| opendevreview | Merged openstack/openstack-ansible master: Bump upstream SHAs on master https://review.opendev.org/c/openstack/openstack-ansible/+/906148 | 03:22 |
|---|---|---|
| noonedeadpunk | mornings | 08:16 |
| noonedeadpunk | ok, yes, excludelist is just absent in aio | 08:16 |
| jrosser | morning | 08:44 |
| jrosser | did i make some mistake in the tempest vars stuff? | 08:44 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Fix condition of excludelist file removal https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/907180 | 08:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_ironic master: Allow to extend default ironic_driver_types https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/907155 | 08:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_ironic master: Fix a typo in pxe_redfish definition https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/906353 | 08:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_magnum master: Adopt for usage openstack_resources role https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/901185 | 08:48 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_magnum master: Move insecure param to keystone_auth section https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/905110 | 08:48 |
| jrosser | oh doh :( | 08:49 |
| jrosser | so close to being right | 08:49 |
| noonedeadpunk | it was suuuuper close :D | 08:59 |
| opendevreview | Merged openstack/ansible-hardening master: Use replace module instead of lineinfile for disabling dynamic motd https://review.opendev.org/c/openstack/ansible-hardening/+/907095 | 09:37 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-hardening stable/2023.2: Use replace module instead of lineinfile for disabling dynamic motd https://review.opendev.org/c/openstack/ansible-hardening/+/906894 | 09:38 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-hardening stable/2023.1: Use replace module instead of lineinfile for disabling dynamic motd https://review.opendev.org/c/openstack/ansible-hardening/+/906895 | 09:38 |
| jrosser | did you see ceph-ansible got deprecated again | 10:11 |
| kleini | and what should be the alternative? cephadm? | 10:16 |
| jrosser | it depends what the question is really | 10:19 |
| jrosser | to make a ceph cluster next to openstack-ansible then really any approach is fine | 10:19 |
| jrosser | but openstack-ansible itself running ceph-ansible to make a production deployment, we said generally "don't do that, it's for CI/development" | 10:21 |
| jrosser | the deb/rpm are all available for the newer releases anyway | 10:22 |
| kleini | so I need to move my staging Ceph deployment from ceph-ansible away to something else. But ceph-ansible was easy to use in Ansible which does for me the Heat stack deployment for the staging environment. | 10:24 |
| jrosser | you can go Q->R just by updating the packages on a ceph-ansible deployment | 10:25 |
| jrosser | that is the recommended upgrade approach | 10:25 |
| jrosser | not actually using the ansible at all | 10:26 |
| jrosser | and i think for my use, it will be less effort just now to fork ceph-ansible and fix up anything that breaks fresh Reef installs than migrate to anything else | 10:27 |
| frickler | jrosser: I didn't see that, thx for the headsup | 10:27 |
| frickler | and we are considering the same, maybe do a fork under the opendev umbrella? | 10:27 |
| jrosser | i would really like to clean up ceph-ansible - do you use it on bare metal, or containers? | 10:27 |
| noonedeadpunk | jrosser: yeah, you've posted that day or so ago | 10:27 |
| frickler | jrosser: containers running on bare-metal | 10:28 |
| noonedeadpunk | ops repo ?:D | 10:28 |
| jrosser | frickler: one of my team is working right now on getting making it work with Reef/Jammy | 10:28 |
| jrosser | and we were just getting all sorted out to make the required PR :/ | 10:28 |
| noonedeadpunk | we can create a repo for that actually | 10:29 |
| jrosser | frickler: the only thing we found was small areas of the Reef dashboard that actually require cephadm or they don't work | 10:30 |
| frickler | jrosser: I didn't look at reef yet, but that sounds pretty promising. pretty sure some other deployers will also be interested | 10:36 |
| jrosser | i agree | 10:36 |
| jrosser | it would be pretty nice to convert it to a collection | 10:41 |
