| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_nova stable/2023.1: Evaluate my_ip address once https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/908699 | 10:48 |
|---|---|---|
| jrosser | noonedeadpunk: i got to busy to reply to the ML but there is also pretty much confusion with 172.x and 10.x addresses in that post | 11:50 |
| jrosser | like there are two different environments being mixed up | 11:50 |
| noonedeadpunk | might be.... | 11:50 |
| noonedeadpunk | it's overall some mess frankly speaking | 11:51 |
| jrosser | indeed, perhaps just join irc and be interactive | 11:51 |
| noonedeadpunk | actually... I was thinking about how we intend rabbitmq mgmt to be used at all? | 11:51 |
| noonedeadpunk | as it's only on private frontend, right? | 11:51 |
| jrosser | should be internal only i think? | 11:52 |
| noonedeadpunk | oh, well, yes | 11:53 |
| noonedeadpunk | but then why we kinda add it to haproxy at all... | 11:54 |
| noonedeadpunk | I was always confused a bit about what we do with rabbitmq management | 11:54 |
| jrosser | well each rabbit provides a dashboard there | 11:54 |
| jrosser | and you can use any of them | 11:54 |
| noonedeadpunk | as I think out of the box it also doesn't have enough permissions for any meaningfull monitoring either... | 11:55 |
| noonedeadpunk | jrosser: but are all dashboards the same? | 11:55 |
| noonedeadpunk | they don't have anything isntance-specific? | 11:55 |
| * noonedeadpunk barely used this dashboard as SOCKS needed to get to the network | 11:56 | |
| jrosser | andrew uses this all the time for making sure upgrades are not going badly | 11:57 |
| andrewbonney | I'm just double checking, but I'm pretty sure they all have the same info. Certainly when looking at queues you can see the ones from remote nodes too | 11:57 |
| noonedeadpunk | ok, gotcha | 11:58 |
| noonedeadpunk | and you have some vpn/proxy established to openstack mgmt network or also just some SSH tunnel? | 11:58 |
| andrewbonney | Yes, nothing obvious which isn't replicated | 11:58 |
| jrosser | we have a squid http proxy that sits between the office network and the mgmt network | 11:59 |
| noonedeadpunk | mhm, got it | 11:59 |
| jrosser | it's actually two layers deep to get across a couple of dmz | 11:59 |
| jrosser | and some auth needed on the inner one | 11:59 |
| jrosser | you can setup rules in the first squid to decide which second one you use depending on some fqdn match | 11:59 |
| jrosser | thats how we get into a number of different deployments as they each have a unique dns subdomain | 12:00 |
| jrosser | a super nice side effect of that is you don't need to populate the office network with any dns info about the hosts in the deployment at all | 12:01 |
| noonedeadpunk | mhm, ok, I see. we jsut have 0 obvious connectivity to the management network. | 12:01 |
| noonedeadpunk | yeah, that is actually interesting... | 12:01 |
| jrosser | it's just the first proxy saying "use this other proxy for anything *.region1.example.com" | 12:02 |
| jrosser | and then inside that deployment things need to resolve, but it makes things much more separate/simple | 12:02 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Allow general purpose resources to be created during setup-openstack https://review.opendev.org/c/openstack/openstack-ansible/+/909411 | 12:09 |
| opendevreview | Elod Illes proposed openstack/openstack-ansible-ops stable/train: Remove grafana/roles/grafana submodule https://review.opendev.org/c/openstack/openstack-ansible-ops/+/909655 | 12:53 |
| jrosser | maybe we just need to noop the jobs on that ^ | 13:16 |
| gokhan__ | hello folks, after running setup hosts, containers on different hosts can not ping themselves. Only they can ping after host reboot. I am using jammy. what can be the reason of this issue. I encounter this problem very often | 13:18 |
| gokhan__ | there is no error in lxc-dnsmasq service | 13:18 |
| gokhan__ | I have deployed antelope 27.4.4 | 13:19 |
| gokhan__ | sorry 27.4.0 | 13:20 |
| noonedeadpunk | gokhan__: so that's reproducible? Like each time you run setup-hosts your containers loose connection? | 13:25 |
| gokhan__ | noonedeadpunk, no not like that, at first run | 13:26 |
