Currently running 6.4 rc7a

If i go into docker here is a list of the network types dropdown

http://prntscr.com/g2ioc4

Bridge is the traditional docker bridge that was there before 6.4 rcx

br0 is my main ethernet bridge as 192.168.1.100 with gateway 192.168.1.1

eth1 is my secondary ethernet setup as 192.168.2.99 with gateway as 192.168.2.1 (i changed it to 192.168.1.1 just to see if it would change anything but same behavior) (i've also tried eth1 in bridge mode as br1 no change)

 

I had a custom docker network setup on my eth0 before switching over to 6.4 and it was working great with all my containers so i decided to switch to 6.4 to have a friendlier ui for managing those.

If i add a container to br0 i can connect to the container web ui but there is no dns resolution from the container

If i add a container to eth1/br1 the container starts up but nothing can access it nor does it have any dns resolution.

If i add a container to Bridge everything works as it originally did.

I can also delete eth1 docker network and recreate it but it has the same issues.

Also tried using the extra parameters "--network eth1 --ip 192.168.2.2" like before with network type as 'none' and that no longer works correctly either.

 

Any ideas? i was fighting with this all weekend decided to throw all my containers back onto Bridge for the time being so everything would work correctly.

 

When creating custom networks (br0, br1) there are no dns settings included. At the moment the only way to add dns settings is by including them in the "Extra parameters".

 

Which dockers require dns?

 

Currently i am using nzbhydra, sonarr, radarr, ombi. not sure if any of the other containers i use need dns but those are the ones i know for sure.

Would this be the correct extra parameter to add in for dns? "--dns=IP_ADDRESS"

yes

Maybe its more than just that, i cannot even ping my router from the docker container but i access the web interface (when set to br0 in this case). While setting back to Bridge I can ping etc

Did you set a fixed IP address for the container?

 

If not you likely need to configure an IP range under Docker settings (need to stop the docker service to do so), which is used to hand out addresses to the containers. Make sure it doesn't conflict with your DHCP server. Docker is unaware of what is already in use.

 

52 minutes ago, bonienl said:

When creating custom networks (br0, br1) there are no dns settings included. At the moment the only way to add dns settings is by including them in the "Extra parameters".

 

Which dockers require dns?

 

 

When it is done this way, does the container still acquire DNS entries from the host and append these?  Or, will the container use only the ones provided via 'Extra Parameters'?

 

John

@bonienl i'm setting the ip address manually and it's an unused address.

4 minutes ago, johnodon said:

When it is done this way, does the container still acquire DNS entries from the host and append these?  Or, will the container use only the ones provided via 'Extra Parameters'?

 

My understanding of the Docker documentation is that it will only use the entries defined by --dns. Haven't tested it though.

2 minutes ago, baneofserenity said:

@bonienl i'm setting the ip address manually and it's an unused address.

 

You will need to use the Fixed IP address field.

Just now, bonienl said:

 

You will need to use the Fixed IP address field.

Yes i have tried using that field and used the --ip=address extra param as well just in case.

# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name="Nginx" --net="br0" --ip="10.0.1.80" -e TZ="Europe/Berlin" -e HOST_OS="unRAID" -e "TCP_PORT_80"="80" -e "TCP_PORT_443"="443" -e "PUID"="99" -e "PGID"="100" -e "VERSION"="latest" -v "/mnt/cache/appdata/Nginx":"/config":rw linuxserver/nginx

Above the docker create command when I create a container with br0 and fixed IP address (10.0.1.80). Have a look and compare to yours.

/usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name="plexpy" --net="br0" --ip="192.168.1.139" -e TZ="America/New_York" -e HOST_OS="unRAID" -e "TCP_PORT_8181"="8181" -e "PUID"="99" -e "PGID"="100" -v "/mnt/user/appdata/plexmediaserver/Library/Application Support/Plex Media Server/Logs/":"/logs":ro -v "/mnt/user/appdata/plexpy":"/config":rw --dns=8.8.4.4 linuxserver/plexpy

 

No real noticeable differences. 

I am using Plexpy too. When changing this container to a unique address it can communicate with the gateway, but it can't communicate with Plex due to a security restriction that Docker has built in (containers in a custom network can not talk to the host address).

 

In other words keep plexpy as 'host' network (same network type as plex).

 

This was working fine before i upgraded to 6.4 when setting up the extra network manually. I'm just using plexpy to mess around with, it is the easiest container for me to mess with without bringing anything "important" down.

Any manual configuration - if still existing - will interfere with the implementation of 6.4. Did you delete your manual settings?

 

They seem to have gotten overwritten during the upgrade, if i make any manual changes, rebooting the server will replace them.

They will interfere. Remove all manual remnants and test again.

 

br0 hasn't been touched, there is nothing to remove.

Anything in your 'go' file which creates macvlan?

 

/boot/config/go ?

#!/bin/bash
# Start the Management Utility
/usr/local/sbin/emhttp -p 59049 &

nope

12 minutes ago, baneofserenity said:

They seem to have gotten overwritten during the upgrade

 

So how / where are these settings done previously?

 

I follow this guide, 

 

What is the output of

docker network ls