February 2024 - openstack-discuss

[Neutron][OVN] Flat routed provider network has external traffic broadcasted to all VMs
by Daniel Failing 09 Feb '24

09 Feb '24

Dear all, I'm currently debugging a (testing) deployment with neutron/OVN, currently most ports in external flat network. So far I'm happy everything can communicate, but we see a strong performance degradation with increasing number of VMs/ports hosted on a hypervisor. The setup is similar to the one described in the routed provider networks docs [1]. For routing between the different segments, we have physical routers, that are not managed by neutron. Debugging the issue showed traffic originating from the DC router to VMs go perfectly as unicast traffic through OVN. Communicating from the VM to other networks via the DC router, however, will broadcast the traffic to all attached ports on the internal bridge. Is there a way to tell neutron/OVN to use a specific MAC address for the router and forward the traffic through the patch port instead of broadcasting it? I could not find any documentation of whether I am missing a neutron port or some other configuration on the hypervisor. Just to make sure: The VM and DC routers have the right ARP information of itself, the traffic that is duplicated is from VM to router/external network. (via default route to router) As far as I could see, this issue does not appear with private networks, since all MAC addresses are known to OVN/Neutron. Versions: Nova/Neutron: Yoga ovs-vswitchd (Open vSwitch) 3.1.4 ovn-northd 22.12.1 kernel 5.14.0-284.30.1.el9_2.x86_64 Best, Daniel [1] https://docs.openstack.org/neutron/latest/admin/config-routed-networks.html

2 4

[manila] multi az replication
by garcetto 08 Feb '24

08 Feb '24

good afternoon, i am using zed, on streched ceph as backend on multiple az. i would like to emulate aws efs multiple nfs mount points for each user tenant (one on each different az) using manila, is it possible? ideally i would like to expose to each users tenant a manila vm as nfs exporter to custormer' instances, and as backend (of manila) use a cinder volume on ceph (rbd? cephfs? nfs?) my problem is how can i create say 3 different manila service vms (one in each az) with different frontend ips all pointing to same rbd/cephfs volume? am i doing wrong? thank you.

2 1

[neutron] Drivers meeting cancelled tomorrow
by Brian Haley 08 Feb '24

08 Feb '24

Hello Neutrinos: Due to the lack of agenda, Friday's drivers meeting is cancelled. Have a nice weekend! -Brian

1 0

[kolla] [Ironic] Is ironic_cleaning_network is tenant network?
by Jayesh Chaudhari 08 Feb '24

08 Feb '24

Hi, I am trying to implement Ironic in my kolla-ansible yoga setup. I am following below document https://docs.openstack.org/kolla-ansible/yoga/reference/bare-metal/ironic-g… Here in ironic_cleaning_network and dnsmasq_dhcp_range is from tenant network?

2 1

[manila] share replication still experimental?
by garcetto 08 Feb '24

08 Feb '24

good evening, is still experimental even in zed or later? https://docs.openstack.org/manila/latest/admin/shared-file-systems-share-re… thank you.

2 1

[magnum] Is magnum support region_name in keystone auth config?
by Seongsoo Cho 08 Feb '24

08 Feb '24

Hi magnum team. It seems that a magnum does not support region_name which is in the keystone_auth and keystone_authtoken section in magnum conf. The internal keystone client of keystoneauth1 only uses the lowest UUID of endpoints for keystone url. [1] And also KeystoneClientV3 's get_auth[2] and get_legacy_auth[3] also don't support region_name as an argument. So, My question is how can I make magnum only use the keystone URL written in the config? Best regards Seongsoo Cho [1] https://github.com/openstack/keystoneauth/blob/master/keystoneauth1/access/… [2] https://github.com/openstack/magnum/blob/master/magnum/common/keystone.py#L… [3] https://github.com/openstack/magnum/blob/master/magnum/common/keystone.py#L… -- Seongsoo Cho OpenStack Korea User Group / Community Leader IRC #seongsoocho

2 2

about magnum capi for production
by Nguyễn Hữu Khôi 07 Feb '24

07 Feb '24

Hello guys. Could you share with me some experience about using Magnum CAPI in production? https://github.com/vexxhost/magnum-cluster-api Thank you so much. Nguyen Huu Khoi

