December 2024 - openstack-discuss

[cinder][ceph rbd] change volume type
by Tony Liu 12 Dec '24

12 Dec '24

Hi, With Antelope, old volume type is on replicated pool and new volume type is on EC pool. Change volume type doesn't work, volume type remains and no error returns. Is such change supported? I also tried change volume type from one replicated pool to another replicated pool. Doesn't work either. Am I missing anything? Thanks! Tony

2 1

[Openstack-ansible][Rabbitmq] Rabbitmq role is installing latest version of rabbitmq even after setting package version in variable file
by Danish Khan 12 Dec '24

12 Dec '24

Dear All, I have upgraded few of my openstack clusters from Ussuri to Wallaby successfully. But it seems rabbitmq repo is changed and now rabbitmq is getting its packages from: https://dl.cloudsmith.io/public/rabbitmq/rabbitmq-server/deb/ubuntu Everything was working fine until I face below issue while running rabbitmq playbook: TASK [rabbitmq_server : Apply rabbitmq policies] ******************************************************************************************************* failed: [rabbit_mq_container] (item={'name': 'HA', 'pattern': '^(?!(amq\\.)|(.*_fanout_)|(reply_)).*', 'priority': 0, 'tags': 'ha-mode=all'}) => {"ansible_loop_var": "item", "changed": false, "cmd": "/usr/sbin/rabbitmqctl -q -n '' set_policy -p / HA '^(?!(amq\\.)|(.*_fanout_)|(reply_)).*' '{\"ha-mode\": \"all\"}' --priority 0", "item": {"name": "HA", "pattern": "^(?!(amq\\.)|(.*_fanout_)|(reply_)).*", "priority": 0, "tags": "ha-mode=all"}, "msg": "Error:\nValidation failed\n\n[{<<\"ha-mode\">>,<<\"all\">>}] are not recognised policy settings", "rc": 69, "stderr": "Error:\nValidation failed\n\n[{<<\"ha-mode\">>,<<\"all\">>}] are not recognised policy settings\n\n", "stderr_lines": ["Error:", "Validation failed", "", "[{<<\"ha-mode\">>,<<\"all\">>}] are not recognised policy settings", ""], "stdout": "", "stdout_lines": []} It seems ha-mode=all is deprecated in latest version of Rabbitmq (4.0.4). I tried to install 3.8 or some other compatible version of rabbitmq but anyway the latest version of rabbitmq is getting installed. I have updated below entries of file /etc/ansible/roles/rabbitmq_server/vars/debian.yml : _rabbitmq_package_version: "3.8.14-1" and _rabbitmq_erlang_version_spec: "{{ (rabbitmq_install_method == 'external_repo') | ternary('1:23.3.4.15-1', '1:22.*') }}" Can someone please help me with this? I guess either I need to find a away to install specific version of rabbitmq to fox rabbitmq policy issue or I need to update parameters of command while applying policy. Regards, Danish Khan

2 2

[KOLLA-Ansible] default values for cpu_allocation_ratio
by Oliver Weinmann 12 Dec '24

12 Dec '24

Hi all, what are the default values for: cpu_allocation_ratio ram_allocation_ratio disk_allocation_ratio when using kolla-ansible? I can't seem to find any clear statement about it. Currently I assume they are 1.0 for each. Cheers, Oliver

2 2

[magnum] Cluster type (vm, ubuntu, kubernetes) not supported
by Satish Patel 12 Dec '24

12 Dec '24

Folks, I am trying to create a template for magnum but getting the following error. I am running the 2024.1 release. Anything changed or missing here? openstack image create ${IMAGE_NAME} --disk-format=qcow2 --container-format=bare --property os_distro=${OS_DISTRO} --file=${IMAGE_NAME}.qcow2; \ openstack coe cluster template create \ --image $(openstack image show ${IMAGE_NAME} -c id -f value) \ --external-network public \ --dns-nameserver 8.8.8.8 \ --master-lb-enabled \ --master-flavor gen.c4-m8-d30 \ --flavor gen.c4-m8-d30 \ --network-driver calico \ --docker-storage-driver overlay2 \ --coe kubernetes \ --label kube_tag=${version} \ k8s-${version}; Cluster type (vm, ubuntu, kubernetes) not supported (HTTP 400) (Request-ID: req-7c7f27b2-d67c-4f5b-aacf-01754b35bba8)

