[00:26] thumper: whenever.... https://github.com/juju/juju/pull/11668 [01:29] * thumper looks [01:29] * thumper looks briefly, and decides to look in more detail loater [08:21] 'juju status' (with 2.7.6 installed from snap) freezes apparently forever, is there a trick to it? [08:23] Chipaca, that normally happens if it can't communicate with the controller [08:24] Chipaca, there are other times when this can happen also as well, but that would be where I'd look first [08:25] stickupkid: it's just a local thing talking to an lxd, does that make a difference? [08:25] Chipaca, same scenario. What does `lxc list` say [08:27] stickupkid: so, this started happening yesterday, and as this is only used by me for testing stuff, i purged juju and lxd and reinstalled; 'lxc list' lists nothing, but 'juju bootsrap localhost overlord' tells me 'ERROR controller "overlord" already exists' [08:27] * Chipaca is probably breaking all kinds of stuff in his ignorance [08:28] Chipaca, so you don't have a controller, juju thinks there still is one. You have two options... clean up the ./local/share/juju yaml files or run `juju controllers --format=json | jq -r ".controllers | keys[]" | xargs -I% juju kill-controller -y -t 0 %` [08:29] it assumes you have `jq` installed. [08:29] why would the juju snap keep state outside of ~/snap :-/ [08:29] sigh [08:29] --classic [08:30] Chipaca, we don't support confinement yet, we do have it on our roadmap... priorities I guess ;-) [08:30] i know, but you're just making harder for yourselves :) [08:30] anyway, thanks, that's exactly the problem [08:30] i'll nuke that and start afresh \o/ [08:31] stickupkid: thanks! [08:32] stickupkid: juju unregister [08:32] will get rid of an orphaned controller entry [08:32] no direct editing of controllers.yaml needed [08:32] wallyworld, haha, knew there was another command [08:33] FWIW the jq pipeline didn't work so I went for removing ~/.local/share/juju [08:33] Unable to open API: open connection timed out [08:33] WARNING unknown config field "lxd-snap-channel" [08:33] ERROR getting controller environ: getting environ using bootstrap config from client store: Get https://10.160.185.1:8443/1.0: Unable to connect to: 10.160.185.1:8443 [08:33] ~ [08:34] achilleasa, this is new --> WARNING unknown config field "lxd-snap-channel" - when trying to remove a controller [08:44] stickupkid: that's a side-effect of how we access these settings in the workers. TLDR: we pop them off the map when accessed and complain about remaining entries. We do that here (https://github.com/juju/juju/blob/develop/worker/provisioner/container_initialisation.go#L207) for this field but this (https://github.com/juju/juju/blob/develop/container/lxd/manager.go#L78) doesn't hence the warning. [08:45] we could add a dummy pop in the latter file to silence the warning but I didn't add it as part of my commit because it looked weird. We can do that as a drive-by fix in any future PR [08:59] manadart, ping [09:00] stickupkid: Pong; just starting a call. [09:00] manadart, ping when you're free === arif-ali_ is now known as arif-ali [10:38] stickupkid: I'm in Daily. === StoneTable is now known as aisrael === hml_ is now known as hml === exsdev0 is now known as exsdev [12:38] hey guys [12:38] manadart: what is the current status on spaces and manual machines? [12:38] whats the default username/password for maas and juju? [12:38] for ssh [12:38] im locked out atm [12:40] achilleasa: Currently unsupported; on the this cycle of work to enable. It just requires population of subnets from the machine NICs. [12:51] manadart: if i got into rescue mode can i get to a console? [12:54] Eryn_1983_FL: There's no default. Juju sets up it's own key in authorized_keys, which allows `juju ssh `. [12:55] i dont has ssh to juju [12:55] Or `juju ssh ` for that matter. [12:56] Eryn_1983_FL: You have no client connectivity to your Juju controller? [12:56] i got the web gui [12:59] just dont understand why my keys are not working [13:29] time to reinstall again === dannf is now known as dannf` [13:35] quick CR https://github.com/juju/utils/pull/312 anyone? needed to fix the broken tests on develop === narindergupta is now known as narinderguptamac [14:35] manadart, https://bugs.launchpad.net/juju/+bug/1882097 [14:35] Bug #1882097: Needless brute forcing with finding AZ in space topology in the provisioner task [14:40] stickupkid: výborne [15:25] Can someone review and run the "straightforward" QA steps for https://github.com/juju/juju/pull/11670? [15:44] achilleasa, I'm stealling your PR for now (sorry for the noise) [15:47] no worries. I want to force-push a fix for one of the tests. Let me know when you need a new !!build!! in there ;-) [15:51] achilleasa, do it when ever I'm breaking stuff atm === verterok` is now known as verterok [19:16] has something changed in 2.8 that may affect deploying bundles? seems slower going through all bundle lines [19:20] really slower [21:18] pmatulis: yes, the multiwatcher code was refactored, and it does slightly more now [21:18] pmatulis: I'm currently looking into it to see what can be done [21:19] this is the basis of a watcher that bundle deploys use === hallback_ is now known as hallback [22:44] thumper, ok thanks. what else does it do? more validation? [23:00] pmatulis: there shouldn't be a measurable difference between 2.7 and 2.8. If there is, we need to work out why [23:00] pmatulis: if you were able to show how different it is, that would be very helpful [23:00] https://github.com/juju/collections/pull/4 [23:08] thumper, ok, will do