The filename, directory name, or volume label syntax is incorrect.: unknown

[jsturtevant]

We saw an error during the k8s e2e tests which usually passes. This looks like it might be an issue in hcsshim/containerd:

May  4 11:42:01.242: INFO: At 2022-05-04 11:41:57 +0000 UTC - event for pod-host-path-test: {kubelet capz-conf-jkltp} Failed: Error: 
failed to create containerd task: failed to create shim task: hcs::CreateComputeSystem test-container-2: The filename, directory name, 
or volume label syntax is incorrect.: unknown

https://prow.k8s.io/view/gs/kubernetes-jenkins/logs/ci-kubernetes-e2e-capz-master-containerd-windows/1521811902055845888

[jsturtevant]

The test that failed is https://github.com/kubernetes/kubernetes/blob/f25c0e5f094485f710e80377fbb833d3eeed7492/test/e2e/common/storage/host_path.go#L67

[jsturtevant]

the test typically passes, this is the first failure I've seen: https://testgrid.k8s.io/sig-windows-signal#capz-windows-containerd-master&include-filter-by-regex=HostPath%20should%20support%20r%2Fw

[kevpar]

A couple questions:

• Does the failure correlate to moving to a new containerd or shim version?
• Does the test log the CRI request sent anywhere?

[jsturtevant]

• Does the failure correlate to moving to a new containerd or shim version?

no, we currently have the containerd version set to 1.6.2

• Does the test log the CRI request sent anywhere?

we collect containerd and hcsshim logs but it looks like the log got trucated. I found the pod hostpath-1199/pod-host-path-tes in kubelet logs: https://storage.googleapis.com/kubernetes-jenkins/logs/ci-kubernetes-e2e-capz-master-containerd-windows/1521811902055845888/artifacts/clusters/capz-conf-e0mz2c/machines/capz-conf-e0mz2c-md-win-7978795f5b-scxpv/kubelet.log

but the containerd logs where truncated: https://storage.googleapis.com/kubernetes-jenkins/logs/ci-kubernetes-e2e-capz-master-containerd-windows/1521811902055845888/artifacts/clusters/capz-conf-e0mz2c/kube-system/containerd-logger-vbs22/containerd-logger.log