Chapter 5. Known issues

Because of changes to the image pull mechanism in 1.10, authentication does not work when pulling container images from private registries such as registry.access.redhat.com. The workaround is to copy the cluster pull secret to the pod namespace. Then, you must either link the pull secret to the default service account or add the pull secret explicitly to the peer pod manifest.

Jira:KATA-4107

Using container CPU resource limits to increase the number of available CPUs for a pod fails if the requested CPUs are offline. If the functionality is available, you can diagnose CPU resource issues by running the oc rsh <pod> command to access a pod and then running the lscpu command:

$ lscpu

Example output:

CPU(s):                                16
On-line CPU(s) list:             0-12,14,15
Off-line CPU(s) list:            13

The list of offline CPUs is unpredictable and can change from run to run.

To work around this problem, use a pod annotation to request additional CPUs as in the following example:

metadata:
  annotations:
    io.katacontainers.config.hypervisor.default_vcpus: "16"

Jira:KATA-1376

You cannot use the sizeLimit parameter in the pod specification to expand ephemeral volumes because the volume size default is 50% of the memory assigned to the sandboxed container.

To work around this problem, change the size by remounting the volume. For example, if the memory assigned to the sandboxed container is 6 GB and the ephemeral volume is mounted at /var/lib/containers, you can increase the size of this volume beyond the 3 GB default by running the following command:

$ mount -o remount,size=4G /var/lib/containers

Note that the mount command needs to run inside the pod. You can either have this as part of the pod manifest itself or you can start a shell session in the pod by running oc rsh and execute the mount command.

Jira:KATA-2579