Open Stack

Oliver,

	Thanks for the feedback!  Changing the kube_tag and container_runtime is inline in what I have done so far.  I agree this will successfully build a cluster.  However, it leaves all the other image and chart tags at the original level.  For example, that leaves us running the cinder csi plugin with tag “v1.23.0” where I think it would probably be more appropriate to run at v1.24.6 (the latest) or at least one of the v1.24.x versions.  Going through and just increasing each tag and other version type parameter should be possible, but I was hoping that somewhere this is being tracked already so that the community might have an idea on what good supported combinations are.

From what I can see in the existing heat templates for 2023.1, these are the default values that you will get if you don’t specify a label override … this is probably not a complete list because I just pulled out the _tag values.  It’s probably also necessary to track at least containerd_version, kube_dashboard_version

metrics_server_chart_tag: v3.7.0
traefik_ingress_controller_tag: v1.7.28
kube_tag: v1.23.3-rancher1
master_kube_tag: v1.23.3-rancher1
minion_kube_tag: v1.23.3-rancher1
cloud_provider_tag: v1.23.1
etcd_tag: v3.4.6
coredns_tag: 1.6.6
flannel_tag: v0.15.1
flannel_cni_tag: v0.3.0
metrics_scraper_tag: v1.0.4
calico_tag: v3.21.2
calico_kube_controllers_tag: v1.0.3
octavia_ingress_controller_tag: v1.18.0
prometheus_tag: v1.8.2
grafana_tag: 5.1.5
heat_container_agent_tag: wallaby-stable-1
k8s_keystone_auth_tag: v1.18.0
prometheus_operator_chart_tag: v8.12.13
prometheus_adapter_chart_tag: 1.4.0
tiller_tag: "v2.16.7"
helm_client_tag: "v3.2.1"
magnum_auto_healer_tag: v1.18.0
cinder_csi_plugin_tag: v1.23.0
csi_attacher_tag: v3.3.0
csi_provisioner_tag: v3.0.0
csi_snapshotter_tag: v4.2.1
csi_resizer_tag: v1.3.0
csi_node_driver_registrar_tag: v2.4.0
csi_liveness_probe_tag: v2.5.0
node_problem_detector_tag: v0.6.2
nginx_ingress_controller_tag: 0.32.0
nginx_ingress_controller_chart_tag: 4.0.17
draino_tag: abf028a
autoscaler_tag: v1.18.1

Any help is appreciated.

