December 2024 - openstack-discuss

[placement] isolating aggregates on kolla-ansible
by przemyslaw.kuczynski＠ochk.pl 18 Dec '24

18 Dec '24

Did someone manage to configure Windows host aggregate ? I tried same is in https://docs.openstack.org/nova/2024.2/reference/isolate-aggregates.html 1. Created 2 aggregates in 2 AZ , added 1 host to each 2. Set the scheduler.enable_isolated_aggregate_filtering config option to true in nova.conf and restart the nova-scheduler service cat /etc/kolla-pilot/config/nova.conf [libvirt] hw_machine_type=x86_64=q35 [scheduler] enable_isolated_aggregate_filtering = True 3. Add trait CUSTOM_LICENSED_WINDOWS to the resource providers for kvm01 and kvm03 in the Placement service.openstack resource provider list +--------------------------------------+-------+------------+--------------------------------------+----------------------+ | uuid | name | generation | root_provider_uuid | parent_provider_uuid | +--------------------------------------+-------+------------+--------------------------------------+----------------------+ | 36c0ae68-cc5f-4219-94bb-e46bcd3311d9 | kvm01 | 257 | 36c0ae68-cc5f-4219-94bb-e46bcd3311d9 | None | | 8abb4b5e-0407-405d-97d1-7a03a9a06cbe | kvm03 | 129 | 8abb4b5e-0407-405d-97d1-7a03a9a06cbe | None | | 14ed0962-d8ce-4bed-b90a-243b3a069f5d | kvm02 | 1096 | 14ed0962-d8ce-4bed-b90a-243b3a069f5d | None | | 8714da8e-7e9b-4823-93bb-df819c143e99 | kvm04 | 87 | 8714da8e-7e9b-4823-93bb-df819c143e99 | None | +--------------------------------------+-------+------------+--------------------------------------+----------------------+ openstack --os-placement-api-version 1.6 trait create CUSTOM_LICENSED_WINDOWStraits=$(openstack --os-placement-api-version 1.6 resource provider trait list -f value 36c0ae68-cc5f-4219-94bb-e46bcd3311d9 | sed 's/^/--trait /')openstack --os-placement-api-version 1.6 resource provider trait set $traits --trait CUSTOM_LICENSED_WINDOWS 36c0ae68-cc5f-4219-94bb-e46bcd3311d9traits=$(openstack --os-placement-api-version 1.6 resource provider trait list -f value 8abb4b5e-0407-405d-97d1-7a03a9a06cbe | sed 's/^/--trait /') openstack --os-placement-api-version 1.6 resource provider trait set $traits --trait CUSTOM_LICENSED_WINDOWS 8abb4b5e-0407-405d-97d1-7a03a9a06cbe openstack --os-compute-api-version 2.53 aggregate set --property trait:CUSTOM_LICENSED_WINDOWS=required WindowsW1 openstack --os-compute-api-version 2.53 aggregate set --property trait:CUSTOM_LICENSED_WINDOWS=required WindowsW2 Flavor metadata trait:CUSTOM_LICENSED_WINDOWS with value required nova.exception.NoValidHost: No valid host was found. There are not enough hosts available. Got no allocation candidates from the Placement API.

1 0

#eventlet-removal Thank you for 2024
by Herve Beraud 18 Dec '24

18 Dec '24

You are no doubt aware that we are currently facing a real challenge. We are living in a complex situation where we have to engage resources to exit from a problematic technology, eventlet, without especially gaining new features. This migration, although without immediate direct functional value, represente for us a considerable investment in time and resources. But such context is not far different from the Python 2/Python 3 migration. Remember that era. The core Python maintainers found themselves in a difficult situation: continue to maintain their code on Python 2, which had become insecure and unsupported, or go through the effort of completely migrating their codebase and toolchain to Python 3. All the community had to make a choice. Third-party libraries, frameworks were also on the edge case, to be upgraded or replaced with Python 3 compatible equivalents. At the time of the decision, these efforts appeared to represent a "non-value-added burden". These efforts did not bring any new functionality, did not respond to any new customer need, and even slowed down other more innovative developments. The Python maintainers could have chosen the easy way. They could have chosen to continue to support Python 2 again for years, they could have chosen to live in statu-quo, but that's not what happened [1]. They chose to fight the problem and they chose to engage strong efforts in this migration. They ensured the security and the perenniality of the Python programming language. Once the migration was completed, the benefits quickly became apparent. The innovation cadence increased. The community united behind one uniq major version. Better performances. More innovations. And last but not the least, a simpler life for the distros maintainers and for end-users, and so a reinforced confidence in Python by our industry. Python is now the number 1 technology, leaving Java and C in the dust [2]. This is the lesson we must learn. Necessary actions without immediate direct functional value can ultimately hide treasures of innovation. Engagement and dedication is the key to a sustainable future. It guarantees the long-term viability of a project, reduces risks, and opens the way to faster and more peaceful innovation. Today, I can tell you that I'm proud that our community engaged the same way. I'm proud to see all the milestones we reached during this exciting year. I'm proud to observe so much engagement coming from every side, from everybody, and from all our companies. That's a strong signal of trust from our industry. Our companies support this initiative, our companies support open source, our companies invest in the future of OpenStack. On behalf of this community initiative, let me thank you very much for all your efforts and engagements in 2024. Let me wish you happy holidays and enjoy your loved ones. [1] https://devguide.python.org/versions/ [2] https://www.tiobe.com/tiobe-index/ -- Hervé Beraud Senior Software Engineer at Red Hat irc: hberaud https://github.com/4383/

