[07:20] <andras-kovacs> Hi! Do you have any idea how can I disable cloud-init's resolv-conf module? I've added the manage_resolv_conf: false to my cloud.cfg but seems like cloud-init doesn't care about it. ☹︎ Thank you! (I'm using cloud-init-18.5)
[09:17] <pmjdebruijn> .wc
[09:29] <tribaal> account ran out of funds
[09:29] <tribaal> bouncer died
[09:30] <tribaal> oops, wrong chan. Apologies :)
[10:54] <meena> tribaal: you don't have to apologize to anyone for *that*
[10:54] <tribaal> hehe
[10:56] <tribaal> Well, you'd think that working for a public cloud you would somehow be more disciplined with regards to paying your cloud bills... :)
[10:57] <meena> imagine if you ran out of money, and the utilities company shut off your water and electricity tho, that would suck.
[10:58] <meena> not saying an irc bouncer is as important as electricity, but we've been selling people on the idea that cloud is a commodity, but we treat it … differently.
[11:31] <tribaal> meena: that's a good point :) In this case however the blame sits squarely on myself and my email filtering rules: one of my account is "free" and antoher account is "paid" to test things as a normal user would, and my email rules dismissed the many  warning emails we usually send our users :)
[13:15] <AnhVoMSFT> rharper blackboxsw is there anyway to ask cloud-init to apply network config again through command line post boot?
[13:17] <AnhVoMSFT> (as in query metadata source, get networking information, render network config, apply it)
[13:37] <Odd_Bloke> andras-kovacs: What behaviour are you trying to disable?
[13:40] <Odd_Bloke> meena: The one time I've paid for electricity up-front (by topping up a card which plugs into the meter), that's exactly what happened when you ran out.
[14:07] <meena> Odd_Bloke: i thought you lived in a more civilised country…
[14:10] <Odd_Bloke> meena: Well, the advantage of the system is that the utility company doesn't know who you are (you literally go to a corner shop with the card and can pay cash to top it up).
[14:10] <Odd_Bloke> The flip side is that they don't have anyone to chase for unpaid bills, so it does turn off when you don't have money on it.
[14:29] <Odd_Bloke> Goneri: Travis failed to report results on #268 for some reason, looks like you have CI failures though: https://travis-ci.org/github/canonical/cloud-init/builds/665860962
[14:29] <Goneri> Odd_Bloke, thanks.
[14:29] <Odd_Bloke> meena: (But yes, I do believe that utilities should be publicly-owned and ~free. ^_^)
[14:47] <Goneri> meena, https://github.com/canonical/cloud-init/pull/269
[14:47] <meena> looking
[15:01] <rharper> AnhVoMSFT: on the cli for re-rendering the network config to the system; that's not there at the moment. it wouldn't be too hard to do though; we have a sub-command 'cloud-init devel net-convert'  which could get extended to take a datasource name, and optional flag to attempt to read from object cache
[15:39] <Smithx10> Hello fellow cloud-init'rs.    I was thinking of extending cloud-init to support void-linux
[15:39] <Smithx10> it uses rc.d,  what is a good way to get started.... just look at distros/ ?
[16:14] <cloaked1> powersj: still waiting for an answer on https://discourse.maas.io/t/how-to-mount-nfs-volume-via-preseed/1404 :)  Hopefully, the person who can answer the question becomes available today
[16:25] <blackboxsw> Smithx10:  I think cloudinit.distros would be a good place to start.  NetBSD was just added a couple weeks ago. https://github.com/canonical/cloud-init/pull/62
[16:26] <Smithx10> Yeah,  it would be nice to see how non systemd systems interface
[16:26] <Smithx10> Thanks
[16:54] <rharper> cloaked1: I just replied, let me know if you have any more questions in the discourse thread, I'll reply there
[16:55] <cloaked1> Thank you rharper. Taking a look.
