| *** rpittau|afk is now known as rpittau | 07:15 | |
| noonedeadpunk | morning | 08:25 |
|---|---|---|
| *** sshnaidm|afk is now known as sshnaidm | 08:50 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Enable senlin tempest plugin when senlin is deployed https://review.opendev.org/c/openstack/openstack-ansible/+/754105 | 09:05 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_senlin master: DNM - test patch for senlin tempest testing https://review.opendev.org/c/openstack/openstack-ansible-os_senlin/+/754045 | 09:07 |
| opendevreview | Furkan Çelik proposed openstack/openstack-ansible-os_monasca master: Re-activate Monasca role for Openstack-Ansible Monasca role has been retired for a while. I want to activate and maintain this role. Now that Monasca can be installed by OSA. https://review.opendev.org/c/openstack/openstack-ansible-os_monasca/+/796616 | 09:42 |
| opendevreview | Furkan Çelik proposed openstack/openstack-ansible-os_monasca master: Re-activate Monasca role for Openstack-Ansible https://review.opendev.org/c/openstack/openstack-ansible-os_monasca/+/796616 | 09:45 |
| furkance | Lets add Monasca :) | 09:46 |
| opendevreview | Furkan Çelik proposed openstack/openstack-ansible-os_monasca master: Re-activate Monasca role for Openstack-Ansible https://review.opendev.org/c/openstack/openstack-ansible-os_monasca/+/796616 | 09:47 |
| opendevreview | Furkan Çelik proposed openstack/openstack-ansible-os_monasca-agent master: Re-activate Monasca Agent role for Openstack-Ansible https://review.opendev.org/c/openstack/openstack-ansible-os_monasca-agent/+/796620 | 09:51 |
| opendevreview | Merged openstack/openstack-ansible-galera_server master: Prevent mariadb from OOM killing https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/790426 | 11:03 |
| snadge | tempest neutron plugin wont install on ussuri with centos 7, because it depends on python 3 and 2.7 is installed | 11:04 |
| tosky | I would be surprised if ussuri worked at all on centos 7 | 11:05 |
| snadge | it is the last version that is supported | 11:06 |
| tosky | snadge: https://governance.openstack.org/tc/reference/runtimes/ussuri.html | 11:07 |
| tosky | the last version of the operating system or of openstack? | 11:07 |
| snadge | the last version of openstack-ansible that is supported on centos 7 | 11:07 |
| snadge | thats what I thought anyway | 11:08 |
| snadge | CentOS 7 supports only source installs, installation of OpenStack with distro packages is not supported. | 11:08 |
| tosky | I'm not really into openstack-ansible, but I guess it can be at most train: https://governance.openstack.org/tc/reference/runtimes/train.html | 11:09 |
| snadge | i think im trying to do a distro install | 11:09 |
| snadge | https://docs.openstack.org/project-deploy-guide/openstack-ansible/ussuri/overview.html | 11:09 |
| tosky | please note that while openstack-ansible itself may deploy an ussuri system on CentOS Linux 7, various services may not work later as they have not been tested on that environment | 11:10 |
| snadge | so far it is only this tempest neutron plugin | 11:14 |
| snadge | so i am electing to just not install tempest | 11:14 |
| snadge | default is source install via pip .. so it is technically supported, but yes i understand it is pushing it a bit | 11:14 |
| snadge | we are having problems with drivers on EL8 | 11:15 |
| jrosser | hold on | 11:18 |
| jrosser | OSA supports a full python3 deployment on centos-7 | 11:19 |
| jrosser | in fact we had full support for centos-7 / python3 for ussuri | 11:20 |
| snadge | yeah so that plugin is probably just a bug | 11:20 |
| jrosser | noonedeadpunk: do you see this for stable/victoria? https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/794158 | 11:21 |
| jrosser | on focal | 11:21 |
| opendevreview | Furkan Çelik proposed openstack/openstack-ansible-os_monasca-agent master: Re-activate Monasca Agent role for Openstack-Ansible https://review.opendev.org/c/openstack/openstack-ansible-os_monasca-agent/+/796620 | 11:22 |
| noonedeadpunk | jrosser: um, can't really recall, but wasn't running tempest against V lately... | 11:23 |
| jrosser | snadge: if you have an example for whats failing with your neutron plugin please make a paste | 11:25 |
| noonedeadpunk | but I can imagine it's applicable | 11:25 |
| jrosser | noonedeadpunk: i was a bit unsure with the commit message if it was fixing something in or out of CI tbh | 11:27 |
| snadge | jrosser: https://pastebin.com/JcG7JLfS | 11:28 |
