November 2022 - openstack-discuss

Re: [neutron-dynamic-routing] BGPspeaker LOOP
by Roberto Bartzen Acosta 22 Nov '22

22 Nov '22

Thanks Sean. I believe that the os-ken driver is not coded to honor this premise because the below function can learn new paths from peer update messages. https://opendev.org/openstack/os-ken/src/branch/master/os_ken/services/prot… Em qua., 16 de nov. de 2022 às 09:14, Sean Mooney <smooney(a)redhat.com> escreveu: > On Wed, 2022-11-16 at 09:05 -0300, Roberto Bartzen Acosta wrote: > > Sorry for the mistake, I meant, the bgpspeaker should only "*advertise*!" > > and not "learn" the AS_PATHs via BGP. > yes that used to be the scope of that project to advertise only and not > learn > so i would geuss either that has change recently and they broke backward > compaitbly > or they have refactord it to use an external bgp speaker like frr and it > learns by default > > i dont really see anything here > https://github.com/openstack/neutron-dynamic-routing/commits/master > im not really familar with the internals of the project but i dont see any > code to learn routs form > > > https://github.com/openstack/neutron-dynamic-routing/blob/master/neutron_dy… > > it just has code for advertizing and withdrawing routes. > > > > > > > Em qua., 16 de nov. de 2022 às 08:57, Sean Mooney <smooney(a)redhat.com> > > escreveu: > > > > > On Wed, 2022-11-16 at 08:43 -0300, Roberto Bartzen Acosta wrote: > > > > Hey folks, > > > > > > > > Please, I have a question here, the bgpspeaker should only "learn" > and > > > not > > > > "advertise" the AS_PATHs via BGP, right? > > > > > > its been a while since i looked at it but in the past it did not > supprot > > > learning at all > > > > > > it just advertised the routes for the neutron netowrks > > > > > > > > > > > In my tests, I can see that it is learning routes from BGP neighbors. > > > This > > > > behavior can cause an AS_PATH loop because the bgpspeaker learns > back its > > > > own advertised routes, and I see a message like this in the logs: > > > > > > > > 2022-11-11 19:45:41.967 7220 ERROR bgpspeaker.peer [-] AS_PATH on > UPDATE > > > > message has loops. Ignoring this message: > > > > > > > > BGPUpdate(len=91,nlri=[],path_attributes=[BGPPathAttributeMpReachNLRI(afi=2,flags=144,length=46,next_hop='2001:db7:1::1',nlri=[IP6AddrPrefix(addr='2001:db9:1234::',length=64)],safi=1,type=14), > > > > BGPPathAttributeOrigin(flags=64,length=1,type=1,value=0), > > > > BGPPathAttributeAsPath(flags=80,length=10,type=2,value=[[65001, > > > > > > > > 65000]])],total_path_attribute_len=68,type=2,withdrawn_routes=[],withdrawn_routes_len=0) > > > > > > > > This can be fixed by suppressing the neighbor route advertisement > (using > > > > route-map export), but have I misunderstood how > neutron-dymanic-routing > > > > works or do we have a possible bug here? > > > > > > > > Regards > > > > > > > > > > > > > > -- _‘Esta mensagem é direcionada apenas para os endereços constantes no cabeçalho inicial. Se você não está listado nos endereços constantes no cabeçalho, pedimos-lhe que desconsidere completamente o conteúdo dessa mensagem e cuja cópia, encaminhamento e/ou execução das ações citadas estão imediatamente anuladas e proibidas’._ * **‘Apesar do Magazine Luiza tomar todas as precauções razoáveis para assegurar que nenhum vírus esteja presente nesse e-mail, a empresa não poderá aceitar a responsabilidade por quaisquer perdas ou danos causados por esse e-mail ou por seus anexos’.*

2 3

[barbican][release] Transition Queens, Rocky & Stein to EOL
by Grzegorz Grasza 22 Nov '22

22 Nov '22

Hi all, The barbican team no longer plans to maintain branches older than Train. The patch to mark these branches as EOL is in review [1], but according to the steps to achieve this [2], we need first to announce it. Please send your feedback by leaving a comment on the patch. I will unblock it next week. / Greg [1] https://review.opendev.org/c/openstack/releases/+/862515 [2] https://docs.openstack.org/project-team-guide/stable-branches.html#end-of-l…

