Network weirdness

[heisenfig]

Something just doesn't seem right to me with networking in my Unraid and hoping to get some feedback from others.

 

BACKGROUND

My LAN is 192.168.5.0/24

Unraid server is on 192.168.5.11

Under Network settings, Interface Extra shows br0

Routing table

 

 

OBSERVATIONS

For the most part, docker containers in HOST or BRIDGED mode work just fine.  However, I can only get Nginx Proxy Manager Official to work in "Custom : br0" mode.   I gave it address 192.168.5.101, which is outside of my DHCP pool.  Ok.. looks like that just adds it to my LAN as a different IP.

 

Problem is that any proxy configured in Nginx can't point to a container that is in HOST or BRIDGED mode.  It just won't load the page, gives BAD GATEWAY error.  If I move the target container to "Custom : br0", then it works fine.  It also works to use a VM as the target (ie, it can't point to the unraid host ip for anything.)  I have to give those containers that I move to "Custom : br0" it's own IP, as unraid won't let me assign to the same IP as Nginx even though the listening ports are different.

 

Now, the containers I had to move to "Custom : br0" for Nginx to access, themselves cannot access other containers that are HOST or BRIDGED (ie. MySQL container).  So I had to move that "Custom : br0" too.

 

Additionally, other containers that are HOST or BRIDGED cannot access the containers that are set to "Custom : br0", (ie. adminer for managing MySql).  Even trying to do a curl from the unraid host terminal cannot access the websites that are on "Custom : br0" even though they are on the same LAN network.  It gives "No route to host".   I can access them from other other PC's on the LAN.

 

Do I just have something setup wrong concerning the "Custom :br0" network?  It just seams really weird that they can't communicate directly with the unraid host IP like that.

 

Thanks in Advance.

James

 

 

[Vr2Io]

I haven't use proxy, but in longtime ago make some test with ipvlan and all work. Could you try macvlan / ipvlan have different ?

 

 

1 hour ago, heisenfig said:

Additionally, other containers that are HOST or BRIDGED cannot access the containers that are set to "Custom : br0"

This could be normal, I can't ping custom ethx.y Container to HOST ( I haven't docker in bridge ) and I use macvlan ( macvtap ). 

Edited June 2 by Vr2Io

[_cjd_]

nginx needs to proxy to an externally reachable IP - furthermore, it'll need to be a unique port. it's just a proxy, not a router. In your setup, those services should be 192.168.5.11:PORT in either host or bridge mode for the nginx settings. Is that not working? I have esphome set up in host mode and it's working via nginx.

 

I will note - I'm using macvlan for docker and most services get a lan/vlan assignment and a dedicated unique IP. Most services are only proxied so I can map :443 to the random port stuff runs on (esphome is :6052 for example).

[rjdipcord]

As a suggestion, take a look into inter-container networks within docker. You can actually route traffic between containers using dedicated virtual networks within docker. If you do it this way, you completely mitigate Layer 2 and Layer 3 routing for your services, thus reducing bandwidth on your LAN.

[heisenfig]

Not sure what macvlan means.  I had an older version of nginx proxy server running in just mode and it seemed fine, but couldn't get the location feature to work, so tried the official version from the apps page.  That's the one that won't run as host or bridged.  If I try, can't even load it's own ui.

Edited June 2 by heisenfig

[heisenfig]

Well..  i went back to this version which works in "host" mode and got it working with the proxy targets setup as "bridge".  3 custom websites and custom docker registry.  Everything working perfect not.  Not sure why the 'Official' one is so troublesome.  Now I have nothing in "custom br0" mode.

 

 

Edited June 2 by heisenfig