| opendevreview | Merged openstack/ansible-role-pki stable/2025.1: Use internal _pki_ca_dirs in vars https://review.opendev.org/c/openstack/ansible-role-pki/+/959107 | 03:29 |
|---|---|---|
| opendevreview | OpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata https://review.opendev.org/c/openstack/openstack-ansible/+/958974 | 03:49 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Make charset and collation configurables https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/959075 | 04:52 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_magnum master: Remove jammy and rocky-9 CI jobs https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/950699 | 04:58 |
| opendevreview | Merged openstack/openstack-ansible-os_octavia stable/2025.1: Add default values to environment lookups https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/958939 | 05:44 |
| gokhan | hello noonedeadpunk, I am trying upgrade, newly added commit throw error about installed_rabbitmq is undefined. https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/958925 | 07:11 |
| gokhan | I didn't find installed_rabbitmq fact. | 07:11 |
| noonedeadpunk | huh | 07:23 |
| noonedeadpunk | and you're running epoxy now? | 07:24 |
| noonedeadpunk | the fact is expected to be set here: https://opendev.org/openstack/openstack-ansible-rabbitmq_server/src/branch/master/tasks/rabbitmq_upgrade_check.yml#L48-L50 | 07:24 |
| noonedeadpunk | I see 2 includes of rabbitmq_feature_flags, and both of them are after this task from what I see | 07:26 |
| noonedeadpunk | so really would be interesting to see the play output in this case | 07:27 |
| noonedeadpunk | gokhan: ^ | 07:28 |
| gokhan | noonedeadpunk, ahh sorry when ı work on rabbitmq, I have removed upgrade check in rabbitmq_install playbook, I added again. it worked. I am testing epoxy upgrade | 07:37 |
| noonedeadpunk | ++ gotcha | 07:48 |
| noonedeadpunk | we should have named it with underscore to show internal nature of the variable... | 07:48 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Add debian trixie job definitions https://review.opendev.org/c/openstack/openstack-ansible/+/954616 | 14:50 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Use management address when available https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/959338 | 15:44 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Use management address when available https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/959338 | 15:44 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Use management address when available https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/959339 | 15:48 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts master: Use management_address instead of ansible_host for hosts files https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/957239 | 15:49 |
| f0o_ | is there a reason why https://opendev.org/openstack/openstack-ansible/src/branch/master/inventory/group_vars/galera_all.yml#L59-L60 doesnt define all galera nodes as backends? Galera is multi-master-able if I recall right | 16:14 |
| jrosser | f0o_: https://docs.openstack.org/large-scale/journey/configure/database.html | 16:24 |
| f0o_ | jrosser: it doesnt really go into it either, it is merely stamped off as 'This is required because OpenStack does not handle well the deadlocks when writing to different nodes in parallel.' without much as to _why_ it does not handle it. | 16:27 |
| f0o_ | my immediate mind goes into lock propagation issues but that would imply that openstack creates locks, does queries and releases locks in distinct queries/connections rather than using statements/transactions etc | 16:27 |
| f0o_ | just curious :) | 16:28 |
| *** f0o_ is now known as f0o | 16:28 | |
| noonedeadpunk | f0o: so while you can write to multiple, only one actually does process writes | 16:34 |
| noonedeadpunk | so you still better do write to the correct one | 16:34 |
| noonedeadpunk | it's just failovering way more smoothly then classic replication | 16:35 |
| noonedeadpunk | but eventually I need get to ProxySQL implementation, which I almost made 3 or 4 years ago but accidentally wiped all the progress.... | 16:35 |
| f0o | heh | 16:36 |
| noonedeadpunk | git reset --hard in a wrong directory /o\ | 16:36 |
| f0o | I just randomly found that while investigating some smaller nuisance of occasional `Got timeout reading communication packets` | 16:36 |
| noonedeadpunk | this is likely just pooling race condition | 16:37 |
| f0o | but I think that's just a natural state of running mariadb through haproxy since eventually the connection will just get terminated | 16:37 |
| noonedeadpunk | in terms that pool timeout could be == to connection timeout | 16:37 |
| noonedeadpunk | or smth in line with that | 16:37 |
| f0o | yeah I dont think this is an actual issue, just the spark to made me look into how the galera playbook actually works - instead of just using it blindly | 16:38 |
| noonedeadpunk | but ideally pool should be having slightly lower timeout, then it should be closing them normally | 16:38 |
| f0o | haproxy has 5000s 😂 | 16:38 |
| noonedeadpunk | or that | 16:38 |
| noonedeadpunk | but it should not affect anything, as before doing request oslo checks for connection aliveness and re-establish it if it's wrong | 16:39 |
| f0o | neutron doesnt | 16:39 |
| noonedeadpunk | (bbut we really need proxysql) | 16:39 |
| f0o | we've had sporadic issues were neutron is presented with a `wsrep: node not ready` error and parts of neutron-server just stop working | 16:40 |
| f0o | agents dont check in, ports cant be created. but any GET operation will just work fine | 16:40 |
| f0o | hence, 'parts'... no clue why | 16:40 |
| f0o | we just trigger a systemctl restart neutron-server when we parse it in the logs and it solves it | 16:41 |
| f0o | but its been a while since we had this, it just popped into my head and sparked intereset | 16:42 |
| f0o | also I cant spell today, being on sickleave is not doing me any good | 16:43 |
| noonedeadpunk | that sounds more like rpc issues tbh | 16:47 |
| noonedeadpunk | rather then db... but not sure | 16:47 |
| opendevreview | Dmitriy Chubinidze proposed openstack/openstack-ansible master: Fix missing security-hardening by switching to FQCN https://review.opendev.org/c/openstack/openstack-ansible/+/959398 | 23:21 |
| opendevreview | Dmitriy Chubinidze proposed openstack/openstack-ansible master: Fix missing security-hardening by switching to FQCN https://review.opendev.org/c/openstack/openstack-ansible/+/959398 | 23:22 |
| opendevreview | Dmitriy Chubinidze proposed openstack/openstack-ansible master: Fix missing security-hardening by switching to FQCN https://review.opendev.org/c/openstack/openstack-ansible/+/959398 | 23:27 |
| opendevreview | Dmitriy Chubinidze proposed openstack/openstack-ansible master: Document deployment host security hardening https://review.opendev.org/c/openstack/openstack-ansible/+/959400 | 23:35 |
| opendevreview | Dmitriy Chubinidze proposed openstack/openstack-ansible master: Document deployment host security hardening https://review.opendev.org/c/openstack/openstack-ansible/+/959400 | 23:45 |
| opendevreview | Dmitriy Chubinidze proposed openstack/openstack-ansible master: [doc] Fix missing security-hardening by switching to FQCN https://review.opendev.org/c/openstack/openstack-ansible/+/959398 | 23:47 |
| opendevreview | Dmitriy Chubinidze proposed openstack/openstack-ansible master: [doc] Document deployment host security hardening https://review.opendev.org/c/openstack/openstack-ansible/+/959400 | 23:47 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!