| noonedeadpunk | o/ | 07:09 |
|---|---|---|
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Remove the get_md5 parameter from ansible stat tasks https://review.opendev.org/c/openstack/openstack-ansible/+/925564 | 07:27 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-tests master: Remove the get_md5 parameter from ansible stat tasks https://review.opendev.org/c/openstack/openstack-ansible-tests/+/925565 | 07:30 |
| opendevreview | Merged openstack/openstack-ansible-plugins stable/2024.1: Updated from OpenStack Ansible Tests https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/925005 | 09:25 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-repo_server master: Ensure default SSL config is removed for EL https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/925571 | 09:49 |
| noonedeadpunk | this fixes TLS jobs ^ | 09:49 |
| noonedeadpunk | andrewbonney: I can recall you was working on some tag for upgrade process to ensure that switch to quorum queues will happen alter on, but I don't see upgrade docs mentioning that. Was you proposing patch there and I jsut can't find it or it was not proposed yet? | 10:35 |
| andrewbonney | I believe it was all merged, but let me have a search | 10:35 |
| noonedeadpunk | ah - https://docs.openstack.org/openstack-ansible/latest/admin/maintenance-tasks.html#migrate-between-ha-and-quorum-queues | 10:36 |
| noonedeadpunk | ok, makes sense, sorry :D | 10:36 |
| andrewbonney | Yeah that's the one, and the release notes for C mention 'Please check RabbitMQ maintenance documentation for more infromation on how to switch...' | 10:37 |
| andrewbonney | I haven't actually had chance to run that against our C staging deployment since upgrading, so that was based upon testing in an AIO | 10:37 |
| noonedeadpunk | will try to run today :) | 10:40 |
| andrewbonney | Have you done any C upgrades yet btw? | 10:43 |
| noonedeadpunk | no, not yet | 10:47 |
| noonedeadpunk | gonna play in close to prod test env | 10:48 |
| andrewbonney | I guess from our experience when you do, watch out for https://bugs.launchpad.net/nova/+bug/2065403 and https://bugs.launchpad.net/cinder/+bug/2070475 | 10:48 |
| noonedeadpunk | yeah, these are reasonable ones... | 10:52 |
| noonedeadpunk | I think we're hitting these pretty much all times | 10:52 |
| noonedeadpunk | specifically for deleted ones | 10:52 |
| noonedeadpunk | but thanks for pointing to that - will obviously save quite some time :) | 10:53 |
| platta | I’m working on deploying an AIO lab, and I want to make it so that the provider network is my physical network. I’ve got my physical network configured with a range of IP addresses it won’t hand out and I want OpenStack to use that as its range for Floating IPs. Having trouble finding my way in the documentation, I haven’t seen any | 18:40 |
| platta | network-related documentation specific to the AIO use case. Would appreciate any documentation links I might be missing or other guidance. | 18:40 |
| jrosser | platta: what we call the "AIO" is a special case where the whole deployment is behind a single IP address on the physical interface, it is primarily for testing, development and as a CI test harness | 18:51 |
| jrosser | having said that, it would be completely possible for you to create a single node configuration that was very similar to the openstack-ansible "AIO" which behaves as you want | 18:52 |
| platta | Aha, so instead of trying to bend “AIO” to be what I want, perform a “non-AIO” deployment and just configure it to be a single node, etc. | 18:54 |
| jrosser | i would take a quick look at this https://github.com/openstack/openstack-ansible/blob/master/tests/roles/bootstrap-host/tasks/prepare_networking.yml | 18:54 |
| jrosser | when you follow the quickstart guide to build an "AIO", that is actually what happens to the networking on the host, in particular see how it NATs the provider network behind a single IP on whatever interface you have | 18:55 |
| jrosser | so thats the part that you don't want | 18:55 |
| platta | That’s a great start for me. Thanks for the information! | 18:56 |
| jrosser | but if you were to take the configuration that is in /etc/openstack_deploy/openstack_user_config.yml, and (with care) some of the user_variables that are autogenerated for the AIO you would be very close to a more useful real world single node config | 18:57 |
| jrosser | the piece of work you would need to do is replicate the host networking setup, similar to what is done in the prepare_networking.yml ansible i linked, but decide how you are going to deal with the provider network specifically | 18:57 |
| jrosser | it is generally easier if this is a different interface to the one you ssh in on / access horizon | 18:58 |
| platta | I have to do some playing with that. The physical host only has one network interface. | 19:00 |
| jrosser | it's likely possible, just needs some care | 19:01 |
| jrosser | just FYI there is most activity here office hours EU timezone, if you get stuck please ask | 19:01 |
| platta | Thank you, I really appreciate it. I’m going to start by reading the documentation more in depth, but I’ll be back if I run into trouble. | 19:02 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!