| opendevreview | Merged openstack/openstack-ansible-ops master: Fix Grafana deployment https://review.opendev.org/c/openstack/openstack-ansible-ops/+/935520 | 10:07 |
|---|---|---|
| opendevreview | Merged openstack/openstack-ansible-os_nova master: Change ordering of /etc/ operations to improve upgrades https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/934941 | 11:39 |
| opendevreview | Merged openstack/openstack-ansible stable/2023.1: Deny access to any paths including /. for console proxies. https://review.opendev.org/c/openstack/openstack-ansible/+/935516 | 11:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Re-enable uWSGI as default for Neutron https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/935642 | 12:04 |
| opendevreview | Merged openstack/openstack-ansible-os_keystone stable/2023.2: federation: ensure cloud credentials are found on utility host https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/934534 | 12:14 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/2023.2: Bump SHAs for 2023.2 https://review.opendev.org/c/openstack/openstack-ansible/+/935647 | 12:32 |
| * jrosser looks at the ironic role | 12:44 | |
| jrosser | i wonder if the effort for doing nginx->apache there is really so different from adding a skeleton apache role to the plugins repo | 12:45 |
| jrosser | noonedeadpunk: did you think about this apache role at all? there seems a ton of stuff with all the install and mpm setup that some sort of "apache base" code would be helpful | 12:46 |
| noonedeadpunk | I tried to look for something suitable in galaxy but eventually failed to find one that would support multiple quite different vhosts | 12:48 |
| noonedeadpunk | and I'd say we should do the role in it's own repo as well.... | 12:49 |
| jrosser | perhaps we can start with something simple, just to do the install and all the boilerplate | 12:49 |
| jrosser | its like a wtf number of lines of code just to take what we put into the repo server and transplant it into the ironic role | 12:49 |
| noonedeadpunk | I beleive we can extract some base thing from ... repo or skyline... | 12:49 |
| jrosser | for simply just serving | 12:50 |
| jrosser | /httpboot | 12:50 |
| jrosser | i'm unsure what the most effective thing to do is here..... | 12:51 |
| noonedeadpunk | for my worklog apache role is kind of "for later this cycle" | 12:51 |
| noonedeadpunk | common role makes total sense | 12:51 |
| noonedeadpunk | I tried to use this one and push some proposals, but as you can see - it didn't go well https://github.com/geerlingguy/ansible-role-apache/pull/256 | 12:52 |
| noonedeadpunk | but I was actually thinking of taking it as "base" of a sort | 12:52 |
| noonedeadpunk | but quite some refactoring would be needed if to do from scratch | 12:53 |
| jrosser | so you think that taking the starting point from the todays repo server may not be the right starting point? | 12:54 |
| noonedeadpunk | and right now I'm quite urged to finalize bgp & ovn thing I left almost year ago :( | 12:54 |
| jrosser | :) | 12:54 |
| jrosser | so i was motivated to fix up the ironic role, and when i look its just a bit 8-O | 12:54 |
| noonedeadpunk | We need to have a way of supplying very arbitrary vhost configs | 12:54 |
| jrosser | well, thats why i was thinking to just ignore that totally to start with | 12:55 |
| jrosser | and do the basic setup and install | 12:55 |
| noonedeadpunk | yeah, same when I tried to deal with mpm stuff - each role does things in a very unique way | 12:55 |
| noonedeadpunk | also EL vs DEB is quite a challenging thing... | 12:56 |
| jrosser | yeah i saw, which was another reason to factor it out as it's just complexity++ | 12:56 |
| noonedeadpunk | but yeah, maybe factoring out bits from repo would be enough... | 12:57 |
