| noonedeadpunk | nixbuilder: that is really weird... I'd potentially check/verify consistency of fernet tokens | 07:24 |
|---|---|---|
| noonedeadpunk | inside keystone | 07:24 |
| noonedeadpunk | as seems that while 1 keystone backend do accept token - the other one is not | 07:25 |
| noonedeadpunk | or indeed - there's smth fishy with one of glance backends.... | 07:25 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_horizon master: Address RemovedInDjango40Warning https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/912107 | 11:46 |
| noonedeadpunk | this should fix horizon ^ | 11:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_horizon master: Do not change mode of files recursively https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/911420 | 11:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_horizon master: Detect OVN VPNaaS installation https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/911421 | 11:47 |
| noonedeadpunk | ugh, no, there's more... | 11:52 |
| opendevreview | Merged openstack/openstack-ansible-plugins stable/zed: Ensure consistent ordering of network_mappings https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/904734 | 11:56 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_horizon master: Address Django Deprecations for 4.1 https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/912107 | 12:00 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_horizon master: Do not change mode of files recursively https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/911420 | 12:00 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_horizon master: Detect OVN VPNaaS installation https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/911421 | 12:00 |
| noonedeadpunk | updated https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/912107 and now it should be way better... | 12:01 |
| nixbuilder | noonedeadpunk: Do you think a re-install of glance may fix this? | 12:17 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_heat master: Grant proper privileges to admin user for testing purposes https://review.opendev.org/c/openstack/openstack-ansible-os_heat/+/912108 | 12:17 |
| noonedeadpunk | nixbuilder: I'm not sure it's really glance at all? | 12:18 |
| noonedeadpunk | as it totally could be keystone as well | 12:18 |
| noonedeadpunk | like some kind of firewall preventing sync of fernet keys | 12:19 |
| nixbuilder | noonedeadpunk: After reading about fernet tokens I tried to run keystone-manage but I get a command not found! | 12:23 |
| noonedeadpunk | huh | 12:23 |
| noonedeadpunk | well | 12:23 |
| noonedeadpunk | have you tried it from venv? | 12:23 |
| noonedeadpunk | just in case | 12:23 |
| nixbuilder | noonedeadpunk: Don't know how to do that... this is a bare metal install. | 12:24 |
| noonedeadpunk | well, still | 12:24 |
| nixbuilder | OK... I found it. | 12:25 |
| nixbuilder | noonedeadpunk:/openstack/venvs/keystone-27.4.0/bin/keystone-manage | 12:26 |
| noonedeadpunk | yeah | 12:26 |
| noonedeadpunk | and then there should be regular cron job | 12:46 |
| noonedeadpunk | which uses ssh to dostribute them | 12:46 |
| noonedeadpunk | but I guess easiest way to check might be disabling all keystone backends except one and see if issue fades awat | 12:47 |
| opendevreview | Merged openstack/openstack-ansible stable/zed: Revert "Drop upgrade jobs for Zed" https://review.opendev.org/c/openstack/openstack-ansible/+/911602 | 12:49 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Bump ansible version to 2.15.9 https://review.opendev.org/c/openstack/openstack-ansible/+/905619 | 13:08 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Do not configure RPC for Neutron with OVN driver https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/907245 | 13:09 |
| noonedeadpunk | jrosser: sorry for the ping in a weekday, but maybe you know from top of your head - why in the world I'm getting `The requested functionality is not supported` on console URL request right after installing ironic and running nova role? | 13:13 |
