[01:46] <timClicks> I've just published a beta-quality plugin: juju-remove https://discourse.jujucharms.com/t/new-plugin-juju-remove/2318
[01:46] <timClicks> hopefully this will prevent some papercuts
[01:53] <wallyworld> timClicks: curious, was that in response to user feedback?
[01:54] <wallyworld> hpidcock: for sometime today, a small 2.7 fix https://github.com/juju/juju/pull/10863
[01:54] <hpidcock> happy to do it now, just got myself a coffee
[01:54] <timClicks> wallyworld: no, was definitely in the scratch own itch
[01:54] <timClicks> category
[01:55] <timClicks> I kept tripping over it getting annoyed with juju that there wasn't a single remove command
[01:56] <timClicks> I also wanted an excuse to play around with pylibjuju
[01:56] <wallyworld> one reason is that remove-machine and remove-relation both accept integers
[01:56] <wallyworld> so which one is intended when "2" is supplied
[01:57] <wallyworld> but yeah, a lot of times you know what you want to remove
[01:57] <timClicks> most of the time that's not a problem
[01:57] <wallyworld> except when it is :-)
[01:57] <timClicks> and when it is, the command asks
[01:57] <wallyworld> so if i type juju remove 4
[01:58] <wallyworld> how does it know relation 4 or machine 4
[01:58] <timClicks> right now, it'll treat it as a machine
[01:58] <timClicks> but there is some logic that I haven't committed to extract all named entities and try to find duplicates
[01:59] <wallyworld> might be worth adding 0/lxc/3 to the examples
[01:59] <timClicks> patch welcome
[01:59] <timClicks> there's a todo in the code
[01:59] <timClicks> i decided to pull back from implementing everything and seeing if people like it
[02:07] <anastasiamac> wallyworld: PTAL https://github.com/juju/juju/pull/10864
[02:16] <wallyworld> anastasiamac: i think warnings are ok for list/show. explanation in the PR
[02:16] <anastasiamac> wallyworld: k but i disagree with ur other stmts ... do u have a min?
[02:17] <wallyworld> sure
[02:17] <anastasiamac> stdup?
[02:17] <wallyworld> ok
[02:34] <hpidcock> wallyworld: https://github.com/juju/juju/pull/10863 LGTM
[02:35] <wallyworld> looking
[02:35] <wallyworld> oh
[02:35] <wallyworld> thanks
[02:35] <hpidcock> yeah can you PR your own PR plz :P
[03:59] <wallyworld> anastasiamac: here's some cherry picks from develop into 2.7 rc2 https://github.com/juju/juju/pull/10865
[04:49] <hpidcock> wallyworld: no rush https://github.com/juju/version/pull/6 but need this for CAAS upgrade code (so we can match version-hash tag in dockerhub)
[04:50] <wallyworld> looking
[04:54] <kelvinliu_> wallyworld: this pr is for adding content interface consumer side to Juju, thanks! https://github.com/juju/juju/pull/10857, also updated microk8s side.
[05:04] <wallyworld> kelvinliu_: gr8, looking, just need coffee real quick
[05:10] <wallyworld> kelvinliu_: do we need the install hook anymore? isn't having the connect-peers script sufficient?
[05:15] <kelvinliu_> wallyworld: just went to get some coffee. im not sure if the connect gets created automatically if microk8s was installed first, then Juju.
[05:15] <kelvinliu_> coz I tested by manually creating the connection
[05:17] <kelvinliu_> as we don't pull if the images are there, it might be ok even if the two hooks are all fired,
[05:18] <wallyworld> kelvinliu_: the way i thought the content interface worked is that it would always trigger regardless of install order. i do think the install hookis now redundant
[05:18] <wallyworld> can you do a quick test without the install hook to confirm?
[05:19] <kelvinliu_> not sure how to test this.
[05:19] <wallyworld> delete both. install in different orders. check the hook logs and the microk8s registry
[05:20] <kelvinliu_> snap connect juju:peers microk8s:juju-info
[05:20] <kelvinliu_> no, because we manually connect them
[05:21] <wallyworld> what is "we"
[05:21] <wallyworld> the connection should be done by the snapd infrastructure automatically
[05:21] <kelvinliu_> we -> tester
[05:21] <wallyworld> can we have a HO?
[05:21] <kelvinliu_> hangout?
[05:43] <kelvinliu_> wallyworld: no.. i have to rename the snap and rebuild them, coz the snap name has to be uniq. no namespace concept here.
[05:44] <wallyworld> yup, flat namespace
[06:06] <kelvinliu_> wallyworld: https://forum.snapcraft.io/t/approval-for-fwupd-classic-snap/5755/11
[06:07] <kelvinliu_> wallyworld: confined snap can't use plug/slots..
