| *** raukadah is now known as chkumar|rover | 04:45 | |
| *** ysandeep|out is now known as ysandeep | 05:46 | |
| *** ysandeep is now known as ysandeep|lunch | 07:30 | |
| opendevreview | Merged openstack/openstack-ansible master: Replace glance_nfs_client https://review.opendev.org/c/openstack/openstack-ansible/+/837551 | 10:09 |
|---|---|---|
| opendevreview | Merged openstack/openstack-ansible master: Add ability to define bridge type for containers https://review.opendev.org/c/openstack/openstack-ansible/+/837742 | 10:09 |
| *** ysandeep|lunch is now known as ysandeep | 10:14 | |
| *** ysandeep is now known as ysandeep|sick | 10:16 | |
| starborn | Time appropiate greetings. Is there any reason why the OpenStack-Ansible project was not present at the summit or did I just miss you guys? | 11:10 |
| mgariepy | good morning everyone | 12:20 |
| damiandabrowski[m] | hey! | 12:22 |
| mgariepy | so how was the summit? | 12:23 |
| damiandabrowski[m] | starborn: some of us were there ;) including me, noonedeadpunk, spatel and admin1 | 12:23 |
| damiandabrowski[m] | mgariepy: it was my first summit so it's hard to compare, I think it was quite small but in overall it was ok ;) | 12:24 |
| mgariepy | cool | 12:25 |
| mgariepy | i went to 2 in the past . one in boston and one in vancouver. | 12:25 |
| mgariepy | the one in vancouver was a lot smaller than the boston one. | 12:26 |
| starborn | @damiandabrowski[m] I see. I remembered the forum session in Denver 2019 and was expecting somehting similar in Berlin but maybe next year in Vancouver :) | 12:28 |
| damiandabrowski[m] | yeah, looking forward to Vancouver ;) | 12:37 |
| damiandabrowski[m] | PS. on public cloud SIG forum we had an interesting discussion about interoperability, here's an etherpad https://etherpad.opendev.org/p/berlin-summit-future-public-cloud-sig | 12:37 |
| jamesdenton | mornin' all | 13:33 |
| spatel | morning :) | 13:33 |
| jamesdenton | :) | 13:34 |
| spatel | after long time.. | 13:34 |
| jamesdenton | how was the summit? invigorating? | 13:34 |
| spatel | It was great! but i was expecting more from ovn :) they had only 2 session of ovn and it was all triple0 oriented | 13:35 |
| jamesdenton | :| | 13:36 |
| spatel | what is your plan to migrate LXB/OVS to OVN? | 13:38 |
| jamesdenton | i will have to look at my notes, i have not yet done it in production | 14:02 |
| admin1 | i will be brute forcing ovn on top of ovs this week :D | 14:09 |
| admin1 | for the sake of knowledge | 14:09 |
| spatel | admin1 go for that | 14:13 |
| admin1 | try1 is to try to brute force ovn on top of ovs and see what breaks .. so that something useful comes out to write migration scripts | 14:15 |
| admin1 | if redhat has automated it, then there is a way | 14:15 |
| mgariepy | admin1, ? https://github.com/openstack/neutron/blob/master/tools/ovn_migration/tripleo_environment/playbooks/ovn-migration.yml | 14:20 |
| admin1 | yep | 14:24 |
| admin1 | https://github.com/openstack/neutron/blob/master/tools/ovn_migration/tripleo_environment/ovn_migration.sh -- this was the demo basically .. | 14:26 |
| mgariepy | ok | 14:27 |
| mgariepy | meh the steps whatever the deploytool is must be quite similar. | 14:27 |
| spatel | migration process is not that hard just because ovn is just extension of ovs. (i am going to give it a shot) | 14:39 |
| spatel | we have to force ourself to use OVN otherwise we won't able to learn. I believe next summit will be more about OVN | 14:41 |
| jamesdenton | admin1 spatel here are my raw notes. very raw. https://paste.opendev.org/show/bAy6MUYoH2AeDLsSrInC/ | 14:47 |