| jrosser | that would make integrating it into specific environments a whole lot easier | 10:41 |
| frickler | jrosser: so what is your idea, rather do a fork on github or under the osa umbrella or a dedicated project in opendev? | 10:41 |
| jrosser | i've not really thought about it tbh | 10:41 |
| jrosser | we already have a fork on github for historical reasons | 10:42 |
| jrosser | like taking forever to get anything merged | 10:42 |
| jrosser | the scope is likley wider that just OSA really | 10:43 |
| frickler | jrosser: ok so maybe discuss this internally and let me know if you want help with setting up something in opendev | 10:51 |
| jrosser | frickler: looks like my held node might be ready, https://github.com/jrosser.keys | 11:41 |
| frickler | jrosser: ack, root@199.204.45.65 | 12:07 |
| farbod | Hi Guys. I stuck at `openstack-ansible setup-infrastructure.yml` in this step https://paste.opendev.org/show/bpB0Wkpltap9dCEsaYt2/ | 12:07 |
| farbod | What can i do? | 12:07 |
| farbod | Also i had a question. Should we configure disks for Ceph OSDs when using OSA? | 12:10 |
| farbod | Here is the full logs: https://paste.opendev.org/show/bdlZUc8FfJ7jP4TwTUMY/ | 12:13 |
| noonedeadpunk | farbod: well. It's kinda tricky. In fact, we jsut have a "layer" right now to connect osa with ceph-ansible project | 12:33 |
| noonedeadpunk | with that we usally suggest to manage these 2 independently | 12:33 |
| noonedeadpunk | but yes, by output it looks like there're no OSDs for ceph were found/configured | 12:34 |
| noonedeadpunk | I personally not 100% aware of ceph-ansible code, but I assume there should be some kind of variable to define which drives should be used as OSDs | 12:35 |
| farbod | So its better to deploy Ceph independently? | 12:37 |
| farbod | What do you suggest? cephadm or ceph-ansile? | 12:37 |
| farbod | And also how can i configure my self deployed ceph(I mean connecting it to ceph) when configuring OSA? | 12:37 |
| jrosser | farbod: there is an ansible role we use to connect openstack components into the ceph cluster https://docs.openstack.org/openstack-ansible-ceph_client/latest/ | 12:39 |
| noonedeadpunk | farbod: I guess cephadm at this point, as ceph-ansible got deprecated | 12:39 |
| jrosser | farbod: when you deploy OSA with "its own" ceph via ceph-ansible there is some automatic setting of things like cinder / glance to use RBD backends, you will have to make some of this config manually when the ceph cluster is external to openstack-ansible | 12:42 |
| jrosser | farbod: but in the simplest case you set this variable to point to the monitors of the external ceph https://github.com/openstack/openstack-ansible-ceph_client/blob/master/defaults/main.yml#L50 | 12:43 |
| kleini | noonedeadpunk, thank you very much for your hint regarding filter_tenant_id. post-filtering works now for my setup. | 12:49 |
| noonedeadpunk | sweet! | 12:49 |
| jrosser | well thats pretty sad, you can't just specify name: <blah> state: absent on openstack.cloud.coe_cluster in order to delete one | 12:59 |
| jrosser | it wants all the parameters you would have needed to create it as well | 12:59 |
| noonedeadpunk | So, original linkage to template made sense then? | 13:02 |
| noonedeadpunk | but it's weird requirement for a module.... | 13:02 |
| jrosser | not really, just seems wierd to not be able to delete a thing given only the name | 13:02 |
| jrosser | module parameter validation not understanding the difference between present/absent state i assume | 13:03 |
| jrosser | oh no actaully, it passes bogus data (like zero node count) for the things you leave out | 13:03 |
| jrosser | and magnum is 400 at you in reply | 13:03 |
| noonedeadpunk | ugh | 13:04 |
| jrosser | i would hope that this is some pattern baked in fundamentally to all those modules (present -> lots of options) (absent -> name|uuid) | 13:05 |
| 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 | 13:14 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: WIP - Bootstrapping playbook https://review.opendev.org/c/openstack/openstack-ansible-ops/+/902178 | 13:14 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Add role to install and run sonobouy k8s validation tests https://review.opendev.org/c/openstack/openstack-ansible-ops/+/906054 | 13:14 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Add playbook to run functional test of magnum capi driver https://review.opendev.org/c/openstack/openstack-ansible-ops/+/906361 | 13:14 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Add hook playbook install and test magnum capi driver https://review.opendev.org/c/openstack/openstack-ansible-ops/+/906363 | 13:14 |
