remote machine learning address

[summX]

Having an existential crisis here: unraid docker, set machine learning host to IP of that host, set machine learning in immich admin panel to 0.0.0.0:3003. Worked this morning. Then I was forced to redo my unraid setup and now I´m stuck with "invalid address" in the log and can´t connect to the machine learning host.
I am not certain why immich needs this kind of "double" entry of values and one does not know what exactly to place where.

[bo0tzz]

0.0.0.0 is not a valid IP address to reach something on - it refers to "every interface" when binding a service to a port on a system. You need to figure out the actual IP of the remote host.

[summX]

0.0.0.0 is not a valid IP address to reach something on - it refers to "every interface" when binding a service to a port on a system. You need to figure out the actual IP of the remote host.

strangely it was working. I had the IP address of my remote learning machine in docker settings and 0.0.0.0:3003 set in immich ui admin panel. Anyway. So I need to set docker settings to remote host IP AND immich admin panel? this still does not work as I get ERROR: invalid address ('IP', 3003) and it won´t connect.

I can test with telnet ip 3003 and it works. It was working before.

[summX]

It now works with 0.0.0.00 as MACHINE_LEARNING_HOST and IP of host in admin panel.

[dbatkinsmith]

Does this still work for you, or has it changed with the 1.106.x updates? I'm getting invalid address errors no matter what I try.

[summX]

Does this still work for you, or has it changed with the 1.106.x updates? I'm getting invalid address errors no matter what I try.

So far it does. I find those settings confusing nonetheless.