Comment 14 for bug 1951245

For the duplex logs, we see some evictions due to disk pressure during the apply, but the pci-irq-affinity-agent pod manages to run in controller-0 after a while:

2021-11-18T10:22:17Z pci-irq-affinity-agent-prj2q Pod Successfully assigned openstack/pci-irq-affinity-agent-prj2q to controller-1 Scheduled Normal
2021-11-18T10:22:17Z pci-irq-affinity-agent-nq688 Pod Stopping container pci-irq-affinity-agent Killing Normal
2021-11-18T10:22:17Z pci-irq-affinity-agent-k85d7 Pod Successfully assigned openstack/pci-irq-affinity-agent-k85d7 to controller-0 Scheduled Normal
2021-11-18T10:22:17Z pci-irq-affinity-agent DaemonSet Created pod: pci-irq-affinity-agent-k85d7 SuccessfulCreate Normal
2021-11-18T10:22:17Z pci-irq-affinity-agent-prj2q Pod MountVolume.SetUp failed for volume "keyring" : hostPath type check failed: /opt/platform/.keyring/21.12/python_keyring/crypted_pass.cfg is not a file FailedMount Warning
2021-11-18T10:22:17Z pci-irq-affinity-agent DaemonSet Created pod: pci-irq-affinity-agent-prj2q SuccessfulCreate Normal
2021-11-18T10:22:18Z pci-irq-affinity-agent-k85d7 Pod Add eth0 [172.16.192.111/32] AddedInterface Normal
2021-11-18T10:22:18Z pci-irq-affinity-agent-k85d7 Pod Started container pci-irq-affinity-agent Started Normal
2021-11-18T10:22:18Z pci-irq-affinity-agent-k85d7 Pod Created container pci-irq-affinity-agent Created Normal
2021-11-18T10:22:18Z pci-irq-affinity-agent-k85d7 Pod Container image "registry.local:9001/docker.io/starlingx/stx-pci-irq-affinity-agent:master-centos-stable-20211116T032946Z.0" already present on machine Pulled Normal

But what calls my attention here is that on controller-1 it fails because it's unable to mount the keyring volume, which is needed for the agent to get information to communicate with the openstack services, in this line:

2021-11-18T10:22:17Z pci-irq-affinity-agent-prj2q Pod MountVolume.SetUp failed for volume "keyring" : hostPath type check failed: /opt/platform/.keyring/21.12/python_keyring/crypted_pass.cfg is not a file FailedMount Warning

This missing file may cause other problems other than only the openstack apply failure, and this file should not be missing on one of the hosts. the problem that was happening yesterday that was causing the pull image failure is not happening anymore after the fix, since the pod was raised on the standard setup and on one of the controllers in the duplex setup.