[01:47] kelvinliu_: or babbageclunk: no rush, a small Python PR https://github.com/juju/charm-helpers/pull/385 [01:49] wallyworld: lgtm [01:50] yay, ty [01:50] lgtm as well [02:27] hi babbageclunk can i get ur a few mins to help me to understand about raftlease plz? [02:28] kelvinliu_: sure - in standup? [02:28] babbageclunk: yup thx [03:19] thumper: not urgent, here's a PR which uses "function" terminology with the v3 feature flag for the actions CLI https://github.com/juju/juju/pull/10722 [03:19] so now we have functions and "call" and tasks [03:19] at least on the CLI [03:20] charms still have actions.yaml etc [03:25] wallyworld: do you have the bug for the caas peer relation? [03:26] thumper: https://bugs.launchpad.net/juju/+bug/1818230 [03:26] ta [03:26] Bug #1818230: k8s charm fails to access peer relation in peer relation-joined hook [03:39] kelvinliu_: how did you and babbageclunk get on with the raft/clock thing? [03:40] wallyworld: we r in standup. mind join us? [03:42] sure [04:41] kelvinliu: https://github.com/juju/juju/pull/10723 [04:43] wallyworld: lgtm thx! [04:43] ty [04:45] babbageclunk: hmmmm, maybe my X1 Extreme does seem a little quieter now.... [04:45] lol [04:45] even with magnificant Goland running [04:52] babbageclunk's sounded like it was taking off at the sprint [04:52] for other reasons [04:57] bloody gnome. Actually, that seems to have been better lately - maybe they fixed the bug [04:58] did you have it when using i3? [05:06] no, but there were too many other things I couldn't do with i3 so I've switched back [05:07] :O [05:11] wallyworld: should we error if the crd scope was cluster scope or just always overwrite to Namespaced peacefully? === nammn_de_ is now known as nammn_de [10:14] Anyone able to review https://github.com/juju/juju/pull/10684 ? [10:18] manadart: I will trade you for https://github.com/juju/juju/pull/10725 [10:18] achilleasa: OK. [10:43] achilleasa: My patch landed that changes the uniter around NetworksForRelation. This conflicts with your patch. Can you pull it down and fix? [10:44] manadart: sure thing. I will rebase and force-push [10:59] manadart: ready [10:59] achilleasa: I see it. Ta. [11:01] rick_h: if you around, can we HO before/after daily? Have Some small questions regarding caas and the pr we were talking before setting adm [11:35] achilleasa: Reviewed. [11:41] manadart: should the settings block return an error for nil or maybe simply skip over nil entries? [11:42] achilleasa: I say return. If [0] is nil the write() call could panic. [11:42] nammn_de: morning, sure thing [11:42] nammn_de: meet you in daily? [11:42] manadart: that's a valid point! [11:47] rick_h: was having lunch, heading over daily [13:30] does `leader_set` stringify all values passed to it? ie: if I pass it a boolean, will I get back a string? [13:32] rick_h: maybe you know? ^ [13:33] icey: I'd expect so as it's a simple key/value vs types [13:33] :-/ [13:33] icey: since it's just bash data not sure what the python library is doing for type handling across the wire there [13:34] e.g. in a bash hook/etc it's just key=value [13:34] rick_h: apparently, stringing it up (https://github.com/juju/charm-helpers/blob/669821489497a547a768f686a2fadf88d2d5f2b2/charmhelpers/core/hookenv.py#L1121) :-/ [13:35] icey: yea, that's what I expected [14:16] rick_h manadart hml regarding trying to upgrade a caas controller with juju. I tried to with " juju upgrade --build-agent" returns --build-agent is not supported for k8s model. Any idea how i can test those upgrades for caas else? [14:17] nammn_de: actually not sure tbh [14:17] they work through the operator pod but I'm not familiar with the dev scenario around using it [14:18] nammn_de: and a heads up, if you want to ping the team you can use the guild highlight nick [14:18] ahhh now thats good to know [14:18] probably need to sync with hpidcock wallyworld or kelvin once about caas [14:19] nammn_de: yea, but make sure to spread the word as honestly I *should* know :( [14:20] rick_h: will do :D [14:20] nammn_de: I *think* you would need to do something like "DOCKER_USERNAME= make microk8s-operator-update" [14:20] Then just "juju upgrade-controller" [14:21] manadart: tahnks gonna try that out! Do we have that somewhere documented in case in run into bad things? [14:23] nammn_de: According to https://discourse.jujucharms.com/t/whats-new-in-juju-k8s-for-2-6/1431 it's just upgrade-controller, but that will assume that there is a new official operator image on dockerhub. [14:24] So I am guessing for local wrangling you need to have built the new image and somehow let juju know where to get it from. [14:24] manadart: right, that's the trick. Doing that with "my own operator" is the part I've not tried at all [14:43] manadart: got it thanks [14:46] achilleasa: I pushed that change to my upgrade patch too. [14:47] manadart: looking [14:52] manadart: much cleaner now! thanks for the change. Doing QA [15:23] manadart: the machineaddresses collection has no spaceid anymore (addresses looks as expected). That's what I should be seeing right? [15:25] achilleasa: Machine addresses has to-date never had space information (it only comes from the provider into addresses), so that is correct. [15:25] manadart: PR approved [15:25] achilleasa: Thanks. [16:12] manadart: apparently accessing the unit from within Flush breaks a whole lot of uniter tests (in a different package).... :-( [16:20] hml: got a few min to help me with the uniter test mess? [16:20] achilleasa: sure [16:20] hml: daily? [16:21] achilleasa: omw [16:44] hml: removing the hardcoded stuff fixed the test... [16:44] achilleasa: hahahahaha awesome [16:45] I will replace the other occurrences as well [17:01] hml: hopefully CI will be happy with my changes. I will wait for Joe to take a look before I land it so my PR is expected to land on Monday morn. Is that OK for your rebase work? [17:02] achilleasa: that should be okay. [18:41] when i use option `--metadata-source` with the 'bootstrap' command i don't see anywhere that this value is exposed. should it not show up in the output to 'show-controller' command?