4 7

Re: [nova][security-sig] Communicating SSH host keys on
by Jeremy Stanley 12 Dec '24

12 Dec '24

first boot (was: Removing md5 in password injection) Reply-To: In-Reply-To: <8a163fa2-f071-4744-aae3-2182c048351c(a)redhat.com> On 2024-12-11 16:30:40 +0000 (+0000), Sean Mooney wrote: [...] > > Tangential, sorry, but could this sort of mechanism also be used to > > provide an out-of-band verification of SSH hostkeys generated on > > first boot for *nix type guests? We used to splat them to the > > primary terminal and then try to scrape them from the console log, > > but having a structured API response would be a lot cleaner (or did > > something similar already get added and I missed it?). [...] > i see the merrit in supprot some kind of reporting for limited > trusted boot type use cases such as providing the ssh host key > fingerprint so you can validate that the vm is the one you expect > it to be. > > im not sure how i would feel about allowing arbiarty metadata to > be store this way. > > is there something beyond the host key that would be useful for > you to record on boot? Nothing else comes to mind. Basically it's the desire to avoid ToFU style finger-crossing that the SSH host key you get on initially connecting to a newly booted server instance is really coming from that instance and not a MitM or honeypot. It's possible in most providers to create an authenticated backchannel by abusing the console log, as I described, but that has its own challenges. I suppose there might be alternative management protocols similar to SSH which rely on asymmetric key algorithms to authenticate hosts, but I'm not personally familiar with any (none with the near ubiquity of SSH at any rate). > perhaps we can try and capture this is a spec or separate thread > and work out what an mvp would look like. there would basically be > two part, one extednign the metadata api to support recording the > host key fingerprint somewhere, and then making that discoverable > via the main api perhaps in server show. That would make sense, and it would probably also be a good idea to try to involve e.g. a cloud-init maintainer in any such design discussion, as this would need client-side integration eventually anyway in whatever's going to communicate the key back to nova. > is we allowed read write access to all instance metadata we would > would not actually need to modify the main api as you can already > list metadata on an instance but im worried that could be abused > so i don't know if that is a good idea. Yes, that sounds like a security nightmare. > the nova metadata api is not intended a a high performance key > value store so i dont necessarily think it would be a good idea to > extned this to a generic crud api but we certainly could provide > limited support for targeted uscease liek the one you raise. Agreed, I suppose the field could be given a more generic name in case certain kinds of guests wanted to store functionally similar things in it, but as I said I don't know what else that would end up being anyway. -- Jeremy Stanley

1 1

Some questions about ironic
by Johnny Hao 12 Dec '24

12 Dec '24

Hi Buddies, I am new to ironic, after some study, I have several questions about it, could anyone help to answer? Thanks a lot. 1. According to the workflow chart below, what's the communication method between nova-compute and Ironic API? HTTP API call or RPC call through mq like rabbitmq? So are the communications between Ironic Conductor and other components of Neutron, Glance and Cinder/Swift etc? 2. Where do Ironic services run (Ironic API/Conductor/Drivers and DB)? In the controller node? Or a separate node is also supported? ** -- Best regards QingFeng Hao

2 1

[nova] riscv64 emulation support
by James Page 11 Dec '24

11 Dec '24

Hi Nova Team We've had some initial support for enabling RISCV64 support in Nova in a review for sometime which I refreshed this week: https://review.opendev.org/c/openstack/nova/+/889137 Sean suggested that this really needs a blueprint and discussion at the next Nova meeting (which I can't make) so proposing discussion via ML instead: https://blueprints.launchpad.net/nova/+spec/riscv64-emulation-support I'm also going to look at the nova-emulation job to see if that can be extended to cover RISCV64 as well - there are a few nuances with dependent packages (opensbi) in Ubuntu 24.04 but I think we can come up with something workable for testing. Review and comments are appreciated. Thanks James

