Description of problem: One of the containers fails to get created with the error "Manifest does not match provided manifest digest" in some e2e tests we are working on fixing (https://bugzilla.redhat.com/show_bug.cgi?id=1695202). https://bugzilla.redhat.com/show_bug.cgi?id=1669096 may be related/dupe? Since CI is running on master, I assume this cluster has that fix already. The image for the container is pulled from quay, unlike most images used in e2e tests, if that matters.

Version-Release number of selected component (if applicable):
0.0.1-2019-04-10-154258
https://storage.googleapis.com/origin-ci-test/pr-logs/pull/22446/pull-ci-openshift-origin-master-e2e-aws/6793/artifacts/e2e-aws/clusterversion.json

How reproducible: It happened twice (out of ~six potential tries, depending on scheduling randomness) that an image was "corrupted" on a node, quay.io/k8scsi/hostpathplugin:v0.4.1 on ip-10-0-170-145.ec2.internal and quay.io/k8scsi/hostpathplugin:v1.0.0 on ip-10-0-149-210.ec2.internal (see
https://github.com/openshift/origin/pull/22446#issuecomment-482278918 and https://openshift-gce-devel.appspot.com/build/origin-ci-test/pr-logs/pull/22446/pull-ci-openshift-origin-master-e2e-aws/6793).

Steps to Reproduce:
1.Run the csi-hostpath e2e tests on a 4.1 cluster
2.Some may fail because the "hostpath" container of the "csi-hostpathplugin-" pod gets the "Manifest does not match provided manifest digest" error

Actual results: pods fail w/ "Manifest does not match provided manifest digest"


Expected results: pods run


Additional info: I'll try to reproduce on a cluster I have ssh access to tomorrow, hoping there is an easy answer to this in the meantime. Thx