[06:08] <wallyworld> oh
[06:08] <kelvinliu_> snapstore complains   - confinement 'classic' not allowed with plugs/slots when i was trying to push the snap..
[06:09] <wallyworld> fark, ok. i'll email zygmund
[06:10] <kelvinliu_> good to catch this now or next release will be interrupted..
[06:10] <kelvinliu_> well, next edge push, probably not release.
[06:15] <wallyworld> yeah, i've emailed him so we'll see what he says
[06:16] <kelvinliu_> wallyworld: I will keep those two PRs  not merged for now
[06:17] <wallyworld> yup, ty
[06:21] <kelvinliu_> np
[06:50] <wallyworld> kelvinliu_: what exact error did you get from the store? can you paste me the command used and error message?
[06:53] <kelvinliu_> wallyworld: https://pastebin.ubuntu.com/p/Tv4ZdqcgzZ/
[06:54] <kelvinliu_> wallyworld: here is the error return from snapstore
[06:56] <wallyworld> ty
[07:02] <kelvinliu_> np
[10:34] <stickupkid> achilleasa, https://github.com/juju/juju/pull/10862/commits/176c41e5f73709778f2802e4afefb89b0a41c1b0
[10:39] <achilleasa> stickupkid: nice!
[10:42] <stickupkid> achilleasa, review done
[10:46] <achilleasa> stickupkid: regarding the ec2 prefix, doesn't main.sh include all files? or does each test-suite run in a sub-shell and we can't have name clashes for functions?
[10:46] <stickupkid> achilleasa, it only includes files in "includes" everything else is included as required
[10:47] <stickupkid> achilleasa, i'd need to check
[10:47] <achilleasa> I wasn't 100% sure hence the prefix
[10:47] <stickupkid> achilleasa, it only includes "includes" and the test suite in question
[10:48] <stickupkid> achilleasa, if you run the whole suite it will cause conflict
[10:48] <stickupkid> achilleasa, nobody is going to run the whole test suite tbh, maybe we should block running the whole test suite
[10:49] <stickupkid> achilleasa, it would really take a LONG time, also I'm sure it would overwrite old functions with new ones, so should be fine
[10:49] <achilleasa> is it one CI job / suite the pattern we are currently going for?
[10:50] <stickupkid> achilleasa, one CI job per test suite
[10:50] <stickupkid> otherwise it would take way too long
[10:50] <stickupkid> esp. if we're going to try and have 100s
[10:51] <achilleasa> cool. just checking bec the spaces one relies on the presence of a floating nic for each test :D
[10:51] <stickupkid> achilleasa,  you will have to add a jenkins job though
[10:51] <stickupkid> achilleasa, i.e. new suite == jenkins job
[10:52] <achilleasa> stickupkid: yeap; I will leave that for next CI day ;-)
[11:06] <nammn_de> manadart: im currently looking into adding the "list-commits" command. Currently our mongo model does not have something similiar, right?  Like a inc. id number for each commit.  Following the doc I would add one, wdyt?
[11:06] <nammn_de> manadart: ohhh it could be that we have one `_id` : <longchars: <number:>
[11:08] <manadart> nammn_de: You want "GenerationId"
[11:08] <nammn_de> manadart: but afaict the generationid is not unique, right? It "only" tells me whether and how much it has been commited
[11:08] <nammn_de> the doc spec tells me somehing like: juju show-commit 3
[11:09] <nammn_de> is expected
[11:09] <manadart> nammn_de: It is unique, monotonically increasing.
[11:11] <nammn_de> manadart: maybe we are talking about something different. Thats my output from mongo https://pastebin.canonical.com/p/s44W68zccy/
[11:12] <manadart> nammn_de: It will be 0 for any branches that are in flight, or that were aborted. It starts at 1.
[11:12] <nammn_de> manadart: time for a quick ho just to make sure?
[11:13] <manadart> nammn_de: In daily.
[13:17] <hml> stickupkid: related to pr 10862, are we going to require if you  use -l that the controller is the active one?
[13:26] <stickupkid> hml, can you show me the output of juju show-controller --format=json
[13:27] <hml> stickupkid: https://pastebin.canonical.com/p/QvZBNZwyqs/
[13:30] <stickupkid> hml, we don't offically support localmaas, or anything but lxd and aws
[13:30] <hml> stickupkid: right, but i’m not trying to use that controller for the tests
[13:31] <stickupkid> the show-controller thinks you are
[13:31] <stickupkid>  juju show-controller <local-controller> --format=json
[13:31] <hml> stickupkid: but i used the -l flag in the cli cmd
[13:32] <stickupkid> can you pastebin that with the local-controller name?
