The OpenStackDataPlaneDeployment we apply has 

spec.nodeTemplate.ansible.ansibleVars.edpm_kernel_args: "default_hugepagesz=1GB hugepagesz=1G hugepages=232 hugepagesz=2M hugepages=4096 isolcpus=2-39,42-79 iommu=pt intel_iommu=on" 

but the nodes are not rebooted by the reboot-os playbook and therefore /boot/grub2/grub.cfg that provides the arguments to the kernel is not updated.{}

This results in subsequent playbooks, such as configure-ovs-dpdk and configure-network, to fail.

Expected behavior

• The nodes should be rebooted after new kernel arguments are set in grub.cfg and before proceeding to configuring networking, etc.

Screenshots

• Attached Image

Bug impact

• Deployment fails unless a manual workaround is applied

Known workaround

• We rebooted the hosts manually during the reboot-os playbook and increased the fail count from 6 to 12 to provide a wider window for them to became again available but this is obviously not a convenient, scalable and overall acceptable handling.

Additional context

The OpenStackDataPlaneNodeSet that is applied:

std-nodeset.yaml

The output of must-gather:

oc adm must-gather  --image-stream=openshift/must-gather  --image=registry.redhat.io/rhoso-operators/openstack-must-gather-rhel9:1.0
...
Reprinting Cluster State:
When opening a support case, bugzilla, or issue please include the following summary data along with any other requested information:
ClusterID: d2b89b89-712e-4fac-86b9-edca27ec2623
ClientVersion: 4.18.32
ClusterVersion: Stable at "4.18.32"
ClusterOperators:
        All healthy and stable