When attempting to download content from a private repository, the Aria Automation Orchestrator is observing an error message similar to:
<date> info Looking in indexes: https://<private repository domain>/<private repository uri>/<date> error WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ProxyError('Cannot connect to proxy.', OSError('Tunnel connection failed: 502 Bad Gateway'))':/<private repository uri>/
Aria Automation/Orchestrator 8.18.x
The private repository may be missing from the internal proxy's user proxy exclude list.
Recommendation:
take a non-memory quiesced snapshot of all nodes in vCenter before making any changes.
SSH to the Aria Automation/Orchestrator appliance (just 1 if in a cluster is enough).
Run 'vracli proxy show' to retrieve the current settings for the internal proxy, save/copy these settings some.
The section and domains of interest are:
"user-proxy-exclude": "docker-registry.prelude.svc.cluster.local,localhost,.cluster.local,192.,vra-k8s.local,kubernetes,kubernetes.default.svc.cluster.local,<my domain 1>,<another domain 2>,<some other repository domain X>,...",
Apart from the environment's internal domains and other existing private or external domains, add the <private repository domain> that is failing to the list as follows:
vracli proxy set --proxy-exclude "docker-registry.prelude.svc.cluster.local,localhost,.cluster.local,192.,vra-k8s.local,kubernetes,kubernetes.default.svc.cluster.local,<my domain 1>,<another domain 2>,<some other repository domain X>,<the private repository domain>,..."
Run:
vracli proxy apply
Running the 'apply' command above will restart the proxy pods and update the list across all nodes in the cluster.
Reversing this change consists of the same commands, only excluding from the list.<the private repository domain>