1 0

[neutron] Hold the rechecks
by Yatin Karel 18 Dec '24

18 Dec '24

With the latest Ubuntu jammy kernel released, jobs running with openvswitch/linux bridge are impacted https://bugs.launchpad.net/neutron/+bug/2091990. Almost all branches are impacted, so hold any recheck until it's fixed. Other projects should also be seeing the issue when running such jobs. -- Thanks and Regards Yatin Karel

1 1

[qa] Team Organization Updates
by Martin Kopec 18 Dec '24

18 Dec '24

Hello everyone, I'm changing my work focus in the new year, therefore, I will have reduced bandwidth for QA and PTL duties for the rest of the Epoxy cycle. As a result, I'd like to announce a few changes: 1. Ghanshyam will take over moderating office hours starting in the new year. To better accommodate his timezone, the office hours will be held *bi-weekly on Wednesdays 18:00 UTC (even weeks).* The first office hour in the new year will be on *Wednesday January 8th*. 2. Ghanshyam will also act as PTL proxy for any emergencies, e.g. patches that require an approval asap. However, feel free to add me as a reviewer to patches that need a final review. 3. Luigi Toscano volunteered to help us with reviews, mainly with Tempest-related ones as he is a long-term contributor to cinder-tempest-plugin and several other projects - https://review.opendev.org/q/owner:ltoscano@redhat.com Thanks, -- Martin Kopec Red Hat EMEA IM: kopecmartin

1 0

[tc][all] No weekly meetings over the holidays (2024-12-24 and 2024-12-31)
by Goutham Pacha Ravi 17 Dec '24

17 Dec '24

Hello Stackers, As we approach the end-of-the-year holidays, please note that the OpenStack Technical Committee weekly meetings [1] previously scheduled for 2024-12-24 and 2024-12-31 are cancelled. We will resume our weekly meetings on 2025-01-07, and I will share the agenda closer to that date. Wishing everyone a happy and restful holiday season! Thank you, Goutham Pacha Ravi (gouthamr) OpenStack TC Chair [1] https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee

1 0

[openstack-ansible] Weekly meetings on Dec 24 and Dec 31 cancelled
by Dmitriy Rabotyagov 17 Dec '24

17 Dec '24

Hi, Last week we decided to cancel the weekly meetings on December 24 and December 31 2024 due to Christmas holidays as many people will not be around for the meeting. This means that the next meeting will happen only on January 7 2025. With that being said, I want to congratulate the community with the upcoming Christmas and New Year and wish to have plenty of fun during holidays! And thank for all the hard work that was put in during 2024.

1 0

[neutron] End of year meeting schedule
by Brian Haley 17 Dec '24

17 Dec '24

Hello Neutrinos: Just wanted to make sure everyone was aware of our meeting schedule for the rest of the year. As discussed in our meeting Tuesday, next week will be the last week of meetings for the year, specifically: December 13th - Drivers meeting cancelled December 16th - Neutron CI meeting December 17th - Neutron meeting December 20th - Drivers meeting cancelled December 23rd - Neutron CI meeting (?) will leave up to Yatin Meetings the week of December 23rd and December 30th are cancelled, starting up again the week of January 6th. As usual, reach out on IRC/gerrit/email with questions for me or the core team, but assume a potentially slow turn-around. I will be offline from December 19th to January 8th, checking-in occasionally to make sure there are no fires, Miguel has offered to lead the team meeting on January 7th in my absence (thanks Miguel!). -Brian, Neutron PTL

