[01:34] <pmatulis> i remember asking about that a year ago. iirc, they are provider-dependent
[01:57] <timClicks> pmatulis: ty
[01:57] <timClicks> that's the conclusion I have come to
[02:28] <wallyworld> thumper: https://github.com/juju/juju/pull/11703
[02:29]  * thumper headdesks at 2.8 branch needing apiserver to test state package
[02:29]  * thumper looks at pr
[02:29] <thumper> igmores
[02:30] <thumper> 15 files?
[02:30] <wallyworld> small changes
[02:32]  * thumper nods
[02:35] <thumper> ah FFS
[02:35]  * thumper wonders about 2.7 all watcher bits
[02:36] <thumper> wallyworld: this isn't going to work
[02:36] <thumper> did you want to jump in a HO to discuss
[02:36] <wallyworld> sure
[02:36] <thumper> 1+1
[03:31] <wallyworld> thumper: tweaked the cache lookup, just need to test with libjuju
[03:32] <wallyworld> and need to fix one unit test
[04:12] <thumper> wallyworld: is it ready for a rereview?
[04:13] <wallyworld> thumper: HO?
[04:27] <thumper> sure
[04:48] <thumper> hpidcock: did your new multichecker land in 2.7?
[04:48] <thumper> hpidcock: I have a perfect usecase for it
[04:48] <hpidcock> thumper: I don't think so
[04:49] <thumper> hpidcock: boo
[04:49] <hpidcock> Feel free to update the juju/testing dep
[04:50] <thumper> hpidcock: hmm...
[04:50] <thumper> hpidcock: may well do
[04:50] <thumper> hangon
[04:50] <thumper> I'm in 2.8
[04:50] <thumper> yay
[04:50] <thumper> hpidcock: what is it called again
[04:51] <hpidcock> jc.NewMultiChecker()
[04:51] <hpidcock> there is an example of its use somewhere in 2.8
[04:52] <thumper> found it
[04:52] <thumper> now I need to work out how to get it to ignore the right field
[04:53] <thumper> hpidcock: got 2 minutes to help out?
[04:55] <thumper> I'm not sure how to map this thing through
[04:55] <hpidcock> sure
[04:55] <hpidcock> standup?
[04:56] <hpidcock> thumper:^
[04:57] <thumper> hpidcock: yeah
[05:03] <wallyworld> thumper: you good with the PR then? i wan tto land it so i can forward port
[05:28] <thumper> wallyworld: sorry, will do another quick pass over
[05:28] <wallyworld> all good
[05:32] <thumper> and you push while I'm looking?
[05:32] <wallyworld> thumper: sorry, there was a go fmt thing
[05:33] <wallyworld> an extra ' '
[05:34] <thumper> approved with a questino
[05:34]  * thumper is done and heading off now
[05:34] <thumper> leaving with 15 failing tests in state package
[05:34] <thumper> get to them next week
[05:34] <thumper> later peeps
[09:05] <achilleasa> manadart: to fix the lingering port issue with ovs we could override the port name via the 'veth.pair' option to a known (based on the container id or similar) value so we can remove it when we destroy the container
[09:05] <manadart> achilleasa: Yep.
[09:46] <stickupkid> yay, managed to replicate the LXD bug, finally
[10:20] <achilleasa> now, we have deterministic iface naming for lxd veths ;-) https://pastebin.canonical.com/p/ysbqGHWJVV/
[10:21] <achilleasa> manadart: stickupkid comments/concerns about format? ^^ (it's "machID-ifaceIdx" where machID has / converted to _ to distinguish from last dash)
[10:22] <stickupkid> achilleasa, i'm never sure if you even need _, but otherwise looks ok to me
[10:23] <stickupkid> achilleasa, i.e. 0lxd2-0
[10:23] <stickupkid> achilleasa, lxd is the separator
[10:23] <stickupkid> achilleasa, like doing `1_,_2_,_3` for csv ;-)
[10:24] <manadart> achilleasa: Agree with stickupkid. This is cool BTW.
[10:25] <manadart> achilleasa: This will apply to all LXD veths created for Juju right? not just connected to OVS?
[10:29] <achilleasa> stickupkid: makes sense; I will just trim it off. manadart yes; I wanted to put a juju prefix but there is a 16 char limit so I tried to keep it shorter
[10:31] <manadart> achilleasa: I love that "ip a" will now give a less opaque picture of what's going on.
[10:31] <manadart> achilleasa: And in the state link-layer data for that matter.
[15:20] <hml> stickupkid: pr review changes made to 11701.  i’m finalizing k8s testing now
[15:41] <stickupkid> hml, been looking
[15:41] <hml> stickupkid: testing my k8s myself
[15:47] <stickupkid> hml, approved again
[15:48] <hml> stickupkid:  cool, ty
[15:49] <hml> stickupkid: ffs: https://pastebin.canonical.com/p/J3Tx9pF6Wc/. will have to fix this
[15:52] <hml> stickupkid: and k8s is now failing to install *sigh
[15:53] <stickupkid> hml, weird
[15:54] <hml>  stickupkid actually no… the LXDProfileRequire doesn’t check machine for KVM or manual etc…. only if the charm has a profile in it
[15:55] <hml> *facepalm*
[15:55] <hml> not sure on the k8s change yet
[16:23] <stickupkid> hml, if you've got 5 minutes before my EOD around https://github.com/juju/juju/pull/11700
[16:24] <hml> stickupkid: had a question for you on there… forgot to click send.  ho?