| jrosser | becasue we see it out if CI here, but the commit suggests that its fixing it *in* CI | 11:28 |
| jrosser | snadge: which branch? | 11:28 |
| noonedeadpunk | jrosser: iirc it was more ci thing, as it happened only in aio/single controller with lxc | 11:29 |
| noonedeadpunk | ah, well, aio outside of the ci was also affected | 11:29 |
| snadge | i swear it was 21.2.6 | 11:30 |
| snadge | but git describe is showing train-em | 11:31 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-haproxy_server master: Replace whitelist with allowlist naming https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/796629 | 11:40 |
| admin1 | where do I set the live_migration_downtime override | 11:58 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Rename black/white list variables https://review.opendev.org/c/openstack/openstack-ansible/+/796636 | 12:05 |
| noonedeadpunk | admin1: I think it's supposed to be in nova.conf? | 12:05 |
| noonedeadpunk | in libvirt section... | 12:05 |
| noonedeadpunk | so should be smth like this http://paste.openstack.org/show/806690/ | 12:06 |
| jrosser | snadge: see this patch for Train https://github.com/openstack/openstack-ansible-os_tempest/commit/eeb3c4285768f2d575d6da5ba692afd685bf3075 | 12:10 |
| jrosser | quite possible the same thing sort of thing is needed for Ussuri to cover Centos-7? | 12:11 |
| noonedeadpunk | ebbex: if you around can you kindly vote on https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/796163 ? | 12:25 |
| *** anbanerj|rover is now known as anbanerj|lunch | 12:25 | |
| opendevreview | Merged openstack/openstack-ansible master: Remove NV distro centos 8 stream job from gates https://review.opendev.org/c/openstack/openstack-ansible/+/795962 | 12:38 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Replace systemd unit overrides with role https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/796041 | 12:40 |
| *** anbanerj|lunch is now known as anbanerj|rover | 13:24 | |
| *** Jonathan[m]5 is now known as mossblaser | 13:28 | |
| yoctozepto | noonedeadpunk: hi; you might want to answer http://lists.openstack.org/pipermail/openstack-discuss/2021-June/023104.html | 14:28 |
| noonedeadpunk | yoctozepto: oh, thanks, I'm sure want!) | 14:29 |
| * noonedeadpunk should stop filtering ML without tags | 14:32 | |
| jrosser | noonedeadpunk: html email-fail i think to the ML | 14:39 |
| noonedeadpunk | doh :( | 14:39 |
| jrosser | wierdly only on the archive though - it's readable in my mail client copy | 14:39 |
| noonedeadpunk | and it's 503 now when I try to access my mailbox ( | 14:40 |
| jrosser | sort of relatedly i was surprised by the reply here http://lists.openstack.org/pipermail/openstack-discuss/2021-June/022982.html | 14:44 |
| noonedeadpunk | `no upgrade path from any of those releases` o_O | 14:45 |
| noonedeadpunk | that's so weird discussion... | 14:46 |
| jrosser | i had similarly wierd conversation with $vendor who told me no-one really used centos for production environments | 14:50 |
| jrosser | which seems to be the same kind of thinking | 14:50 |
| noonedeadpunk | have it been after Stream announced to be a replacement or before?:) | 14:51 |
| jrosser | oh just afterwards :) | 14:52 |
| noonedeadpunk | but yeah, it's weird statements... why you even publish this if it's not supposed to be used in prod by anybody... or at least without huge warning at the beginnig of the README... | 14:53 |
| noonedeadpunk | and eventually, wasn't ooo designed to solve upgrades problem with underclouds? | 14:54 |
| mgariepy | hello spatel | 15:24 |
| frickler | funny reference, confirms my choice to stay away from ooo, rhel and centos as much as possible | 15:27 |
| frickler | and I should also answer to the first mail and mention openstack-chef and make sure it secures the bottom seat :D | 15:27 |
| mgariepy | spatel, http://paste.openstack.org/show/806696/ | 15:30 |
| mgariepy | spatel, with the haproxy config, and the way ovn seems to work what does happen if the leader is not the first server ? | 15:36 |
| spatel | sorry i was dealing with some production issue | 15:37 |
| mgariepy | no worries :) haha | 15:38 |
| spatel | haproxy doesn't care who is leader, if you make request to follower it will send request to leader for write or modification | 15:38 |
| spatel | just like rabbitMQ doing | 15:38 |
| spatel | if you tell your compute node to use ptcp:1.1.1.1:6641,ptcp:2.2.2.2:6641,ptcp:3.3.3.3:6641 then client will automatically figure out who is leader and they will talk to leader only (just like rabbitMQ) | 15:39 |