| *** carloss_ is now known as carloss | 13:15 | |
| noonedeadpunk | oh, I see | 13:20 |
| jrosser | noonedeadpunk: you get that in the ansible output? | 13:20 |
| noonedeadpunk | nah, I get that when trying to `openstack console url show` for VM | 13:21 |
| noonedeadpunk | but I guess issue is that I'm doing that in AIO... | 13:21 |
| jrosser | it should work for a nova vm | 13:22 |
| noonedeadpunk | yeah, I guess I messed up things, and nova-compute acts as ironic one now or smth like that | 13:22 |
| jrosser | we have this deployed in a couple of places so if you want to compare anything give andrewbonney a shout | 13:22 |
| jrosser | ah well, you should have a nova compute on the control plane too | 13:22 |
| jrosser | so for metal AIO this likely makes total mess | 13:23 |
| noonedeadpunk | yeah | 13:23 |
| noonedeadpunk | though ironic compute is in lxc... | 13:23 |
| jrosser | I think this should be ok | 13:23 |
| noonedeadpunk | but likely is completely confused | 13:23 |
| noonedeadpunk | *nova-compute | 13:23 |
| jrosser | imho putting ironic on a AIO should not break nova console for vm | 13:24 |
| jrosser | if it does then we probably have a bug | 13:25 |
| noonedeadpunk | yeah, so I guess nova_virt_type is getting resolved to ironic | 13:25 |
| jrosser | only for the nova compute inside the lxc | 13:25 |
| noonedeadpunk | yeah, `compute_driver = ironic.IronicDriver` | 13:25 |
| jrosser | the one on the host should be just as before? | 13:25 |
| noonedeadpunk | well. it's not | 13:26 |
| jrosser | huh right | 13:26 |
| jrosser | so I think that you should be able to build an ironic only cloud | 13:26 |
| jrosser | and in that case there would only be the control plane nova compute service | 13:27 |
| noonedeadpunk | I have same driver "on metal" and in lxc | 13:27 |
| jrosser | is it a metal AIO btw? | 13:27 |
| noonedeadpunk | no, lxc one | 13:27 |
| noonedeadpunk | I would get it on metal more | 13:27 |
| noonedeadpunk | basically why I pinged, as was o_O | 13:27 |
| jrosser | ok sure so this certainly sounds like a bug | 13:27 |
| noonedeadpunk | like - having vnc in config and all bits. But missed the driver... | 13:27 |
| noonedeadpunk | yeah, will check what leads to that | 13:28 |
| noonedeadpunk | thanks a lot | 13:28 |
| jrosser | there is logic in the nova role which has to understand if it’s doing hypervisor vs. ironic when the compute service is installed | 13:28 |
| noonedeadpunk | so far I found only that? https://opendev.org/openstack/openstack-ansible-os_nova/src/branch/master/tasks/nova_virt_detect.yml#L26-L43 | 13:30 |
| noonedeadpunk | but it's probably not it | 13:30 |
| jrosser | maybe it’s because the metal host is a nova compute, rather than just infra like in multinode | 13:31 |
| noonedeadpunk | anyway, I will find as I know what to look for | 13:31 |
| jrosser | and that means that many group vars are in scope | 13:31 |
| noonedeadpunk | ah, sure.... | 13:35 |
| noonedeadpunk | this was "easy" one | 13:35 |
| noonedeadpunk | ironic_compute.yml is a symlink to ironic-compute_hosts.yml | 13:36 |
| noonedeadpunk | and ironic-compute_hosts inlcudes actual HOST on which container resides | 13:36 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible/src/branch/master/inventory/group_vars/ironic-compute_hosts.yml#L16 | 13:36 |
| jrosser | well - if you remember there was all sorts of madness with env.d and inventory when I first looked at ironic | 13:37 |
| noonedeadpunk | and these are not same things at all: https://paste.openstack.org/show/bpaj4T7k1cTuRvvDWL5S/ | 13:38 |
| noonedeadpunk | yeah, for sure | 13:38 |
| jrosser | I bet this is all similar mess | 13:39 |
| noonedeadpunk | it's not touched for 7y, so... | 13:39 |
| noonedeadpunk | but there's same mess with qemu/kvm actually | 13:40 |
| noonedeadpunk | ok | 13:40 |
| jrosser | we should totally get virtualbmc working for the AIO | 13:40 |
| jrosser | then all this would be revealed in CI that it’s broken | 13:40 |
| noonedeadpunk | I'm going to dive into ironic soonish... | 13:40 |
