- openstack-discuss - lists.openstack.org

[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 09 Feb '24

09 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 09 Feb '24

09 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 09 Feb '24

09 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 09 Feb '24

09 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 09 Feb '24

09 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 08 Feb '24

08 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 08 Feb '24

08 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

[largescale-sig] Next meeting: February 7, 9utc
by Thierry Carrez 07 Feb '24

07 Feb '24

Hi everyone, The Large Scale SIG will be meeting this Wednesday in #openstack-operators on OFTC IRC, at 9UTC, our EU+Asia time. You can doublecheck how that UTC time translates locally at: https://www.timeanddate.com/worldclock/fixedtime.html?iso=20240207T09 Feel free to add topics to the agenda: https://etherpad.opendev.org/p/large-scale-sig-meeting Regards, -- Thierry Carrez

2 2

[Kolla][Octavia] worker unable to connect amphora instance
by Jayesh Chaudhari 07 Feb '24

07 Feb '24

Folks, I have set up kolla-ansible yoga openstack and configured Octavia using VLAN. But when I am trying to create a LB, it is stuck in pending create then eventually failed. And in logs I can see octavia-worker unable to connect to amphora instance. Is there any sanity which I can do to check if my implementation is correct ? Or am I missing something? Please advise. My configuration in global.yml enable_octavia: yes octavia_network_interface: "vlan.2140" octavia_auto_configure: yes octavia_amp_flavor: name: "amphora" is_public: yes vcpus: 2 ram: 1024 disk: 5 octavia_amp_security_groups: mgmt-sec-group: name: "lb-mgmt-sec-grp" rules: - protocol: icmp - protocol: tcp src_port: 22 dst_port: 22 octavia_amp_network: name: lb-mgmt-net provider_network_type: vlan provider_segmentation_id: 2140 provider_physical_network: physnet1 external: false shared: false subnet: name: lb-mgmt-subnet cidr: "10.145.50.128/26" allocation_pool_start: "10.145.50.135" allocation_pool_end: "10.145.50.190" gateway_ip: "10.145.50.129" enable_dhcp: yes octavia_amp_image_tag" "amphora" octavia_loadbalancer_topology: "SINGLE" Thanks, Jayesh

3 8

[tc] Next Technical Committee Meeting Tomorrow, Feb 3, 1800 UTC
by Jay Faulkner 07 Feb '24

07 Feb '24

Hi all, The next scheduled meeting of the Technical Committee is tomorrow. This meeting will be held primarily in Zoom, with minutes kept in IRC. Interested parties can join at https://us06web dot zoom dot us/j/87108541765?pwd=emlXVXg4QUxrUTlLNDZ2TTllWUM3Zz09 (repair the URL first). The proposed agenda ia: Agenda - Roll call - Follow up on tracked action items - Gate health check - Implementation of Unmaintained branch statuses - 2024.1 TC Tracker - https://etherpad.opendev.org/p/tc-2024.1-tracker - Open Discussion and Reviews

1 1

[openstack-helm] Remove Calico chart
by Vladimir Kozhukalov 07 Feb '24

07 Feb '24

Hi everyone, I would like to suggest removing the calico chart from the openstack-helm-infra repository which does not seem worth it to support any more. Tigera provides great tools for managing Calico deployments (helm chart, operator and even plain kubectl manifest). Also there are plenty of other networking solutions on the market and it looks like users can choose on their own the CNI implementation. There have not been many contributions to this chart for quite some time and we don't use this chart in any test jobs. In the deploy-env role we use the upstream Calico manifest [1]. Having said that, I think that we can remove it immediately even without any depreciation period. Opinions are welcome. [1] https://opendev.org/openstack/openstack-helm-infra/src/branch/master/roles/… -- Best regards, Kozhukalov Vladimir

4 3

Glance image import to AWS S3 fails for files larger than 5GB
by Shrishail K 06 Feb '24