[13:32] <hml> stickupkid: it’s in the first pastbin
[13:32] <hml> pastebin
[13:32] <stickupkid> ah - damn it "localhost"
[13:32] <stickupkid> :(
[13:46] <stickupkid> hml, https://github.com/juju/juju/pull/10862/commits/6e63b5b60dd08fd94294261967c58939d12107be
[13:56] <hml> stickupkid: will look after the mtg
[13:56] <stickupkid> hml, ta
[14:34] <rick_h> stickupkid:  did that 2.7 edge building work out?
[14:35] <stickupkid> rick_h, it would seem so https://launchpad.net/~juju-qa-bot/+snap/2.7-edge
[14:35] <rick_h> stickupkid:  cool yea looking at the builds. I thought manadart backported this morning but maybe it's not hit yet
[14:36] <stickupkid> rick_h, let me have a look
[14:36] <rick_h> stickupkid:  all good, it  hit an hour ago but these build every 4 I think?
[14:36] <stickupkid> yeah it is
[14:36] <rick_h> stickupkid:  cool cool, just making sure since we're landing stuff like crazy on there :)
[14:45] <stickupkid> manadart, quick CR for a backport https://github.com/juju/juju/pull/10869
[14:48] <manadart> stickupkid: Yep.
[14:48] <stickupkid> good man
[14:54] <manadart> stickupkid: Ticked like an old cattle dog.
[14:57] <stickupkid> what's up with the windows unit tests
[14:58] <stickupkid> i think we're smashing the windows box to obilvian
[16:00] <manadart> hml: Minor mechanical one: https://github.com/juju/juju/pull/10870
[16:40] <hml> manadart:  trade you: https://github.com/juju/juju/pull/10872  it’s a draft until i get a full set of unit tests run.
[16:40] <hml> manadart: was otp
[16:47] <hml> achilleasa: do you have a few to qa and review ^^ also
[16:50] <achilleasa> hml: sure
[16:52] <stickupkid> rick_h, i'm making a model migration as my first step... I'll add that to trello
[16:54] <rick_h> stickupkid:  sounds like a party
[16:55] <stickupkid> rick_h, it should help me prevent it from breaking whilst i make changes :D
[16:55] <rick_h> stickupkid:  hah sounds good
[17:26] <hml> achilleasa: manadart: looking to see if i can untangle some of the interfaces/mocks/shims between the containerizer and apiserver/provisioner with this.
[17:42] <stickupkid> achilleasa, let isn't available in most shell environments :(
[17:42] <stickupkid> achilleasa, i've updated it to be compatible
[17:43] <achilleasa> stickupkid: really? I have been using let like forever :D... maybe we should just use bash everywhere
[17:43] <stickupkid> achilleasa, i'm adverse to it atm
[17:43] <stickupkid> achilleasa, Constraints Liberate, Liberties Constrain
[17:44] <achilleasa> one shell to rule them all? :
[17:44] <achilleasa> :D
[17:49] <pmatulis> is "workload status" a thing in the juju universe? if so, what does it mean?
[18:05] <stickupkid> anyone want to CR an integration test around model migrations https://github.com/juju/juju/pull/10873
[18:05] <stickupkid> ?
[18:06] <stickupkid> i'm still impressed that we can bootstrap 2 controllers, deploy two charms and run a model migration, with a ton of introspection in under 5 minutes
[18:08] <hml> stickupkid: if i get a chance i’ll take a look, have a queue of reviews going :-)
[20:18] <tvansteenburgh> rick_h: Do you know when --series=focal will be available in juju?
[20:24] <N3tw0rK> what can I do if I have a host stuck in waiting/allocating, the node is booted and waiting so im assuming the agent wasnt installed. Is there a way to force the agent to retry?
[20:38] <rick_h> tvansteenburgh:  mid-cycle
[20:39] <rick_h> tvansteenburgh:  we've agreed to track the next series and have it available by mid-cycle
[20:39] <rick_h> N3tw0rK:  hmmm, you can retry-provisioning but might have to check the cloud-init logs to see what wasn't happy
[20:39] <tvansteenburgh> thanks rick_h
[20:57] <kelvinliu_> wallyworld: a tiny intermittent test failure fix, would be great to take a look when u got time, thanks ! https://github.com/juju/juju/pull/10866
[23:46] <wallyworld> anastasiamac: this PR backports the doc link fixes to 2.7 https://github.com/juju/juju/pull/10875
[23:46] <anastasiamac> wallyworld: k
[23:51] <wallyworld> anastasiamac: i didn't want to solve the world for the 2.7 PR as we are past time to finalise that branch; just a straight backport so that we don't ship with stale urls at least. and we can solve it properly or differently in 2.8
[23:51] <anastasiamac> wallyworld: yep, hence +!
[23:52] <anastasiamac> +1 even :D
[23:52] <wallyworld> ! works for me
[23:52] <anastasiamac> :)