[09:59] hi all [09:59] maybe not the right place to ask but with rackspace I still have trouble creating a new controller [10:00] usnig `metadata-source` (because there own default does not expose any images) [10:00] I end up with [10:00] PNG [10:00] Y [10:00] oops sorry [10:00] caused by: request (https://lon.servers.api.rackspacecloud.com/v2/10001843/servers) returned unexpected status: 403; error info: {"forbidden": {"message": "Policy doesn't allow standard_flavor:create:attach_volume to be performed.", "code": 403}}}]) [10:01] any idea where to look at? [10:32] manadart, i'll fix the vpc issue on that aws script [10:35] manadart, i'm not a fan of creating a VPC, so this might actually be worth it [12:00] manadart, I updated it to re-use the VPC - https://github.com/juju/juju/pull/11412 [12:00] manadart, pure cattle approach [12:01] stickupkid: Great. Are you working as usual today? [12:01] I am [12:02] well I am after Lunch [12:02] haha [12:51] hml, where did we get with the focal stuff? [12:52] stickupkid: the PR is ready… had to make a few changes, so I wanted another set of eye [12:52] eyes [12:52] hml, I shall do it [12:52] stickupkid: ack [12:56] hml, we want this to fail right? ./main.sh -V -S focal -l friday deploy. [12:56] hml, or do we want it to bootstrap a new controller? [12:56] stickupkid: I was copying the behavior of the env var for bootstrap series… it should bootstrap a new controller if controller friday is not focal [12:57] ok, let's fix that [12:57] stickupkid: uh? [12:57] stickupkid: chat at stand up? [12:57] sureo [13:59] hml, approved, I can't approve my own PR though, so can you approve it for me? [13:59] stickupkid: sure… since we’re in agreement now [14:00] hml, yeah... happy with the changes [14:00] stickupkid: tick [14:01] hml, we should think about how to add this to jenkins now [14:01] hml, i'll have a look now [14:03] stickupkid: i think we need to look at series in jenkins over all - there are still some jobs that use trusty by default. we should test it as an LTS, not necessarily by default etc [14:03] ok, job for today then - hml [14:03] we’re overridding the default too often I think, should be more explicit [14:04] so when we want to update the default, it’s easier [14:06] stickupkid: shall we use the default series by “default” and set it for explict tests, rather than having it set as a default in random spots? [14:06] then the question is, which tests should we run for trusty, xenial etc.. and bionic moving forward [14:06] hml, unsure - I'm not sure how to do it yet [14:07] stickupkid: would be a good place to have a test jenkins setup [14:07] :-D [14:08] stickupkid: investigate and compare notes in an hour? [14:09] hml, I'm going to try and do the integration tests now, and see what the fallout it. Then I can check how others do it [14:09] rgr [14:09] hml, question is, do we want to test LTS+non-LTS [14:09] or is LTS alone enough? [14:22] stickupkid: we should still do non-LTS… but like we’ve been, explicitly [14:44] stickupkid: what did you change? === xnox is now known as xnox|sessions [15:04] we only have a 60 minute timeout for the merge job, we're hitting that atm [15:04] stickupkid: why are we hitting it? there is barely anythign running? [15:04] 15:15:01 2020-04-27T14:14:19Z INFO Waiting for restart... [15:04] 15:25:58 go (1.14/stable) 1.14.2 from Michael Hudson-Doyle (mwhudson) installed [15:05] It took 10 minutes to install go [15:05] LOL [15:05] Another 10 minutes to fetch all the dependencies [15:05] stickupkid: so this isn’t waiting on other jobs? [15:05] 20 minutes to install [15:05] hilarious [15:06] as we download/build in containers the go mod cache is pure worthless [15:19] stickupkid: think we have enough time on the replacement deploy bundle integration tests that we can dump the nw-deploy-lxd-sub-profile-bundle-lxd [15:19] yeah [15:43] stickupkid: https://github.com/CanonicalLtd/juju-qa-jenkins/pull/437 [15:43] stickupkid: it has a link to the related juju pr [15:43] hml, tick [15:54] stickupkid: https://github.com/juju/juju/pull/11501. :-D [15:54] hml, it would nice to land this one https://github.com/juju/juju/pull/11412 [15:55] hml, that way we could add it to the integration tests [15:55] hml, focal testing etc [15:55] stickupkid: ack, looking now [15:56] hml, acceptancetests/assess_bootstrap.py <- I wonder if we can remove this if we target more architectures? [15:57] stickupkid: perhaps… maybe series too. [15:57] hml, ah it bootstraps to maas, I've not do that yet [15:58] :-/ almost [16:01] stickupkid: how do you config “ that you [16:01] have a default region that probably isn't us-east-1” [16:02] hml, I've got an .aws credential and I set the default region [16:02] [default] [16:02] output = json [16:02] region = eu-west-1 [16:02] that's my .aws/config and use autoload credentials [16:03] are you going to fix all the install petname pieces? for jenkins :-D [16:04] hml, yeah [16:04] :( [16:07] hml, forgot that everything is cattle in the integration tests https://github.com/CanonicalLtd/juju-qa-jenkins/pull/438 [16:07] stickupkid: tick [16:08] having a very productive day today [16:08] tick this, tick that [16:08] got loads of stuff working [16:09] stickupkid: https://pastebin.canonical.com/p/jVS3VyCXVF/ [16:09] brwhahahahaha [16:09] :-D [16:10] how do I solve this :thinking-face: [16:10] I'm like revinventing juju in tests [16:10] sigh [16:10] but without a type system === narindergupta is now known as narinderguptamac [18:09] can an application's leader unit be set (changed) by the juju client? [18:09] (the docs don't mention it) [18:27] pmatulis: no [18:29] rick_h, alright thanks