://launchpad.net/bugs/1538573> | 15:16 |
| jamespage | thedac, gnuoy, beisner: retro'ed my upload script into designate* charms | 15:23 |
| jamespage | https://jujucharms.com/u/openstack-charmers-next/designate | 15:23 |
| jamespage | https://jujucharms.com/u/openstack-charmers-next/designate-bind | 15:23 |
| magicaltrout | jacekn: for some reason that request is blankly ignored :) I even asked yesterday for you and got no response either...... | 15:23 |
| jamespage | marcoceppi, is the 'charm' binary a go binary? | 15:24 |
| magicaltrout | you must have done something mean in a past life | 15:24 |
| beisner | jamespage, cool. we should probably add both to openstack-charmers @ github for dev, then propose upstream projects from there. | 15:24 |
| beisner | aodh and designate that is | 15:24 |
| jamespage | beisner, we can actually just slurp them direct from lp | 15:24 |
| beisner | jamespage, oh cool. | 15:24 |
| jamespage | beisner, yah - them being git repos and all | 15:24 |
| jamespage | beisner, we should probably sort out some tests first tho | 15:25 |
| beisner | right | 15:25 |
| jamespage | they are a bit cowboy right now but only due to the very early lifecycle they are in | 15:25 |
| jamespage | I'd prefer to push under openstack once we have them in better shape... | 15:25 |
| beisner | ack | 15:25 |
| jamespage | beisner, series in metadata is a must for all of our charms... | 15:26 |
| jamespage | its a single charm upload command and the charm-store does the rest! | 15:26 |
| beisner | jamespage, yes. so, what happens if we define series now, and try to use the charms with 1.25.x ? | 15:26 |
| jamespage | beisner, it all works | 15:26 |
| beisner | well then we should add that metadata now :-) | 15:27 |
| jamespage | beisner, charm store interprets the cs:xxx request and figures out the right thing todo | 15:27 |
| jamespage | beisner, agreed | 15:27 |
| jamespage | beisner, I feel we could move to a post-commit charm upload quite quickly now | 15:27 |
| jamespage | beisner, http://paste.ubuntu.com/15341662/ | 15:28 |
| jamespage | its that simple | 15:28 |
| jacekn | magicaltrout: maybe people are afraid of collectd :) | 15:29 |
| magicaltrout | jacekn: indeed | 15:30 |
| magicaltrout | spam the mailing list as well until you get a response :) | 15:30 |
| magicaltrout | post a commend on the LP tracker as well | 15:30 |
| magicaltrout | someone will be watching somewhere | 15:30 |
| jacekn | I'm also afraid that there is a bigger problem with the review process here | 15:31 |
| magicaltrout | jacekn: i believe the process is being overhauled so its a lot more visible, but you need to ask marcoceppi and/or tvansteenburgh about that stuff | 15:32 |
| jacekn | there should be no need to chase your contributions, it's charmers who should be chasing and encouraging people to do this work | 15:32 |
| tinwood | jamespage, beisner: are you seeing anything odd with ceilometer testing? | 15:32 |
| magicaltrout | it shouldn't take 3 or 4 days of asking to get an answer :) | 15:32 |
| magicaltrout | absolutely jacekn | 15:32 |
| jamespage | tinwood, yes - what do you see? | 15:32 |
| magicaltrout | kick marcoceppi | 15:32 |
| tinwood | TypeError: can't pickle thread.lock objects. | 15:32 |
| jamespage | tinwood, nope never seen that one | 15:33 |
| tinwood | jamespage, it's in ceilometerclient/v2/client.py | 15:33 |
| tinwood | Might be a version thing. | 15:33 |
| * tinwood hmmm | 15:33 | |
| jamespage | maybe... | 15:33 |
| marcoceppi | jamespage: it is, it's being switched to that in xenial | 15:34 |
| magicaltrout | jacekn: my charm went into the review process, got reviewed once then stopped, but I didn't know if I had to do something or not to get it re-reviewed, so its just waiting until I get enough time to stand it back up again | 15:34 |
