[00:04] blackboxsw, http://paste.ubuntu.com/24804339/ === rangerpb is now known as rangerpbzzzz [04:36] rharper, where does the setuptools depends come from ? [04:36] no where, we need to explicitly depend on it [04:37] in el6, it gets pulled in via some other package dep but we BuildRequires we should also just Requires it as well [04:37] we had it fixed earlier today [04:38] i dont know.. i dont think i ever had that in there. [04:38] hrm [04:38] i think that requirements.txt file is getting written with it in there. [04:38] while we do not need it [04:39] test-requirements.txt:setuptools [04:39] [04:39] not sure if that's getting picked up; in any case, we *do* need it since the loader requires it (/usr/bin/cloud-init calls it for pkg_resource) [04:40] hm.. [04:40] test-requirements.txt says [04:40] # Only really needed on older versions of python [04:40] contextlib2 [04:40] setuptools [04:42] I'm just grepping around, I don't know how we got setuptools into the previous build that worked, but it's not installed now via yum install cloud-init [04:42] I thought we agreed it's a requirement as long as we're using setup.py [04:43] no. i think that was pkg_resources [04:43] i forget. [04:43] what a pain [04:44] pkg_resources comes from setuptools [04:56] ok. i added that to spec [04:56] and pushed [05:04] rharper, ^ [05:04] k [05:04] and wih that... /me goes to bed. [05:04] ci is happy with it now too. [05:04] good night [05:04] k [05:12] https://copr.fedorainfracloud.org/coprs/g/cloud-init/cloud-init/build/562783/ [05:12] smoser: that's working on el6 and el7 =) === frickler_ is now known as frickler === sambetts|afk is now known as sambetts [14:36] rharper, i split into 2 merge proposals [14:36] https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/325311 [14:36] https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/325192 [14:36] ok [14:36] ie, cloud-config in one and redhat spec in another. [14:36] ack [14:36] one question... i dont know if we need the centos variant. [14:36] do we? [14:36] * smoser walks down [14:36] yes, for the cloud-config [14:37] we use the distro.variant to set the distro value as well as the default user name (and Gecos values) [16:30] smoser: dumb question: should I expect cloud-init to handle network_data.json when booting with a network data source, rather than config drive? [16:36] larsks, not yet [16:36] really, really want to do that... [16:36] but not at the moment :-( [16:36] smoser: thanks, just wanted to make sure I wasn't crazy. [16:38] larsks, the goal is to make the openstack datasource work like the digital ocean one does. [16:38] we'd identify (via dmi data) that we are running on penstack. [16:38] and then raise an interface with the link local address, and hit the metadata service [16:38] get the network data [16:38] take the nic down [16:38] apply the network_data.json [16:39] That makes sense. [16:49] rharper, [16:49] https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/325192 [16:49] addressed all your things there === sambetts is now known as sambetts|afk [18:00] smoser: ok, trunk needs this http://paste.ubuntu.com/24809440/ [18:02] smoser: http://paste.ubuntu.com/24809461/ look ok ? [18:13] powersj, http://paste.ubuntu.com/24809515/ [18:48] powersj, http://paste.ubuntu.com/24809682/ [20:43] rharper: https://paste.ubuntu.com/24810376/ [20:46] powersj: nice! === cpaelzer_ is now known as cpaelzer [23:03] larsks, we're trying to get spec file going in trunk. it builds happily now in copr and we get rpms out. [23:03] horay! [23:04] but we would also like the things to actually *work*. :) [23:04] and that is causing a problem in that right now we're not getting systemd stuff enabled in the spec file. [23:05] smoser: "not getting enabled" == "you are running systemctl enable but service isn't starting"? Or something else? [23:05] %define use_systemd (0%{?fedora} && 0%{?fedora} >= 18) || (0%{?rhel} && 0%{?rhel} >= 7) || (0%{?suse_version} && 0%{?suse_version} >=1210) [23:06] so, in specfile, that sets a flag to use systemd based on the dist version [23:06] Sure. [23:06] I think we need that [23:06] larsks: does that look sane to add ? [23:07] then we do an if %use_systemd and BuildRequires systemd [23:07] http://paste.ubuntu.com/24811338/ [23:07] rharper: It seems like a reasonable idea. Is it not working? [23:07] spec ends up looking like ^ [23:07] * larsks looks [23:07] larsks: using what smoser posted [23:07] but I thought that rhel7 builder chroots would have systemd in it by default [23:07] if we switch to my post, then we need to buildrequires systemd [23:07] based on version, I think we'll need to do that [23:09] COPR builds pushed by jenkins: https://jenkins.ubuntu.com/server/job/cloud-init-build-rpm/1/console I'll do the testing of the rpm's tomorrow. [23:09] rharper: I can take a closer look later this evening; kids go to sleep by 9pm (us/eastern). [23:09] larsks: sure, thanks [23:09] powersj: \o/ [23:10] Will you be around then? And/or are you using a spec file different from what smoser posted? [23:11] I'll be out though [23:11] larsks, well ^ is in trunk now. [23:11] you can 'make srpm' [23:11] Cool. [23:11] Will do. [23:11] PYVER=python2 make srpm [23:11] one of the two [23:12] it seems odd to me that one would need from cent6 to cent7 people would have to change every package that had systemd scripts [23:12] i guess maybe over time in fedora that just happened. [23:39] blackboxsw, [23:39] return json.dumps(data, indent=1, sort_keys=True, [23:39] separators=(',', ': ')).encode('utf-8') [23:39] format the json with that json.dumps [23:41] thanks [23:49] smoser: larsks: this looks to work http://paste.ubuntu.com/24811524/ [23:49] re: specfile updates