| jrosser | maybe we need eventually seperate apache_base and apache_vhost roles or something | 12:57 |
| noonedeadpunk | I thnk we should be able to deal with tasks_from | 12:58 |
| jrosser | yeah of course | 12:58 |
| jrosser | let me see if i can take the common parts from repo_server | 12:58 |
| noonedeadpunk | as eventually it's gonna be quite alike to haproxy | 12:58 |
| noonedeadpunk | I can ask for a repository for the role today | 12:59 |
| jrosser | ok | 13:00 |
| noonedeadpunk | also good opprtunity to work on murano/senlin deprecation.... | 13:14 |
| kleini | I am stumbling now several times over the same issue with Galera re-deployment. The normal galera-install.yml fails and keeps failing with database out of sync state. I fix that, by placing a .my.cnf with the root database password inside the Galera container. Is my solution right? Is the galera-install.yml supposed to place that .my.cnf file? How should that work correctly? | 13:53 |
| noonedeadpunk | kleini: so eventually we stopped handling `root` password through my.cnf in favor of `admin` user as mariadb dev recommended to not use `root` and leave it alone | 13:56 |
| noonedeadpunk | so it's assumed that `root` should be allowed through socket auth rather then password | 13:56 |
| noonedeadpunk | but we didn't make any upgrade script to restore root user from password to socket auth | 13:56 |
| noonedeadpunk | you can also set `galera_root_user: root` instead of current default of `admin` | 13:57 |
| kleini | okay, then that is the issue. Mysql wants a password for root when using the unix socket | 13:58 |
| noonedeadpunk | that indeed is the case for upgraded envs.... | 13:59 |
| noonedeadpunk | jrosser: ok, I think I got now to understanding about your pointer regarding capi proxy. As now I got that it's needed to access kubernetes API when loadbalancer does not have floating IP assigned, right? | 14:04 |
| kleini | mysql_native_password must be auth_socket | 14:04 |
| noonedeadpunk | I can't recall the command altering user tbh... but yeah, default for root is not `mysql_native_password` for mariadb nowadays | 14:05 |
| jrosser | noonedeadpunk: yes - the capi proxy lets the control plane k8s call to the workload k8s API endpoint when either no fip, or isolated control plane | 14:06 |
| jrosser | and also there is different deployment needed for OVN / not-OVN | 14:07 |
| jrosser | it must be on network nodes for not-OVN, and all computes for OVN | 14:07 |
| noonedeadpunk | but how the heck does it work? As worker cluster has it's own network | 14:07 |
| noonedeadpunk | aha.... | 14:07 |
| jrosser | if you look in the generated haproxy config, the backend server configs are put into the right nertwork namespaces | 14:08 |
| * noonedeadpunk playing in ovn env there | 14:08 | |
| jrosser | there are also env vars, to allow you to always use the proxy, if you want that behaviour instead of ever using the fip | 14:09 |
| jrosser | see here for OVN, you can override the playbook target https://github.com/openstack/openstack-ansible-ops/blob/master/mcapi_vexxhost/playbooks/mcapi_proxy.yml#L23 | 14:10 |
| kleini | sorry, it must not be auth_socket but unix_socket | 14:10 |
| noonedeadpunk | yeah, trying to find in code how it does connect to networks in OVN | 14:11 |
| jrosser | right so afaik on the compute node that has the vm, you do have the namespace you need | 14:11 |
| jrosser | for dhcp or something | 14:11 |
| noonedeadpunk | the octavia VM? Or it doesn't go through the LB then? | 14:12 |
| jrosser | this is nothing to do with octavia | 14:12 |
| noonedeadpunk | iirc there's only metadata namespaces.... | 14:12 |
| noonedeadpunk | aha | 14:12 |
| jrosser | right, so this leverages those | 14:12 |
| noonedeadpunk | so it doesn't spawn own namespaces, but rely on existing ones | 14:12 |
| jrosser | correct | 14:12 |