06 Feb '24

Hi Team, I have a use case where I need to replicate images from file store to S3 store. While image creation directly on S3 works fine with large files over 5G. The same fails when attempted via image-import with copy-image method. The failure is due to AWS rejecting a single object upload operation where the object size is over 5GB. What I see from logs is that direct image creation on S3 uses multipart upload when images are large. However, I fail to see the same happening when image import was used to copy an image from file to s3. Is this a known issue? Or Am I missing some configs related to image-import? Openstack version: Antelope Client version: 6.4.0 Thanks, Shrishail

3 14

[puppet] EOL plan of wallaby, xena and yoga
by Takashi Kajinami 06 Feb '24

06 Feb '24

Hello, I'm sending this as an early announcement of my current plan about these old stable branches in Puppet OpenStack project. Because we've not see any interest to backport changes to these stable branches and we have limited capacity to maintain our project, I'd propose transitioning EOL these old stable branches by the following steps. * stable/yoga is transitioned to unmaintained/yoga but we may not work on maintaining it now (which eventually means it's abandoned right after its transition to unmaintained status). We'll move it to EOL at the next cadence, when zed is transitioned to unmaintained). * wallaby and xena will become immediate candidates for EOL so we'll just move the EOL process forward. If you have any concern, then please let me know. We'll formally work on these processes in the release patches. [1] https://docs.openstack.org/project-team-guide/stable-branches.html#unmainta… -- Takashi Kajinami irc: tkajinam github: https://github.com/kajinamit launchpad: https://launchpad.net/~kajinamit

1 0

[puppet] Retiring puppet-qdr
by Takashi Kajinami 06 Feb '24

06 Feb '24

Hello, Let me restore my old proposal[1] about puppet-qdr. # I lost access to my old emails so I'm starting a new thread. In the previous mail I proposed deprecation instead of retirement, because the module has been used by TripleO. However I noticed that TripleO is now under the process of complete retirement[2]. So, I'd like to revise my proposal, and propose retiring the module completely. Please let me know if you have any concern with this module retired. Unless I hear any objections, I'll start the official retirement process. [1] https://lists.openstack.org/pipermail/openstack-discuss/2023-August/034742.… [2] https://review.opendev.org/c/openstack/governance/+/905145 Thank you, Takashi Kajinami irc: tkajinam

2 2

[neutron] Bugs deputy report January 29th to February 4th
by Miguel Lavalle 06 Feb '24

06 Feb '24

Here's this week's bugs deputy report: Critical ====== https://bugs.launchpad.net/neutron/+bug/2051831 [ovn] neutron-ovn-tempest-slow job fail tests relying on FIP. Assignee ralonsoh. Proposed fix: https://review.opendev.org/c/openstack/neutron/+/907312 High ==== https://bugs.launchpad.net/neutron/+bug/2051729 issue dhcp cleaning stale devices process when enable action. Assignee Sahid Orentino. Proposed fix: https://review.opendev.org/c/openstack/neutron/+/907250 https://bugs.launchpad.net/neutron/+bug/2051928 tests - Python 3.12 - TypeError: Object of type _SentinelObject is not JSON serializable. Needs an owner https://bugs.launchpad.net/neutron/+bug/2051935 [OVN] SNAT only happens for subnets directly connected to a router. Owner haleyb Medium ====== https://bugs.launchpad.net/neutron/+bug/2051510 [OVN] Audit the L2/L3 extensions support. Assignee: https://bugs.launchpad.net/neutron/+bug/2051869 OVS agent DHCP extension fail if not set subnet gateway_ip. Assigned to Liu Yulong https://bugs.launchpad.net/neutron/+bug/2052367 neutron-keepalived-state-change run neutron-rootwrap-daemon anyway regardless of config. Assigned to Liu Yulong. Proposed fix https://review.opendev.org/c/openstack/neutron/+/653378 Need follow up: ============ https://bugs.launchpad.net/neutron/+bug/2051863 Race condition during VM creation - could not open network device tapXXX (No such device). Liu Yulong created an associated LP bug for Nova, but might require a follow from us in the future. Incomplete ========= https://bugs.launchpad.net/neutron/+bug/2052299 neutron-tempest-plugin - HaRoutersTest.test_convert_legacy_router Test "ha" Check Error

