| jrosser | morning | 10:47 |
|---|---|---|
| jrosser | i think i might have come up with a way to extend OSA with extra playbooks defined as user_variables | 10:48 |
| jrosser | `- import_playbook: "{{ hook_playbook | default('noop.yml') }}"` | 10:48 |
| jrosser | this finally lets us call out to a collection or somewhere else during setup_*.yml to add extras | 10:49 |
| noonedeadpunk | jrosser: oh, actually that might work indeed | 13:00 |
| noonedeadpunk | though ofc it'd better to have that as a list.. | 13:00 |
| noonedeadpunk | But then I assume this should be some kind of template deployed | 13:01 |
| jrosser | I don’t think that’s possible | 13:01 |
| noonedeadpunk | or well | 13:01 |
| jrosser | a list | 13:01 |
| noonedeadpunk | can you import a playbook from imported playbook? | 13:01 |
| jrosser | I’ll try | 13:02 |
| jrosser | but it certainly works pointing at a playbook in a collection | 13:02 |
| noonedeadpunk | As we can have some openstack_deploy/pre-openstack.yml and post-openstack.yml for example | 13:02 |
| noonedeadpunk | where user can add any set and amount of other imports? | 13:02 |
| jrosser | let’s check that, but hopefully | 13:03 |
| noonedeadpunk | but kinda default is most tricky thing indeed | 13:03 |
| noonedeadpunk | like what to do when these doesn't exist basically | 13:03 |
| jrosser | ah well my noop.yml sorts that | 13:04 |
| jrosser | it’s a play against localhost with no tasks | 13:04 |
| noonedeadpunk | yeah, true, I think that might be quite sweet solution then... | 13:20 |
| jrosser | my use case was to try to make the cluster api patch less terrible | 13:24 |
| jrosser | and I think we should be able to put pretty much everything into an ops repo collection now | 13:25 |
| noonedeadpunk | that sounds really great | 13:46 |
| mgariepy | is there a way to have quota on gpus ? | 15:30 |
| noonedeadpunk | mgariepy: I think you would need to use cyborg for that | 16:27 |
| noonedeadpunk | or well, depends what you mean :) | 16:27 |
| noonedeadpunk | like don't allow tenant X to use more then Y GPUs? | 16:27 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: WIP - test Vexxhost CAPI driver for magnum https://review.opendev.org/c/openstack/openstack-ansible/+/893240 | 17:24 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: WIP - Add collection to deploy magnum cluster-api with vexxhost driver https://review.opendev.org/c/openstack/openstack-ansible-ops/+/901450 | 18:02 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: WIP - Bootstrapping playbook https://review.opendev.org/c/openstack/openstack-ansible-ops/+/902178 | 18:02 |
| jrosser | noonedeadpunk: ^ this is just an idea at the moment for how to put everything in the collection | 18:03 |
| jrosser | mcapi_vexxhost/playbooks/bootstrap.yml is some kind of playbook to copy all the config into place for the scenario | 18:03 |
| jrosser | but issue remains, how we know to call that, like a zuul job var, or $something-else | 18:04 |
| noonedeadpunk | oh, well, that pluggable ;) | 18:05 |
| jrosser | and where to do it, zuul pre playbook vs bootstrap-aio | 18:05 |
| jrosser | wonder if you have any good ideas | 18:05 |
| jrosser | becasue it wants to work in and out of CI | 18:05 |
| jrosser | it is kind of neat, becasue then the collection in the ops repo carries all the needed example files | 18:06 |
| jrosser | and chicken / egg of course that you cant use the collection playbook via osa_ops.mcapi_vexxhost.bootstrap until the collection is installed :) | 18:08 |
| jrosser | perhaps the zuul pre playbook is the right place | 18:10 |
| jrosser | because in a local deployment you'd have to create/copy some config into place manually anyway | 18:11 |
| noonedeadpunk | Well, I guess it should be somewhere right before setup-hosts actually but after bootstrap? | 18:11 |
| jrosser | to get the collections, the data needs to be there at bootstrap-ansible | 18:12 |
| jrosser | like user-collection-requirements | 18:12 |
| noonedeadpunk | ah, the chiken-egg you've mentioned | 18:12 |
| jrosser | quite | 18:13 |
| jrosser | hmm anyway, not in a rush but if you have any inspiration then would be nice | 18:13 |
| noonedeadpunk | I'm a bit /o\ mode with HSM now.... | 18:22 |
| jrosser | sounds unpleasant | 18:22 |
| noonedeadpunk | Was very close from getting partition wiped | 18:22 |
| noonedeadpunk | Just some luck and that it's not an HSM but Thales DPoD service that have some different defaults for destructive policy changes | 18:23 |
| noonedeadpunk | So /o\ | 18:23 |
| noonedeadpunk | And now it fails with 0x80001604 CKR_BUSY which I have no idea what it means | 18:27 |
| noonedeadpunk | I really love type of docs, where they reffer all error codes, map them with some other name, but never explain wtf and why | 18:28 |
| jrosser | i tried to make the yubihsm work wit barbican | 18:32 |
| jrosser | and the whole pkcs11 interface is just a world of wtf | 18:33 |
| noonedeadpunk | it really is | 18:33 |
| opendevreview | Danila Balagansky proposed openstack/openstack-ansible-os_neutron master: Neutron OVN plugin AZ support https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/858271 | 19:06 |
| noonedeadpunk | what does `0/0` mean o_O | 19:17 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: WIP - test Vexxhost CAPI driver for magnum https://review.opendev.org/c/openstack/openstack-ansible/+/893240 | 19:40 |
| *** cloudnull0 is now known as cloudnull | 20:07 | |
| -opendevstatus- NOTICE: The Gerrit service on review.opendev.org will be restarting momentarily for a patch update to address a recently observed regression preventing some changes from merging | 21:08 | |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: WIP - Add collection to deploy magnum cluster-api with vexxhost driver https://review.opendev.org/c/openstack/openstack-ansible-ops/+/901450 | 22:34 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!