Thank you,

Jay

> On Aug 14, 2023, at 4:59 AM, Oliver Weinmann <oliver.weinmann at me.com> wrote:
> 
> Hi Jay,
> 
> K8s 1.24 needs containerd so you need to add an additional label. You also need to cange the hyperkube prefix.
> 
> labels:
>   container_runtime: containerd
>   hyperkube_prefix: docker.io/rancher/ <http://docker.io/rancher/>
>   kube_tag: v1.24.16-rancher1
> 
> The following template works just fine for me under Antelope 2023.1 deployed with Kolla-Ansible:
> 
> openstack coe cluster template create k8s-flan-small-37-v1.24.16-containerd --image Fedora-CoreOS-37 --keypair mykey --external-network public --fixed-network demo-net --fixed-subnet demo-subnet --dns-nameserver <ip_of_your_dns> --flavor m1.kubernetes.small --master-flavor m1.kubernetes.small --volume-driver cinder --docker-volume-size 10 --network-driver flannel --docker-storage-driver overlay2 --coe kubernetes --labels kube_tag=v1.24.16-rancher1,hyperkube_prefix=docker.io/rancher/,container_runtime=containerd <http://docker.io/rancher/,container_runtime=containerd>
> 
> (2023.1) [vagrant at seed ~]$ openstack coe cluster template show k8s-flan-small-37-v1.24.16-containerd -f yaml
> insecure_registry: '-'
> labels:
>   container_runtime: containerd
>   hyperkube_prefix: docker.io/rancher/ <http://docker.io/rancher/>
>   kube_tag: v1.24.16-rancher1
> updated_at: '2023-08-14T07:25:09+00:00'
> floating_ip_enabled: true
> fixed_subnet: demo-subnet
> master_flavor_id: m1.kubernetes.small
> uuid: bce946ef-6cf7-4153-b858-72b943c499a2
> no_proxy: '-'
> https_proxy: '-'
> tls_disabled: false
> keypair_id: mykey
> public: false
> http_proxy: '-'
> docker_volume_size: 10
> server_type: vm
> external_network_id: 60335752-0c01-40b0-b152-365b23576309
> cluster_distro: fedora-coreos
> image_id: Fedora-CoreOS-37
> volume_driver: cinder
> registry_enabled: false
> docker_storage_driver: overlay2
> apiserver_port: '-'
> name: k8s-flan-small-37-v1.24.16-containerd
> created_at: '2023-08-10T15:16:00+00:00'
> network_driver: flannel
> fixed_network: demo-net
> coe: kubernetes
> flavor_id: m1.kubernetes.small
> master_lb_enabled: false
> dns_nameserver:
> hidden: false
> tags: '-'
> 
> (2023.1) [vagrant at seed ~]$ kubectl get nodes -o wide
> NAME                               STATUS   ROLES    AGE   VERSION    INTERNAL-IP   EXTERNAL-IP    OS-IMAGE                        KERNEL-VERSION          CONTAINER-RUNTIME
> containerd-j5ob3gd2dvqo-master-0   Ready    master   78m   v1.24.16   10.0.0.125    192.168.4.134   Fedora CoreOS 37.20221127.3.0   6.0.9-300.fc37.x86_64   containerd://1.4.4 <>
> containerd-j5ob3gd2dvqo-node-0     Ready    <none>   71m   v1.24.16   10.0.0.71     192.168.4.132   Fedora CoreOS 37.20221127.3.0   6.0.9-300.fc37.x86_64   containerd://1.4.4 <>
> 
> (2023.1) [vagrant at seed ~]$ kubectl get pods -A
> NAMESPACE     NAME                                          READY   STATUS    RESTARTS   AGE
> kube-system   coredns-7f4bcd98d7-9vjzk                      1/1     Running   0          77m
> kube-system   coredns-7f4bcd98d7-wnc7h                      1/1     Running   0          77m
> kube-system   csi-cinder-controllerplugin-dc7889b4f-vk4jd   6/6     Running   0          77m
> kube-system   csi-cinder-nodeplugin-l8dqq                   3/3     Running   0          71m
> kube-system   csi-cinder-nodeplugin-zdjg6                   3/3     Running   0          77m
> kube-system   dashboard-metrics-scraper-7866c78b8-d66mg     1/1     Running   0          77m
> kube-system   k8s-keystone-auth-c9xjs                       1/1     Running   0          77m
> kube-system   kube-dns-autoscaler-8f9cf4c99-kq6j5           1/1     Running   0          77m
> kube-system   kube-flannel-ds-qbw9l                         1/1     Running   0          71m
> kube-system   kube-flannel-ds-xrbmp                         1/1     Running   0          77m
> kube-system   kubernetes-dashboard-d78dc6f78-2qklq          1/1     Running   0          77m
> kube-system   magnum-metrics-server-564c9cdd6d-2rxpc        1/1     Running   0          77m
> kube-system   npd-8xkbk                                     1/1     Running   0          70m
> kube-system   openstack-cloud-controller-manager-46qsl      1/1     Running   0          78m
> 
> I have a blog where I blog about OpenStack related stuff. One of my older blog posts is also about deploying K8s via Magnum:
> 
> https://www.roksblog.de/deploy-kubernetes-clusters-in-openstack-within-minutes-with-magnum/
> 
> Cheers,
> Oliver
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.openstack.org/pipermail/openstack-discuss/attachments/20230814/497c4df6/attachment-0001.htm>