=== harlowja is now known as harlowja_away [07:51] Reading backlog, I really feel I should note that I see more Hash sum mismatches from Ubuntu mirrors than I do any sort of problem with PyPI. :p [08:32] jetole: Using something like https://gist.github.com/smoser/1278651 on local image build services. [08:35] Have edited it slightly so it can read in the userdata it serves from a local file, but that's about it. [13:23] ndonegan, patches are welcome to that :) [13:26] smoser: Will see if I can put what I've done up as a proper project on github. [13:27] It's even setup to install cleanly onto a Centos 6 box and just work. [13:27] (as rpm that is) [13:29] ndonegan, nice. for cloud-init 2 , i want to have sort of a set of datasource mimickers that we can easily test against. [13:31] We've purposely disabled all data-sources except for EC2, and None which is just setup to report if EC2 fails. [13:34] ndonegan, nice. [13:34] you shouldnt have to disalbe other servcies. [13:35] other sources [13:36] We have a security team who'd prefer it to be strictly defined where the data is coming from ;) [13:36] And we had some interesting issues with Config Drive. [13:36] (Although some of that was due to faulty deployment) [13:50] that makes sense. [13:50] so going forward, i'd like for nocloud to fit your bill [13:50] would there be somethign we could do that would make it work better for you ? === alexpilotti_ is now known as alexpilotti [16:00] harmw, i just /join'ed #cirros [16:36] no shit === harlowja_away is now known as harlowja === alexpilotti_ is now known as alexpilotti [18:25] ndonegan: that's cool [19:06] alright https://review.openstack.org/#/c/170242/ seems useful; is that ok to just clone over smoser or do i need to do something else (license wise?) [19:07] we have to ge tlicense header set right [19:07] k [19:08] should i just set it to the apache one? [19:09] or do we need to contact 'juerg.haefliger@hp.com' (the other editor i think of that file) [19:09] no we dont. [19:09] well, i dont :) [19:10] because i'm acting as canonical. and juerg signed CLA, so canonical has right to do that. [19:10] hold on [19:11] k [19:11] holding horses [19:28] harlowja, https://review.openstack.org/170249 [19:29] take that one, then change header to match [19:29] k [19:30] done [19:35] alright, let's see; what else should we take over [19:35] https://review.openstack.org/#/c/170252/ (the yaml stuff) [19:35] and maybe the templating stuff [19:36] those seem like generally useful [19:39] so .. wrt pulling stuff over. [19:40] i think largely lets wait a bit. [19:40] 2 things i am afraid of from just pulling stuff without tihnking [19:40] a.) LOG verbosity [19:41] b.) string translation looseness [19:41] string.decode() + translation.encode() + looseness [19:42] k [19:43] don't fear josh is here! [19:43] lol [19:45] ok, https://review.openstack.org/170257 (template stuff) and safeyaml and url helping [19:45] all i move over [19:45] all seem generally like we'll need them anyway... [19:46] * harlowja killed cheetah though [19:46] sorry cheetah [19:53] poor cheetah [19:53] not many tears are shed i think [19:53] ha [19:54] what do you think about ditching requests. [19:54] for? [19:54] it generally doesn't seem useful. urllib2 or urllib3 maybe ? [19:54] we went to requests for https sanity. [19:54] i had urllib2 [19:54] urllib3 maybe [19:54] but iknow that urllib3 does the right thing in pytho3n [19:54] *hate urllib2 [19:55] i dont knwo tha trequests really buys us antyhign. [19:55] sure [19:55] urllib3 powers requests though; so idk [19:55] right [19:55] looks like we could jsut use it though [19:57] i guess i'd be ok with it, urllib3 seems fine [19:57] nothing to crazy like urllib2, lol [19:57] look at https://github.com/stackforge/cloud-init [19:57] why does that say "http://openstack.org" ? [19:57] :-/ [19:57] unsure [19:57] my guess infra folks put that there? [19:57] i really want to be careful about dependencies. [19:58] sure [19:58] do u want to jump on #openstack-infra and ask why? thats there [19:58] sure [19:58] "Cross-platform instance initialization http://openstack.org" -> "Cross-platform instance initialization"... [19:59] https://github.com/stackforge/anvil also has that; and a few others [19:59] so i guess its just a common template or something [19:59] although weird [20:04] smoser the mordered using it was a joke, apparently he was complaining about it recently... [20:05] smoser http://paste.ubuntu.com/10726145/ [20:05] blah blah, lol [20:06] https://review.openstack.org/#/c/165914/ ... [20:06] stupid stuff, lol [20:06] :) [20:07] hehe, let's see here [20:08] http://paste.ubuntu.com/10726170/ ... [20:09] smoser ^ [20:09] more fun fun [20:09] lol [20:09] i'm pretty sure its because hp cloud is using 0.6.3 (which isn't fully functional i think with newer openstacks) [20:10] http://eavesdrop.openstack.org/irclogs/%23openstack-infra/%23openstack-infra.2015-03-19.log [20:10] search '2015-03-19T17:14:19' [20:10] and read from there, ha [20:10] then i don't need to paste [20:10] lol [20:11] odd. [20:11] i'm not sure what they made instead :-/ [20:11] let's see what he says, ha [20:12] i dont know why you'd mount the config drive and leave it mounted. that doesnt seem to make much senes. [20:12] ya, u got me :-/ [20:23] bb food; smoser u got it covered :-P [20:23] will read in a little [20:23] fun fun [20:23] lol [20:27] thanks. [20:27] :) [20:27] i didn't eed such discussion. [20:27] i want to make cloud-init not suck for sure. === ijw_ is now known as ijw [22:13] smoser agreed; but some of these statements from all those people are all conflated [22:14] and putting blame in the wrong place.... [22:14] packaging sucks, its all your fault [22:14] i packaged my images in weird ways, its all your fault... [22:14] blah blah