| magicaltrout | so I know what you mean | 15:34 |
| thedac | jamespage: narindergupta: \o/ Thanks for getting nuage charms promulgated. That was a long time coming. | 15:34 |
| marcoceppi | jacekn: there's a huge problem with the review process, it's being overhauled with a public beta of the new process coming out in a few weeks | 15:34 |
| jamespage | narindergupta, what's your lp id? | 15:35 |
| narindergupta | thedac: thanks for your patience as well i know it was long dur | 15:35 |
| thedac | no problem | 15:35 |
| narindergupta | jamespage: narindegupta | 15:35 |
| jacekn | marcoceppi: cool! Maybe in the mean time nominate one person a day and put them in the topic here so that people know who to chase? | 15:35 |
| marcoceppi | jacekn: what do you mean? | 15:36 |
| jamespage | narindergupta, thedac: I created https://launchpad.net/~nuage-charmers | 15:36 |
| jamespage | that's the team under which the charms are uploaded and published on the charm store | 15:36 |
| thedac | great | 15:36 |
| jacekn | marcoceppi: so instead of me asking every day and getting silence I could hilight somebody directly and get a response, even if that response is "your review is in the queue" | 15:36 |
| narindergupta | jamespage: thanks | 15:36 |
| jamespage | so adding people to that will allow future charm upload's | 15:36 |
| narindergupta | jamespage: gotch you | 15:37 |
| jamespage | thedac, narindergupta: they are one of the first partners I've promugated this way ... | 15:37 |
| jamespage | its neat | 15:37 |
| thedac | Sounds like a much better process | 15:37 |
| narindergupta | jamespage: yeah looks cool | 15:37 |
| jamespage | narindergupta, if they want to move they charm development out of launchpad that's ok - the requirement for bzr branches for charm injestion is going away | 15:37 |
| narindergupta | jamespage: is it possible to move other partners also on same process specially plumgrid? | 15:37 |
| jamespage | narindergupta, yes | 15:38 |
| narindergupta | jamespage: please look into it whenever gets time. Also it seems nuage and plumgrid wants to stick to launchpad for now. May be in future they might move to github | 15:39 |
| jamespage | narindergupta, they can do git on launchpad if they prefer - and this is only vcs | 15:39 |
| jamespage | not bug tracking etc... | 15:39 |
| narindergupta | jamespage: gotch you and will inform them. | 15:40 |
| jamespage | narindergupta, I made you an admin of nuage-charmers so you can add the nuage guys | 15:41 |
| narindergupta | jamespage: sure will ask the team who else to be involved and add them. | 15:41 |
| tinwood | jamespage, is ceilometer charm written against the v2 client? | 15:41 |
| jamespage | tinwood, tbh you're outside of my knowledge now | 15:42 |
| jamespage | it might be - the tests? | 15:42 |
| tinwood | oh it's definitely tests. Whose our ceilometer specialist? :) | 15:42 |
| marcoceppi | jacekn: well, the queue is open, http://review.juju.solutions | 15:42 |
| jamespage | tinwood, sounds like you might be now :-) | 15:43 |
| * tinwood lol | 15:43 | |
| tinwood | jamespage, I will dig into it. | 15:43 |
| jamespage | tinwood, tbh its always been a peripheral charm - not that many users user it | 15:43 |
| jamespage | we don't on serverstack | 15:43 |
| tinwood | ah, I see. I don't think it'll be a big change, just need to work out what's going on and whether it needs to be backwards compatible. | 15:44 |
| jamespage | lazyPower, you gonna use charm upload for nexenta as well? | 15:44 |
| lazyPower | Yep! | 15:44 |
| lazyPower | Now that all the pipeline bits are there, there's almost no reason not to :) | 15:45 |
| jamespage | indeed | 15:45 |
| beisner | tinwood, ceilometer charm tests seem ok. lmk if you see troubles therein. | 15:46 |