| jrosser | frickler: i am done with the held node, can we go again with a hold on the same job? | 13:16 |
| frickler | jrosser: set up, do you need to change anything or just do a recheck? | 14:01 |
| jrosser | frickler: i've updated the depends-on patches so it should just need a recheck | 14:02 |
| jrosser | is it right that X-Auth-Token gets printed unredacted in debug logs? | 14:09 |
| jrosser | like this https://zuul.opendev.org/t/openstack/build/5b8fdb4272c646cf8d72df0d6ad6616b/log/logs/host/magnum-api.service.journal-10-25-08.log.txt#110 | 14:11 |
| noonedeadpunk | that looks quite bad.... | 14:32 |
| jrosser | seems to be just glanceclient | 14:33 |
| jrosser | for novaclient i see things like "X-Auth-Token: {SHA256}7ee335fad4314c128bc8c15b40425f288674156cda1181374e86fbbebb6e3680" | 14:35 |
| spatel | noonedeadpunk jrosser admin1 Here is the trove DBaaS blog - https://satishdotpatel.github.io/trove-dbaas-openstack/ | 14:44 |
| spatel | I will create next one with dedicated rabbitMQ way.. | 14:44 |
| noonedeadpunk | #startmeeting openstack_ansible_meeting | 15:00 |
| opendevmeet | Meeting started Tue Jan 30 15:00:16 2024 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. | 15:00 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 15:00 |
| opendevmeet | The meeting name has been set to 'openstack_ansible_meeting' | 15:00 |
| noonedeadpunk | #topic rollcall | 15:00 |
| noonedeadpunk | o/ | 15:00 |
| damiandabrowski | hi! | 15:00 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Fix condition of excludelist file removal https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/907180 | 15:04 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Fix condition of excludelist file removal https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/907180 | 15:05 |
| noonedeadpunk | #topic office hours | 15:05 |
| noonedeadpunk | So we got couple of roles, like ironic and magnum due to this small nit ^ | 15:05 |
| noonedeadpunk | *broken | 15:05 |
| noonedeadpunk | *failing CI | 15:05 |
| * noonedeadpunk needs to get himself together | 15:05 | |
| noonedeadpunk | So I don't actually have many updates for the previous week | 15:06 |
| noonedeadpunk | I'd eventually need to go through bug reports, as while most of them got some first reply, I'm not sure they all were properly triaged | 15:07 |
| noonedeadpunk | Hopefully to have some time on Friday for that | 15:07 |
| noonedeadpunk | I guess main 2 things that are in progress - openstack_resources and capi driver for magnum | 15:07 |
| noonedeadpunk | I was not able to work on openstack_resources to implement suggested improvements for image management. | 15:07 |
| noonedeadpunk | It also might be worth doing that in follow-up, not sure | 15:08 |
| noonedeadpunk | for magnum capi I see some quite good progress, but apparently work is still in progress. Likely very close to the finish line | 15:09 |
| jrosser | o/ hello | 15:13 |
| jrosser | yes for cluster api it is very close | 15:14 |
| jrosser | working in my AIO locally, just need to allow some more time for the cluster to create in the CI node i think | 15:14 |
| jrosser | but there are a *ton* of dependant patches | 15:14 |
| jrosser | some of which are good to go, some of which might need some discussion about the best approach | 15:14 |
| noonedeadpunk | despite beeing super short on time this season, I'm ready to check through things when they're ready :) | 15:16 |
| noonedeadpunk | I can recall there were some "volunteers" to test it out as well | 15:16 |
| jrosser | yeah, that would be good | 15:18 |
| jrosser | getting the AIO going is a little subtle though so i need to document that | 15:18 |
| jrosser | some of the dependancies are generally moving stuff out of openstack-ansible repo into plugins collection | 15:19 |
| jrosser | those are low-hanging fruit | 15:19 |
| noonedeadpunk | actually moving to things to plugins, I was wondering what's up with https://review.opendev.org/q/topic:%22osa/collectify%22 ? | 15:21 |
