January 2025 - openstack-discuss - lists.openstack.org

[manila] Feb 5th and 6th mid cycle
by Carlos Silva 30 Jan '25

30 Jan '25

Hello, zorillas and interested stackers, Manila's 2025.1 Epoxy mid cycle is right around the corner. We will meet next on February 5th and 6th, from 14 to 16 UTC on meetpad [1]. Please add your topics to the mid cycle etherpad [2]. We will use the second day of the mid cycle for collaborative review sessions, so in case you are proposing a feature or would like to have reviewers attention in a specific change, please bring it up and walk us through your change. [1] https://meetpad.opendev.org/manila-epoxy-mid-cycle [2] https://etherpad.opendev.org/p/manila-epoxy-mid-cycle Looking forward to it, carloss

1 0

OS-EXT-SRV-ATTR incomplete in compute api
by Jan 30 Jan '25

30 Jan '25

Hi, is there a reason why some member variables of the nova instance object are returned by the api and some not. For example the "os_type" value is not returned but the "kernel_id" (OS-EXT-SRV-ATTR:kernel_id) is. I have a use case where it would be very helpful to get the "os_type" via the api and python sdk. Is there a chance that a patch implementing this would be accepted? Regards, Jan

3 4

Skip upgrades
by engineer2024 29 Jan '25

29 Jan '25

Is there a way to upgrade openstack older versions like ussuri , victoria, by migrating just the nova, neutron dbs, schema to the latest version by skipping one level at once and then install the latest venvs against the latest migrated galera dbs. On the compute nodes, if we preserve just the var lib nova instances and on top of it reinstall the compute and neutron agents, will it work ? I am asking becos, I don't know what r the dependencies among the major nova versions. Are the db schema and data, the only diffs from one version to another, or is there any thing else ?

5 5

[keystone] Weekly meeting canceled 29-Jan-25
by Dave Wilde 29 Jan '25

29 Jan '25

Hi folks, Apologies for the late notice but I need to cancel the Keystone weekly meeting today due to a scheduling conflict. Please reach out either here or via IRC if you need anything. Thanks!! /Dave

1 0

[rabbitmq] Need help about RabbitMQ quorum HA mode
by İzzettin Erdem 29 Jan '25

29 Jan '25

Hi, In kolla-ansible yoga environment, rabbitmq queues are set to quorum by default instead of classical mirroring, as it seems they are not replicated (per queue has +0 quorum). We tried to find a solution to handle it in configs, but we couldn't find it and we set queues to be replicated by using CLI commands (rabbitmq-queues grow). Now, my questions are: 1- Why do all rabbitmq queues come as non-replicated by default? 2- How can we set HA mode to quorum queues?

5 5

Nova
by engineer2024 29 Jan '25

29 Jan '25

Does nova support flavors where vcpus of a single flavor can be in mixed modes where some of them are pinned and the rest dedicated ? Say, a flavor has 5 vcpus, I want 2 of them to be of dedicated type and the remaining 3 vcpus to be shared

2 2

[neutron][upgrades] IPv6 Prefix Delegation 'dibbler' driver removal
by Brian Haley 28 Jan '25

28 Jan '25

Hi again Stackers, In 2023, during the Bobcat cycle, the Neutron team changed the IPv6 Prefix Delegation (IPv6 PD) 'dibbler' driver to be an experimental feature [0]. The reason was due to it (dibbler) becoming unmaintained/retired upstream, as well as the low level of testing support we had in Neutron for the feature. So what does this mean? We (the neutron cores) have come to the conclusion that it is time to take the next step and remove the driver from the tree [1]. The framework to support IPv6 PD is *not* being removed, but as this is the only in-tree driver the functionality will no longer be available to end-users. We plan on doing this in the current Epoxy cycle. We realize that there may be deployments using IPv6 PD, although we have not to my knowledge ever received a bug against it, so we think any usage is very small. If there is desire to add an in-tree driver back to the code base, there is an open OVN-related bug for that at [2], we have just not had the interest from downstream customers to warrant the investment in this area. Thanks for reading, -Brian, Neutron PTL [0] https://review.opendev.org/c/openstack/neutron/+/879030 [1] https://review.opendev.org/c/openstack/neutron/+/934283 [2] https://bugs.launchpad.net/neutron/+bug/1895972

1 0

[Octavia][magnum] change timeout value
by Satish Patel 28 Jan '25

28 Jan '25

Folks, I am running magnum-cluster-api (mcapi) with openstack to spin up k8s but now I want to change octavia LB timeout. Is there a way to tell magnum to change the timeout value of LB? loadbalancer.openstack.org/timeout-client-data: "120000" loadbalancer.openstack.org/timeout-member-data: "120000"

1 0

[Nova] Nova-compute service flapping on Antelope
by Jakub Darmach 28 Jan '25

28 Jan '25

Hello, Recently I encountered an interesting issue - nova-compute service started to temporarily lose Rabbit connectivity to regain it after a few seconds on Antelope. The issue was replicated on the test environment, without specific steps to replicate though - it just starts to happen after some time. I pasted a log example in the bug I opened [1]. First we can see rabbit logging closing AMQP connection, followed by nova-compute reporting rabbit server being unreachable, with next message being successful reconnection. Restarting nova-compute services helps temporarily, the issue starts to manifest after a few days. Initially disconnects once every few hours, each time in shorter intervals, to the point it happens every minute or so. Did anyone encounter something similar? [1] https://bugs.launchpad.net/nova/+bug/2092297 Pozdrawiam / Best regards, *Jakub Darmach*

3 2

[administrivia] Bounce processing disabled for now
by Jeremy Stanley 28 Jan '25

28 Jan '25

Due to repeated challenges with random messages being rejected by subscribers' mail providers (from DMARC invalidations, lines too long to be forwarded, content misidentified as spam, et cetera), the time required for list administrators to manage the resulting extra work is greater than the benefit we get from Mailman being able to figure out when subscriber addresses become defunct. As such, I've disabled the "bounce processing" feature for the time being, until we can work out solutions to make it less of a burden on our list admins and less confusing/failure prone for the list subscribers. If we do decide to give it another try at some point in the future, after some of the current challenges with it get ironed out, I'll let the list know. -- Jeremy Stanley

1 0