[07:08] <Odd_Bloke> smoser: Do we have a good way of specifying apt mirrors for different distros via vendor-data?
[08:26] <Odd_Bloke> harlowja: Tut, tut, only 25% coverage of url_helper.py.
[08:36] <Odd_Bloke> Oh, no, tox wasn't recreating a virtualenv to include new test deps.
[08:36] <Odd_Bloke> harlowja: Tut, tut, only 81% coverage of url_helper.py. ^_^
[09:07] <Odd_Bloke> Oh, hmph, the Windows changes have completely broken the coverage reporting.
[09:07] <Odd_Bloke> To ensure that we're getting accurate numbers (including modules not touched by tests), nose imports all the modules.
[09:07] <Odd_Bloke> As you might imagine, this goes poorly.
[09:23] <openstackgerrit> Daniel Watkins proposed stackforge/cloud-init: Don't use --cover-inclusive; Windows modules break it.  https://review.openstack.org/190557
[09:24] <Odd_Bloke> harlowja: claudiupopa: ^ fixes the coverage breakage.
[09:24] <Odd_Bloke> I'm looking at a fix in nose, which will allow us to reintroduce it.
[09:24] <Odd_Bloke> But, for now, there's no way around it.
[10:24] <Odd_Bloke> https://github.com/nose-devs/nose/pull/924 is the nose fix.
[15:06] <smoser> Odd_Bloke, i'd have to look.
[15:06] <smoser> not really here today, just came in for a call.
[15:07] <Odd_Bloke> smoser: No worries; enjoy your absence. :)
[15:12] <a-tal> hey, does anyone know how I can configure cloud-init to use a metadata version endpoint that's not 2009-04-04 ?
[15:14] <Odd_Bloke> a-tal: Which data source are you using?
[15:14] <a-tal> ec2
[15:16] <Odd_Bloke> a-tal: I don't think you can.
[15:16] <Odd_Bloke> a-tal: But even if you could, cloud-init wouldn't know what to do with anything more recent than that without code changes.
[15:16] <Odd_Bloke> a-tal: Is there something specific that you need?
[15:17] <a-tal> so, it's not the real ec2, but it does respond to http://169.254.169.254/latest/meta-data/instance-id and the other endpoints using latest. you can configure the datasource metadata_urls; but that's the path/version seems to be hardcoded?
[15:18] <a-tal> s/that's//
[15:18] <Odd_Bloke> a-tal: Yeah, we hard-code the version because we know we'll work against that version.
[15:19] <Odd_Bloke> a-tal: EC2 could completely change where everything is located in 2015-06-11, but we'd still be fine because we're pinned at that version.
[15:19] <Odd_Bloke> (Where 'that version' is 2009-04-04')
[15:20] <a-tal> makes sense
[15:22] <Odd_Bloke> So your fake EC2 will have to provide those URLs and, crucially, actually do what the API should do in that version.
[15:23] <Odd_Bloke> I would imagine that's not much of a problem (though I haven't looked at it at all), I suspect the parts of the API that cloud-init exercises aren't changed a great deal.
[15:23] <a-tal> yeah. i think i'll have to try to get a rewrite rule added
[15:23] <a-tal> hopefully nothing explodes :)
[15:24] <Odd_Bloke> :)
[16:10] <ByPasS> Odd_Bloke : hi again :)
[16:11] <Odd_Bloke> ByPasS: Hola!
[16:11] <ByPasS> Odd_Bloke : I realized that I had libvirt injection set to true which well from what Ive read is not needed and should be set to false for cloud-init admin password injection
[16:15] <ByPasS> Odd_Bloke : back to square one :)
[16:16] <ByPasS> Odd_Bloke : so I disabled libvirt password and key injection (keys are injected according to the log)
[16:16] <ByPasS> Odd_Bloke : which metadata url exactly you wanted me to paste ?
[16:21] <Odd_Bloke> ByPasS: So how are you expecting the password to be set at the moment?
[16:22] <ByPasS> Odd_Bloke : via cloud-init + nova metadata
[16:22] <ByPasS> Odd_Bloke : leet me try something I found a probable cause brb
[16:23] <Odd_Bloke> Sure. :)
[16:53] <frickler> how are chances to get manage_resolv_conf working on Ubuntu? we have a case where we want to override nameservers for a specific subset of instances only
[16:53] <frickler> see https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1394224 and https://answers.launchpad.net/ubuntu/+source/cloud-init/+question/234041
[16:55]  * frickler waves to smoser a bit ashamed about not having responded yet regarding cirros :D
[18:41] <harlowja> Odd_Bloke only 81% i want 181%
[18:41] <harlowja> lol
[18:41] <harlowja> harder faster, stronger, lol
[18:41] <harlowja> or else fired
[18:41] <harlowja> ha
[19:03] <openstackgerrit> Merged stackforge/cloud-init: Don't use --cover-inclusive; Windows modules break it.  https://review.openstack.org/190557