stan | How do you guys plan for DR? Openstack itself gives very little option to back my vm to another openstack cluster. | 01:40 |
---|---|---|
mnasiadka | thanks frickler | 04:46 |
opendevreview | Merged openstack/kolla master: keystone: Fix requests-kerberos installation https://review.opendev.org/c/openstack/kolla/+/927461 | 05:51 |
opendevreview | Michal Nasiadka proposed openstack/kolla stable/2024.1: keystone: Fix requests-kerberos installation https://review.opendev.org/c/openstack/kolla/+/927773 | 06:02 |
opendevreview | Michal Nasiadka proposed openstack/kolla stable/2023.1: keystone: Fix requests-kerberos installation https://review.opendev.org/c/openstack/kolla/+/927774 | 06:02 |
opendevreview | Michal Nasiadka proposed openstack/kolla stable/2023.2: keystone: Fix requests-kerberos installation https://review.opendev.org/c/openstack/kolla/+/927775 | 06:03 |
SvenKieske | morning o/ | 07:13 |
sylvr | morning ! :) | 07:14 |
SvenKieske | I hope all the people who can vote already voted on the technical committee election? :) | 07:38 |
opendevreview | Roman Krček proposed openstack/kolla-ansible master: Don't notify handlers during copy-cert https://review.opendev.org/c/openstack/kolla-ansible/+/924651 | 07:41 |
opendevreview | Roman Krček proposed openstack/kolla-ansible master: Rewrite kolla-ansible CLI to python https://review.opendev.org/c/openstack/kolla-ansible/+/923110 | 07:44 |
opendevreview | Roman Krček proposed openstack/kolla-ansible master: Refactor dev mode https://review.opendev.org/c/openstack/kolla-ansible/+/925714 | 07:50 |
opendevreview | Merged openstack/kolla master: Fix symlinking of openstack-base source https://review.opendev.org/c/openstack/kolla/+/924759 | 08:22 |
opendevreview | Roman Krček proposed openstack/kolla-ansible master: Reintroduce kolla-ansible check https://review.opendev.org/c/openstack/kolla-ansible/+/599735 | 08:30 |
kevko | frickler: what does 'jinxed it means' ? :D What I did ? :D | 08:35 |
kevko | morning | 08:37 |
opendevreview | Bartosz Bezak proposed openstack/kolla master: Add ovn-bgp-agent / FRR / Horizon BGPVPN dashboard https://review.opendev.org/c/openstack/kolla/+/891617 | 08:55 |
frickler | kevko: you talked bad about the CI and it reacted promptly with a gate failure ;) | 08:55 |
frickler | something is jinxed when it is the target of a bad spell from a witch or wizard. or at least I think so, beware I'm not a native speaker either | 08:58 |
kevko | frickler: aaah :D ok | 08:59 |
kevko | frickler: lastly i deployed default kolla setup with quorum queues and it works nice ...but whole stack will stop work after few hours ...because out of ram ..and OOM | 09:00 |
sylvr | Hello ! I'd like to have my monitoring host to be the same as my control host. I modified etc/kayobe/kolla.yml with : | 09:05 |
sylvr | ``` | 09:05 |
sylvr | kolla_overcloud_inventory_top_level_group_map: | 09:05 |
sylvr | control: | 09:05 |
sylvr | groups: | 09:05 |
sylvr | - controllers | 09:05 |
sylvr | - monitoring | 09:05 |
sylvr | ``` | 09:05 |
sylvr | which according to the comment in the file and the doc should map the kayobe "control" group to the kolla-ansible controllers and monitoring hosts, but I don't find any open http ports(on the kolla_VIP) allowing access to grafana, did I get the mapping the wrong way ? Thanks :) | 09:05 |
sylvr | kolla-ansible controllers and monitoring groups* | 09:11 |
SvenKieske | this is so on point (thus I advise to always "source $foo" instead of ". $foo" in bash scripts): https://morizbuesing.com/blog/greppability-code-metric/ | 09:12 |
SvenKieske | would be nice to get this merged: https://review.opendev.org/c/openstack/kolla-ansible/+/915403 | 09:34 |
opendevreview | Michal Nasiadka proposed openstack/kolla stable/2024.1: Fix symlinking of openstack-base source https://review.opendev.org/c/openstack/kolla/+/927798 | 09:43 |
kevko | SvenKieske: btw, we lastly forgot to discuss https://review.opendev.org/c/openstack/kolla/+/926744 vs https://review.opendev.org/c/openstack/kolla-ansible/+/926632 << My opinion is to just make images simpler than add some special role to fetch additional information to just render into some templates ... symlinks are normal way how to handle such | 10:10 |
kevko | stuff | 10:10 |
opendevreview | Merged openstack/kolla stable/2023.2: keystone: Fix requests-kerberos installation https://review.opendev.org/c/openstack/kolla/+/927775 | 10:11 |
opendevreview | Merged openstack/kolla stable/2024.1: keystone: Fix requests-kerberos installation https://review.opendev.org/c/openstack/kolla/+/927773 | 10:11 |
opendevreview | Merged openstack/kolla stable/2023.1: keystone: Fix requests-kerberos installation https://review.opendev.org/c/openstack/kolla/+/927774 | 10:13 |
kevko | It seems to me that it's getting quieter around here :D | 11:51 |
mnasiadka | It’s just because I’m in Korea | 12:10 |
kevko | oh ! vacation ? :) | 12:41 |
kevko | Hope that not in North Korea :D | 12:41 |
SvenKieske | openinfra asia | 12:45 |
frickler | https://2024.openinfraasia.org/ | 12:45 |
opendevreview | Michal Nasiadka proposed openstack/kolla stable/2024.1: Fix symlinking of openstack-base source https://review.opendev.org/c/openstack/kolla/+/927798 | 12:59 |
kevko | meh, i would like to go also :/ | 13:12 |
kevko | I need to ask my bosses :D | 13:12 |
kevko | folks, it's now second day i am wondering about oslo_concurrency and lock_path ...i think we are not using it in kolla in good way | 13:13 |
