[00:02] wallyworld: plz review the fix for a bug I realised I'd left in the relation cleanup (occurred to me while working on application cleanup) https://github.com/juju/juju/pull/10623 [00:02] will do, just finishing anoher [00:03] thanks! [00:09] hpidcock: jump in standup a bit early? [00:14] Sorry was getting coffee [00:14] 1 sec [00:34] hpidcock: lgtm [00:34] wallyworld: I'll make those changes and get it landed [01:39] if anyone has a few minutes (hopefully less), I would appreciate it if you could deploy and road test the new hello-juju charm [01:39] juju deploy hello-juju [02:01] timClicks: you gonna do a k8s version :-) [02:02] wallyworld: when microk8s is working again [02:02] works now [02:02] oh cool [02:02] just need to add yourself to the microk8s group [02:08] timClicks: https://discuss.kubernetes.io/t/explicit-use-of-sudo-in-microk8s-cli/7605 [02:08] juju won't work without that step [02:09] so we need to document that - and/or have the CLI prompt to do it when juju and microk8s arte used together [04:36] kelvinliu: here's a fix for the attach resource issu in 2.6 https://github.com/juju/juju/pull/10624 [04:56] wallyworld: sorry meant to mention that i've updated the microk8s-cloud docs [04:56] wallyworld: https://jaas.ai/docs/microk8s-cloud [04:58] timClicks: you'll want to mentioned the newgrp command otherwise those steps won't work without a logout and login [05:00] wallyworld: will do (just added the instruction from the link above) [05:08] wallyworld: just back from lunch, looking now [05:12] no worries [05:13] ty [05:14] wallyworld: so the url querystring includes fingerprint? [05:16] it in the header [05:18] i saw it's in query url.Values [05:19] yeah, that sounds right. i didn't change any of that [05:20] ok, lgtm, thanks [05:20] ty === exsdev0 is now known as exsdev [08:24] jam, any thoughts on this, before I continue on? https://github.com/juju/description/pull/62#issuecomment-530721114 [08:27] jam, it's still a WIP, but good feedback early... [12:32] manadart: can you take a look at https://github.com/juju/juju/pull/10625? [12:34] achilleasa: Yep. === exsdev0 is now known as exsdev [18:58] Anyone know if spaces are supported on the AWS provider? [18:59] thedac: you can manually create them in juju… [19:00] juju will act according to the user created spaces, but aws won’t have explicit knowledge of them. [19:00] interesting. OK [19:02] thedac: only maas and aws support spaces. [19:02] ack [20:47] thedac: what are you thinking of poking at? [21:11] rick_h: we were looking at https://bugs.launchpad.net/vault-charm/+bug/1843809 and I wanted to rule out AWS space support as the problem. They are related but it is not juju's fault ;) [21:11] Bug #1843809: vault-kv relation is broken when not using network spaces [21:21] babbageclunk: can you +1 this merge from 2.6, with the relation cleanup fixes plus a few others? https://github.com/juju/juju/pull/10627 [21:21] wallyworld: yup yup [21:23] wallyworld: weird - there's been a typo introduced in environs/config/config.go - ContainerInheritPropertiesKey -> ContainerInheritProperiesKey [21:24] not sure which pr it was from [21:24] oh jeez, ok. there was a conflict in the merge [21:24] i chose what looked like the right one [21:24] good pickup [21:25] I mean, it might be right depending on what else is in the code [21:25] (but we should probably still fix it) [21:26] babbageclunk: i think the typo was fixed in develop [21:26] as a driveby [21:26] oh right [21:26] yeah that would do it [21:26] i' [21:26] i'll just fix the typo in my PR [21:27] as long as it still compiles I guess! [21:27] yeah, it doe snow [21:27] it certainly do. But not in Brisbane. [21:27] was in the middle of doing that while you were revuewing [21:27] lol [21:30] approved [21:30] wallyworld: ^ === exsdev0 is now known as exsdev [21:57] babbageclunk: huh, the forward port has lint errors that are not picked up in 2.6 [21:57] maybe the develop linter has more stringent settings? [21:58] yup [22:00] wallyworld: https://github.com/juju/juju/pull/10628 The fix for remove-application --force after a stuck remove-application [22:00] ok, looking after i fix the lint issues [22:00] Sorry, got distracted by the raft stuff at the end of the day yesterday [22:01] thanks [22:01] babbageclunk: one is on this line, scope, role, unitName, err := unpackScopeKey(doc.Key) [22:01] i think we can just return err there right [22:01] ? looking [22:02] oh right - yup, just forgot to check the error [22:02] no worries, fixing [22:02] thanks ineffassign! [22:06] babbageclunk: i think the query should be alive or dying to exclude dead units [22:06] +1 with that change [22:07] once this lands we should ask for this to be included in 2.6.9 testing [22:07] I'm not sure about that - if the machine agent is down then the unit could be dead but it'll never be removed, so the application will still be stuck [22:08] I'm going to try it now [22:09] wallyworld: ^ [22:12] babbageclunk: ok [22:13] if i'm wrong, that's fine, just land [22:18] wallyworld: ok - worth checking though [22:45] wallyworld: yeah, that's what happens - I've added a comment to clarify. [22:45] ty [22:53] timClicks: you going to land your older PRs? [22:54] wallyworld: keen to, but wasn't/am not sure what the best window was/is [22:55] 4 weeks ago :-) [22:55] the azure one is targetted at 2.6 but i think we can now target to develop [22:55] and just land anytime [22:57] there's now a juju call command for juju v3 which might need mproved help text in addition to run-action (which is now deprecated) [22:57] maybe that PR coud also look at juju call [23:01] babbageclunk: my PR to forward port 2.6 had an intermittent error so will pull in your change before trying again [23:02] wallyworld: ah, right [23:34] wallyworld: remind me of the trick for bootstrapping on openstack when there's no metadata?