| noonedeadpunk | ok, that does make sense | 14:12 |
| noonedeadpunk | jsut trying to asses if I'd prefer to hide in horizon menu to spawn without fip or rolling out proxy is safe enough :D | 14:13 |
| jrosser | well do consider that allowed_cidr in the cluster template is ignored | 14:14 |
| jrosser | we patched magnum-cluster-api to fail creation if a FIP was enabled | 14:14 |
| jrosser | https://github.com/vexxhost/magnum-cluster-api/issues/454 | 14:15 |
| noonedeadpunk | also AZs are not respected.... | 14:15 |
| noonedeadpunk | when there's more then 1 AZ supplied one in Horizon is simply ignored as well | 14:15 |
| jrosser | i would say that for use cases that are not private, FIP is not safe | 14:15 |
| jrosser | but that depends on local circumstances i guess | 14:15 |
| noonedeadpunk | yeah. true | 14:15 |
| jrosser | so we have ours set to always proxy | 14:15 |
| noonedeadpunk | how you do that? as I don't see variables in the role... | 14:17 |
| noonedeadpunk | `os.getenv("PROXY_ALWAYS", 0) == 0` | 14:17 |
| noonedeadpunk | found it | 14:17 |
| jrosser | https://paste.opendev.org/show/bSsnR74uaiydKGEKzHFA/ | 14:18 |
| noonedeadpunk | once I get it working, will try to document it :) | 14:18 |
| noonedeadpunk | and move that var to defaults.... | 14:19 |
| noonedeadpunk | okay, perfect, thanks! | 14:19 |
| jrosser | it might be that POD_IP is mandatory and i have forgotton it from the code/vars | 14:19 |
| jrosser | this origianted from a service running in k8s of course so it needed a little work to be OK outside that | 14:20 |
| jrosser | then we fixed the port to allow iptables to be possible, rather than random port selection | 14:20 |
| jrosser | https://github.com/bbc/magnum-cluster-api/commit/5fdc755aca57d0dc662ff994b74367d99c6874ae | 14:23 |
| kleini | many thanks, by changing the root user in Galera I don't have issues with galera-install.yml anymore | 14:24 |
| jrosser | kleini: this all came about because the old code adjusted the permissions of the root user, and with some new release of mariadb that had really strange side effects resulting in brokenness | 14:29 |
| jrosser | the advice was to never touch the root user, which is how the `admin` user came to exist now | 14:30 |
| jrosser | it does sound like we miss a step somewhere in an upgrade script though? | 14:31 |
| kleini | ALTER USER 'root'@'localhost' IDENTIFIED WITH unix_socket; GRANT ALL ON *.* TO 'root'@'localhost' WITH GRANT OPTION; DROP USER 'root'@'127.0.0.1'; DROP USER 'root'@'::1'; DROP USER 'root'@'%'; | 14:44 |
| kleini | that was necessary for my database initially created with Stein or Train. | 14:44 |
| noonedeadpunk | yeah, I had same issue back then, but already can't recall when it was. I think on 2023.1 when upgrading ubuntu from 20.04 to 22.04 | 14:56 |
| noonedeadpunk | so yes, I think we've missed covering that back then | 14:56 |
| noonedeadpunk | but also I probably won't dare to adjust root user with playbook anyway | 14:56 |
| noonedeadpunk | as who knows if there were some other usecases in deployment with it... | 14:57 |
| noonedeadpunk | #startmeeting openstack_ansible_meeting | 15:00 |
| opendevmeet | Meeting started Tue Nov 19 15:00:27 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 |
| NeilHanlon | o/ | 15:00 |
| noonedeadpunk | o/ | 15:00 |
| NeilHanlon | hello from SuperComputing 2024 | 15:00 |
| noonedeadpunk | wow, sounds quite thrilling | 15:00 |
| noonedeadpunk | #topic office hours | 15:02 |
| noonedeadpunk | so, we have landed a CI change to address intermittent rocky issues | 15:02 |
| noonedeadpunk | until infra mirrors will be in place | 15:02 |
| noonedeadpunk | so instead of mirrorlist we're using a specific mirror | 15:03 |
| * jrosser has other meeting :( | 15:04 | |
| noonedeadpunk | http://dl.rockylinux.org/ to be specific | 15:04 |