| noonedeadpunk | Like I was thinking it can be trivial to move all playbooks to plugins while get some benefits (potentially) | 15:21 |
| noonedeadpunk | And that should not take dramatical time to move | 15:22 |
| jrosser | yeah so some of those things actually are needed for the capi work | 15:22 |
| jrosser | but i have done "whatever it takes" so interested in feedback | 15:23 |
| noonedeadpunk | but I can recall not all of these worked like expected? | 15:23 |
| noonedeadpunk | I just though of creating playbooks like openstack.osa.nova, openstack.osa.glance, etc and move content from integrated repo | 15:24 |
| jrosser | right - particularly the one to set vars for source installs i think | 15:24 |
| noonedeadpunk | while leaving just import_playbook: openstack.osa.glance for compatability | 15:24 |
| noonedeadpunk | in the integrated | 15:25 |
| noonedeadpunk | oh | 15:25 |
| noonedeadpunk | another thing we might want to land "early" | 15:25 |
| noonedeadpunk | I had couple of patches to the inventory script... | 15:26 |
| jrosser | yeah | 15:26 |
| jrosser | factoring out playbooks/common-* would be a really good start | 15:26 |
| noonedeadpunk | so this might be useful for your capi hussle I assume: https://review.opendev.org/c/openstack/openstack-ansible/+/905433/2 | 15:27 |
| noonedeadpunk | as I can recall you said before about inconsistent hostnames/inventory names for containers | 15:27 |
| noonedeadpunk | and this https://review.opendev.org/c/openstack/openstack-ansible/+/905438/2 should let us get rid of ANSIBLE_TRANSFORM_INVALID_GROUP_CHARS | 15:28 |
| noonedeadpunk | basically making `-` and `_` interchangable | 15:28 |
| noonedeadpunk | which I dunno if good or terribly bad idea frankly speaking | 15:28 |
| noonedeadpunk | but that means we should be able to jsut rename groups in env.d to be only underscores, and hopefully not breaking anybody during upgrade.... | 15:29 |
| noonedeadpunk | but that might be actually good to leave for the next cycle even.... | 15:30 |
| noonedeadpunk | as I can see how playbooks might be using or relying on current group names | 15:30 |
| jrosser | it does sound like we should try to land a bunch of this stuff as soon as possible | 15:33 |
| jrosser | it’s all quite big and don’t want to leave it too late | 15:33 |
| jrosser | yes capi would benefit from the underscores change | 15:34 |
| jrosser | as there’s a bit of a break/fix cycle keeps happening with external collections there | 15:35 |
| noonedeadpunk | With that we should still keep internal logic of converting hostnames for older deployments... | 15:38 |
| noonedeadpunk | as basically only new containers will be generated this way - inventory won't (or should not) be altered otherwise | 15:38 |
| noonedeadpunk | #endmeeting | 16:00 |
| opendevmeet | Meeting ended Tue Jan 30 16:00:58 2024 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 16:00 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/openstack_ansible_meeting/2024/openstack_ansible_meeting.2024-01-30-15.00.html | 16:00 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/openstack_ansible_meeting/2024/openstack_ansible_meeting.2024-01-30-15.00.txt | 16:00 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/openstack_ansible_meeting/2024/openstack_ansible_meeting.2024-01-30-15.00.log.html | 16:00 |
| jrosser | frickler: would you be able to put my key on the held node again? | 16:35 |
| frickler | jrosser: seems I messed up earlier and deleted the fresh autohold instead of the old one. so this is still the old one that we see, will trigger another recheck now | 16:52 |
| jrosser | is the old node still there? | 16:52 |
| frickler | jrosser: no, I deleted it now | 16:53 |
| jrosser | no problem | 16:53 |
| frickler | I can add access to the new node once the job starts running, assuming it will fail at the end. you just shouldn't interfere with the job too early, then | 16:54 |
| jrosser | that would be great, thanks | 16:55 |
| frickler | jrosser: root@199.19.213.123 | 18:18 |
| 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 | 18:18 |
| jrosser | thankyou | 18:18 |
| 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 | 18:34 |
| 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 | 19:08 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!