| tinwood | thanks beisner - it's probably a difference in versions on the test box vs my bastion. | 15:51 |
| tinwood | beisner, I have on the bastion: python-ceilometerclient 2.3.0-0ubuntu1 | 15:52 |
| cory_fu | Is it always the case that the first leader-elected hook will run before install? | 15:57 |
| === JoseeAntonioR is now known as jose | ||
| deanman | Is there a way to download manually the cloud-trusty lxc image before issuing a deploy --to lxc:x ? | 16:08 |
| marcoceppi | deanman: there is | 16:13 |
| marcoceppi | kind of | 16:13 |
| marcoceppi | well, in general, it's in http://cloud-images.ubuntu.com, but the process for preseeding on a deployed node I'm not 100% on | 16:13 |
| deanman | marcoceppi: when issuing a deploy it's not very informative on what goes on the background and whether is downloading the lxc or something. So i was thinking that maybe if i could download it manually i could present some more feedback to the user... | 16:16 |
| marcoceppi | tvansteenburgh: my stab at updating charmworldlib turned out pretty decently using theblues instead, finishing up tests now | 16:24 |
| tvansteenburgh | marcoceppi: excellent | 16:24 |
| marcoceppi | it's not a perfect match charmworldlib -> libcharmstore, but it's easy enough to patch | 16:24 |
| aisrael | It looks like the base charm layer is installing/upgrading pip outside of what's packaged in trusty, creating a /usr/local/bin/pip that calls python3 instead of python2. Does that sound right? | 16:34 |
| aisrael | I'm running into issues because something I'm installing has a syntax error with python3. I expect pip vs. pip3 similar to the packaged versions. | 16:35 |
| stokachu | wallyworld: new bootstrap for maas built-in is nice | 16:43 |
| stokachu | wallyworld: especially now you can just define your credentials for maas and not a local name | 16:43 |
| pmatulis | with ec2 provider, how do i specify the instance type for the controller? tried '--config instance-type=t2.micro' but it didn't work, and did not complain | 16:46 |
| marcoceppi | aisrael: if you need to, use the venv flag in the basic layer to work around this | 16:52 |
| aisrael | marcoceppi: ack. I'm trying the wheelhouse approach, but I'll look at that next if this doesn't work. | 16:53 |
| aisrael | marcoceppi: in general, though, I'm worried this behavior could be problematic (pip actually being pip3) | 16:54 |
| marcoceppi | aisrael: yes, this is why I always recommend we use virtualenv for all packaging and managing of pip | 16:55 |
| marcoceppi | deb and pip are competing packaing formats, and many people don't care | 16:55 |
| jamespage | tvansteenburgh, hey - can we get the juju-deployer release into https://launchpad.net/~juju/+archive/ubuntu/stable ? | 16:57 |
| jamespage | then I can transition our core bundles over to use git directly rather than lagging on bzr mirrors... | 16:57 |
| aisrael | marcoceppi: That totally sounds like a blog post waiting to be written. ;) | 16:57 |
| marcoceppi | aisrael: I will, getting tired of finding problems because people pip install willy nilly | 16:58 |
| * marcoceppi rage posts | 16:58 | |
| tvansteenburgh | jamespage: yeah. have you used the latest release? | 16:59 |
| === bcsaller_ is now known as bcsaller | ||
| jamespage | tvansteenburgh, yes I put it in xenial last week and have been using it since... | 17:00 |
| tvansteenburgh | jamespage: great, thanks. i'll get it in stable then | 17:00 |
| marcoceppi | tvansteenburgh: lmk if you need help backporting to the ppa | 17:03 |
| * marcoceppi has a script | 17:03 | |
| jamespage | tvansteenburgh, ta | 17:04 |
| tvansteenburgh | marcoceppi: i always just ask frankban to do it. i need to fix a build problem first though | 17:04 |