1 0

[all][openstack-dev][ptls] Migrating devstack jobs to Jammy (Ubuntu LTS 22.04)
by Dmitriy Rabotyagov 21 Nov '22

21 Nov '22

Hi everyone, According to a 2023.1 community-wide goal [1], base-jobs including but not limited to devstack-minimal, devstack, devstack-ipv6, devstack-multinode, tox, will be switched from Ubuntu Focal (20.04) to Jammy (22.04). That will also bring python 3.10 as a default python interpreter for these jobs. Migration will affect most of the projects, so to make it smooth, we will split the process into the following steps: 1. Patches for switching base jobs to Jammy will be proposed and marked as WIP (Work In Progress). It will allow projects to check if their code is compatible without breaking gates. These patches are ready and you can check them out here [2] 2. All projects should ensure that the code is compatible and merge changes if required. To properly track and test these changes you should: * Create a DNM patch that will contain in its commit message: "Depends-On: https://review.opendev.org/c/openstack/tempest/+/861110" to consume new nodesets. You can check the sample patch for Nova below [3] * In case of any job failure due to Ubuntu 22.04 or Python 3.10, land requires changes to support this new distribution * If you have overridden nodeset for your project jobs, ensure that you also have jobs against Ubuntu 22.04 and these jobs set to voting. * Set topic to all related changes to "migrate-to-jammy" * Please use etherpad [4] for tracking patches, bug reports, current status, or any other activity related to this topic 3. On R-18, which is the first 2023.1 milestone that will happen on the 18th of November 2022, base-jobs patches mentioned in step 1 will be merged. Please ensure you have verified compatibility for your projects and landed the required changes if any were needed before this date otherwise, they might fail. Please, do not hesitate to raise any questions or concerns. [1] https://governance.openstack.org/tc/goals/selected/migrate-ci-jobs-to-ubunt… [2] https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/861116 https://review.opendev.org/c/openstack/tempest/+/861110 https://review.opendev.org/c/openstack/devstack/+/860795 [3] https://review.opendev.org/c/openstack/nova/+/861111 [4] https://etherpad.opendev.org/p/migrate-to-jammy

2 2

[neutron] metadata IPv6
by Roberto Bartzen Acosta 21 Nov '22

21 Nov '22

Hey folks, Can you confirm if the metadata should work in an ipv6-only environment? As I understand from this discussion on LP:1460177 <https://bugs.launchpad.net/neutron/+bug/1460177> and the fork of the discussion in many opendev reviews #315604 <https://review.opendev.org/c/openstack/neutron-specs/+/315604>, #738205 <https://review.opendev.org/c/openstack/neutron-lib/+/738205> #745705 <https://review.opendev.org/c/openstack/neutron/+/745705>, ..., it seems like it should work. However, this comment in the openstack doc [1] has me questioning if it really works. *"There are no provisions for an IPv6-based metadata service similar to what is provided for IPv4. In the case of dual-stacked guests though it is always possible to use the IPv4 metadata service instead. IPv6-only guests will have to use another method for metadata injection such as using a configuration drive, which is described in the Nova documentation on config-drive <https://docs.openstack.org/nova/latest/user/config-drive.html>."* Is anyone using metadata in an ipv6-only Openstack setup? Regards, Roberto [1] https://docs.openstack.org/neutron/latest/admin/config-ipv6.html#configurin… -- _‘Esta mensagem é direcionada apenas para os endereços constantes no cabeçalho inicial. Se você não está listado nos endereços constantes no cabeçalho, pedimos-lhe que desconsidere completamente o conteúdo dessa mensagem e cuja cópia, encaminhamento e/ou execução das ações citadas estão imediatamente anuladas e proibidas’._ * **‘Apesar do Magazine Luiza tomar todas as precauções razoáveis para assegurar que nenhum vírus esteja presente nesse e-mail, a empresa não poderá aceitar a responsabilidade por quaisquer perdas ou danos causados por esse e-mail ou por seus anexos’.*

5 8

Unable to access Internet from an instance and accessing instance using floating-point IPs from external network
by vincent lee 21 Nov '22