| noonedeadpunk | NeilHanlon: that's smth we can trust and bother someone from your team if it's desynced, right?:) | 15:05 |
| noonedeadpunk | also it seems that with switch to mirror.mariadb.org centos is broken | 15:07 |
| noonedeadpunk | #link https://zuul.opendev.org/t/openstack/build/5d488593758d4945b8e3efeaa9625ab3/log/logs/etc/host/yum.repos.d/MariaDB.repo.txt | 15:07 |
| noonedeadpunk | as repo doesn't have `centos-9` path - https://mirror.mariadb.org/yum/11.4/ | 15:08 |
| noonedeadpunk | it's centos/9 instead.... | 15:08 |
| noonedeadpunk | and then it's not amd64 but x86_64 /o\ | 15:08 |
| mgariepy | hello | 15:08 |
| noonedeadpunk | o/ | 15:09 |
| noonedeadpunk | eventually all distros except rocky and alma support both ways | 15:10 |
| noonedeadpunk | and I wonder if we can do just `rhel` for both centos and rocky.... | 15:11 |
| noonedeadpunk | so smth to look at at least... | 15:11 |
| noonedeadpunk | we've discussed today shortly need for apache role and repo for it | 15:11 |
| noonedeadpunk | with this I was having a question - how to name it? | 15:12 |
| noonedeadpunk | ansible-role-apache? | 15:12 |
| noonedeadpunk | ansible-role-httpd? | 15:12 |
| noonedeadpunk | I'm inclined to the second option as apache is confusing/broad name | 15:12 |
| noonedeadpunk | but not sure | 15:12 |
| mgariepy | either suits me not much to say for or against. | 15:13 |
| noonedeadpunk | jrosser: any opinion? | 15:14 |
| noonedeadpunk | on ansible-role-(apache|httpd) | 15:14 |
| noonedeadpunk | with asking the new repo will also work on sunsetting murano/senlin. and probably reviving watcher | 15:15 |
| noonedeadpunk | Today also looked/briefly tested Neutron uWSGI mode, which seems to be finally fixed for 2024.2 | 15:17 |
| noonedeadpunk | and also for OVN | 15:17 |
| noonedeadpunk | #link https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/935642 | 15:17 |
| noonedeadpunk | it doesn't seem to work for distro path though.... | 15:18 |
| noonedeadpunk | there're simply no such services so far: https://zuul.opendev.org/t/openstack/build/26fc4af095b34c6eaa8d47772bfc3af4/log/logs/host/neutron-ovn-maintenance-worker.service.journal-12-56-05.log.txt | 15:20 |
| noonedeadpunk | could be missing some packages.... | 15:20 |
| noonedeadpunk | but well... jammy is also having old packages iirc.... | 15:21 |
| noonedeadpunk | it's still 2024.1 or so | 15:21 |
| noonedeadpunk | #link https://zuul.opendev.org/t/openstack/build/26fc4af095b34c6eaa8d47772bfc3af4/log/logs/etc/host/apt/sources.list.d/uca.sources.txt#34 | 15:21 |
| noonedeadpunk | likely we need to add the noble distro job as well | 15:22 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Re-enable uWSGI as default for Neutron https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/935642 | 15:25 |
| noonedeadpunk | eventually... distro method is marked as unsupported for 2024.2 for jammy anyway... | 15:25 |
| noonedeadpunk | but will double-check on that | 15:28 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Switch Ubuntu distro jobs to 24.04 https://review.opendev.org/c/openstack/openstack-ansible/+/935664 | 15:34 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Re-enable uWSGI as default for Neutron https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/935642 | 15:34 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Use rhel packages for all EL-based distros https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/935665 | 15:41 |
| noonedeadpunk | I am also failing to work on rabbitmq force_bootstrap flag and further refactoring of haproxy role | 15:41 |
| noonedeadpunk | with that time for branching 2024.2 is coming really rapidly | 15:42 |
| jrosser | ansible-role-httpd is fine | 15:42 |
| noonedeadpunk | so I guess I would suggest to land what we have today and branch. and backport whatever we decide we need after branching | 15:42 |
