[03:19] <babbageclunk> wallyworld: quick review? Once that's landed I'll tag another juju-restore release. https://github.com/juju/juju-restore/pull/16
[03:20] <wallyworld> sure, just otp will try and do concurrently
[03:28] <wallyworld> babbageclunk: lgtm, ty
[03:28] <babbageclunk> wallyworld: nice one
[04:19] <babbageclunk> hey, anyone have any idea why merge commits in juju-restore are full of \rs? (I guess it's a setup problem with the bot?)
[04:19] <babbageclunk> https://github.com/juju/juju-restore/commit/e3f234891e55368ac906959c775cf25563f6b754
[04:20] <babbageclunk> I thought it might be because of the pull request template but it doesn't seem to have them
[04:28] <hpidcock> "pr_descr=$(echo ${ghprbPullLongDescription} | sed $'s/\\r\\n/ \\\\\\n/g')"
[04:31] <hpidcock> babbageclunk: https://api.github.com/repos/juju/juju-restore/pulls/14
[04:31] <hpidcock> should have worked
[04:34] <babbageclunk> :/
[04:34] <babbageclunk> thanks hpidcock - I'll check against the juju one
[04:35] <babbageclunk> love that quoting
[04:47] <babbageclunk> hpidcock: what does that $' syntax mean? It's totally ungooglable
[04:50] <babbageclunk> oh no, it's googlable as "bash dollar quote"
[04:52] <babbageclunk> weirdly, the line from the juju/juju bot uses $"
[04:52] <babbageclunk> pr_descr=$(echo ${ghprbPullLongDescription} | sed $"s/\\\r//g")
[04:54] <babbageclunk> from what I can find $" doesn't actually do anything, assuming the locale's C
[05:23] <babbageclunk> man, that is some super-weird shell-quoting madness.
[05:25] <hpidcock> shell fun
[08:11] <manadart> stickupkid: Any special reason to go from `//go:generate mockgen...` to `//go:generate go run github.com/golang/mock/mockgen...`?
[11:41] <hpidcock> manadart: it will use the go.mod versioned mockgen
[11:41] <manadart> hpidcock: Yeah, figured that since.
[11:41] <hpidcock> nice
[11:41] <hpidcock> tools.go in the root dir pins it as a dependency
[12:54] <achilleasa> Did we rename other pre-existing go:generate bits?
[13:34] <hpidcock> achilleasa: just mockgen at the moment
[14:56] <stickupkid_> hml, manadart petevg, I was working on this yesterday and is required for testing focal for 2.8, can somebody shepherd it through? - https://github.com/juju/juju/pull/11497
[14:57] <hml> stickupkid_: sure.
[14:57] <stickupkid_> ta
[14:57] <hml> stickupkid_: what’s left?  review and qa?
[14:57] <stickupkid_> test
[14:57] <stickupkid_> haha
[14:57] <hml> stickupkid_: rgr
[14:57] <hml> :-)
[14:57] <stickupkid_> was meant to do that today, but never got around to it
[14:57] <petevg> thx, hml!
[14:57] <hml> stickupkid_: no worries
[14:58] <stickupkid_> hml, thanks, later \o
[18:29] <flxfoo> hi all,
[18:31] <flxfoo> What does it costs to boorstrap a new controller (with custom image ID, simplestream) and migrate the a current model to new controller, especially in turns of availability (we are talking about a dozen of containers) thanks
[18:33] <rick_h> flxfoo:  so when you migrate the model the actual units running don't go anywhere. It's just the db that tracks state that's moving.
[18:34] <rick_h> flxfoo:  so the cost would be a new controller node (t3 medium ish sized?)
[18:36] <flxfoo> rick_h: thanks, so no service downtime
[18:37] <rick_h> flxfoo:  no, just the downtime during transition in controller/changing the model
[18:37] <flxfoo> rick_h: sorry, can you elaborate that?
[18:42] <rick_h> flxfoo:  so while the model is migrating, there's safety checks going on to make sure all the units running can reach the new controller, that the db is copied over correctly, etc.
[18:43] <rick_h> flxfoo:  so while that's taking place you can't juju deploy new things into that mode, or make config changes in that model, etc
[18:43] <flxfoo> rick_h: ok ok thanks for the details.... :)
[18:45] <rick_h> np