| psymin | aio deployed | 01:14 |
|---|---|---|
| hamburgler | Is there any recommendations for setting network QOS on East West traffic before private networks are created? Or is it best to look at doing this on the flavor side? | 02:59 |
| manuvakery1 | Hi I just deployed Openstack with neutron ml2-ovn (1 controller and 2 compute).. Playbook succeded but neutron server failed to start with following error. | 04:03 |
| manuvakery1 | https://www.irccloud.com/pastebin/9E8YXwXM/neutron_server_dependancy_issue | 04:04 |
| manuvakery1 | using openstack-ansible stable/xena branch, any help would be great | 04:05 |
| manuvakery1 | installing openvswitch-common fixed the issue | 05:03 |
| jrosser | manuvakery1: in openstack-ansible stable/xena, OVN would have been highly experimental | 07:15 |
| jrosser | it would be much much better to be using stable/zed where OVN is the default, and has lots of testing | 07:18 |
| manuvakery1 | jrosser: ok , thanks | 07:27 |
| jrosser | it’s still likely a bug really, but lots and lots of work was done on OVN in the later OSA releases | 07:38 |
| manuvakery1 | jrosser: I will try with zed | 08:07 |
| noonedeadpunk | mornings | 08:22 |
| jrosser | hello | 08:23 |
| noonedeadpunk | manuvakery1: I think you can get OVN on Xena if really needed, but I think that upgrade to Zed won't go really well as we didn't make an upgrade path | 08:23 |
| noonedeadpunk | I think this is main reason why we won't recommend OVN before Z | 08:24 |
| noonedeadpunk | Damn, what a hassle I had yestarday.... Offset time on controllers do sooooooo much mess... | 08:24 |
| noonedeadpunk | and it's so not easy to spot | 08:25 |
| jrosser | we completely wedged up a ceph cluster with wrong time | 08:26 |
| jrosser | that was what broke first | 08:26 |
| noonedeadpunk | oh yes, for ceph that is super crucial.... | 08:27 |
| noonedeadpunk | but it was like - 1 controller marks nova/neutron/cinder agents as active on the itself, while rest are down, and others marks this one as down and all rest are up. So depending on when in time you do API reuqest to check agent status you see completely different picture | 08:28 |
| manuvakery1 | noonedeadpunk: I was just testing it in a Vagrant env. I'll rebuild with zed | 08:28 |
| jrosser | regarding timekeeping, hardware is getting sooooo tiny these days - i have this for example https://www.oscilloquartz.com/en/products-and-services/ptp-grandmaster-clocks/sfp-pluggable-ptp-grandmasters/osa-5401-series | 08:29 |
| noonedeadpunk | and with that agents just die randomly as they go crazy about mesaging | 08:29 |
| noonedeadpunk | oh yes, looks quite amazing | 08:31 |
| jrosser | no idea how they fit entire computer and gps receiver in an sfp | 08:32 |
| noonedeadpunk | and without extra power | 08:33 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Workaround ovs bug that resets hostname with add command https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/881582 | 08:51 |
| noonedeadpunk | one more vote is needed on https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/881582?tab=change-view-tab-header-zuul-results-summary to unblock the gates | 13:06 |
| mgariepy | does it needs backport ? | 13:32 |
| mgariepy | hopefully they fix the uca pkg at some point.. | 13:41 |
| noonedeadpunk | nope it does not | 14:01 |
| noonedeadpunk | it's affecting only antelope uca at the moment | 14:01 |
| mgariepy | k | 14:02 |
| opendevreview | Merged openstack/openstack-ansible-os_neutron master: Workaround ovs bug that resets hostname with add command https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/881582 | 15:46 |
| NeilHanlon | woah jrosser, that's crazy (re: SFP GNSS receiver) | 16:42 |
| jrosser | NeilHanlon: it's incredible isnt it..... first time i saw that sort of system-in-sfp was these https://www.embrionix.com/product/subfamily/sdi-to-ip | 16:57 |
| jrosser | those are professional video interface <> IP converters in SFP form factor, so computer + pretty large FPGA | 16:58 |
| jrosser | you take a regular spf+ switch, push a load of those it and it obseletes a whole class of niche broadcast hardware | 16:59 |
| NeilHanlon | damn. broadcast is something I've always been intrigued by | 17:23 |
| NeilHanlon | nothing i've done has ever required precision time, at least, not high precision | 17:23 |
| jrosser | one of the biggest challenges is the legacy stuff | 17:25 |
| jrosser | where everything is assumed to be synchronous and has absurdly small latency requirements and absurdly small RX packet buffers for anything IP | 17:26 |
| jrosser | i guess these sfp things are a by-product of the assembly miniaturisation required for smartphones | 17:28 |
