[11:47] ~/join #ubuntu-release === meetingology` is now known as meetingology [16:13] smoser: https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/348704 is up with your comments addressed [16:13] just tested on azure to make sure it works [16:22] blackboxsw: its perfectly fine to tell me to go away when i suggest thing like additional functions that just get called from other functions. [16:22] i just find that easier to mock the method '_is_platform_viable' than mock.patchObject [16:22] heh. [16:23] I agree with testability. I actually forgot to address your separate network_config logic into a standalone function. [16:23] do you think that 'is_platform_viable' should be a class method ? [16:23] i guess there may well be some datasource that needs to look at system config in order to figure that out. [16:23] smoser: at some point yes as the superclass will be able to walk through is_viable check before get/crawl_metadata processing once subclasses all have it. [16:24] it's an evolution though. I don't mind taking small steps to get there (and simpler test coverage) [16:24] the unittests in azure are notably bad as there are so many mocks/setup [16:24] decoupling from the datasource makes it easier as you said [16:24] simpler test == more certainty and better test coverage [16:25] ok. i'm fine with instance method for now. [16:26] smoser: instance method for network_config? or are you talking about _is_platform_viable [16:26] blackboxsw: maybe the diff just dint updat e? [16:26] https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/348704 [16:26] shows <<<<< [16:26] i was talking about _is_platform_viable [16:26] smoser: yeah that diff is out of step, I've tried --force etc. [16:26] please do move the convert out to a methoc [16:26] function [16:26] wahtever [16:27] ok network_config -> function [16:27] (/me probably used 'method' confusingly above) [16:27] yeah [16:27] yeah I got what you meant [16:52] smoser: just pushed separate _parse_network_config function for Azure branch [17:12] blackboxsw: did you ping about sru release ? [17:12] smoser: I did, yesterday I pinged sil2100, this morning I pinged RAOF [17:12] smoser: no response to either ping https://pastebin.ubuntu.com/p/DDYsvBmd8Y/ [17:13] same comment set both days [17:13] shall we ping robie? [17:13] :-( [20:08] smoser: I'll add net-convert format option type for azure. good idea (from a test/validation perspective) [20:13] <[42]> is there an easy way to skip the first boot config? i'm trying to migrate a vm without reinstall to use cloud-init for network config [20:13] <[42]> without applying anything else [20:17] <[42]> basically how do i mark it as "first boot already run"? [20:42] [42]: what is it that you do not want to run [20:43] <[42]> i definitely want to keep my ssh hostkey [20:43] <[42]> and i don't need user generation [20:43] but you need cloud-init to read networking information and apply it? [20:43] smoser: pushed net_convert --kind to support azure-imds https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/348704 [20:43] <[42]> yes [20:43] what cloud platform? [20:43] <[42]> kvm on proxmox [20:45] [42]: well, its not really a common / supported option [20:45] but if you look in /var/lib/cloud/instance/sem [20:45] you'll see a bunch of files [20:46] <[42]> doesn't exist before first run [20:46] that represent markers for things in /etc/cloud/cloud.cfg 'cloud_init_modules', 'cloud_config_modules', 'cloud_final_modules' [20:47] hm.. ignore the first part of that statement. [20:47] but you can probably basically go and comment out anything youthink you might not want to have run/re-run in /etc/cloud/cloud.cfg [20:48] <[42]> i guess that's a better option to just comment out everything but network for now [20:50] added test instructions to https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/348704 for the new cli [20:50] <[42]> what's /etc/cloud/templates for? [20:51] files from there are rendered into other places. [20:52] blackboxsw: annoying [20:52] i think launchpad went to lunch permenantly on updaitng the diff there. [20:53] smoser: I can remove the branch and resubmit and it'd be fine. shall I do that (it's lose our comment review history though) [20:53] *it'll lose* [20:53] well you can just reject it [20:53] and then link to it from the other [20:53] then its not delted. [20:53] but maybe you forgot to push ? [20:53] <[42]> which module sets the network config? [20:53] [42]: it doesnt happen in a module [20:54] <[42]> where does it happen? [20:54] so it will happen if cloud-init thinks its a new instance. [20:54] <[42]> okay [20:54] from cloud-init init (possibly --local ) [20:54] smoser I forgot I rebased. just hit --force on the MP [20:55] should see 1d429c3cb514b35b84efc40533f5935ec2abdf33 committsh [20:56] blackboxsw: ? [20:56] im confused [20:56] <[42]> why does it explicitly add `post-up ifup eth0:1` to the interfaces config? [20:56] <[42]> at least when i used it in the past it would automatically start anyways [20:57] what is "it" ? [20:58] <[42]> `cloud-init init` [21:00] well... iirc it was related to getting some static routes to be guaranteed aplied. [21:00] intrfaces is kind of messy. [21:00] can you show what the input netowork config was ? [21:03] blackboxsw: i still see <<<< [21:03] (line 455 of visual diff ?) [21:04] <[42]> smoser: https://gist.github.com/Nothing4You/fadcc21d17a35da106842fdd7101ebf1 [21:04] smoser: I just hit resubmit in LP. https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/352639 [21:04] it should clear up the diff [21:05] I don't see it when I merge into master [21:05] the diff is devoid of <<<<<< [21:05] weird, new merge proposal has it too [21:07] your maybe_remove_ubuntu_network_config_scripts is in _get_data [21:07] were you going to move that to activate ? or is that too late. [21:08] i have to run though. i'll lok more tomrrorw. [21:10] smoser: last try https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/352660 [21:10] smoser: activate is too late [21:11] it happens in init stage instead of init-local when we need to render netplan [21:11] and if we don't remove /etc/netplan/99-azure-hotplug.yaml cloud-init's netplan will collide [21:11] in init-local timeframe [21:12] [42]: fwiw, you must hvea a downlevel cloud-init. newre cluod-init will render that with multiple 'iface' entries [21:12] smoser: woo hoo last merge proposal listed doesn't contain merge conflict markers [21:12] <[42]> smoser: 0.7.9 [21:12] <[42]> debian stable [21:13] http://paste.ubuntu.com/p/ddyXzmBwxs/ [21:13] [42]: 0.7.9 is old :-( [21:13] i really do have to run. [21:13] later [21:13] <[42]> are there pre-built packages for debian? [21:14] <[42]> thanks smoser