[14:05] <Odd_Bloke> Goneri: I've labelled #298 as wip, give me a shout when you want that undone.
[14:05] <Goneri> Odd_Bloke, sure, thanks :-)
[15:07] <AnhVoMSFT> Gaffel which cloud provider is this? You can also deploy a new VM, attach the OS disk that you used for deploying the Centos VM as a datadisk, and get the logs that way
[15:08] <Gaffel> AnhVoMSFT, libvirt
[15:09] <AnhVoMSFT> Can you attach the disk to another VM and extract the logs that way?
[15:48] <Gaffel> AnhVoMSFT, I got around the problems and I'm able to get it working.
[21:14] <Odd_Bloke> powersj: So I realised that we can't exercise the new URL handling code via user-data, only via cloud.cfg{,.d/*} modification.  AFAICT (via grep), we don't have any tests that currently modify cloud.cfg.  Am I missing something that already exists?
[21:15] <powersj> Odd_Bloke, ah, you are correct we do not have tests that do that today
[21:18] <rharper> Odd_Bloke: well, you can write_files a new cfg entry; and then use runcmd to exec cloud-init with the module you want in question
[21:18] <rharper> Odd_Bloke: the alternative is changes to the image build section where we're building the snapshot image (inject deb, upgrade) and it seems like we'd want to optionally include cloud-cfg
[21:20] <Odd_Bloke> I think the latter would be the path to go; this probably _shouldn't_ be the only test we have that exercises cloud.cfg. :p
[21:21] <rharper> yeah, makes sense
[21:51] <Odd_Bloke> Upon digging a bit further, I think we snapshot the image once for an entire test run, so I don't think we can do this per-test by changing the snapshot phase.
[21:55] <Odd_Bloke> I've captured what I've found in the card and will leave it at that for now.
[22:33] <blackboxsw> rharper: has all of Odd_Bloke's invalid URL handling landed as of https://github.com/canonical/cloud-init/commit/1bbc4908ff7a2be19483811b3b6fee6ebc916235 ?
[22:37] <powersj> that's what it sounded like in standup
[22:38] <powersj> 291 and 296 I think were the #s