21 Nov '22

Hi all, I have an OpenStack deployment using kolla-ansible <https://docs.openstack.org/kolla-ansible/latest/user/quickstart.html> with the Yoga version. Our network setup is provided in the attachment ( *network.jpg*). The configuration file (*globals.yml)* and inventory ( *multinode*) are also attached to this email. After the deployment steps, I try to run the script *init-runonce* to create the images and the networks by modifying the external network parameters according to our network setup. I am successful in launching instances from the horizon dashboard. However, the instances were not able to connect to the Internet. I have tried attaching the floating point IPs to the instance, where the IPs are successfully allocated from the range I have specified in the* init-runonce* script. When I launched the instance and tried to ping the external network, it failed. It will be very helpful if anyone can advise me to resolve this issue. I am currently using Ubuntu 20.04.05 LTS version on the controller and compute nodes. Best regards, Vincent

2 1

In Need of Reviewer for Proposal
by Juan Pablo Suazo 21 Nov '22

21 Nov '22

Hello All, I'm in need of a reviewer to take a look at my kolla-ansible proposal, which consists of the modification of a couple of cinder tasks to support placing multiple ceph conf files in a cinder/ceph/ directory to make integrating multiple volume services easier. I have already received insight from another reviewer ( radoslaw.piliszek(a)gmail.com) and made changes to comply with conventions and fix bugs. I couldn't tend to this proposal for a while, and in that time my reviewer has stopped reviewing patches altogether so I'm unable to get the approval for merging. All Zuul tasks are currently successful and the change has been tested locally to ensure it works correctly, so the review should be pretty straight forward. Best wishes, *Juan Pablo Suazo* Cloud Developer Trainee jsuazo(a)whitestack.com

2 2

[TripleO] Intermittent gate failure - centos-9-content-provider jobs
by Ronelle Landy 21 Nov '22

21 Nov '22

Hello All, We are investigating an intermittent check/gate failure on content provider jobs related to DNS and mirror access/resolution: tripleo-ci-centos-9-content-provider <https://zuul.opendev.org/t/openstack/build/8a3afc815e8743568acf3906c92ee203> and the -zed and -wallaby versions. Details are in the related Launchpad bug: https://bugs.launchpad.net/tripleo/+bug/1997202. We will update the list as we know more. Thanks, TripleO CI Team.

1 0

[neutron] Bug deputy report (week starting on Nov-14-2022)
by Bernard Cafarelli 21 Nov '22