1 0

[publiccloud-sig] Reminder - next meeting February 7th - 0800 UTC
by Felix Kronlage-Dammers 06 Feb '24

06 Feb '24

Hi everyone, This Wednesday (February 7th) the next meeting of the public cloud sig is going to happen. Since the majority of participants in the sessions started to like the videocalls, we’ll be doing this as a videocall again. We’ll be meeting at 0800 UTC here: https://conf.scs.koeln:8443/OIF-public-cloud-sig See also here for all other details: https://wiki.openstack.org/wiki/PublicCloudSIG Some of us met at FOSDEM last weekend, so we can share a bit about our discussions and findings there :) See you Wednesday! felix -- Felix Kronlage-Dammers Product Owner IaaS & Operations Sovereign Cloud Stack Sovereign Cloud Stack — standardized, built and operated by many Ein Projekt der Open Source Business Alliance - Bundesverband für digitale Souveränität e.V. Tel.: +49-30-206539-205 | Matrix: @fkronlage:matrix.org | fkr(a)osb-alliance.com

1 0

[storlets] Transition wallaby, xena and yoga to EOL
by Takashi Kajinami 06 Feb '24

06 Feb '24

Hello, As you know, stable/yoga already reached its end of maintenance phase and is now being transitioned to the unmaintained branch. Because we've not seen any attempt to backport changes to it and even older branches, I'd propose retiring yoga and older branches (wallaby and xena) instead of keeping inactive branches. Please let me know if you have any concern. -- Takashi Kajinami irc: tkajinam github: https://github.com/kajinamit launchpad: https://launchpad.net/~kajinamit

1 1

[fwaas][devstack] help configuring dashboard
by garcetto 05 Feb '24

05 Feb '24

good morning, is there some tutorial on how to configure fwaas web dashboard on devstack? thank you.

2 1

[nova] Help with support to SGX
by Winicius Allan 05 Feb '24

05 Feb '24

Hello o/ Anyone already work with Nova to provide VMs with SGX support?

6 7

nova-compute down and no data in log file
by russell.stather＠ignitiongroup.co.za 05 Feb '24

05 Feb '24

nova-compute showing as enabled but down. Log files are not being written to. Service is showing as running. This is a juju installation. Have tried pausing and resuming the nova-compute, and rabbitmq, and nova-cloud-controller.

1 0

[cinder] volume type properties / extra_specs
by Josephine Seifert 05 Feb '24

05 Feb '24

Hi, I was experimenting with the extra_specs of volume types for a project. And while I was able to set and unset properties freely, the volumes created from volume types with random properties could not be created. They always went into Error state. When I removed the properties from the volume type, volumes could be created again. Is it meant to behave like this? So in other words: Are custom properties not allowed for volume types? And if so, why am I allowed to set custom properties? I would expect an Error to happen in the volume type set --property ... process and not when creating a volume. greetings. Josephine (Luzi)

2 2

[oslo] Feature freeze.
by Daniel Bengtsson 04 Feb '24

04 Feb '24

Hi there, Just a friendly reminder that oslo's feature freeze will happen in 2 weeks (R-7 - Feb 12 - Feb 16). To know more about the freeze, we have a policy[1] that discusses it. The main thing is that if you have features you want to get into Oslo libraries this cycle, please make sure they merge by Friday. After that we will need to go through the FFE process and there is no guarantee we can land them. Feel free to ping us on IRC if you need code review. [1] http://specs.openstack.org/openstack/oslo-specs/specs/policy/feature-freeze… -- Daniel Bengtsson Software Engineer

3 2