| spatel | jamesdenton Thanks this is good info to start something. | 14:51 |
| spatel | what are you doing here with IRONIC ? | 14:51 |
| spatel | or this is general notes of OVN | 14:52 |
| jamesdenton | these were my notes from LXB -> OVN migration, sorry | 14:52 |
| jamesdenton | i have not gone through the process again to clean up/verify | 14:53 |
| jamesdenton | i guess i did it back in Oct 2021 | 14:53 |
| *** dviroel is now known as dviroel|lunch | 14:54 | |
| spatel | okie | 14:57 |
| *** dviroel|lunch is now known as dviroel | 16:06 | |
| opendevreview | Jesse Pretorius proposed openstack/ansible-config_template master: Remove orphaned releasenotes https://review.opendev.org/c/openstack/ansible-config_template/+/845609 | 16:52 |
| odyssey4me | jrosser_ noonedeadpunk ^ :) there he goes... doing cleanup things again ;) | 16:53 |
| jamesdenton | :o | 16:59 |
| noonedeadpunk | \o/ | 17:04 |
| admin1 | minor upgrade from 24.2.0 -> 24.3.0 fails on Fail if galera_cluster_name doesnt match provided value ... suggestions ? | 17:10 |
| admin1 | or workarounds | 17:11 |
| noonedeadpunk | are all backends healthy? | 17:12 |
| admin1 | yeah | 17:12 |
| noonedeadpunk | And what;s output of mysql --silent --skip-column-names --connect-timeout=10 -e 'SHOW VARIABLES LIKE "wsrep_cluster_name";'? | 17:13 |
| admin1 | just ran openstack-ansible galera-install.yml -e galera_ignore_cluster_state=true -e galera_upgrade=true to brute force it :D | 17:14 |
| admin1 | but i have another cluster to do the same .. will run it at that | 17:14 |
| noonedeadpunk | well... that could be the way... | 17:14 |
| noonedeadpunk | but I'd be quite cautious if things don't match | 17:14 |
| noonedeadpunk | as then cluster may fall apart udring upgrade | 17:15 |
| noonedeadpunk | or end up empty | 17:15 |
| admin1 | actually this has been coming up since 21.x tag ... | 17:15 |
| admin1 | at least for me .. | 17:15 |
| noonedeadpunk | huh | 17:15 |
| noonedeadpunk | interesting | 17:15 |
| admin1 | i just brute forced my way ahead :D | 17:15 |
| admin1 | i actually have 3-4 more prod clusters to upgrade .. will report on those | 17:16 |
| goldenfr_ | @jrosser_ Are you around? Thanks for the help on Friday, you were right it seems ansible-inventory isn't working correctly. It won't load any of my old inventory files or even a newly generated one | 19:04 |
| goldenfr_ | This is what I get, but I am clueless as to what is causing this: https://paste.opendev.org/show/bp9iurLkvcNpCe9E1L62/ | 19:04 |
| mgariepy | what does /opt/openstack-ansible/test_inventory.json contains? | 19:13 |
| goldenfr_ | it contains an inventory I generated | 19:14 |
| mgariepy | is it valid json ? | 19:14 |
| goldenfr_ | I get the same result using openstack_inventory.json | 19:14 |
| goldenfr_ | yes, they check out ok using --check | 19:15 |
| noonedeadpunk | well, you can't use json as inventory input for ansible | 19:15 |
| goldenfr_ | with inventory/dynamic_inventory.py | 19:15 |
| goldenfr_ | oh so I'm doing it wrong then | 19:15 |
| noonedeadpunk | openstack_inventory.json is just cache file which used by dynamic_inventory.py. So it can be 1. ini file. 2. yaml file. 3. executable that returns json | 19:16 |
| noonedeadpunk | in our case dynamic_inventory.py just uses openstack_inventory.json as cache and returns it when it doesn't need to be adjusted | 19:16 |
| noonedeadpunk | but you can't feed json directly | 19:17 |
| goldenfr_ | ah ok, this is my problem, none of the openstack commands work anymore: https://paste.opendev.org/show/bFmFQarBe89z8tQ76DEL/ | 19:17 |
