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

Openstack kolla networking setup
by matek7323＠interia.pl 10 Jan '25

10 Jan '25

I am setting up an Openstack Kolla environment with the goal of moving all virtualizations into one place At the moment I have 3 management controllers and 5 compute node for virtualization + CEPH I have there such settings as eth0 management network 172.16.50.0 and eth1 as external neutron 91.1.1.0/24 as public address And now I want to achieve two things: 1) How to make the vmka get the public address directly without natting it on the principle. Private address --> 10.224.1.15 --> public 91.1.1.15 2) Leave the option of nating after through the vrouter so that the machine assigns a private address and then a public IP address To all these servers for now the situation looks like this Port eth0 = vlan 750 access on the switch - MGMT port eth1 = vlan 888 access on switch - VLAN to operator so that I can broadcast public IP address network: version: 2 ethernets: eth0: match: macaddress: “***********” addresses: - “172.16.50.11/24” nameservers: addresses: - 8.8.8.8 set-name: eth0 routes: - to: default via: 172.16.50.254 eth1: match: macaddress: “**********” dhcp4: false dhcp6: false set-name: eth1 Each of the servers has the same config and now how to configure this to adjust the config of the coll, the servers as well as the openstack to be able to work natively and on the principle of assigning IP addresses directly to the machine by vrouter Translated with DeepL.com (free version)

1 0

[cinder][kubernetes] pvc stuck in reserved state
by Eugen Block 10 Jan '25

10 Jan '25

Hi, we have a couple of k8s clusters in our openstack environment (Victoria) and are facing some problems when pods get moved to different worker nodes. The PVCs get stuck in "reserved" state, and they only get attached after resetting the volumes to "available" state. First I thought it was due to the missing multiattach volume type which seems to work fine in our customer's cluster. I also found the "duplicate connector" error messages in the nova-compute.log: > Unable to update attachment.(Invalid volume: duplicate connectors > detected on volume But after adding the multiattach volume type and retyping the affected volumes, the PVCs still get stuck in "reserved" state while I don't see any error messages on the hypervisors anymore. So it seems that something has changed/improved, but now I don't have a trace where to look next. Could this be a cinder-csi driver issue or something? Our cinder backend is ceph (Pacific). Not sure which information could help so please ask for anything I can provide. Any pointers are highly appreciated! Thanks! Eugen

1 2

Third Party CI using Zuul v3 and nodepool - request for set up
by Roquesalane, Jean-Pierre 10 Jan '25

10 Jan '25

Hello OpenStack lovers! First, let me wish you all the best for this upcoming year. All our Cis are currently broken and we recently lost the team responsible for maintaining it. Unfortunately, there is a very poor documentation, and I guess we must rebuild something. I'm currently looking at Software Factory but can't get my hands on a starting point. I have a system deployed but would appreciate if you can lead me to people who have configured Zuul and nodepool. Our previous system was based on Jenkins and devstack-gate as a side note. I'll document the whole process as a help for future users. Thank you. Jean-Pierre Roquesalane SW System Sr Principal Engineer - OpenStack Champion Dell Technologies | Security and Ecosystem Software Mobile +33 6 21 86 79 04<tel:1-613-314-8106> Email : jeanpierre.roquesalane(a)dell.com<mailto:jeanpierre.roquesalane@dell.com> [cid:image001.png@01DB62C5.98BB1B40] Dell SAS, Si?ge Social 1 rond point Benjamin Franklin 34000 Montpellier, France. Capital 1,946,545.26 Euros. RCS Montpellier 351 528 229 - APE 4651Z. TVA Intracommunautaire FR 20 351 528 229. SIRET 351 528 229 00088. Vat Numbers: FR 20351528229 (France) / IT00001709997 (Italy) / ESN0012622G (Spain) Internal Use - Confidential

1 0

LXC containers
by engineer2024 10 Jan '25

10 Jan '25

I am trying to setup OSA branch 29.2.0 caracal. Whenever I try creating lxc containers using setup-hosts playbook, the new containers of version ubuntu 22 , as is the host, are missing openssh-server packages. The file /etc/ansible/roles/openstack_hosts/vars/ubuntu-22.04.yml, doesn't mention the ssh package in the distro packages. How to get ssh pkgs installed in the lxc containers ? Please help with this issue

2 1

LXC networking
by engineer2024 10 Jan '25

10 Jan '25

I have observed that in the OSA setups, on the controller nodes, each lxc container has two networks inside , eth0, eth1 interfaces. Ideally eth1 should be attached to br-mgmt bridge via veth pair. When I do 'ip link show type veth' on the host node, it is listing both the veth interfaces for each lxc container, but it is not listing the other member of that veth pair. --- # ip link show type veth 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000 link/ether fe:b6:db:27:e4:a7 brd ff:ff:ff:ff:ff:ff link-netnsid 0 9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 link/ether fe:a8:30:7a:ac:3f brd ff:ff:ff:ff:ff:ff link-netnsid 0 --- --- # ip link show type veth | grep 9f9add59 72: 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000 73: 9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 # --- But I can't find the other member of the veth pair, for example, for the interface 9f9add59_eth1, in the ip list. But the containers are running and reachable on the br-mgmt network. How to find out the other end of a veth pair, which is connected to the bridge ?

2 3

[blazar] reservation quota
by Winicius Allan 09 Jan '25

09 Jan '25

Is there a way to determine the quota of a project that the users can reserve? Regards.