| frankban | tvansteenburgh: I think we decided to just copy the daily built deployer package when you are ready for a new release and you have QAed the deb | 17:05 |
| tvansteenburgh | frankban: yes that's correct! | 17:06 |
| marcoceppi | frankban: where's that built? | 17:06 |
| cholcombe | how do you specify the amazon machine image when deploying ? | 17:06 |
| tvansteenburgh | https://launchpad.net/~ahasenack/+archive/ubuntu/juju-deployer-daily | 17:06 |
| tvansteenburgh | marcoceppi: ^ | 17:07 |
| frankban | marcoceppi: there | 17:07 |
| rick_h__ | cholcombe: the instance type? | 17:07 |
| marcoceppi | cool | 17:07 |
| cholcombe | rick_h__, yeah i want to deploy on xenial to test | 17:07 |
| rick_h__ | cholcombe: oh series? | 17:07 |
| rick_h__ | cholcombe: --series= | 17:07 |
| cholcombe | ah cool | 17:07 |
| rick_h__ | cholcombe: on 2.0? | 17:07 |
| rick_h__ | cholcombe: or 1.25? | 17:07 |
| cholcombe | no i'm 1.25 | 17:07 |
| rick_h__ | cholcombe: so there you need to local charm it and do the series directory/etc | 17:08 |
| cholcombe | looks like maybe juju set-env "default-series=xenial" ? | 17:08 |
| stokachu | so how do i bootstrap with juju 2.0 and define a different default model? | 17:22 |
| stokachu | I was thinking something like: juju bootstrap maas-production:prod1 maas/192.168.10.114 --debug --upload-tools | 17:23 |
| rick_h__ | stokachu: 'default model'? | 17:23 |
| stokachu | so when you define the controller it creates a model with the same name | 17:23 |
| rick_h__ | stokachu: oh, that's coming in the future code | 17:23 |
| rick_h__ | stokachu: I thought it was in beta2 to go out today but I'm told it didn't make it | 17:23 |
| stokachu | ah ok | 17:23 |
| stokachu | i would like to have like maas-production:development,staging,production models | 17:23 |
| rick_h__ | stokachu: soon, (beta3?) the default model will always be called admin, and that's the main controller model. Juju will also create a model called 'default' which you are auto put into | 17:24 |
| rick_h__ | stokachu: and to change that you can pass an arg that's something like '--default-name' | 17:24 |
| rick_h__ | stokachu: so you'd have models "admin" and your custom name passed on the flag | 17:24 |
| stokachu | so there will always be a default model for admin? | 17:24 |
| rick_h__ | stokachu: there will always be an "admin" model as that's technically the 'controller' | 17:24 |
| rick_h__ | stokachu: but folks shouldn't mess with it much, but we expect folks to do things common to all models there | 17:25 |
| stokachu | yea i think that's a bit confusing, having a 'controller' but a model that is also the 'controller' | 17:25 |
| rick_h__ | stokachu: yea, it's an implementation detail that's leaked that we'll work on over time | 17:25 |
| stokachu | ok cool | 17:25 |
| rick_h__ | stokachu: but yea, you can't remove-model admin | 17:26 |
| rick_h__ | stokachu: because that's your controller | 17:26 |
| stokachu | ah right, yea i think once the definition of the admin model is more clearer that'll help the confusion | 17:26 |
| rick_h__ | stokachu: yep | 17:26 |
| rick_h__ | that's the hope | 17:26 |
| stokachu | is the --default-name something that was part of the changes in beta2 that didnt make it? | 17:27 |
| stokachu | im running from master as well | 17:27 |
| tvansteenburgh | ahasenack: ping | 17:35 |
| ahasenack | tvansteenburgh: hi | 17:37 |
| tvansteenburgh | ahasenack: hey, i'm trying to figure out why juju-deployer daily build started failing. i can see the test failure in the build log, but the confusing bit is when i look in older (successful) build logs, there's no indication that `make test` was ever run | 17:39 |
| ahasenack | tvansteenburgh: I think, if I remember correctly, that tests might be skipped if it detects it's in a launchpad buildd | 17:39 |
