[04:17] new (draft) guide up for anyone interested in what happens under the hood when `juju deploy` is executed (https://discourse.jujucharms.com/t/2209). critique very welcome - please add comments to the thread in discourse. [07:16] wallyworld: looks like the snap build didn't work because the built tarball wasn't ready yet? I [07:16] oops [07:16] I'm kicking it off again [17:28] CR anyone, this ensures that we can reuse a bootstrapped controller without having to create one https://github.com/juju/juju/pull/10729 [21:18] babbageclunk: any luck narrowing down the user permission? good to get confirmation that the approach works. we can close that other PR now right? [21:19] wallyworld: I don't think it can be any narrower unfortunately - because the extend task has no target I think it's the top level permission that gets applied. [21:19] yes, I'll close Pedro's PR [21:20] ok. i guess it's only relevant if they use a root disk size constraint and it's out of our hands. were you able to find the api to query to see if that perm has been granted? [21:20] bit of a head-desk moment when we realised it was creating the VMs with 4M of ram. But good to have it solved. [21:21] yeah [21:21] always use units of measure [21:21] There's a govc command that lists permissions, so I'm cribbing from that, [21:22] \o/ [23:53] wallyworld: as discussed yesterday when U get a chance PTAL https://github.com/juju/juju/pull/10730 - renaming option to --client-only and introducing --controller-only [23:53] ok