Operator mirroring for disconnected install in OpenShift 4 ignores the operator index image
Issue
- After mirroring a selection of four operators for a disconnected install they consume a lot of disk space.
The operators were mirrored using the following commands to a local mirror registry:
opm index prune \
-f registry.redhat.io/redhat/redhat-operator-index:v4.8 \
-p kubernetes-nmstate-operator,local-storage-operator,elasticsearch-operator,cluster-logging \
-t local-registry:8443/redhat/redhat-operator-index:v4.8
oc adm catalog mirror -a ./pullsecret.json --insecure \
local-registry:8443/redhat/redhat-operator-index:v4.8 \
local-registry:8443 \
--index-filter-by-os='linux/amd64'
info: Mirroring completed in 2h28m41.06s (4.634MB/s)
After the mirroring was complete, the mirror was using over 70GiB of disk space.
Environment
- Red Hat OpenShift Container Platform [RHOCP]
- 4.8 or higher
- podman version 1.9.3+
- opm version 1.18.0+
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.