[01:29] stub: can you ping me when you are around, I have a (hopefully simple) postgresql question for you [02:41] Can someone help me find more info about an issue I'm having: (Very) Approximately 50% of deploys I'm testing with the local provider, result in 7 of 8 machines all coming up successfully, but 1 machine remaining pending/allocating. If I check the machine log, that one machine is unable to connect to the api apparently because the juju-generated CA is rejected: http://paste.ubuntu.com/11071952/ [02:55] I'll create a bug after lunch and destroy the env again, but if there's anything else I can put on the bug report while it's still up, that'd be great to know. [03:21] noodles775: hmm... sounds bad [03:21] yes, bug please [03:48] thumper: done (jfyi) https://bugs.launchpad.net/juju-core/+bug/1453644 [03:48] Bug #1453644: One unit remains pending with local provider [04:27] noodles775: ta [06:09] good morning, is charm-helper-sh package abandoned?! if yes does anyone know why? [06:11] schkovich: its depreciated in favor of the `charmhelpers` package, which exposes a lot of the functionality through the 'ch' command [06:11] thanks lazyPower :) [06:12] schkovich: however, the `ch` command is largely undocumented at this point, but marcoceppi may have published them while I wasn't looking. [06:12] them being, updated docs on how to use it :) [06:13] hmhhh... but it is python package. am i on the wrong page? [06:20] lazyPower: I can find only projects and packages with dash eg charm-helpers and all of those are python and not bash helpers/libraries [06:21] is bash deprecated in favour of python ;) [09:28] ls -lhrt [10:45] jamespage, ooi are you using juju 1.24 ? I seem to be seeing juju-dpeloyer explosions with it [10:46] gnuoy, not yet [10:46] ah, ok. [10:48] mgz, do you use juju-deployer in any of your juju testing? [10:48] gnuoy, https://code.launchpad.net/~james-page/charms/trusty/neutron-api/kilo-dvr-stable/+merge/258750 [10:48] jamespage, ta [10:50] gnuoy: yup [10:52] gnuoy: eg http://reports.vapour.ws/releases/2627/job/aws-deployer-bundle/attempt/314 [10:53] mgz, Bug #1453760 [10:53] Bug #1453760: "u'Error': u'status for key "" not found" before adding relations when using juju 1.24 [10:57] gnuoy: could well be a juju bug, not clear what's triggered it though [10:58] mgz, is the # in the unit name suspicious ? [10:58] { u'Error': u'status for key "u#keystone/0" not found', [10:59] yeah, it *looks* like an interal tag has leaked out somehow [10:59] but I'm not sure I'm reading that correctly [11:00] gnuoy: see bug 1437266 for comparison [11:00] Bug #1437266: Bootstrap node occasionally panicing with "not a valid unit name" [11:02] mgz, ah, i think mine is probably a dupe of that one [11:02] well, not as written, but seems likely the cause is the same [11:35] mgz, would it be possible to get Bug #1434458 triaged ? It's alsp killing random deployments [11:35] Bug #1434458: juju-deployer address,port split not working with ipv6 [11:40] gnuoy: who have we got actually working on jujuclient? [11:41] gnuoy: in other words, should you or I just fix this? [11:41] seems like bumping the bug around won't do anything [11:41] mgz, oh, I assumed juju-core handled it. [11:41] mgz, I can prep a branch [11:41] I will review it [12:58] Hi, I tried to deploy landscape-dense-maas bundle by dragging and dropping it to the canvas, it gives "An error occurred while deploying the bundle: ". Anyone know where in my configuration might be wrong? [12:59] catbus11: atm a colocated bundle (like the landscape-dense) doesn't work but we'll have releases of the juju gui and juju quickstart that do supported the colocation in the bundle this week. [13:00] rick_h_: What does colocated bundle mean? [13:01] ah, services deployed in the same host? I guess? [13:01] catbus11: the services are colocated on the same machine (using lxc containers or the like) [13:03] rick_h_: thanks. [13:34] wallyworld, if you're about have an environment with Bug #1451283 [13:34] Bug #1451283: deployer sometimes fails with a unit status not