[2.5] Update hyperkube base to using iptables >1.8.5

[mitulshah-suse]

Update hyperkube base to use ubuntu 22.04 for version 1.19
Update KDM with the above changes
Related issue #35709

[vivek-shilimkar]

Fresh Install checks :

Rancher Server :

• Rancher version: v2.5.16
• Docker install

Downstream Cluster info

• Custom cluster
• 4 nodes: 2 worker, 1 etcd, 1 control plane
• k8s: v1.19.16-rancher2-1

Validation Steps :

• KDM pointing dev-v2.5 branch.
• Provisioned a k8s: v1.19.16-rancher2-1 downstream custom cluster.
• Ran Network checks
• Within the custom cluster deploy a workload in the default namespace.
• From explorer --> pods --> in the kube-system namespace --> Execute shell in one of the Canal pods.
• Change the container to kube-flannel, and run the dmesg command.
• Ran Canal checks.

Result :
Checked the dmesg logs in the canal pod and there are no segfault errors.

---

Upgrade Checks :

Rancher Server :

• Rancher version: v2.5.16
• Docker install

Validation Steps:

• KDM pointing to release-v2.5
• Provisioned k8s v1.19.16-rancher1-6 4 nodes (1-cp, 2-w, 1-etcd)
• Deployed workload in default ns.
• Ran pre-upgrade checks
• Changed the KDM to dev-v2.5 branch.
• Upgrade cluster from k8s v1.19.16-rancher1-6 to k8s v1.19.16-rancher2-1
• From explorer --> pods --> in the kube-system namespace --> Execute shell in one of the Canal pods.
• Change the container to kube-flannel, and run the dmesg command.
• Ran post-upgrade checks

Result :
Checked the dmesg logs in the canal pod and there are no segfault errors.

[vivek-shilimkar]

Reopening to test the cluster provisioning with standalone RKE.

[vivek-shilimkar]

Validated RKE standalone cluster provisioning with RKE v1.2.23-rc1.
Cluster provisioning was successful and kube-flannel shows no segfault errors.

Hence, the RKE checks passed.

[kinarashah]

Fix released in KDM with rancher/kontainer-driver-metadata#970 and in RKE v1.2.23.