| noonedeadpunk | as I don't see too much of backports, except if you wanna land apache/httpd role this cycle jrosser | 15:43 |
| jrosser | if we can get something simple for a base config i think that would be beneficial | 15:48 |
| jrosser | i have already cut/paste from the repo server into a temporary repo and have some time in a AIO now to integrate it | 15:49 |
| noonedeadpunk | aha, ok... then question is - how fast repo will be created... | 15:49 |
| jrosser | even if we leave generation of vhost till some later time | 15:49 |
| noonedeadpunk | ok, let's try | 15:49 |
| noonedeadpunk | we heave release deadline in 2 weeks | 15:50 |
| jrosser | hmmm this is maybe somewhat optimistic | 15:50 |
| noonedeadpunk | yeah | 15:50 |
| jrosser | what is for sure is that ironic role is broken for metal deploys | 15:50 |
| jrosser | on rocky/centos | 15:50 |
| jrosser | so perhaps there is some quick fix there | 15:51 |
| jrosser | as an alternative | 15:51 |
| noonedeadpunk | quick fix sounds like what we need now.... | 15:51 |
| jrosser | example is here https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/933010 | 15:52 |
| jrosser | and fairly obvious fail to re-use port 80 https://zuul.opendev.org/t/openstack/build/52b7f07374cb4f22832cf22fe74c3593/log/logs/host/nginx.service.journal-14-26-18.log.txt | 15:53 |
| noonedeadpunk | I guess that's could be just a different path? https://opendev.org/openstack/openstack-ansible-os_ironic/src/branch/master/tasks/ironic_conductor_post_install.yml#L102-L108 | 15:56 |
| noonedeadpunk | ah https://zuul.opendev.org/t/openstack/build/196b2de293234c00a9aadba5868a0afc/log/logs/etc/host/nginx/nginx.conf.txt#39-40 | 15:56 |
| noonedeadpunk | maybe some kind of lineinfile to replace 80 with management_address.... | 15:57 |
| noonedeadpunk | ubuntu simply doesn't contain server{} section in nginx.conf | 15:58 |
| jrosser | https://github.com/openstack/openstack-ansible-os_ironic/blob/8980eacd6736a0df3f6259b03a17793494ecb2cb/tasks/ironic_conductor_post_install.yml#L102-L108 | 15:58 |
| noonedeadpunk | it works for ubuntu, yeah. but on EL default is part of /etc/nginx/nginx.conf | 15:59 |
| jrosser | is this that we need to handle the path to the conf file differently on ubuntu/rh? | 15:59 |
| jrosser | oh you mean that the conf file is merged | 16:00 |
| noonedeadpunk | yeah | 16:00 |
| noonedeadpunk | #endmeeting | 16:01 |
| opendevmeet | Meeting ended Tue Nov 19 16:01:54 2024 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 16:01 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/openstack_ansible_meeting/2024/openstack_ansible_meeting.2024-11-19-15.00.html | 16:01 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/openstack_ansible_meeting/2024/openstack_ansible_meeting.2024-11-19-15.00.txt | 16:01 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/openstack_ansible_meeting/2024/openstack_ansible_meeting.2024-11-19-15.00.log.html | 16:01 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Retire murano/senlin/sahara https://review.opendev.org/c/openstack/openstack-ansible/+/935670 | 16:12 |
| jrosser | that really is made hard to not have the default server on rh nginx | 16:23 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_senlin master: Retire Senlin role https://review.opendev.org/c/openstack/openstack-ansible-os_senlin/+/935673 | 16:24 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_murano master: Retire Murano role https://review.opendev.org/c/openstack/openstack-ansible-os_murano/+/935674 | 16:26 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_sahara master: Retire Sahara role https://review.opendev.org/c/openstack/openstack-ansible-os_sahara/+/935675 | 16:27 |
| noonedeadpunk | they've tried their best | 16:35 |
| noonedeadpunk | actually, I think it's about same in apache? | 16:35 |
