=== howbazaar is now known as thumper [06:15] Good Morning! [08:24] achilleasa, why bother with this? https://github.com/juju/juju/pull/11984/files#diff-656c47fa69165a72c0003590bb9d2a1aR806 [08:24] achilleasa, isn't this a 2.9 thing anyway? [08:28] so imaginge you have a 2.9 client, a 2.8 controller and this bundle: https://pastebin.canonical.com/p/zYNkK9dTpf/ [08:29] ah, fair [08:29] you got this bundle from a 2.9 controller running "juju expose foo" [08:29] back to code reviewing [08:29] so basically, this equivalent to a juju expose on a 2.8 controller so we shouldn't complain about it [08:30] yeah got it [08:30] :) [08:30] backwards compat is hard :D [08:30] dude, in hell atm as well [08:30] (also, it might be beneficial to review each commit separately to see how they all fit together) [08:31] achilleasa, I know you want to do this lazily, but if you end up with multiple params, this is expensive https://github.com/juju/juju/pull/11984/files#diff-fe9a84a5a30e2b9e38fac2299b55442eR1472 [08:32] no wait [08:32] i didn't read that right [08:32] d'ho [08:32] ;-) [08:35] I'm starting to think our help text is way too long [08:35] juju man expose [08:40] or juju -H (or -hv) [08:40] yeah, either way [08:40] adding flags to help is very verbose [08:41] info juju might be a more standard-y way of reading manual [08:54] see juju bootstrap for overload [09:57] stickupkid: got a few min to chat about shared facade code? (networking common) [09:57] sure [10:03] I think that I may have missed a step or misconfigured something here https://docs.openstack.org/project-deploy-guide/charm-deployment-guide/latest/app-bridge-interface-configuration.html [10:04] when I run juju run --unit ovn-central/leader 'ovn-nbctl show' I get [10:04] port provnet-940c91fe-5397-4471-bb57-1f956d54be08 [10:04] type: localnet [10:04] tag: 16 [10:04] addresses: ["unknown"] [10:05] which makes me think that I have something misconfigured in ovn-chassis [10:08] or perhaps I also need to configure https://jaas.ai/ovn-dedicated-chassis/3 [10:49] achilleasa, you might know more info? [10:51] unfortunately I am not familiar with ovn-chassis [11:08] stickupkid: one more for you https://github.com/juju/juju/pull/11992 (most changes are in tests though) [11:47] stickupkid: got a question about SpaceInfos [11:47] sure [12:40] stickupkid: looks like auth checks in the firewaller facade are rather lax... https://github.com/juju/juju/blob/develop/apiserver/facades/controller/firewaller/firewaller.go#L410 (also see [12:41] I should ask jam at standup if there is something we can auth against for these kind of calls... [12:47] doh.. that's a private method... but anyway, I need to do some sort of auth for grabbing the space infos [12:48] (also WatchIngressAddressesForRelations seems to skip auth checks) [13:00] stickupkid: btw, do you know if we can have an API invocation that does not expect an argument? [13:08] achilleasa, it always requires one [13:10] achilleasa, I wonder if I could turn schema js into some sort of OpenAPI thing [13:14] juju 3 feature? :D [13:15] I ... think I can inject a read-model access check [14:26] stickupkid: you have a WIP commit in your PR ;-) [14:27] I'll clean that up in a sec [14:28] achilleasa, smashed it out [16:55] I have a nice openstack deployment where everything is green [17:00] when launch an instance, it says that that the instance power state is running, but there is no console log, no vnc console, no icmp, no ssh. this leads me to believe that the machine is not building properly, I would expect to see a libvirt.xml file in /var/lib/nova/instances/uuid, but all that is there is an empty console.log [17:04] My assumption is that there are a number of different configuration issues that could cause this behavior, I'm just wondering if anyone has run into anything similar, that could point me in the right direction === tychicus_ is now known as tychicus [18:53] tychicus, i would start with `openstack server show `. there should be a message, and it sometimes points you to in a certain direction (e.g. "look at neutron logs"). regardless, i would then determine the hypervisor involved and look at its nova-compute.log [18:54] it would also be useful to know what guide you used to deploy the cloud. the exact steps would also be very helpful. as would a description of the environment [18:55] and i see you are also conversing in #openstack-charms