2 2

Vault cannot authorize approle
by Andy Speagle 17 Dec '24

17 Dec '24

Hey Team, I'm working with charmed openstack and it looks like I let my certs expire stupidly... and now I'm seeing this from vault: Vault cannot authorize approle Further info from the logs indicate this: unit-vault-0: 08:42:04 WARNING unit.vault/0.juju-log InternalServerError: Unable to authorize approle. This may indicate failure to communicate with the database unit-vault-0: 08:42:04 ERROR unit.vault/0.juju-log Traceback (most recent call last): File "/var/lib/juju/agents/unit-vault- 0/charm/reactive/vault_handlers.py", line 896, in client_approle_authorized vault.get_local_client() File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/tenacity/__init__.py", line 339, in wrapped_f return self(f, *args, **kw) File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/tenacity/__init__.py", line 430, in __call__ do = self.iter(retry_state=retry_state) File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/tenacity/__init__.py", line 378, in iter raise retry_exc.reraise() File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/tenacity/__init__.py", line 206, in reraise raise self.last_attempt.result() File "/usr/lib/python3.10/concurrent/futures/_base.py", line 451, in result return self.__get_result() File "/usr/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result raise self._exception File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/tenacity/__init__.py", line 433, in __call__ result = fn(*args, **kwargs) File "/var/lib/juju/agents/unit-vault-0/charm/lib/charm/vault.py", line 258, in get_local_client client.auth_approle(app_role_id) File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/hvac/v1/__init__.py", line 2072, in auth_approle return self.auth('/v1/auth/{0}/login'.format(mount_point), json=params, use_token=use_token) File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/hvac/v1/__init__.py", line 1726, in auth return self._adapter.auth( File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/hvac/adapters.py", line 159, in auth response = self.post(url, **kwargs).json() File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/hvac/adapters.py", line 103, in post return self.request('post', url, **kwargs) File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/hvac/adapters.py", line 233, in request utils.raise_for_error(response.status_code, text, errors=errors) File "/var/lib/juju/agents/unit-vault-0/.venv/lib/python3.10/site- packages/hvac/utils.py", line 39, in raise_for_error raise exceptions.InternalServerError(message, errors=errors) hvac.exceptions.InternalServerError: internal error The database seems to be up and accessible... so, I gotta believe it's a cert issue... so, I feel like this might be a chicken-egg issue where it can't connect to the DB due to certs... but then can't issue certs because it can't connect to the DB. Any thoughts? ----- Andy Speagle

2 4

[neutron] bug deputy report for week of 2024-12-09+
by Ihar Hrachyshka 16 Dec '24

16 Dec '24

Hi all, Bugs ------- Gate issues ========= 1. fullstack failures in l3ha agent, placeholder: https://bugs.launchpad.net/neutron/+bug/2091351 May be related to https://bugs.launchpad.net/neutron/+bug/2083609 There's a fix for the latter: https://review.opendev.org/c/openstack/neutron/+/937758 2. functional suite OOMing when CPU number is low: https://bugs.launchpad.net/neutron/+bug/2091855 There's a fix to split the suite in serial runs: https://review.opendev.org/c/openstack/neutron/+/937759 The bug alludes that memory consumption could be improved; I'd recommend to limit the scope of this bug to fixing the CI failure and let the memory concerns be addressed elsewhere / by another owner if needed. 3. tempest failure in stable branch in psql job: https://bugs.launchpad.net/neutron/+bug/2091711 No fix; no owner. API / policy ========= All assigned to Slawek; not sure if he really wants to own all of them. 1. no quota on the number of tags breaks api: https://bugs.launchpad.net/neutron/+bug/2091410 DoS when too many tags are created; CVE adjacent. Need quotas for tags? 2. policy mechanism allows to work around some limitations on external connectivity when fwaas is used: https://bugs.launchpad.net/neutron/+bug/2091536 Not sure if it has a security angle or affects anything but fwaas. 3. custom policy rule failing due to missing fields: https://bugs.launchpad.net/neutron/+bug/2091493 Needs owner ========== 1. Unicast leakage on flat network: https://bugs.launchpad.net/neutron/+bug/2091211 The next step would be to set up a 2-compute devenv setup with flat and see if ml2/ovs openvswitch driver can reproduce. Additional tests to functional / fullstack / tempest validating leakage is not happening could be helpful too. Already fixed ========== 1. 500 when backend doesn't support vlan-transparent: https://bugs.launchpad.net/neutron/+bug/2091613 2. pylint change broke ovn agent delete: https://bugs.launchpad.net/neutron/+bug/2091071 New RFEs -------------- 1. use unicast mode for keepalived to avoid vxlan flooding when mcast is off: https://bugs.launchpad.net/neutron/+bug/2091599 Ihar

1 0

[Neutron] DHCP port status inactive
by Vinicius Afonso 16 Dec '24

16 Dec '24

Hello, I am new to Openstack environment and I need help with a Neutron problem. When I create a new network, the DHCP port launches with status inactive. Even though I tried to create a VM, the DHCP generate a port for the machine, However it takes a few minutes to give an error: "Build of instance ed6c8835-f2d4-4730-9638-74ded388361b aborted: Failed to allocate the network(s), not rescheduling." The only alternative solution to overcome this issue is to change the DHCP agent of the network. If I change it, the new DHCP port will have an active status and then I can launch a VM. Nothing conclusive was found in the logs, just the error to bind the port. It doesn't say anything to why the port is inactive, just that the port was created and with status active. Things that I already tried: -> See nova and neutron logs -> Restart neutron-server, neutron-dhcp-agent, neutron-openvswitch-agent -> Change manually the DHCP agent Anyone had the same problem as this one? Any tips to troubleshoot? Any help is welcome! Thanks.

2 1