| noonedeadpunk | ah, no | 16:36 |
| jrosser | maybe we could place a systemd dropin to override with `ExecStart=/usr/sbin/nginx -c /path/to/our/own/config` | 17:30 |
| noonedeadpunk | hehe, we can do that, sure | 17:31 |
| noonedeadpunk | won't actually some kind of `replace` work? or issue is that they do have multiple records for v4 and v6? | 17:31 |
| jrosser | there is two records yes, but we could replace | 17:31 |
| jrosser | is that safe when the package is upgraded? | 17:32 |
| noonedeadpunk | shouldn't a current version be kept by default if upgraded with ansible? | 17:33 |
| noonedeadpunk | dunno how it's in rhel though... | 17:33 |
| noonedeadpunk | hm. we've discussed apache role on some ptg, didn't we? | 17:51 |
| jrosser | perhaps we did | 18:04 |
| jrosser | for now i am making a rocky9 aio and i think that the systemd drop in will be the most robust | 18:04 |
| noonedeadpunk | I've proposed patches for makingapache repo in the meanwhile | 18:05 |
| noonedeadpunk | hm, seems smth is off even with http://dl.rockylinux.org/ | 18:08 |
| jrosser | i'm also just having my second odd failure here doing a manual aio | 18:09 |
| noonedeadpunk | https://zuul.opendev.org/t/openstack/build/6cf414ee1e344890af64f38ece47fb9c | 18:09 |
| noonedeadpunk | and https://zuul.opendev.org/t/openstack/build/3d849db128914961b2a641ec1cd03d16 | 18:09 |
| jrosser | oh dear | 18:09 |
| noonedeadpunk | here is netcat https://zuul.opendev.org/t/openstack/build/7fa360a9baf2409eb934650caf1d5147 | 18:13 |
| noonedeadpunk | NeilHanlon: any guess wtf going on with rocky mirrors?:) | 18:13 |
| jrosser | this is not right https://paste.opendev.org/show/bprusvHe3Jg7Yp8SZhSL/ | 18:15 |
| noonedeadpunk | are you testing with https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/934060 ? | 18:17 |
| noonedeadpunk | so somehow this condition is not passing? | 18:20 |
| jrosser | i just am trying master | 18:20 |
| noonedeadpunk | but we don't have 4.0 merged there.... | 18:20 |
| noonedeadpunk | `stdout': '4.0.3'` | 18:21 |
| noonedeadpunk | o_O | 18:21 |
| jrosser | right | 18:21 |
| noonedeadpunk | um... | 18:22 |
| jrosser | https://paste.opendev.org/show/bqtANv0pXO6J5oVPwJRu/ | 18:25 |
| jrosser | thats almost like only the contents of the system repo were known at the point it installed rabbitmq-server | 18:32 |
| jrosser | i just removed the package, re-ran the rabbitmq playbook and the version is now as it should be | 18:33 |
| noonedeadpunk | so we have some kind of order issue? | 18:38 |
| noonedeadpunk | I'm also amazed how bravely on "stable" release they've bumped rabbitmq to 4.0 | 18:39 |
| noonedeadpunk | as happening this to me on production - I would be very-very-very unhappy and confused.... | 18:39 |
| jrosser | i booted a rocky 9 vm - perhaps not the newest image | 18:42 |
| jrosser | just ran setup-everything | 18:42 |
| jrosser | that thats what i got | 18:42 |
| jrosser | i'm perhaps not knowing enough rocky stuff to understand if the older 9.x image is a reasonable thing to use | 18:43 |
| noonedeadpunk | yeah, dunno what to say. that's indeed super confusing | 18:43 |
| noonedeadpunk | I think it shouldn't matter if it doesn't have rabbit pre-installed. | 18:43 |
| noonedeadpunk | but still it's likely being installed inside the role | 18:44 |
| noonedeadpunk | could be as some dependency.... | 18:44 |
| jrosser | yep - i can try again tomorrow but for now i fixed it and carry on | 18:44 |
| noonedeadpunk | would be reely good to understand if this is reproducible or not... | 18:53 |
| noonedeadpunk | as https://zuul.opendev.org/t/openstack/build/72a3705c1352473bb8582a29a3572eea/log/logs/openstack/aio1-rabbit-mq-container-0199806c/dnf.rpm.log.txt says it should be fine..... | 18:54 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!