So here is a brief summary of my setup:
I’ve set up a pull-through cache container registry reachable through local network only reverse proxy and I added to Hostname Database (in Global Configuration for Network) that it can be reachable via localhost (127.0.0.1), entry was essentially 127.0.0.1 registry.local.mydomain.redacted.
I also have an HTTP Proxy setup as of late because either my own gov or iX/StorJ began to block access to the update service.
Since when you set up HTTP Proxy, docker pulls get routed too, and my proxy is rather slow, I wanted pulls to get routed through the registry cache, which doesn’t use HTTP Proxy.
Here is the problem:
If I try to add the local registry through Docker Registries with HTTP Proxy enabled, I get an error.
But if I remove/disable HTTP Proxy, it gets added just fine.
I don’t believe this is correct behavior that localhost (127.0.0.1) gets routed through a proxy, and I am inclined to open a bug report for this on Jira, but am I correct to think so?
Should I instead open a feature request asking for the equivalent of NO_PROXY envar to be added instead, or maybe for more granular control of what services should get routed through HTTP Proxy?
Also, is there perhaps a way to set NO_PROXY in TrueNAS in a way that would persist across updates?