Serviceability pod in Cloud Pak for Security times out on ROKS

[nastacio]

Describe the bug I am following the post-installation instructions on the install guide to configure the OpenLDAP server: https://www.ibm.com/docs/en/cloud-paks/cp-security/1.9?topic=support-installing-cpctl-utility-access-actions

This instruction takes around 5-7 minutes and fails:

oc cp $POD:/opt/bin/<operatingsystem>/cpctl ./cpctl && chmod +x ./cpctl

I can see it slowly copying the file locally, to the tune of 200Kb/second, but after 30Mb (out of 40Mb) are copied, the oc cp command shows:

(this took ~5 minutes)

oc cp ${pod_name}:/opt/bin/linux/cpctl /tmp/cpctl --request-timeout=15m 
tar: Removing leading `/' from member names
Dropping out copy after 0 retries
error: unexpected EOF

This problem does not happen on AWS-based clusters, where the copy takes only a few seconds.

I tried it about a dozen times in a row on ROKS, the problem always happens.

I suspect some ROKS-specific problem in the way it sets up its network.

To Reproduce Steps to reproduce the behavior:

1. (see description)

Expected behavior The download of the command should not timeout.

Screenshots If applicable, add screenshots to help explain your problem.

Additional context Add any other context about the problem here.