| opendevreview | Merged openstack/ansible-role-systemd_networkd master: Ensure filename is always defined for network/netdev https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/903186 | 13:22 |
|---|---|---|
| opendevreview | Merged openstack/ansible-role-systemd_networkd master: Fix defenition of multiple static routes for network https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/903232 | 13:22 |
| kleini_ | Is somewhere an example, how to create a new user in Galera cluster? openstack.osa.db_setup seems to require to create some database always. Prometheus mysqld exporter just needs a user with certain privileges. | 13:51 |
| *** kleini_ is now known as kleini | 13:51 | |
| jrosser | kleini: only if you pass it a list of databases? pass an empty list and it should skip that task | 13:53 |
| jrosser | oh | 13:53 |
| jrosser | no that’s not going to work is it | 13:54 |
| opendevreview | Merged openstack/openstack-ansible-os_cinder master: Restart cinder-purge-deleted service only on abnormal exit https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/900347 | 13:55 |
| kleini | https://opendev.org/openstack/openstack-ansible-plugins/src/branch/master/roles/db_setup/tasks/main.yml#L41 <- it always iterates _oslodb_databases, its conditions and names | 13:56 |
| jrosser | galera_additional_users looks more useful | 13:56 |
| kleini | and I need to change privileges to '*.*:PROCESS,REPLICATION_CLIENT,SELECT' | 13:56 |
| kleini | thanks, jrosser. that looks helpful | 13:57 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_cinder stable/2023.2: Restart cinder-purge-deleted service only on abnormal exit https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/903696 | 14:05 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_cinder stable/2023.1: Restart cinder-purge-deleted service only on abnormal exit https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/903697 | 14:06 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_networkd stable/2023.2: Ensure filename is always defined for network/netdev https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/903698 | 14:06 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_networkd stable/2023.2: Fix defenition of multiple static routes for network https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/903754 | 14:08 |
| deflated | Hi all, is there a way to override a repository that's added to I presume all containers? My deployment is failing as the ceph repo that is added doesnt include the correct packages but i have a private one that does, ive tried in /etc/ansible/roles/ceph/vars and it works (adding another file for that repo) but it seems by default it adds a ceph.list with a different repo that i cant seem to find a way | 14:39 |
| deflated | to override | 14:39 |
| kleini | jrosser, would it be useful to have a Prometheus mysqld_exporter playbook within openstack-ansible-ops? I think, I will finish that shortly. | 14:40 |
| noonedeadpunk | deflated: I"m prettu sure that should be possible | 14:45 |
| noonedeadpunk | But I assume you're reffering to ceph-anisble here? | 14:46 |
| jrosser | it could be ceph_client perhaps | 14:47 |
| noonedeadpunk | oh..... | 14:48 |
| jrosser | but I think it’s important to have a bit more understanding of the deployment for a proper answer | 14:48 |
| noonedeadpunk | we have that in debian.yml: https://opendev.org/openstack/openstack-ansible-ceph_client/src/branch/master/vars/debian.yml#L49 | 14:48 |
| noonedeadpunk | which means there's kinda no way to override these? | 14:48 |
| noonedeadpunk | As vars are included during runtime, so they override whatever is even in user_*.yml? | 14:49 |
| noonedeadpunk | (extra-vars) | 14:49 |
| noonedeadpunk | huh, or not... | 14:49 |
| deflated | its a bookworm deployment with ceph reef already deployed, thats where the issue comes in as ceph doesnt yet have the repo up for it | 14:50 |
| noonedeadpunk | etra-vars should still work... | 14:50 |
| jrosser | deflated: we need to know if you’re deploying ceph with OSA | 14:50 |
| jrosser | or if this is an osa you want to integrate with some other ceph | 14:50 |
| deflated | no its already deployed via cephadm | 14:50 |
| noonedeadpunk | aha, so it's only client | 14:51 |
| jrosser | ok | 14:51 |
| noonedeadpunk | have you tried to define ceph_apt_repo_url in your user_variables? | 14:51 |
| deflated | i was looking for a function like that, i couldnt find it but i imagine that will do the job, i had tried editing /ceph_client/vars/debian.yml originally just to see if it would work | 14:52 |
| jrosser | then thing to do will stuff like this | 14:53 |
| jrosser | is to go through the defaults/main.yml for the ansible role | 14:53 |
| jrosser | as that’s really the “external interface” of things that you can override in your user_variables | 14:54 |
| deflated | so ceph_apt_repo_url: url in user_variables and another run, i'll get that tried out now, thanks so much for that variable | 14:54 |
| jrosser | same is true for all the roles in OSA | 14:54 |
| noonedeadpunk | jrosser: the problem that in ceph_ansible we don't expose a lot to defaults. | 14:55 |
| noonedeadpunk | so it is really limiting options of how you can override these. Like you can't use group_vars 100% for ceph_apt_repo_url | 14:55 |
| jrosser | right | 14:56 |
| noonedeadpunk | So we probably should fix that... | 14:56 |
| jrosser | I guess this is partly because our ceph testing is excusively on Ubuntu with the upstream packages | 14:56 |
| noonedeadpunk | I was looking couple of times, but then just let it be until the usecase / bug report won't be filled... | 14:56 |
