[03:54] <hpidcock> kelvinliu: babbageclunk: https://github.com/juju/juju/pull/10446 I think it's finally ready for review.
[03:55] <hpidcock> ended up trimming a lot that I was doing out that was not needed.
[03:56] <timClicks_> 90% sure that I've re-enabled resource constraints (root-risk, num cpus) on vsphere
[03:56] <babbageclunk> hpidcock: I've got a call at 4 but I'll look after that
[03:57] <hpidcock> no rush :)
[04:01] <kelvinliu> hpidcock: yep looking
[04:45] <timClicks> babbageclunk: hey we have dozens of models on vsphere/braixen - should I clean them up?
[04:46] <babbageclunk> timClicks: can you see who created them in the UI?
[04:46] <babbageclunk> definitely clean them up unless they were created by k8s-test
[04:47] <timClicks> no I don't think it's possible to see the creator
[04:47] <timClicks> perhaps I'll try to find the age
[04:47] <timClicks> I think that we have a CI job that's not cleaning up after itself
[04:47] <babbageclunk> that's really weird - there's no concept of an owner?
[04:47] <babbageclunk> no, think it's just accretion of testing
[04:48] <babbageclunk> (rather than anything automated)
[04:49] <timClicks> there probably is, but I can't find it
[04:50] <timClicks> I'll delete anything that's older than 3 months
[04:53] <babbageclunk> I couldn't find it either - I think that's very likely fine
[08:36] <stickupkid> jam, https://github.com/juju/juju/pull/10453#discussion_r306690098 responded
[08:37] <jam> stickupkid: works for me
[08:41] <jam> manadart: fwiw, I talked with tim a bit today about the issues with the cache, I don't think he realized the extent of coherency issues. We'll likely talk the most next week
[08:41] <manadart> jam: Yep, just had 1:1 with him.
[08:42] <jam> gotcha
[09:14] <achilleasa> Can I get a CR on https://github.com/juju/juju/pull/10460? It is changing stuff in the deploy code so please take your time and try to break it ;-)
[09:25] <achilleasa> also, I seem to be getting lots of intermittent test failures (unrelated tests)...
[10:05] <stickupkid> turns out pylibjuju doesn't remove the sodding LXD profile when it's done with it's tests and I've ended up with 1000 profiles :|
[10:16] <achilleasa> stickupkid: try this: lxc profile list | grep juju | cut -d' ' -f2 | xargs -L 1 lxc profile delete
[10:17] <stickupkid> achilleasa, yeah did that :D
[10:17] <stickupkid> achilleasa, i need to fix pylibjuju tbh :D
[10:17] <stickupkid> add another task to my never ending list
[10:57] <stickupkid> why pylibjuju goes mental and creates lots of models
[10:57] <stickupkid> "juju models --format=json | jq '.models | .[] | .name' | xargs -I % echo % | grep -v -E "controller|default" | xargs -I % juju destroy-model -y --force %"
[11:35] <stickupkid> embedded stucts don't work in python :|
[11:51] <stickupkid> gr8 the schema gen has a bug where it doesn't get all params/methods from structs that have embedded values in them
[11:51] <stickupkid> sigh
[12:02] <manadart> stickupkid: You can't just nest dicts?
[12:03] <manadart> Need a review - just dead code removal: https://github.com/juju/juju/pull/10462
[12:04] <stickupkid> manadart, turns out it got dumped into kwargs, even though it should be mapped to the field in the object in python
[12:04] <stickupkid> manadart, that was rather painful to work out
[12:04] <stickupkid> manadart, in otherwords the typed objects just don't work with nested objects
[12:25] <hml> manadart: approved
[12:26] <manadart> hml: Ta.
[12:28] <achilleasa> can someone please verify for me if running the following command in 'juju/apiserver' on the 2.6 branch fails? go test -check.v -check.f TestToolsUploadedSecurely
[12:31] <manadart> achilleasa: Pass here.
[12:32] <achilleasa> manadart: this is what I get: https://pastebin.canonical.com/p/DmJWcFzjP2/ (with go 1.12.6)
[12:33] <achilleasa> actually there are lots of tests in that package that fail with the same error...
[12:33] <achilleasa> HEAD is ebae7c6df409ff82ccdcf385ac91ca4334df1106
[12:36] <manadart> On the current HEAD (58d60964ab) I got 1 failure for model not found in cache (intermittent). 1.12.5.
[12:38] <hml> manadart:  with rebased to develop, it became too easy to resolve the todos in the pr,  https://github.com/juju/juju/pull/10443  .  could you pls take a quick look at the last commit for sanity check.
[12:38] <achilleasa> I got the error on CI for my PR but it seems unrelated so I also tried against 2.6 and still get it... odd
[12:39] <achilleasa> manadart: did you make dep before running the test?
[12:39] <manadart> achilleasa: Yes.
[12:39] <hml> achilleasa: that bug was a pita yesterday trying to merge
[12:39] <manadart> hml: Stand by.
[12:44] <hml> achilleasa: interesting, it’s passing for me with 1.12.7
[12:44] <achilleasa> running on my laptop now to double-check
[12:50] <achilleasa> ok, so it passes on my laptop... weird...
[12:53] <manadart> hml: Left one comment.
[12:53] <hml> manadart: rgr
[15:31] <hml> manadart:  ping
[15:31] <manadart> hml: pong
[15:32] <hml> manadart:  have time for a quick ho?  on spacename vs spaceid in subnets?
[15:32] <manadart> hml: Yep.
[16:54] <stickupkid> i'm going to spend tomorrow breaking my CMR PR into smaller ones, so they're easier to land