| noonedeadpunk | um... don't know then... I think I in fact never faced an issue myself... | 13:27 |
| noonedeadpunk | so hard to tell | 13:27 |
| noonedeadpunk | sounds like some settings are rolled but not applied.... | 13:28 |
| noonedeadpunk | like `net.ipv4.ip_forward`? | 13:28 |
| noonedeadpunk | but I never seen it here frankly speaking. At least not during controllers OS re-setup... | 13:29 |
| gokhan__ | may be but I am not sure. I had this issue also on victoria. Maybe this issue is related with my env. | 13:29 |
| noonedeadpunk | this sounds pretty much weird.... | 13:31 |
| noonedeadpunk | and you're using netplan for configuring bridges and things? | 13:31 |
| gokhan__ | yes I am using netplan | 13:32 |
| noonedeadpunk | just thinking if it could be some old ifupdown that may not reload/apply networking properly | 13:35 |
| opendevreview | Merged openstack/openstack-ansible-ops stable/train: Remove grafana/roles/grafana submodule https://review.opendev.org/c/openstack/openstack-ansible-ops/+/909655 | 13:35 |
| noonedeadpunk | As then there could be just couple of things: 1. firewall 2. disabled ip_forward 3. interfaces missing on bridge | 13:35 |
| gokhan__ | there is ufw on jammy but it is inactive | 13:42 |
| gokhan__ | ip forward is enabled on sysctl.conf | 13:42 |
| gokhan__ | ı have doubts on netplan only | 13:43 |
| ThiagoCMC | Hey folks! I just experimented with the latest Ceph Ansible (the stable-8.0 branch) on Ubuntu 22.04, alongside Ceph 18, using the UCA Bobcat repositories. I'm excited to share that it's working well (though maybe you already knew that, lol)! I'm also hopeful it'll be compatible with Ubuntu 24.04 when the time comes (I'll test it soon). In addition, I've been looking into the new cephadm (also reading about cephadm-ansible) and stumbled | 14:07 |
| ThiagoCMC | upon this: https://github.com/canonical/ceph-containers - It looks like there are Ubuntu images available for cephadm! Have any of you tried it out, or have you only experimented with the CentOS version so far? | 14:07 |
| opendevreview | Merged openstack/openstack-ansible-plugins master: Add override for gluster host used for bootstrap operations https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/908981 | 14:12 |
| noonedeadpunk | ThiagoCMC: great news! | 14:13 |
| ThiagoCMC | ^_^ | 14:13 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add ovn-bgp-agent to source install requirements https://review.opendev.org/c/openstack/openstack-ansible/+/909694 | 16:32 |
| admin1 | ThiagoCMC, i use cephadm on ubuntu/debian .. | 16:52 |
| admin1 | i have dropped off "rpm's" completely | 16:52 |
| noonedeadpunk | admin1: how you've managed doing that, given that cephadm igaes can be only CentOS Stream ones ? :D | 16:55 |
| noonedeadpunk | unless you build your own ones with deb inside ofc | 16:55 |
| noonedeadpunk | and keep internal registry for these. | 16:55 |
| admin1 | i meant the host nodes :) | 17:03 |
| admin1 | i don't know/care what the podman images are :) | 17:03 |
| noonedeadpunk | well. given how often they get borked due to base packaging regressions.... | 17:04 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Add support for ovn-bgp-agent deployment https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/909780 | 17:54 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Add support for ovn-bgp-agent deployment https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/909780 | 17:55 |
| maxim | with the ceph packages from proxmox repo its stable on baremetal, inside vm and also incus. Though i was only testing them for 3 weeks each iteration | 18:13 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible master: [doc] Use bootstrap node override for gluster primary upgrade https://review.opendev.org/c/openstack/openstack-ansible/+/908982 | 19:00 |
| opendevreview | Merged openstack/openstack-ansible master: [doc] Use bootstrap node override for gluster primary upgrade https://review.opendev.org/c/openstack/openstack-ansible/+/908982 | 19:39 |
| ThiagoCMC | admin1, but is the container running Ceph itself also Ubuntu, or CentOS? | 22:45 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!