found error [13:34] s/have/I have/ [13:36] hey evilnickveitch [13:37] jcastro, hey yourself [13:37] on the getting started page [13:37] I think we should get rid of the links to each provider, since they're already on the sidebar [13:37] it just gives us 2 places to update each time there is a new provider [13:38] yes, that's a fair point [13:38] I'll send a PR after I send one for dreamhost [13:38] mgz, https://code.launchpad.net/~gnuoy/python-jujuclient/ipv6/+merge/258770 [13:39] gnuoy: looking [13:40] mgz, I haven't had much joy with test_jujuclient.py. I've done a manualy deploy with ipv4 and ipv6 but that's the extent of my testing [13:42] what kind of not joy? [13:42] hm, ipaddress module is 3.3 or later [13:45] mgz, I tried it against the openstack provider and got failures=5, errors=2 and I don't believe they're related to my branch. I assume it makes some ec2 assumptions [13:45] is I see [13:46] gnuoy: what I'd do in cases like that, [13:46] is writea "split port" function and just add unit tests for that [13:47] not worry about making the whole suite green on my machine [13:47] mgz, ok, can do [13:48] like, I'm pretty sure your use of re.search is wrong [13:48] and the following line is just .strip("[]") spelt funny [13:51] gnuoy: like, would your unit test pass with the input "[2001:db8::1]:80" [13:52] mgz, yes I understand. I believe it would but I'll write the unit test to prove [14:12] mgz, mp updated [14:13] gnuoy: ta, looks fine, do you want pytho nits? [14:14] mgz, yes pls [14:18] gnuoy: done [14:19] mgz, very much appreciated, thank you [14:20] lazyPower: it's `chlp` [14:21] ah - right. [14:21] 2am though - i'm not surprised I got it wrong. [14:21] thanks for just fixing. we'll need to find out what the release process is too - looks from the log like dpb1 did one [14:24] andreas and I have the latest builds in a PPA. I don't have access to push to pypi, that is just hazmat (from what I know) [14:26] Same for python-jujuclient btw. I run both on my machine all the time. [14:27] https://launchpad.net/~ahasenack/+archive/ubuntu/python-jujuclient, https://launchpad.net/~ahasenack/+archive/ubuntu/juju-deployer-daily [14:27] dpb1: thanks! [14:28] dpb1: several people have pypi access [14:29] dpb1: tvansteenburgh and frankban for example, let me know if you want it as well [14:30] hazmat: tvansteenburgh is enough, IMO [14:30] * tvansteenburgh winces [14:35] mgz, nits crushed [14:40] hazmat, I believe I've addressed the nits, any chance you could land my branch ? [14:41] gnuoy: i can't atm (on corporate network) i can merge it latter this evening, if tvansteenburgh doesn't beat me to it. [14:41] hazmat, thanks, that'd be great. [14:46] gnuoy, hazmat: merged [14:46] tvansteenburgh, thanks! [15:49] any JuJu AWS experts about? Trying to add a custom endpoint (for eucalyptus) by modifying aws.go, but when running juju bootstrap, I get the error: ERROR index file has no data for cloud {emea-az-1 http://109.104.120.3:8773/services/compute} not found [15:49] what other files do I need to add the endpoint to to test this? [15:56] aisrael: You pinged me last week [15:57] stub: Hey. I was testing the postgresql charm at the time, but then you moved it to work-in-progress. [15:58] aisrael: Yeah, I'm not sure I can get the tests running sanely without much more restructuring. I'm thinking of doing that as part of a major rewrite now I have cool new juju features to work with. [15:58] lazyPower, ping [15:58] mwenning: pong [15:59] stub: ack. If you need eyes on it, let me know. [15:59] aisrael: Maybe I can get some of them back. I'll be looking at it again this week. [15:59] Just got an email from Jean-Francois Joly, is everything set up for them? any roadblocks? [16:00] mwenning: they're at the point they need review from the ~openstack team [16:00] ~openstack-charmers [16:00] lazyPower, who's that [16:00] https://launchpad.net/~openstack-charmers [16:01] oi *facepalm* i saw your reply and didnt verify... === urulama_ is now known as urulama__ === mwhudson_ is now known as mwhudson [21:49] kirkland: ping