[01:28] <hpidcock> wallyworld: I added https://bugs.launchpad.net/juju/+bug/1870458 for 2.9-beta1, something we can't forget about but isn't too urgent.
[01:28] <mup> Bug #1870458: extend k8s remote hook context with sh and python <juju:Triaged> <https://launchpad.net/bugs/1870458>
[01:29] <wallyworld> ta
[01:29] <hpidcock> also added the k8s application cleanup bug to 2.8-beta1
[01:29] <wallyworld> excellent
[03:12] <wallyworld> kelvinliu: small one https://github.com/juju/juju/pull/11398
[03:12] <kelvinliu> looking
[03:16] <kelvinliu> nice! wallyworld
[03:16] <wallyworld> it was manadart's excellent suggestion
[03:25] <kelvinliu> initially, k8s controller supposed to only support k8s models but now it's not the case anymore.
[03:32] <wallyworld> kelvinliu: correct. but even for k8s models, there could be another cluster added as a separate cloud
[03:36] <wallyworld> babbageclunk: any chance to look at the add-model bug?
[03:37] <kelvinliu> that's right
[03:37] <babbageclunk> wallyworld: I'll look at it now
[03:38] <wallyworld> babbageclunk: whereever you get to let me know so i can progress it. thanks for helping!
[05:02] <wallyworld> hpidcock: did you raise a bug or create a card for the remote exec path issue?
[05:02] <hpidcock> ahh will do
[05:02] <hpidcock> thanks
[05:02] <wallyworld> ta, add to 2.8-beta1 milestone please
[06:45] <wallyworld> manadart: or stickupkid__ or achilleasa: a vert small PR to fix a long standing, subtle add-model bug https://github.com/juju/juju/pull/11399
[07:38] <manadart> wallyworld: That's approved.
[07:43] <wallyworld> manadart: awesome, ty
[08:43] <stickupkid> manadart, https://github.com/juju/juju/pull/11388 <- it wasn't as easy as I first thought, see
[09:11] <stickupkid> manadart, agreed, I'll fix it up
[09:11] <stickupkid> spent most of my time fixing the client, will go back to the original problem space
[09:53] <wallyworld> achilleasa: did you consider max-unitagent-state-size instead of max-uniter-state-size? unitagent is what end users think about. uniter is of little meaning outside juju dev and is more of an implementation artefact. compare to how we present status - Agent Status, Workload Status etc
[09:53] <wallyworld> max-agent-state-size even
[09:53] <wallyworld> since we want to combine the agents eventually
[09:58] <achilleasa> wallyworld: the uniter-state-size was added for symmetry and is probably something operators never have to worry about or tweak (although you can set it to 0 as an escape hatch if something goes wrong). It is there more as a sanity check for ensuring that the uniter does not suddenly try to persist an odd amount of data
[09:58] <achilleasa> wallyworld: happy to rename if you think it's ambiguous
[09:59] <wallyworld> not some muh ambiguous but not how we expose that particular juju component elsewhere. everywhere else we talk externally about the "unit agent"
[10:00] <wallyworld> the uniter is just what we chose to call the package implementing the main aspect of the unit agent
[10:00] <wallyworld> it has no meaning to users
[10:00] <wallyworld> and since we want to combine unit agent and machine agent to just a single agent, just "agent" makes sense
[10:01] <achilleasa> wallyworld: makes sense. I will open a small PR to rename to max-agent-state-size
[10:02] <wallyworld> thank you :-) sorry for the late input
[10:02] <achilleasa> wallyworld: no prob. Small renames are no biggie as opposed to omg-you-implemented-this-wrong! :D
[10:03] <wallyworld> that i have no opinion on :-) haven't read the code in detail, but i'm sure you did a greta job :-)
[10:03] <wallyworld> greta lol. *great
[10:41] <achilleasa> manadart: or stickupkid small PR to rename the config option for the max uniter state size: https://github.com/juju/juju/pull/11401 ^
[11:22] <stickupkid> manadart, I updated and added some tests
[12:11] <manadart> stickupkid: https://github.com/juju/juju/pull/11402. Was deep in this one, will grab a bite and review yours.
[12:14] <stickupkid> manadart, I had to rebase my changes anyway
[15:12] <achilleasa> hml: any tips for testing the meterstatus worker? IIRC you had something metrics-related in the QA steps for one of your PRs
[15:13] <hml> achilleasa:  i would have to look at the QA steps in that pr.  :-)
[15:13] <achilleasa> hml: asking bec I can't find it :D
[15:13] <hml> achilleasa:  sure, let me look
[15:16] <hml> achilleasa:  i’m thinking it’s broken with the stdout change:   juju collect-metrics keystone/0
[15:16] <hml> failed to collect metrics: could not read stdout
[15:16] <SpecialK|Canon> Hello! If I have Charm source on Launchpad that I'm actively developing, what's the easiest way for me to be able to use it in JaaS / jujucharms.com please?
[15:17] <hml> achilleasa:  here is the pr https://github.com/juju/juju/pull/11318
[15:17] <SpecialK|Canon> Or do I need to get it into the charmstore as a prerequisite?
[15:18] <hml> achilleasa:  got out output from metrics —all like i should have
[15:18] <hml> achilleasa:  i’m still… need to deploy a specific charm first
[15:22] <achilleasa> hml: https://paste.ubuntu.com/p/xJC9FnmjkX/ seems like it's working :D
[15:22] <achilleasa> now I need to test what happens when you move from 2.6 -> 2.7
[15:23] <achilleasa> er... 2.7 yo 2.8
[15:23] <hml> achilleasa:  i’d leave it for a bit… gather data then due the upgrade
[15:26] <achilleasa> hml: seems like it fires immediately after the hook comes online. So I will bootstrap 2.7, verify the state file is there and then upgrade to verify the state got migrated
[15:26] <hml> achilleasa:  rgr
[15:48] <manadart> achilleasa, hml, stickupkid: I messed the LXD thing up: Here's the fix: https://github.com/juju/juju/pull/11404
[15:56] <manadart> If anyone approves it, please go ahead and merge.
[16:12] <stickupkid> manadart, nps, bridge vs bridged :)
[18:01] <achilleasa> hml: the meter-status state PR is ready for review; please have a look and I will address any feedback on Monday: https://github.com/juju/juju/pull/11405
[18:02] <hml> achilleasa:  sure.  might need a break from the UT in relations.  ha!
[18:03] <achilleasa> I also have a separate PR for renaming SetStateArg.State -> SetStateArg.CharmState (and everywhere we use this; also a bit of naming cleanup in the uniter context to avoid confusion as to what "state" we are caching). I will push this on Monday though
[18:03] <hml> achilleasa:  ack
[18:03]  * achilleasa waves