| jrosser | so it never really needs to be away from the default | 14:56 |
| noonedeadpunk | and extra vars still work... | 14:57 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-plugins stable/2023.2: Update TOX_CONSTRAINTS_FILE for stable/2023.2 https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/903059 | 17:32 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-plugins stable/2023.2: Update TOX_CONSTRAINTS_FILE for stable/2023.2 https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/903059 | 17:32 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/2023.2: [doc] Remove references of older releases https://review.opendev.org/c/openstack/openstack-ansible/+/903778 | 17:40 |
| opendevreview | Merged openstack/openstack-ansible-plugins stable/2023.2: Fix building release notes for the project https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/902887 | 17:40 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_networkd stable/2023.1: Ensure filename is always defined for network/netdev https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/903699 | 17:55 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_networkd stable/2023.1: Fix defenition of multiple static routes for network https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/903779 | 17:57 |
| mgariepy | anyone here use cyborg ? | 18:22 |
| mgariepy | or have any way to manage some kind of quota over gpu/vgpu ? | 18:22 |
| opendevreview | Merged openstack/openstack-ansible-plugins stable/2023.2: Update TOX_CONSTRAINTS_FILE for stable/2023.2 https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/903059 | 18:30 |
| opendevreview | Merged openstack/openstack-ansible-tests stable/2023.2: Update TOX_CONSTRAINTS_FILE for stable/2023.2 https://review.opendev.org/c/openstack/openstack-ansible-tests/+/903068 | 18:53 |
| spatel | anybody using this driver to deploy k8s? - https://vexxhost.github.io/magnum-cluster-api/user/getting-started/#creating | 20:25 |
| jrosser | spatel: yes, I have patches you can try for the whole thing with OSA | 20:26 |
| spatel | How does this thing work? | 20:26 |
| jrosser | well cluster_api is a thing that runs in k8s and understands how to manage the lifecycle of other k8s clusters | 20:27 |
| jrosser | the thing you linked is a driver for magnum to make it use cluster_api | 20:27 |
| spatel | I am deploying it with my lab kolla-ansible and getting this error https://paste.opendev.org/show/bfiowewnIEQQyfpg8w22/ | 20:27 |
| jrosser | well I can’t really help you with kolla | 20:28 |
| spatel | I know but just trying to understand workload here.. | 20:28 |
| jrosser | but magnum conductor needs the control plane k8s cluster admin credentials | 20:28 |
| spatel | Do i need initial k8s to deploy magnum with capi ? | 20:29 |
| jrosser | and that error is that it doesn’t have it | 20:29 |
| jrosser | yes | 20:29 |
| jrosser | I believe that kolla leave that up to you | 20:29 |
| spatel | what really? | 20:29 |
| jrosser | but my patches to osa deploy that for you | 20:29 |
| spatel | where do you create that cluster? | 20:29 |
| jrosser | so…… | 20:29 |
| jrosser | in lxc control plane in my patches | 20:29 |
| jrosser | I have make this all work in osa AIO | 20:30 |
| spatel | I will try with OSA but just explain me how this thing required k8s to run k8s :O | 20:30 |
| spatel | what is the initial k8s cluster requirement? | 20:31 |
| jrosser | it’s pretty much lots of moving parts | 20:31 |
| jrosser | in what way? | 20:31 |
| spatel | In my case do I need to deploy k8s outside kolla right? | 20:33 |
| spatel | How that k8s will talk to my magnum ? | 20:33 |
| jrosser | ok hold on | 20:33 |
| jrosser | I made a really nice diagram that shows all the parts | 20:33 |
| spatel | I am very confused with entire workflow :( | 20:33 |
| jrosser | right I am sure | 20:34 |
| spatel | I love to see your diagram | 20:34 |
| jrosser | diagram is in my work laptop | 20:34 |
| spatel | :( | 20:34 |
| jrosser | can’t do it now unfortunately | 20:34 |
| jrosser | but I will defiantly share it | 20:34 |
| spatel | Thank you :) | 20:34 |
| spatel | can you explain me briefly what is the use of initial k8s cluster? | 20:35 |
| spatel | Why that is not mentioned anywhere in offical doc - https://vexxhost.github.io/magnum-cluster-api/user/getting-started/ | 20:36 |
| jrosser | because that’s really cluster_api stuff | 20:36 |
| jrosser | which runs in k8s always | 20:36 |
| spatel | hmm | 20:38 |
| spatel | How does outside k8s will talk to magnum containers? | 20:38 |
| spatel | Do I need to keep them on same network? | 20:38 |
| jrosser | cluster api YMHA’s drivers for openstack/azure/AWS/whatever | 20:39 |
| spatel | that is where my brain fart :( | 20:39 |
| jrosser | argh I’m on my phone sorry this is too hard now | 20:39 |
| jrosser | another time? | 20:39 |
| mgariepy | i'm also interested can you ping me as well when you start talking about it ? D | 20:40 |
| spatel | Sure!! I know :) send me diagram when you have available. Thank you for the reply | 20:40 |
| spatel | mgariepy :) | 20:40 |
| mgariepy | i have more and more pressure for k8s stuff.. id like to see what it does and what it doesnt | 20:41 |
| spatel | mgariepy same here, my company want to run huge workload on k8s so I am all in love with k8s | 20:44 |
| spatel | but k8s doesn't love me :( | 20:44 |
| spatel | so freaking confusing when I read more about it | 20:44 |
| mgariepy | lol | 20:44 |
| opendevreview | Merged openstack/openstack-ansible-os_ironic stable/2023.2: Updated from OpenStack Ansible Tests https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/903071 | 22:04 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!