| *** gmann is now known as gmann_pto | 01:16 | |
| frickler | exoceph: not sure what you exactly want to do, but you should always build your own images anyway | 06:33 |
|---|---|---|
| bbezak | morning | 07:03 |
| bbezak | frickler: kevko could you please take a look at backports of rabbimtq queue manager related changes? It is needed for upgrades to Epoxy, therefore are a prerequisite for the release | 07:05 |
| bbezak | of Epoxy | 07:06 |
| frickler | bbezak: which are these? can you add RP+2 if it is that urgent? I was thinking most of those shouldn't be backported | 07:20 |
| frickler | bbezak: also what's the status of prometheus v3? iiuc we are lacking an upgrade path? | 07:20 |
| bbezak | sure. I thought I already added RP2 :) - https://review.opendev.org/c/openstack/kolla-ansible/+/945810 | 07:21 |
| bbezak | and 2024.2 counterpart | 07:21 |
| bbezak | prometheus v3 is being done by mr @cityofships. but I think he is a bit busy. I'll check with him | 07:23 |
| bbezak | (which I'm doing on daily basis btw) | 07:26 |
| exoceph | frickler I'm testing the ml2 arista driver and l3 plugin. When I'm ready to deploy prod I'll definitely build the images and uploaded them to a central registry. | 07:38 |
| opendevreview | Verification of a change to openstack/kolla-ansible stable/2024.2 failed: Fix Nova and Cinder external ceph enabled check https://review.opendev.org/c/openstack/kolla-ansible/+/946944 | 08:31 |
| opendevreview | Marek Buch proposed openstack/kolla-ansible master: Add support for volume label filtering https://review.opendev.org/c/openstack/kolla-ansible/+/947116 | 08:41 |
| opendevreview | Merged openstack/kolla-ansible stable/2023.2: zuul: Bump base job timeout to 2.5h https://review.opendev.org/c/openstack/kolla-ansible/+/945166 | 08:57 |
| opendevreview | Merged openstack/kolla-ansible stable/2024.2: Document dev mode with custom repository https://review.opendev.org/c/openstack/kolla-ansible/+/946172 | 08:57 |
| opendevreview | Merged openstack/kolla-ansible stable/2024.1: Document dev mode with custom repository https://review.opendev.org/c/openstack/kolla-ansible/+/946173 | 08:57 |
| opendevreview | Merged openstack/kolla-ansible stable/2023.2: Document dev mode with custom repository https://review.opendev.org/c/openstack/kolla-ansible/+/946174 | 08:57 |
| exoceph | I'm trying to deploy on 3 nodes but I have have ha currently disabled (had it enabled previously but changed it). RabbitMQ has failed start every time because it's waiting to peer but kolla-ansible never spins up RabbitMQ on the other control hosts. I even destroyed the cluster through kolla-ansible and it's still trying to peer? | 09:16 |
| opendevreview | Verification of a change to openstack/kolla-ansible master failed: ironic: change enable_ironic_neutron_agent default to no https://review.opendev.org/c/openstack/kolla-ansible/+/938819 | 09:47 |
| opendevreview | Verification of a change to openstack/kolla-ansible stable/2024.2 failed: Fix Nova and Cinder external ceph enabled check https://review.opendev.org/c/openstack/kolla-ansible/+/946944 | 10:55 |
| frickler | bbezak: I'm not convinced that the qm things should be backported? won't this break existing deployments? and if it is needed for upgrades to epoxy, why aren't our upgrade jobs broken? | 11:31 |
| bbezak | because we added that in the upgrade in CI here - https://review.opendev.org/c/openstack/kolla-ansible/+/927497 | 11:39 |
| bbezak | mattcrees: FYI | 11:39 |
| bbezak | better link frickler https://review.opendev.org/c/openstack/kolla-ansible/+/927497/48/tests/upgrade.sh | 11:47 |
| frickler | bbezak: MattCrees[m]: but that exactly proves my point, doesn't it? anyone running 2024.1 currently (e.g. 18.6.0), will be broken by 18.7.0 unless they perform the same migration? | 12:11 |