| qwebirc99852 | Hello | 18:02 |
| NeilHanlon | hiya qwebirc99852 | 18:12 |
| qwebirc99852 | I'm Getting this following error when in heat role: | 18:19 |
| qwebirc99852 | https://storyboard.openstack.org/#!/story/2010579 | 18:19 |
| qwebirc99852 | I receive this exception: | 18:25 |
| qwebirc99852 | "msg": "Must specify either a user or a group" | 18:25 |
| qwebirc99852 | I don't know if threre are some variables that is missing on default. | 18:28 |
| jrosser | qwebirc99852: did you try the patch on the storyboard bug? | 18:45 |
| jrosser | also, this is the irc channel for the openstack-ansible deployment tool (https://opendev.org/openstack/openstack-ansible), not so much the ansible modules | 18:46 |
| qwebirc99852 | I'm using openstack-ansible to deploy | 18:51 |
| jrosser | this is your own playbook though? | 18:53 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible stable/yoga: [DNM] Do not define cluster template labels https://review.opendev.org/c/openstack/openstack-ansible/+/881660 | 19:04 |
| jrosser | qwebirc99852: i think you might need to give a bit more info - i don't recognise the ansible task in the storyboard bug being from openstack-ansible | 19:06 |
| qwebirc99852 | In this role: /etc/ansible/ansible_collections/openstack/osa/roles/service_setup/tasks/setup_roles.yml | 19:15 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible master: Deprecate certbot-auto https://review.opendev.org/c/openstack/openstack-ansible/+/881579 | 19:16 |
| qwebirc99852 | The "msg" exception happed when utility container try to run the following file: /openstack/venvs/utility-26.1.0/lib/python3.10/site-packages/openstack/cloud/_identity.py:1192 | 19:17 |
| qwebirc99852 | This is the code where i'm getting error: | 19:18 |
| jrosser | that is a different path to the one here though? https://storyboard.openstack.org/#!/story/2010579 | 19:18 |
| jrosser | /usr/local/lib/python3.9/site-packages/openstack/cloud/_identity.py | 19:19 |
| qwebirc99852 | Yeah, this issue looks like my problem. But is different, letme write here the specific code where i'm getting error | 19:20 |
| qwebirc99852 | I'm getting "Must specify either a user or a group" in this code https://pastebin.com/mF77eJBa | 19:22 |
| jrosser | can you paste the actual ansible output of the task which fails? | 19:24 |
| damiandabrowski | jrosser: I'd appreciate a review here if you have some time tomorrow ;) so I can continue patching other roles | 19:30 |
| damiandabrowski | https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/876429 | 19:30 |
| qwebirc99852 | Yeah, this is the output. https://pastebin.com/LnTxXWGy | 19:33 |
| jrosser | qwebirc99852: is this a new deployment or an upgrade? | 19:52 |
| qwebirc99852 | New Deployment | 19:52 |
| admin1 | its the API returning that its not seeing the parameter that you passed .. may something changed .. openstack ansible has nothing to do with it though .. . its best to put this in the heat channel | 19:53 |
| jrosser | admin1: the error is from the ansible collection | 19:54 |
| jrosser | which is not the code of openstack-ansible, but we call the module | 19:54 |
| jrosser | and are responsible for making sure the right version of the collection & openstacksdk are present | 19:54 |
| qwebirc99852 | Oh ok, understand. | 19:55 |
| qwebirc99852 | So, thanks for your help guys | 19:55 |
| jrosser | qwebirc99852: this should work though, so it is of interest | 19:56 |
| admin1 | unless some api changed in heat and it does not see the way data is passed | 19:56 |
| jrosser | it's not code in heat :) | 19:56 |
| jrosser | anyway | 19:57 |
| admin1 | sorry .. i read the heat wrong | 19:57 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_heat stable/zed: DNM - test heat deployment for zed https://review.opendev.org/c/openstack/openstack-ansible-os_heat/+/881759 | 20:03 |
| jrosser | qwebirc99852: it's kind if late here and i can't think of anything obvious, but that patch i just made ^^ will run the heat CI tests for Zed to check if they pass or something has broken | 20:04 |
| opendevreview | Merged openstack/openstack-ansible-os_blazar master: Add uWSGI support to blazar https://review.opendev.org/c/openstack/openstack-ansible-os_blazar/+/880651 | 20:12 |
| qwebirc99852 | Thanks | 20:17 |
| opendevreview | Merged openstack/openstack-ansible master: Add support for TLS backends https://review.opendev.org/c/openstack/openstack-ansible/+/879085 | 21:01 |
| opendevreview | Merged openstack/openstack-ansible master: Fix dstat run in gates https://review.opendev.org/c/openstack/openstack-ansible/+/879355 | 21:33 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!