-
Feature
-
Resolution: Unresolved
-
Undefined
-
None
-
None
-
None
-
False
-
None
-
False
-
Not Selected
I tried installing the mirror registry, following the instructions from product docs:
https://docs.redhat.com/en/documentation/openshift_container_platform/4.17/html-single/disconnected_environments/index#creating-mirror-registry
The result was a very cryptic error message:
$ ping -c1 registry.lab.example.com PING servera.lab.example.com (172.25.250.10) 56(84) bytes of data. 64 bytes from servera.lab.example.com (172.25.250.10): icmp_seq=1 ttl=64 time=0.059 ms ... $ sudo mkdir /var/quay $ sudo ./mirror-registry install --quayHostname registry.lab.example.com --quayRoot /var/quay ... INFO[2024-11-14 13:51:10] Install has begun INFO[2024-11-14 13:51:10] Found execution environment at /home/registry/temp-tar/execution-environment.tar INFO[2024-11-14 13:51:10] Loading execution environment from execution-environment.tar ERRO[2024-11-14 13:51:10] An error occurred: exit status 127
The commands above prove I had everything right: using an existing path, a valid FQDN, and with full sudo access.
Yes I was missing an obvious requirement – the server didn't have podman nor the container-tools installed. But product docs never state that podman was a prerequisite, and the instalation program could check that and other prerequisites and provide a clear and explicit error message.
Or, being based on an Ansible playbook, they could include a few extra tasks to install podman and anything else required by the mirror registry.
To make things worse, product docs also do not state any way to troubleshoot mirror registry installation errors.
I agree it wouldn't be hard to find the --verbose option and retrying using it, as I did, but again it's something that could be stated on docs and on the error message ("try again with --verbose to get troubleshooting information") for the benefit of customers who may not be very knowledgeable in Red Hat products.