opendevreview | Verification of a change to openstack/kolla-ansible stable/yoga failed: [CI] Improve Cinder testing robustness https://review.opendev.org/c/openstack/kolla-ansible/+/860831 | 06:34 |
---|---|---|
opendevreview | Merged openstack/kolla-ansible stable/yoga: [CI] Improve Cinder testing robustness https://review.opendev.org/c/openstack/kolla-ansible/+/860831 | 09:32 |
opendevreview | Merged openstack/kolla master: Bump Opensearch and Opensearch Dashboards to version 2.5 https://review.opendev.org/c/openstack/kolla/+/872903 | 10:48 |
opendevreview | Maksim Malchuk proposed openstack/kayobe stable/yoga: Support configuring VLANs with systemd-networkd syntax https://review.opendev.org/c/openstack/kayobe/+/873373 | 14:29 |
opendevreview | Maksim Malchuk proposed openstack/kayobe stable/xena: Support configuring VLANs with systemd-networkd syntax https://review.opendev.org/c/openstack/kayobe/+/873374 | 14:29 |
opendevreview | Merged openstack/kolla-ansible master: Default neutron_tls_proxy and glance_tls_proxy to haproxy_tag https://review.opendev.org/c/openstack/kolla-ansible/+/871986 | 14:47 |
guesswhat[m] | Did anyone notice that Cinder backend in Glance is broken ? And Cinder is still enabled as one of default backends.. | 16:14 |
kevko | guesswhat[m]: could you describe little bit more into detail ? | 16:28 |
spatel | guesswhat[m] i just deploy cinder backend with ceph and all good | 17:07 |
spatel | kevko I have question related RabbitMQ polices, why kolla doesn't have /etc/kolla/config override based to adjust policy? | 17:08 |
kevko | spatel: w8 a minute ..going o check | 17:08 |
kevko | spatel: which version are u using ? | 17:10 |
spatel | Latest one zed release | 17:10 |
spatel | zed-15.0.0 | 17:11 |
spatel | This file has policies option /opt/venv-kolla/share/kolla-ansible/ansible/roles/rabbitmq/templates/definitions.json.j2 | 17:12 |
spatel | but its NULL "policies":[] | 17:12 |
spatel | That means no policy of HA | 17:12 |
kevko | spatel: https://github.com/openstack/kolla-ansible/blob/1b24449892580f6a553d0991755a05c19eb6b02b/ansible/roles/rabbitmq/tasks/config.yml#L100-L116 | 17:13 |
kevko | spatel: you can provide your own definitions.json if you want | 17:13 |
spatel | inside /etc/kolla/config/rabbitmq/definitions.json ? | 17:14 |
kevko | spatel: no, you will take definitions.json.j2 ...copy content and place to /etc/kolla/config/rabbitmq/definitions.json ..and define whatewer you want ...where is problem ? | 17:15 |
kevko | spatel: but ha-policy was removed because of oslo.messaging durable queus ..but i am not expert for rbmq | 17:15 |
spatel | Like this correct - https://paste.opendev.org/show/b5V1jGCqM3QAs54bxcTC/ | 17:16 |
spatel | my problem is when i reboot on of controller node making whole rabbitMQ cluster bad. I thought may be because i don't have HA so trying to put HA and see | 17:17 |
spatel | kevko Am i missing something? where is monitoring user? - https://paste.opendev.org/show/bAxIJIAXtmjsyc0HeJ3q/ | 19:35 |
kevko | spatel: there is no monitoring user if i am correct | 19:35 |
spatel | how do i login in GUI? | 19:35 |
spatel | or its disabled default? | 19:36 |
spatel | In /etc/kolla/passwords.yml i can see rabbitmq_monitoring_password: | 19:36 |
kevko | spatel: hmm, it looks like there should be if i am checking code | 19:38 |
spatel | i found this - https://review.opendev.org/c/openstack/kolla-ansible/+/517386/ | 19:38 |
kevko | spatel: you have empty rabbitmq_monitoring_user defined | 19:41 |
spatel | let say i add rabbitmq_monitoring_user: foo in global.yml | 19:41 |
spatel | Do i need to set password or it will use from password.yml? | 19:42 |
kevko | spatel: then you will have user configured | 19:42 |
spatel | oh so this is boolean variable? rabbitmq_monitoring_user: True or False? | 19:43 |
kevko | spatel: the name of user | 19:43 |
spatel | and password? | 19:43 |
kevko | passwords.yml -> rabbitmq_monitoring_password | 19:43 |
spatel | cool! trying now | 19:45 |
spatel | should i be using deploy or reconfigure option? - kolla-ansible -i multinode reconfigure -t rabbitmq | 19:45 |
spatel | it always confused me what to use for what type of config | 19:45 |
kevko | spatel: in rabbitmq it is same | 19:45 |
spatel | What is the best way to re-deploy rabbitmq cluster in kolla (Example if running cluster messed up or crash and want to redeploy?) | 19:46 |
kevko | spatel: delete containers, delete volumes, deploy again | 19:47 |
spatel | docker delete volumes ? | 19:47 |
kevko | yup | 19:47 |
spatel | i did rm -rf /var/lib/docker/volume/rabbitmq/_data/* | 19:47 |
spatel | it works!!! rabbitmq_monitoring_user: | 19:48 |
spatel | Thank you!! | 19:50 |
kevko | spatel: better is docker volume rm rabbitmq | 19:55 |
spatel | i like that instead of delete :) | 19:55 |
kevko | because some roles checking if volume exist and based on this tasks are doing something | 19:55 |
spatel | but anyway goal is to wipe out data | 19:55 |
kevko | as I said , some roles are checking if volume exist and if yes ..kolla think that this is "bootstrapped" node and doing something else | 19:56 |
kevko | for example mariaddb | 19:56 |
spatel | Hmm | 19:56 |
kevko | spatel: example -> https://github.com/openstack/kolla-ansible/blob/df12f2ce023be3a0e5486757834345e4ae874d6d/ansible/roles/mariadb/tasks/lookup_cluster.yml#L10-L13 | 19:57 |
kevko | spatel: so if you will do the same as you did with rabbitmq (delete just data ..no volume) mariadb redeploy will fail | 19:58 |
spatel | Got it :) | 19:59 |
spatel | Better to remove volume | 20:00 |
kevko | yup | 20:02 |
kevko | but best is just fix it :D | 20:03 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!