| ahasenack | tvansteenburgh: maybe that check failed, and it thought it should run the tests | 17:39 |
| ahasenack | I have to check | 17:40 |
| bbaqar | Hey guys. Any way I can get the CIDR of the maas network in a charm? | 17:54 |
| tvansteenburgh | ahasenack: okay thanks for your time. i just pushed a test fix so we'll see if that fixes the build | 18:00 |
| tvansteenburgh | jamespage: whenever this build succeeds we can get the package moved to stable ^ | 18:03 |
| tvansteenburgh | ahasenack: okay the build is fixed, but the version is behind (should be 0.6.4, not 0.6.1). how does that get updated? | 18:11 |
| ahasenack | tvansteenburgh: it's in the packaging branch | 18:19 |
| ahasenack | tvansteenburgh: let me update that | 18:19 |
| ahasenack | tvansteenburgh: juju-deployer - 0.6.1~bzr168~48~ubuntu16.04.1 <-- that's your build? | 18:20 |
| tvansteenburgh | ahasenack: yep | 18:20 |
| ahasenack | tvansteenburgh: updated, r49 | 18:21 |
| jose | marcoceppi: ping | 18:22 |
| marcoceppi | o/ | 18:22 |
| jose | marcoceppi: do you have a sec to take a look at an ubucon branch? | 18:22 |
| marcoceppi | jose: link me up | 18:23 |
| jose | https://code.launchpad.net/~ubucon-site-developers/ubucon-site/update-source-action | 18:23 |
| tvansteenburgh | ahasenack: ta. for future reference, do i need to contact you for version bumps? | 18:23 |
| ahasenack | tvansteenburgh: I think so, the pkging branch the recipe uses is mine, https://code.launchpad.net/+branch/~ahasenack/juju-deployer/juju-deployer-packaging | 18:23 |
| jose | marcoceppi: so, I think that update-source action is pretty straightforward, but are there any services that need restarting apart from that? | 18:24 |
| jose | marcoceppi: action is here http://bazaar.launchpad.net/~ubucon-site-developers/ubucon-site/update-source-action/view/head:/actions/update-source | 18:24 |
| marcoceppi | jose: that won't work | 18:24 |
| marcoceppi | jose: you have to run an entire migration, etc in addition | 18:24 |
| tvansteenburgh | ahasenack: ok thanks! | 18:24 |
| jose | marcoceppi: the site isn't live-running from the directory? | 18:24 |
| ahasenack | tvansteenburgh: if you want you could copy the recipe to a team-owned account perhaps | 18:24 |
| marcoceppi | jose: it is, but you have to run db migrations and such in addition to just updating the files | 18:25 |
| jose | hmmk I'll dig a bit more into it | 18:25 |
| marcoceppi | jose: you need the event to show up in the reactive framework and to reset the djang.ready or intalled or whatever event | 18:26 |
| jose | ok, thanks | 18:27 |
| tvansteenburgh | ahasenack: can you force a new build so i get the new version? | 18:27 |
| ahasenack | tvansteenburgh: yep | 18:27 |
| tvansteenburgh | ahasenack: ta | 18:27 |
| ahasenack | done: https://code.launchpad.net/~ahasenack/+recipe/juju-deployer-daily | 18:27 |
| ahasenack | tvansteenburgh: can you kick that as well? Just curious | 18:28 |
| tvansteenburgh | ahasenack: i don't see a way to do so | 18:28 |
| ahasenack | ok | 18:29 |
| ahasenack | "Request build(s)" green link just above "Recipe contents"? | 18:29 |
| tvansteenburgh | ahasenack: heh, there it is | 18:29 |
| ahasenack | cool | 18:29 |
| tvansteenburgh | good to know, thanks | 18:30 |
| tvansteenburgh | ahasenack: just noticed that frankban is offline now. can you copy that build to juju/stable, or tell me who can? jamespage has been using it for a week now and is ready to get it in stable | 18:34 |
| ahasenack | wow, that built fast | 18:35 |
| ahasenack | tvansteenburgh: hm, there might be a problem with the version number | 18:35 |
| ahasenack | tvansteenburgh: xenial has 0.6.4-0ubuntu1 | 18:36 |
| ahasenack | that's the same? | 18:36 |