kevko | in 99% of cases it is totally fine ..but i am wondering how it should be | 13:14 |
kevko | frickler: lastly you gave me a bug into my head about oslo locking :/ | 13:22 |
PrzemekK | is this event free ? | 13:27 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: [DNM] Lock path to /dev/shm https://review.opendev.org/c/openstack/kolla-ansible/+/927752 | 13:50 |
SvenKieske | PrzemekK: I don't think you need to pay if you are an openstack contributor in some way. but the travel costs can be quite high I guess. I don't know if there is the possibility to get foundation funding for that | 13:53 |
PrzemekK | ok i was thinking that is also online event. Maybe some will be published to https://www.youtube.com/@OpenInfraFoundation | 14:00 |
PrzemekK | just watching something about rabbit :D https://www.youtube.com/watch?v=M6mR8YpkRu8 | 14:01 |
kevko | SvenKieske: I think we payed for summit when I was in Berlin twice | 14:03 |
SvenKieske | kevko: for the actual tickets? so last openinfra day europe was pretty certainly free if had contributions (commits merged) in the last 6 or 12 months, but I never went to an in person summit, yeah those did cost money afaik in the past. | 14:04 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: [DNM] Lock path to /dev/shm https://review.opendev.org/c/openstack/kolla-ansible/+/927752 | 14:21 |
opendevreview | Martin Hiner proposed openstack/kolla-ansible master: Move to high level client in DockerWorker https://review.opendev.org/c/openstack/kolla-ansible/+/908295 | 14:49 |
kevko | uuuuu, i found very strange stuff in kolla :D | 15:36 |
kevko | some containers has host /dev/shm ..some doesnt't have ..even if ipc_mode is private there is a /dev:/dev mounted ...oh that's mess ...:'( | 15:37 |
kevko | i know that there is a reason somewhere and somewhere no ... but it's mess :( | 15:49 |
opendevreview | Ivan Vnučko proposed openstack/kolla-ansible master: Add backend to kolla_externally_managed_cert behaviour https://review.opendev.org/c/openstack/kolla-ansible/+/927853 | 16:16 |
SvenKieske | kevko: well in general you don't want to expose arbitrary sub trees of /dev/ but I know some container need direct device access, afaik neutron for example. so I'm not sure it's a mess, but maybe badly documented? :) | 17:17 |
kevko | SvenKieske: yeah, i know ..but i am looking around stuff about queue_manager ...and if i turned off manager .. there is fanout queue with some ID ..but everytime old is removed .. when i turn on ... every restart create new queue ...and old is not deleted | 17:27 |
kevko | it's look like a bug | 17:27 |
SvenKieske | no? at least from what you describe this sounds exactly like the behaviour I would expect if queue manager is off? Maybe I misunderstood? If the manager is off, you get random uuid queue names which get recreated (with differen uuid) when removed. | 17:29 |
SvenKieske | if you use queue_manager you should get something like $hostname_%processname_$number or something like that | 17:30 |
SvenKieske | but I have only read docs and code so far, never actually tried it :D | 17:30 |
kevko | SvenKieske: yes ... | 17:33 |
kevko | SvenKieske: but what with the old queue ? | 17:33 |
opendevreview | Merged openstack/kolla stable/2024.1: Fix symlinking of openstack-base source https://review.opendev.org/c/openstack/kolla/+/927798 | 17:35 |
SvenKieske | kevko: ah I see, from which one did you migrate? quorum queues? ha? or plain old non ha queues? | 17:35 |
kevko | SvenKieske: https://ibb.co/CJzTZys | 17:35 |
kevko | SvenKieske: no, I just turned on use_queue_manager = True | 17:36 |
kevko | SvenKieske: on test | 17:36 |
SvenKieske | mhm, classic queues, good question! for the approach where I implemented quorum queues for fanout/transient I made sure to "migrate" everything in our tests/upgrade.sh wait a second.. | 17:36 |
kevko | SvenKieske: I need to go from pc ...because my bageteria will close :D ..but i have a bouncer on irc ..i am very curious | 17:37 |
SvenKieske | https://review.opendev.org/c/openstack/kolla-ansible/+/927497/5/tests/upgrade.sh | 17:37 |
SvenKieske | basically I just reintroduced reconfigure rabbitmq and rabbitmq-reset-state, this leads to downtime of course | 17:38 |
SvenKieske | but at least upgrade CI is green with this :) | 17:38 |
SvenKieske | so maybe a bad solution but it works, should also work for stream I guess | 17:38 |
SvenKieske | I also need to run now! :D | 17:38 |
SvenKieske | this is c&p from our old solution when we migrated from classic/ha queues to quorum queues :) | 17:39 |
SvenKieske | HTH | 17:39 |
SvenKieske | see you tomorrow | 17:39 |
SvenKieske | o/ | 17:39 |
kevko | SvenKieske: what I am trying to say that if you don't have manager turned on ..you restart cider volume for example and you see that queue disappeared and the new one is created with new id ... When I turn on manager ... I can see hostname-processname:number ...but if I restart ..old one persist and new one is created with new number ...did you get | 17:41 |
kevko | it ? | 17:41 |
kevko | This is not about kolla ...this is about the patches around queue manager and Oslo messaging and friends | 17:42 |
kevko | SvenKieske: I am not saying it's not working ...CI is just about functionality ..not about stability | 17:43 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!