| jrosser | well if you get stuck or it looks wierd we have a pretty well working ironic now | 13:40 |
| jrosser | multi arch and all the toys | 13:41 |
| jrosser | but tbh we never really payed attention to the AIO in detail | 13:41 |
| noonedeadpunk | sweet. I would need at least some networking advice I guess. As I can recall you was saying some things about it | 13:41 |
| jrosser | I put a ton of docs in the role | 13:42 |
| noonedeadpunk | I mainly do AIO to present UI/UX more or less | 13:42 |
| noonedeadpunk | ok, yes, true, I will check on them! | 13:42 |
| noonedeadpunk | so just baby steps | 13:42 |
| jrosser | are you planning to use dell hardware? | 13:42 |
| noonedeadpunk | no idea | 13:42 |
| jrosser | well don’t :) | 13:43 |
| noonedeadpunk | yeah, I know it's redfish is just a nighmare | 13:43 |
| jrosser | ^ this is learned the hard and painful way now | 13:43 |
| noonedeadpunk | and here we have mostly solely Dell as of today | 13:43 |
| jrosser | we did write some scripting using the idrac api | 13:43 |
| noonedeadpunk | so when it comes to automate compute provisionment - it's all /o\ even with maas | 13:43 |
| jrosser | to do things like bulk switch to uefi boot which is horrible to do manually | 13:44 |
| noonedeadpunk | but I guess for Ironic thing it would be mostly supermicro | 13:44 |
| noonedeadpunk | SO I was looking at just ipmi 2.0 driver more or less | 13:44 |
| jrosser | yeah there is catch 22 there because if you don’t use the ipmi driver you can’t have the ipmi-sol console | 13:45 |
| jrosser | which is wtf, really | 13:45 |
| noonedeadpunk | oh, well, yeah, I'd be surprised | 13:46 |
| noonedeadpunk | not having it | 13:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Apply ironic virt_type only for ironic_compute group https://review.opendev.org/c/openstack/openstack-ansible/+/912173 | 14:11 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix inventory defenition for Cloudkitty https://review.opendev.org/c/openstack/openstack-ansible/+/912269 | 14:58 |
| noonedeadpunk | IF there're any good ideas on how to handle this less painfully - I'm very open to suggestions ^ | 14:58 |
| noonedeadpunk | we're doing smth very stupid from what I guess with OVN namings also... | 15:08 |
| noonedeadpunk | `flat_networks = dbaas-mgmt,lbaas,vlan` - wtf is vlan doing in flat network.... | 15:08 |
| jrosser | ^ merge my provnet1 patch :) | 15:10 |
| jrosser | otherwise there is like flat = vlan total confusion and you don't know if it's type or name | 15:11 |
| jrosser | sorry physnet1 https://review.opendev.org/c/openstack/openstack-ansible/+/908766 | 15:13 |
| andrewbonney | ^ reviewed that now | 15:17 |
| noonedeadpunk | ah, it was adressing specificlly that :D | 15:20 |
| noonedeadpunk | potentially it would be also nice to replace keystone crontasks with systemd timers... | 15:48 |
| * noonedeadpunk having a working day-off today :D | 15:48 | |
| opendevreview | Merged openstack/openstack-ansible master: Determine if upgrade source branch is stable/ or unmaintained/ https://review.opendev.org/c/openstack/openstack-ansible/+/911583 | 15:51 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Upgrade Gnocchi to 4.6 https://review.opendev.org/c/openstack/openstack-ansible/+/912277 | 15:54 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_cloudkitty master: Enable CloudKitty APIv2 https://review.opendev.org/c/openstack/openstack-ansible-os_cloudkitty/+/912291 | 16:59 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_cloudkitty master: Add variable to configure storage options https://review.opendev.org/c/openstack/openstack-ansible-os_cloudkitty/+/912292 | 17:36 |
| opendevreview | Merged openstack/openstack-ansible master: Always use physnet1 as external network name in AIO and examples https://review.opendev.org/c/openstack/openstack-ansible/+/908766 | 17:40 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_cloudkitty master: Add variable to configure storage options https://review.opendev.org/c/openstack/openstack-ansible-os_cloudkitty/+/912292 | 20:06 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!