[03:46] new introduction to the "localhost" cloud provided by LXD https://juju.is/docs/lxd-cloud [08:13] manadart_, it looked better in the error message https://github.com/juju/charm/pull/310#discussion_r448183798 [08:13] manadart_, the problem with juju error messages is they're crap [08:13] stickupkid: OK. [08:14] manadart_, we really should slap localisation on them, and then you can have better language support [08:14] manadart_, have a localisation of "dev" and it just prints the raw error message out, everything else is much nicer === MmikeM is now known as Mmike [09:22] manadart_: reminder for taking a look at 11781 ;-) [09:23] achilleasa: Literally just typing that I've been looking at it and am doing QA. [09:23] manadart_: haha ok. thanks! === rawr is now known as grumble [10:02] anybody know why `juju model-config` isn't isomorphic? [10:06] stickupkid: isomorphic to what? [10:07] well juju model-config --format=yaml outputs a different type to what it consumes if you pass in a yaml file [10:08] so the type isn't isomorphic, which is really weird [10:10] i don't have anything constructive to add, i was just puzzled by that use of 'isomorphic' :) [10:10] (i'm still very much a juju n00b) [10:11] I'm going to ask again, because yesterday I got no answer [10:11] Is it possible to add a new LXC having an interface on a space that its host didn't originally had this space? [10:12] I can add the vlan/IP on the host manually [10:12] and then try to launch an lxc on this host having the new space [10:12] but then I get "matching subnets to zones: space "new-space-name" not found" [10:13] /j #lxd [10:13] er, oops [10:14] yan0s: maybe ask in #lxd? [10:14] (i joined first to double-check that was where it was at) [10:14] THis is a juju question [10:15] ok, i'll shut up then [10:15] I need the container to knwow about the space so I can use it as a binding in juju [10:16] thanks for the suggestion though [10:38] manadart_: AFAICT, juju loses track of the parent nic for container link-layer devices. However, while implementing Networking it seems like we can pull this info (via another API call) for lxd. Is that something we want? [10:46] achilleasa: Probably, but as per https://github.com/juju/juju/blob/ad9586173d624d35d32bf3fc372297cff7bbae11/apiserver/facades/agent/provisioner/provisioner.go#L1069 we want to change how it is put together - create the container, then get the info for filling in. [10:47] achilleasa: This would be great though, because then the provisioner would not need the networking common API and we could nuke it and move SetObserved... to the machiner. [11:28] anybody know how to set resource-tags via model-config? [11:29] ah ignore me [11:33] nice, resource-tags also break the yaml output like cloudinit-userdata [12:47] manadart_: does this look legit? https://pastebin.canonical.com/p/rvZFKRrXBv/ [12:48] manadart_: ah wait; got the wrong network id; let me do another run [12:59] manadart_: updated version: https://pastebin.canonical.com/p/tQcS84Sfry/ [14:46] achilleasa: https://github.com/juju/juju/pull/11785 [14:52] manadart_, https://github.com/juju/juju/pull/11784 [14:52] CR for this, if possible [14:52] as you know a bit about it [14:53] stickupkid: Yep. [14:56] manadart_: looking [15:25] stickupkid: Requested changes. I have to run; we can talk about it tomorrow if you don't agree. [15:25] pretty sure I do [15:25] :) [15:33] stickupkid: achilleasa quick pr review? https://github.com/juju/juju/pull/11787 [15:38] petevg: ^^ [15:42] manadart_, responded [16:02] manadart_: looks good; need to tweak a broken test. Can you take a look at https://github.com/juju/juju/pull/11788 tomorrow? [16:40] CR anybody https://github.com/juju/juju/pull/11786 === tris- is now known as tris [22:48] Have created a (first draft of) a page for adding a Pull Request workflow back into updating our docs https://discourse.juju.is/t/updating-documentation/3300