[18:52] <Goneri> Odd_Bloke, could you drop the wip label? https://github.com/canonical/cloud-init/pull/147
[18:52] <Goneri> Odd_Bloke, a review is also welcome :-)
[18:52] <blackboxsw> done Goneri
[19:27] <Odd_Bloke> Minor refactoring I found when digging into the archive mirror bug: https://github.com/canonical/cloud-init/pull/271
[19:32] <blackboxsw> Odd_Bloke: rharper thoughts on this approach for Oracle non-initramfs distros? https://github.com/canonical/cloud-init/pull/174#pullrequestreview-379733515
[19:33] <blackboxsw> I think we can avoid going doing EphemeralDHCP duplicate route setup in iscsi environments also by checking connectivity_url on Oracle only.
[19:40] <blackboxsw> landed #271
[19:40] <rharper> blackboxsw: did you see my email about daily ppa build failure ?
[19:40] <blackboxsw> Odd_Bloke: shepherding question, are we waiting another week on robjo's feedback for https://github.com/canonical/cloud-init/pull/160 ?
[19:40] <blackboxsw> rharper: no, looking now
[19:41] <rharper> blackboxsw: thanks, it looks like the ec2 static patches aren't building correctly; they don't apply
[19:41] <robjo> I was just about to look at that when something else popped up. I'll get to it tomorrow
[19:41] <rharper> this is breaking all daily builds since it landed
[19:42] <blackboxsw> robjo: excellent, thanks sir
[21:02] <Odd_Bloke> Goneri: Label removed.  Apologies, thought Chad had done that eariler. :)
[21:03] <Goneri> thanks. I just saw a couple of remaining comments from rharper. Otherwise the PR works fine.
[21:04] <johnsonshi> Hey there, do you know what the syntax for a multiline runcmd is? Because of our specific use case, we cannot rely on the write_files module to write the files.
[21:04] <johnsonshi> What I am trying to do is use the runcmd module to do a here-doc cat
[21:04] <johnsonshi> cat <<EOF >> /some/file
[21:05] <johnsonshi> But it wouldn't be proper YAML syntax if a runcmd command spans multiple lines. I couldn't find examples of runcmd with a multiline command. Thanks as always guys :)
[21:05] <rharper> johnsonshi: have you tried write_files ?
[21:06] <rharper> and you can template that cloud-config;  lemme see if I can work out a multi-line runcmd for you... I know we have a few of them; prepare for some yaml anchors
[21:07] <johnsonshi> For our purpose, we need to use runcmd because we are need to run `sed -i`, and right after that, run `cat <<EOF`
[21:10] <rharper> k
[21:12] <Odd_Bloke> johnsonshi: YAML does support multi-line strings.
[21:14] <Odd_Bloke> johnsonshi: https://yaml-multiline.info/ <-- you want the literal block style, I believe
[21:14] <johnsonshi> Odd_Bloke: Thanks! Yeah I was trying out the other anchor (>) and it was wrong. Turns out (|) is what I needed. Thanks :)
[21:16] <rharper> johnsonshi: https://paste.ubuntu.com/p/jvRWBkHxbN/    we make use of this  a lot in another project where we write quite a bit of shell in yaml;  this lets me write literal shell under my anchor and then just reference the anchor in a runcmd exec of sh or bash -c
[21:17] <johnsonshi> rharper: Thanks. That syntax is neat and fits our purpose exactly. :)
[21:18] <rharper> cool
[21:25] <Odd_Bloke> johnsonshi: Yep, it's one of those things that once you know you know, but if you don't know what you're looking for then it's super hard to find. :)
[21:26] <Odd_Bloke> https://github.com/canonical/cloud-init/pull/272 is the tests that I'll be modifying as I make my changes for the archive mirror selection bug we've been working
[21:26] <Odd_Bloke> Figured I may as well submit them separately, as they're useful regardless of the rest of the work (and I'm almost EOD).
[22:45] <Goneri> meena, could you take a look at this one? https://github.com/canonical/cloud-init/pull/273