21 Nov '22

Hey neutrinos, last bug deputy rotation of the year started: https://wiki.openstack.org/wiki/Network/Meetings#Bug_deputy and I was deputy last week, here are the reported bugs. Overall all bugs have patches or at least assignees Critical * [neutron-lib] Stable py3x jobs failing - https://bugs.launchpad.net/neutron/+bug/1996776 Patch and backports by ralonsoh merged - https://review.opendev.org/q/I3a4a27ec4672f8ea8848d7c04651730dae6f40ff * [CI] "test_live_migration_with_trunk" failing - https://bugs.launchpad.net/neutron/+bug/1997025 Caused by recent os-vif bump, workaround suggested by ralonsoh to block this newer version - https://review.opendev.org/c/openstack/neutron/+/865026 High * With new RBAC enabled (enforce_scope and enforce_new_defaults): 'router:external' field is missing in network list response - https://bugs.launchpad.net/neutron/+bug/1996836 Assigned to slaweq, fixed with https://review.opendev.org/c/openstack/neutron/+/865032 Also a related neutron-lib patch merged https://review.opendev.org/c/openstack/neutron-lib/+/864809 * With new RBAC enabled (enforce_scope and enforce_new_defaults): some security groups aren't visible for admin user - https://bugs.launchpad.net/neutron/+bug/1997089 Also assigned to slaweq, patch in progress https://review.opendev.org/c/openstack/neutron/+/865040 * Metadata service broken after minor neutron update when OVN 21.09+ is used - https://bugs.launchpad.net/neutron/+bug/1997092 Fix by ihrachys merged, backports in progress - https://review.opendev.org/c/openstack/neutron/+/864777 Medium * Unit test failure with Python 3.11 - https://bugs.launchpad.net/neutron/+bug/1996527 Fix by haleyb merged - https://review.opendev.org/c/openstack/neutron/+/864448 * OVN metadata randomly stops working - https://bugs.launchpad.net/neutron/+bug/1996594 Happening with xena and OVS 2.16, otherwiseguy investigating (may be caused by OVS version too low for RAFT) * QoS rules policies do not work for "owners" - https://bugs.launchpad.net/neutron/+bug/1996606 Assigned to ralonsoh * [OVN] support update fixed_ips of metadata port - https://bugs.launchpad.net/neutron/+bug/1996677 Patch in progress by huanghailun - https://review.opendev.org/c/openstack/neutron/+/864715 * Add support for DHCP option 119 (domain-search) for IPv4 in ML2/OVN - https://bugs.launchpad.net/neutron/+bug/1996759 Fix by lucasagomes merged - https://review.opendev.org/c/openstack/neutron/+/864740 * [OVN] Enabling and disabling networking log objects doesn't work as expected - https://bugs.launchpad.net/neutron/+bug/1996780 Patch in progress by elvira - https://review.opendev.org/c/openstack/neutron/+/864152 * [ovn-octavia-provider] HM created at fully populated loadbalancer stuck in PENDING_CREATE - https://bugs.launchpad.net/neutron/+bug/1997094 Assigned to froyo * ML2 When OVN mech driver is enabled dhcp extension is disabled - https://bugs.launchpad.net/neutron/+bug/1997185 When OVN is enabled with another driver (linuxbridge here), DHCP extension is disabled. Fix suggested by author to mark extension as supported in OVN - https://review.opendev.org/c/openstack/neutron/+/865081 Low * Replace the Linux Bridge references with Open vSwitch in the installation manuals - https://bugs.launchpad.net/neutron/+bug/1996772 Discussed in team meeting, this part of the doc is quite outdated Patch by ralonsoh merged - https://review.opendev.org/c/openstack/neutron/+/864748 * Install and configure controller node in Neutron - https://bugs.launchpad.net/neutron/+bug/1996889 Doc fix on domain names case, patch by haleyb - https://review.opendev.org/c/openstack/neutron/+/864921 RFE * add address scope to OVN Southbound/Northbound - https://bugs.launchpad.net/neutron/+bug/1996741 Enhancement to use ovn-bgp-agent with non-SNATed tenant networks Patch in review: https://review.opendev.org/c/openstack/neutron/+/861719 Opinion * The virtual network is broken on the node after neutron-openvswitch-agent is restarted if RPC requests return an error for a while - https://bugs.launchpad.net/neutron/+bug/1996788 ralonsoh++ jumped in on that one, issue on OVS agent start when RPC fails. Not much we can do here probably, but left as opinion for potential further discussion Incomplete * [OpenStack-OVN] Poor network performance - https://bugs.launchpad.net/neutron/+bug/1996593 Reported much lower performance with security groups enabled between ML2/OVS and ML2/OVN, more details asked on deployment setup -- Bernard Cafarelli

1 0

Kolla Ansible Image add python package.
by Nguyễn Hữu Khôi 20 Nov '22

20 Nov '22

Hello guys. I read https://docs.openstack.org/kolla/latest/admin/image-building.html but I would like to know how to install python packages for all images at building time. I would be very glad if I can get some guides. Nguyen Huu Khoi

1 0

OpenStack: Can't create instances using KVM on host.
by Seyed Amir Hossein Ahmadi 19 Nov '22

19 Nov '22

I have a Dockerize installation from Devstack all-in-one. The goal for me is to connect to the host's KVM and create instances there. Nova was configured as follows for this purpose. *# /etc/nova/nova.conf# /etc/nova/nova-cpu.conf# [libvirt]# connection_uri = qemu+ssh://root@172.10.1.1/system <http://root@172.10.1.1/system>* When I try to build the instance, I get the following error. *# Build of instance cdd6f8b4-6dcf-4a43-b96a-fb6166b20235 aborted: Failed to allocate the network(s), not rescheduling.* ovs-vsctl commands cause the error. What is the problem? Does this need to be done differently?

1 0