| noonedeadpunk | Also, osa has inventory.ini, that you can use for any extra stuff that's outside of dynamic inventory | 19:17 |
| noonedeadpunk | and it will be fed to ansible as inventory file by default | 19:17 |
| goldenfr_ | ok, I don't think I've ever had an inventory.ini | 19:17 |
| noonedeadpunk | oh, what version do you runb? | 19:18 |
| goldenfr_ | rocky | 19:18 |
| goldenfr_ | its been running for years | 19:19 |
| noonedeadpunk | `Parsed /opt/openstack-ansible/inventory/inventory.ini inventory source with ini plugin` -> says you have :p | 19:19 |
| goldenfr_ | ohh I was looking in /etc/openstack_deploy | 19:19 |
| noonedeadpunk | hm | 19:19 |
| noonedeadpunk | It indeed should be /etc/openstack_deploy/inventory.ini by defaukt.... | 19:20 |
| noonedeadpunk | I don't remeber what we had in Rocky and this could be broken back then... | 19:20 |
| noonedeadpunk | So. If you jsut run /opt/openstack-ansible/inventory/dynamic_inventory.py it will print out json? | 19:20 |
| goldenfr_ | yes | 19:21 |
| noonedeadpunk | and you got this even for some test dummy playbook as well I believe? | 19:23 |
| goldenfr_ | All the openstack ansible playbooks give me the same error, I didn't test a super simple one | 19:24 |
| noonedeadpunk | So it feels that that's indeed smth with content that dynamic_inventory produced or what's cached/stored in openstack_inventory.json... | 19:25 |
| goldenfr_ | How do I fix that? | 19:25 |
| goldenfr_ | so yea ansible-playbook works fine on a simple ping script but openstack-ansible doesn't | 19:27 |
| noonedeadpunk | yeah, that's due to inventory being broken | 19:28 |
| noonedeadpunk | and I don't have simple answer how to fix it... | 19:28 |
| noonedeadpunk | I think even in Rocky, we were archieving openstack_inventory.json already | 19:28 |
| noonedeadpunk | You can try looking into archives and comparing/restoring from that "backup" | 19:28 |
| noonedeadpunk | But in Rocky there could be a bug with that as I can recall archive becoming super big for some reason... | 19:29 |
| goldenfr_ | yea I have backups of openstack_inventory.json, I tried replacing that file with older ones, but do I need to do something with the dynamic inventory? | 19:29 |
| noonedeadpunk | when you restore from backup, and run dynamic_inventory.py (just manually), is content of openstack_inventory.json changes? | 19:30 |
| noonedeadpunk | And what actually changes? | 19:30 |
| noonedeadpunk | This might give you an idea of what's wrong | 19:31 |
| noonedeadpunk | Because that basically can go to some problem inside openstack_user_config.yml | 19:31 |
| goldenfr_ | ok yea if I run dynamic_inventory.py it does change openstack_inventory.json | 19:34 |
| noonedeadpunk | I think then you should review your openstack_user_config or files inside conf.d | 19:35 |
| noonedeadpunk | they likely bring in these changes | 19:36 |
| goldenfr_ | Thanks for the help! I'll see what I can find | 19:41 |
| admin1 | noonedeadpunk, this is the output for cluster_name in a new one before minor upgrade .. https://gist.github.com/a1git/1e823f1422f2c5b8e023442420e4c884 | 19:53 |
| admin1 | still fails | 19:53 |
| goldenfr_ | woohoo! I restored an old openstack_user_config and it's working again. I still can't figure out what is wrong with the old one, it just has one removed host and an extra line | 19:58 |
| goldenfr_ | Thank you! @noonedeadpunk | 20:01 |
| noonedeadpunk | admin1: sorry, will need to look on that with fresh head | 20:07 |
| noonedeadpunk | goldenfr_: great that you got it working :) | 20:08 |
| *** dviroel is now known as dviroel|out | 21:30 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!