| MattCrees[m] | We're disabling it by default in the backport. The idea is that people will enable it and perform the migration before upgrading to Epoxy. Same as what we did for the Antelope->Caracal RabbitMQ queue migrations | 12:13 |
| opendevreview | Verification of a change to openstack/kolla-ansible master failed: ironic: change enable_ironic_neutron_agent default to no https://review.opendev.org/c/openstack/kolla-ansible/+/938819 | 12:14 |
| frickler | MattCrees[m]: so does that mean that these are not clean backports? I at least didn't see some note about this yet, but I can do some more detailed comparisons | 12:16 |
| MattCrees[m] | Yeah I edited them to disable all the features by default. I updated the commit messages to reflect that, but I'll admit that's not clear right away. | 12:17 |
| MattCrees[m] | Where would you expect a note about this to be, a comment via Gerrit? | 12:17 |
| frickler | MattCrees[m]: in the commit message and maybe in the renos. but maybe I just didn't look closely enough, let me do some more detailed reviews and I'll let you know whether this is fine | 12:19 |
| MattCrees[m] | Ok thank you | 12:20 |
| amir58118 | Hi, I missed the Tuesday and Wednesday PTG sessions, but I noted my interest in the Etherpad to work on the MariaDB 11.4 bump task under Flamingo release. As someone newer to Kolla, I wanted to ask how task assignment usually works and if it's okay for me to take this one. I'm eager to contribute more actively to Kolla and OpenStack in general. | 12:48 |
| opendevreview | Jack Hodgkiss proposed openstack/kolla master: Update prometheus alertmanager to newest available version https://review.opendev.org/c/openstack/kolla/+/947239 | 13:00 |
| opendevreview | Merged openstack/kolla-ansible stable/2024.2: Fix Nova and Cinder external ceph enabled check https://review.opendev.org/c/openstack/kolla-ansible/+/946944 | 13:04 |
| yuval_ | frickler: bbezak: how/when do images for epoxy will be built? is there a schedule pointing for that? | 14:10 |
| opendevreview | Merged openstack/kolla-ansible master: ironic: change enable_ironic_neutron_agent default to no https://review.opendev.org/c/openstack/kolla-ansible/+/938819 | 14:24 |
| nhk | hello. Could I know when kolla ansble 2025.1 was released? | 14:25 |
| frickler | MattCrees[m]: I'm done with the 2024.2 stack now, some minor comments only. I assume they will apply similarly to 2024.1, so I'll wait for you to check those before continuing | 15:05 |
| MattCrees[m] | Cheers frickler, I'll get on it first thing tomorrow morning | 15:06 |
| frickler | yuval_: nhk: kolla 2025.1 wasn't released yet, there is no fixed plan for that yet. iirc the deadline is up to 3 months after the main release | 15:06 |
| yuval_ | Thanks | 15:07 |
| frickler | you can build images from current master that will be quite like what will be released, though | 15:07 |
| nhk | hello, all images from master are from openstack 2025.1 source, am I right? | 15:09 |
| * frickler sighs. we really should stop publishing images completely. just cache them for the CI | 15:11 | |
| nhk | Hope we will official release soon | 15:13 |
| opendevreview | Jack Hodgkiss proposed openstack/kolla master: Update prometheus alertmanager to newest available version https://review.opendev.org/c/openstack/kolla/+/947239 | 15:56 |
| exoceph | I hit my VM a few times and rabbitmq finally starts on all the nodes | 17:04 |
| exoceph | Is a ~12 minute start time normal for first RabbitMQ? | 19:08 |
| exoceph | Finally got a HA cluster working. I moved the controller nodes over to faster cores and it Kolla didn't time out waiting for RabbitMQ (even though it still took ~12 minutes to boot). The other cores weren't even that much slower lol. They're on the same storage too and it's plenty fast. Sharing this since this channel is logged and someone comes across the same issue. | 21:28 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!