6 44

Cinder Bug Report 2024-02-07
by Jon Bernard 07 Feb '24

07 Feb '24

Hello Argonauts, Cinder Bug Meeting Etherpad <https://etherpad.opendev.org/p/cinder-bug-squad-meeting> Undecided - Service cinder-backup failed when Swift with TLS enabled <https://bugs.launchpad.net/cinder/+bug/2051985> - Status: In Progress - hpe xp: wrong copyrights in test scripts <https://bugs.launchpad.net/cinder/+bug/2052409> - Status: In Progress - GRUB menu format error <https://bugs.launchpad.net/cinder/+bug/2052414> - Status: New Thanks, -- Jon

1 0

[cinder][vendors] Third party CI compliance check
by Rajat Dhasmana 07 Feb '24

07 Feb '24

Hello Vendors, This is related to the original discussion on the ML thread[1]. We have noticed that a lot of the drivers don't have their CI reported on a patch in 2024. Based on the last CI report that I checked[2], the number comes out to be 32 though it's not an accurate measure of the inactivity in the project. The following cases are possible: 1. The driver that doesn't have active CI reporting is already marked unsupported 2. The driver is supported but CI is not actively maintained 3. The CI information is not accurate in the wiki page If it's case 2, vendors should look into their CI and fix it else the driver will be marked unsupported in the current or upcoming release. If it's case 3, make sure to have updated CI info in the wiki page. Let's take an example of the HPE driver. The last seen status is shown as: last seen: 2016-08-09 but the driver is actively reporting on patches. The issue here is either: 1. The CI name in the volume driver is not pointing to the correct wiki page[3] 2. The wiki page does not have correct information about CI[4] Vendors should look into updating their CI related information before the midcycle-2, which is going to be held on 14th Feb, 2024 from 1400-1600 UTC[5]. Since we only have the cireport[2] as the source of truth, we will be looking at the vendor drivers that don't have a recent last seen date and start marking them unsupported during the midcycle-2. Any query is welcome on this thread or during the midcycle-2 topic discussion. [1] https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack… [2] http://cinderstats.ivehearditbothways.com/cireport.txt [3] https://opendev.org/openstack/cinder/src/branch/master/cinder/volume/driver… [4] https://wiki.openstack.org/wiki/ThirdPartySystems/HPE_Storage_CI [5] https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack… Thanks Rajat Dhasmana

3 2

[all] SLURP support survey
by Mark Goddard 07 Feb '24

07 Feb '24

Hi, The Kolla project is planning to support SLURP upgrades [1] in the Caracal release cycle, allowing users to upgrade from Antelope to Caracal. There are two areas in which I'd like to request information from the various project teams. # 1. SLURP support We understand that for various reasons there can be a gap between TC resolutions and implementation across projects. Using some rudimentary grepping through various codebases I found that many projects have added the grenade-skip-level job (or a variation of it) to their CI pipelines, and a few mention SLURP in their release notes. There's little mention of it in documentation at this stage. AFAICT the following projects have SLURP upgrade jobs: Cinder, Glance, Ironic, Manila, Neutron, Nova, Octavia While I could not find a SLURP job in these: Barbican, Bifrost, Blazar, Cloudkitty, Designate, Heat, Horizon, Keystone, Magnum, Placement Question: Does your project explicitly support SLURP upgrades? Do you have or are you planning to add testing and documentation for SLURP upgrades? # 2. Rolling SLURP upgrades The TC resolution [1] states that rolling upgrades are optional for SLURP releases. This could affect the upgrade procedure required in Kolla Ansible. Nova mentions in its Antelope release notes that it supports rolling SLURP upgrades, but I could not find any information from other projects. It would be helpful if this was documented for each project. Question: Does your project support or plan to support rolling upgrades? If changes to the upgrade procedure for SLURP are required, are they documented? Thank you in advance for your responses. Apologies if I have missed any obvious source of information on this. Regards, Mark [1] https://governance.openstack.org/tc/resolutions/20220210-release-cadence-ad…

7 7