| spatel | there is two way to deal here, tell client to use haproxy or give client list of IP address of ovn | 15:40 |
| spatel | i like haproxy way so i can add / remove ip and no need to tell client about that | 15:41 |
| mgariepy | i think i would prefer the multiple ips in the config files. | 15:41 |
| mgariepy | lol | 15:41 |
| spatel | sure! either way is fine :) | 15:42 |
| spatel | give list of ips to client or handover single haproxy ip | 15:42 |
| mgariepy | if the first one listen but is out of the cluster for random reason the client will fallback to the other (i suppose) | 15:42 |
| spatel | or write smart haproxy script to detect leader and set that in top list | 15:42 |
| mgariepy | if haproxy is not configured to look a the service correctly it might cause issues. | 15:43 |
| spatel | +1 agreed | 15:43 |
| mgariepy | write a script .. another point that can fail here.. lol | 15:43 |
| jrosser | feels a bit how we use haproxy/tcp currently for galera | 15:45 |
| mgariepy | yes | 15:45 |
| jrosser | it's not really the right tool and does introduce some oddness | 15:45 |
| mgariepy | max scale would be better :D | 15:45 |
| mgariepy | but well if the config option support multiple ips, i think we should use it. just like we do for rabbit or memcache | 15:46 |
| jrosser | it does feel more of a native solution | 15:46 |
| mgariepy | indeed. | 15:47 |
| noonedeadpunk | frickler: oh, I clean forgot about puppet as well... | 16:05 |
| noonedeadpunk | mgariepy: we can't use max scale due to wierd license, that prohibits it's free usage in production system | 16:06 |
| spatel | mgariepy jrosser this is what you need to tell your compute and give all 3 IP address of ovn - http://paste.openstack.org/show/806699/ | 16:07 |
| noonedeadpunk | so we'd use proxysql I guess | 16:07 |
| *** rpittau is now known as rpittau|afk | 16:11 | |
| mgariepy | sad about maxscale. | 16:16 |
| noonedeadpunk | yeah, me too :( | 16:17 |
| noonedeadpunk | `The Licensor hereby grants you the right to copy, modify, create derivative works, redistribute, and make non-production use of the Licensed Work` | 16:19 |
| noonedeadpunk | https://mariadb.com/bsl11/ | 16:19 |
| spatel | jrosser or mgariepy can we get some +2 here so we can have this in master - https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/794266 | 16:50 |
| spatel | jamesdenton you too :) | 16:51 |
| mgariepy | here you go. :) | 16:52 |
| spatel | we can cut the patch to update client with multiple IPs so it will not depend on haproxy | 16:55 |
| spatel | noonedeadpunk do you think we should add small section of OVN cluster documentation here, like validation and removing dirty node etc.. - https://docs.openstack.org/openstack-ansible-os_neutron/latest/app-ovn.html | 16:57 |
| spatel | or any short of clustering related doc for awareness | 16:58 |
| jamesdenton | more docs the better. and nice work on this@ | 17:10 |
| jrosser | spatel: you could use your blog content as a really good starting point | 17:10 |
| jrosser | and add in more practical things like you mention about validating the setup | 17:10 |
| spatel | sure i can put reference doc in that page | 17:11 |
| mgariepy | also add the ssl stuff :D | 17:18 |
| mgariepy | with the pki role it wouldn't be too difficult i think. | 17:18 |
| spatel | mgariepy sure! i will verify that and patch it out | 17:21 |
| mgariepy | i also wonder if we need to pin the version of ovsdb. | 17:27 |
| mgariepy | https://docs.openvswitch.org/en/latest/ref/ovsdb.7/#upgrading-from-version-2-14-and-earlier-to-2-15-and-later | 17:27 |
| mgariepy | anyway let's cross that bridge when we get there. | 17:27 |
| mgariepy | wow. i wish the manual repair wouln't be needed .. lol | 18:02 |
| snadge | im still getting this strange configuration related issue which I haven't been able to pinpoint, maybe it is inventory related, here is the log though: https://pastebin.com/7gGLqKRs | 23:05 |
| snadge | thats from running setup_infrastructure.yml | 23:10 |
| snadge | oh dear.. i forgot to bootstrap which installs python 3.. its just been a bit like that lately | 23:21 |
| snadge | so i was still running from the old bootstrap.. whoops | 23:22 |
| snadge | this is why we don't skip steps ;) | 23:22 |
| snadge | it could have bailed out and said.. hey we need python 3.. and it did that last night when the tempest neutron plugin failed to install | 23:36 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!