| ahasenack | I mean, it's higher than my PPA, and on purpose | 18:36 |
| tvansteenburgh | it's the same | 18:36 |
| ahasenack | tvansteenburgh: that being said, no, I cannot copy. I don't see the juju stable ppa in the target list | 18:37 |
| tvansteenburgh | ahasenack: ok, will wait til tomorrow, thanks | 18:37 |
| tvansteenburgh | jamespage: ^ | 18:37 |
| ahasenack | tvansteenburgh: have you tried sinzui? | 18:38 |
| ahasenack | or any of the juju a folk I bet | 18:38 |
| tvansteenburgh | ahasenack: good idea, will try sinzui | 18:38 |
| ahasenack | s/a/qa/ I meant | 18:38 |
| cholcombe | for systemd with charmhelpers host.service_stop i need to pass an id parameter. It looks like I need to add that to the function | 18:55 |
| cholcombe | wait.. i think i can pass the service name with the id. I'm going to try it | 18:56 |
| wallyworld | stokachu: yeah, the built in maas was a last minute idea which seemed worthwhile. we'll get the credentials stuff sroted and it will be even better | 19:51 |
| tvansteenburgh | jamespage: new deployer is in juju/stable | 19:51 |
| stokachu | wallyworld: yea man sounds awesome | 20:06 |
| stokachu | wallyworld: is there a way via list-models to determine which one is the admin model? | 20:30 |
| stokachu | i want to be able to fitler out the admin model when presenting deployment models to an end user | 20:30 |
| wallyworld | stokachu: for now, it's the one called "admin". we'd have to look at tagging or something to do it better | 20:31 |
| wallyworld | stokachu: wait | 20:31 |
| wallyworld | that will be the case in the next beta | 20:32 |
| stokachu | ah ok | 20:32 |
| wallyworld | for this beta it's the one named after the controller | 20:32 |
| wallyworld | juju bootstrap mycontroller lxd | 20:32 |
| stokachu | right i figured as much, rick_h__ mentioned --default-name will that be exposed for the admin model? | 20:32 |
| wallyworld | the admin model will be mycontroller | 20:32 |
| wallyworld | --default-model will be to name the initial hosted model | 20:33 |
| wallyworld | the admin model will be "admin" | 20:33 |
| stokachu | not the admin model correct? | 20:33 |
| stokachu | ok | 20:33 |
| wallyworld | the default hosted nodel will be "default" | 20:33 |
| stokachu | cool so i can just filter out based on the name | 20:33 |
| wallyworld | yeah | 20:33 |
| stokachu | ok perfect | 20:33 |
| wallyworld | we may do it better with an attribute or something | 20:33 |
| stokachu | can i change the name of default with --default-model? | 20:33 |
| wallyworld | but for now that will work | 20:33 |
| wallyworld | yes | 20:33 |
| wallyworld | in the next beta | 20:34 |
| stokachu | ok cool | 20:34 |
| wallyworld | the nme of the hosted model | 20:34 |
| stokachu | wallyworld: https://paste.ubuntu.com/15343579/ just to clarify what im thinking | 20:35 |
| wallyworld | looking | 20:35 |
| wallyworld | stokachu: yep, looks good to me. the next beta should drop next week | 20:36 |
| stokachu | sweet | 20:36 |
| wallyworld | lte next week | 20:36 |
| stokachu | also, one last thing i promise, did you see my bug on determining the provider type through list-controllers? | 20:36 |
| wallyworld | yeah, we're looking into it | 20:37 |
| stokachu | ok cool thanks man | 20:37 |
| wallyworld | np, might be a nice surprise next beta also :-) | 20:37 |
| stokachu | hah ill keep a lookout | 20:37 |
| wallyworld | stokachu: btw, thanks for the awesome feedback and bugs etc - we're racing against time to get this work done so that external perspective if great | 20:40 |
| stokachu | wallyworld: np! so far i think the experience is great, just small things here and there | 20:41 |
| wallyworld | yeah, we can fix those :-) | 20:42 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!