Filesystem corruption on an Elasticsearch pod on OpenShift 4.x

Issue

• An Elasticsearch instance is stuck on CreateContainerError with the following error in the OpenShift events in the openshift-logging namespace:

  Error: relabel failed /var/lib/kubelet/pods/374138d9-a02e-4470-91a8-b63653c7ccc4/volumes/kubernetes.io~csi/pvc-76158b52-88f2-43ed-ae9b-8bf4175a1bfd/mount: lstat /var/lib/kubelet/pods/374138d9-a02e-4470-91a8-b63653c7ccc4/volumes/kubernetes.io~csi/pvc-76158b52-88f2-43ed-ae9b-8bf4175a1bfd/mount/elasticsearch/data/nodes/0/indices/OcBsYffpRBepn2EDbTrh_A/_state/state-545.st: bad message
  

• This error comes from the OpenShift node kubelet logs:

  Jun 17 12:15:02 node01.example.com hyperkube[1541]: E0617 12:15:02.200829    1541 remote_runtime.go:200] CreateContainer in sandbox "50f97a52935c10c919033166e3e3459fa6f2ccb6998839b084f122af89db7c47" from runtime service failed: rpc error: code = Unknown desc = relabel failed /var/lib/kubelet/pods/c8f949ed-0bcd-432c-9ebf-7df762b5ea7e/volumes/kubernetes.io~csi/pvc-76158b52-88f2-43ed-ae9b-8bf4175a1bfd/mount: lstat /var/lib/kubelet/pods/c8f949ed-0bcd-432c-9ebf-7df762b5ea7e/volumes/kubernetes.io~csi/pvc-76158b52-88f2-43ed-ae9b-8bf4175a1bfd/mount/elasticsearch/data/nodes/0/indices/OcBsYffpRBepn2EDbTrh_A/_state/state-545.st: bad message