2 1

OpenStack metrics and logs
by Timothé Bauge 11 Dec '24

11 Dec '24

Hello Stackers, I would like to know what are you all using to monitor and supervise your Openstack clusters ? We are in the process of setting up our own private cloud based on RedHat OpenStack Platform 17.1, we choose not to go with the RedHat Service Telemetry Platform, and were strongly advise against Ceilometer and aodh. At the moment, we built our own stack based on Prometheus (with the plenty of exporters) for the metrics, Graylog + OpenSearch for the logs, and Grafana for the visualisation. For now, we are only looking to retrieve basic information in order to know if a dysfunction occurs, but the end goal might be to go as far as to be able to count the cpu/mem/disk usage per hour per vm per project, etc. So, what are you all using and how did you implement it ? Best regards, [cid:659cb34d-63ad-4c5a-94e4-734ef252c20b] Timothé BAUGÉ timothe.bauge(a)covage.com<mailto:pr%C3%A9nom.nom@covage.com> COVAGE I Wholesale B2B du Groupe Altitude

3 3

Re: [openstack-helm] [Cinder] Cinder-Scheduler oslo_privsep.daemon.FailedToDropPrivileges: Privsep daemon failed to start
by daniel890723＠gmail.com 11 Dec '24

11 Dec '24

Hello Daniel here I have been test and trying to solve the oslo_privsep.daemon.FailedToDropPrivileges: Privsep daemon failed to start Error on openstack-helm deployed Cinder After the test with kolla-ansible i can confirm the problem is from kubernetes So i perform the test with all Cinder Service with Priviliged true setting in Value.yaml: pod: security_context: volume_usage_audit: pod: runAsUser: 42424 container: cinder_volume_usage_audit: readOnlyRootFilesystem: true privileged: true allowPrivilegeEscalation: true cinder_api: pod: runAsUser: 42424 container: ceph_coordination_volume_perms: runAsUser: 0 readOnlyRootFilesystem: true cinder_api: readOnlyRootFilesystem: true allowPrivilegeEscalation: true privileged: true cinder_scheduler: pod: runAsUser: 42424 container: ceph_coordination_volume_perms: runAsUser: 0 readOnlyRootFilesystem: true cinder_scheduler: capabilities: add: - SYS_ADMIN privileged: true readOnlyRootFilesystem: true allowPrivilegeEscalation: true Also Glance-api container too: pod: security_context: glance: pod: runAsUser: 42424 container: glance_perms: readOnlyRootFilesystem: true runAsUser: 0 ceph_keyring_placement: readOnlyRootFilesystem: true runAsUser: 0 glance_api: privileged: true readOnlyRootFilesystem: true allowPrivilegeEscalation: true nginx: readOnlyRootFilesystem: false runAsUser: 0 With Cinder and Glance container set as priviliged container the problem is gone and i can create volume with image normally So it is confirmed that the ERROR is because the pod permission Also my System Setup: | OS | Ubuntu:22.04 | | Kubernetes | 1.30.4 | | Openstack-helm | 2024.1 | P.S. i haven't test with a complete clean enviroment with only Ceph and Openstack component installed Daniel Lu

3 3

[nova] Specs review day is TODAY and Spec soft freeze is TOMORROW
by Sylvain Bauza 11 Dec '24

11 Dec '24

Hi folks, As we discussed on yesterday's nova meeting, today is our second specs review day and tomorrow (Dec 12th) will be the last day for providing a new spec to be reviewed. Please see our schedule here: https://releases.openstack.org/epoxy/schedule.html Thanks, -Sylvain

1 0