2 1

[neutron] Proposal to deprecate and remove vlan_transparent config option
by Sławek Kapłoński 09 Jan '25

09 Jan '25

Hi, We have config option to enable vlan_transparent api extension currently. We will also have similar config option for vlan_qinq (it is now proposed in [1]). In both cases we have validation by mechanism drivers if option is or is not supported while networks are created. We can also filter out enabled API extensions by mechanism drivers, see [2] for example. So I proposed [3] to remove that config option and keep API extension to be loaded always if enabled ML2 drivers supports it. But before I will propose patch for that I would like to ask if there is maybe some use case for this config option really and maybe some of you somehow depends on it. In such case maybe we should keep it like it is now and don't remove it. Please let me know if you are using somehow this option. [1] https://review.opendev.org/c/openstack/neutron/+/937372 [2] https://github.com/openstack/neutron/blob/559672f777312d00c4358626650cc577e… [3] https://bugs.launchpad.net/neutron/+bug/2092174 -- Slawek Kaplonski Principal Software Engineer Red Hat

3 3

[Tempest] Devstack to configure tempest on a remote node
by Jakub Libosvar 09 Jan '25

09 Jan '25

Hello stackers, I have been attempting to deploy a “spine & leaf” like topology for upstream testing job of the ovn-bgp-agent [1]. The idea is to have an all-in-one devstack VM where all openstack services are running. This VM is plugged to two leaf switches respectively with its two interfaces, and each switch is connected to the “spine” switch. With the leaf switches behaving as route reflected we can validate routes are advertised to the spine. Now in order to validate that, I need to execute the tempest tests on the spine. That means I want to use devstack on the spine VM that calls OpenStack services API on the devstack all-in-one VM. The spine VM should not have any openstack service running. The problem I’m struggling with that the tempest is configured based on the ENABLED_SERVICES for the current box. I tried to hack the spine VM portion by calling configure_tempest with custom ENABLED_SERVICES from a local.sh [2] which seems to configure the tempest.conf correctly but the stack.sh script calls test-config [3] phase after local.sh is invoked, which re-configures tempest.conf back to the state as if no OpenStack services were enabled. I kinda hit the wall and ran out of ideas for hacks/workarounds with the current devstack codebase. I thought one way of solving this would be to decouple ENABLED_SERVICES into two separate variables, one being used to configure services and one for tempest purposes but before I dive into that, I wanted to ask if anybody has any experience doing something similar (configuring only tempest with devstack) or maybe there is already some variable I am not aware of. Thank you for any inputs, ideas or suggestions :) Kuba [1] https://review.opendev.org/c/openstack/ovn-bgp-agent/+/936968 [2] https://review.opendev.org/c/openstack/ovn-bgp-agent/+/936968/54/playbooks/… [3] https://opendev.org/openstack/devstack/src/branch/master/stack.sh#L1482-L15…

2 1

RDO User Survey
by Amy Marrich 09 Jan '25

09 Jan '25

Hi everyone, During today's RDO meeting we had a few topics that we need further information on in order to make decisions. As a result, we have created a 2 question survey[0] to collect the needed information. This survey will remain open for 2 weeks so we can make plans for the Epoxy release with the gathered information. The first question is in regards to CentOS Stream version(s) for the Epoxy release, CentOS Stream 9 only, 10 only, or both versions. The second question is to determine whether future RDO releases will continue to be all releases or only SLURP releases, if SLURP are the only releases folks are using. Thanks, Amy 0 - https://forms.gle/44atURvazYPXVvwZ8

1 0

[Neutron] Bug report last 2 weeks
by Rodolfo Alonso Hernandez 08 Jan '25

08 Jan '25

Hello Neutrinos: This is the bug report of the last 2 weeks. HIGH * Packet Logging Framework (OVN/ML2) ( https://bugs.launchpad.net/neutron/+bug/2092474) *Not assigned*. * with pyroute2 master dvr functional tests failing with route mismatch as returned routes are not from specific table ( https://bugs.launchpad.net/neutron/+bug/2092567) There is an open bug for pyroute2 (https://github.com/svinota/pyroute2/issues/1225) that should fix this problem. * with pyroute2 master dvr functional tests failing as pyroute.NetNS not honoring flags (https://bugs.launchpad.net/neutron/+bug/2092722) There is an open bug for pyroute2 (https://github.com/svinota/pyroute2/issues/1226) that should fix this problem. * Potential GPL violation in master branch due to GPL script embedded into Apache2 python code (https://bugs.launchpad.net/neutron/+bug/2092907) *Not assigned*. * Metadata service does not work when port security is disabled ( https://bugs.launchpad.net/neutron/+bug/2092971) *Not assigned*. MEDIUM * DeprecationWarning: create_subnetpool ( https://bugs.launchpad.net/neutron/+bug/2092657) *Not assigned*. * NoneType' object has no attribute 'registered_rules' ( https://bugs.launchpad.net/neutron/+bug/2092659) *Not assigned*. * OVN: Ports not activated on time when DPDK is used ( https://bugs.launchpad.net/neutron/+bug/2092407) Assigned to Ihar, patch proposed: https://review.opendev.org/c/openstack/neutron/+/938197. INCOMPLETE * deadlock： delete the network ( https://bugs.launchpad.net/neutron/+bug/2092983) Regards.

3 5