4.19.0-okd-scos.1 vSphere IPI Install Fails at Bootstrap

[yurivader]

hi,
I am currently in the process of installing
OKD 4.19.0-okd-scos.1
Platform: vsphere
Install method: IPI

cluster in a disconnected Environment.
Environment: Vsphere 8 vcenter
Sonatype Nexus as Registry Proxy
install-config.yaml.txt
install-config section for proxy;

imageDigestSources:
- mirrors:
  - sonartype-repo.local:11445
  source: quay.io
- mirrors:
  - sonartype-repo.local:11446
  source: registry.k8s.io
- mirrors:
  - sonartype-repo.local:11449
  source: registry.access.redhat.com
- mirrors:
  - sonartype-repo.local:11448
  source: registry.redhat.io

I do not have a direct internet connection from the cluster

on the Bootstrap node the Units
node-image-pull.service and
node-image-finish.service are marked as failed.

node-image-finish.service
fails due to dependency node-image-overlay.service
fails due to dependency node-image-pull.service

when looking at the logs for node-image-pull.service
it fails with error: --create specified but ref coreos/node-image already exists

May 22 06:55:35 localhost.localdomain node-image-pull.sh[23683]: Currently on CoreOS version 9.0.20250510-0
May 22 06:55:35 localhost.localdomain node-image-pull.sh[23683]: Target node image is quay.io/okd/scos-content@sha256:3813e6608a999756931d3d621932af9662>
May 22 06:55:35 localhost.localdomain node-image-pull.sh[23683]: Not in live environment
May 22 06:55:35 localhost.localdomain node-image-pull.sh[23683]: Pulling quay.io/okd/scos-content@sha256:3813e6608a999756931d3d621932af9662860e71a552b26[...]
May 22 06:55:35 localhost.localdomain ostree-containe[23829]: Fetching ostree-unverified-registry:quay.io/okd/scos-content@sha256:3813e6608a999756931d3d[...]
May 22 06:55:35 localhost.localdomain node-image-pull.sh[23829]: No changes in ostree-unverified-registry:quay.io/okd/scos-content@sha256:3813e6608a9997[...]
May 22 06:55:35 localhost.localdomain node-image-pull.sh[23887]: error: --create specified but ref coreos/node-image already exists
May 22 06:55:35 localhost.localdomain systemd[1]: node-image-pull.service: Main process exited, code=exited, status=1/FAILURE
May 22 06:55:35 localhost.localdomain systemd[1]: node-image-pull.service: Failed with result 'exit-code'.
May 22 06:55:35 localhost.localdomain systemd[1]: Failed to start Node Image Pull.

the openshift-installer shows following Log messages after the 15 min timeout:

INFO 2025-05-22 06:55:51.453820384 +0000 UTC m=+0.018762938 image pull 41a16d633e9ad10d435382e947d262b17349467b390d3f835e7db32981105d34 quay.io/okd/scos-release@sha256:26b3364692625d89573b65069f1353a75b7b5c4aa9eef76ac32bba89c14cc5f2
INFO 2025-05-22 06:55:51.558743754 +0000 UTC m=+0.123686312 container remove d972893a899c37e8220d620e4dd432d4cb32bf918c35acae12a48ea97b423db7 (image=quay.io/okd/scos-release@sha256:26b3364692625d89573b65069f1353a75b7b5c4aa9eef76ac32bba89c14cc5f2, name=flamboyant_mcclintock, io.openshift.release=4.19.0-okd-scos.1, io.openshift.release.base-image-digest=sha256:676522d547d87a6946a45b593171417861e637236e13feb339b1dc8430b9457e)
INFO error: unable to read image quay.io/okd/scos-release@sha256:26b3364692625d89573b65069f1353a75b7b5c4aa9eef76ac32bba89c14cc5f2: Get "https://quay.io/v2/": read tcp 10.123.231.40:38552->44.215.52.77:443: read: connection reset by peer

when pulling the Images mentioned via podman pull, they pull allright.

 [root@localhost ~]# podman pull quay.io/okd/scos-release@sha256:26b3364692625d89573b65069f1353a75b7b5c4aa9eef76ac32bba89c14cc5f2
Trying to pull quay.io/okd/scos-release@sha256:26b3364692625d89573b65069f1353a75b7b5c4aa9eef76ac32bba89c14cc5f2...
Getting image source signatures
Copying blob 58c18bd4e79f skipped: already exists
Copying blob fb6485c1d3bf skipped: already exists
Copying blob b0268b2d3348 skipped: already exists
Copying blob 2916c184a58b skipped: already exists
Copying blob 54cfc3a2e459 skipped: already exists
Copying blob 6b2459a9848d skipped: already exists
Copying config 41a16d633e done   |
Writing manifest to image destination
41a16d633e9ad10d435382e947d262b17349467b390d3f835e7db32981105d34

I 've tried different versions of OKD (4.18, 4.19.scos.0)
different cluster images, eg. scos-9.0.20250510-0-vmware.x86_64.ova and scos-9.0.20250515-0-vmware.x86_64.ova

Thank you all in advance for helping me solving this mystery!