Hi, 

I recently changed my docker networking configuration to macvlan so that I could see all of the containers with individual IPs on my router. In order to avoid the issues mentioned in many of the forums, I decided to follow the official unraid guide for setting a separate interface that is docker dedicated. This seemed to work for a day and then it crashed. So after looking around, I found that I still have the macvlan calls in my syslog (relevant info attached). After browsing the forum I came across this post which seemed to suggest that if I turn off the bridging in the dedicated docker interface it would fix the issue. I dd this and now none of the containers are able to access the internet. Community apps, docker updates are both showing errors. Fix common problems says it cannot reach github. 

 

Is there a way to keep bridging off but allow docker to access the internet.?

 

Is this just a waste of time and turning off bridging will not solve the crashing issue?

 

Thanks!

 

nancenas-diagnostics-20230802-1322.zip MacVLAN before disabling bridging.txt

I guess the simpler way to ask this is :

1. Does doecker need bridged access to eth0/br0 in order to get internet if there is internet available on another interface that it has access to. 

 

11 hours ago, rnance said:

Does doecker need bridged access to eth0/br0 in order to get internet if there is internet available on another interface that it has access to. 

Nope.

On 8/2/2023 at 6:58 PM, rnance said:

Community apps, docker updates are both showing errors. Fix common problems says it cannot reach github. 

 

Not a network expert but this suggests other issues, since that would still use eth0/br0, I see that your dedicated NIC is set to DHCP, try using a static IP and don't assign a gateway, since having multiple gateways can cause issues with internet access.

8 hours ago, JorgeB said:

using a static IP and don't assign a gateway, since having multiple gateways can cause issues with internet access.

Unfortunately this didn’t work. I’ve included my updated diagnostics. I am now also getting a two NICs on one network warning from FCP. Still getting no internet access to docker containers or to unraid. I’ve also included a photo of the routing table. All of these were automatically configured. Eth2 is the docker configured interface. Br0/eth0 is the onboard nic. Both are connected to my network switch. 

nancenas-diagnostics-20230808-1401.zip

Although that by itself should not cause a problem, and certainly not worse than having a DHCP address, though FCP would not complain about that, the IP address for the dedicated docker NIC should be set to none, try that, but I think that still doesn't explain the issue, are you using the bridge on eth1? If not disable it.

24 minutes ago, JorgeB said:

Although that by itself should not cause a problem, and certainly not worse than having a DHCP address, though FCP would not complain about that, the IP address for the dedicated docker NIC should be set to none, try that, but I think that still doesn't explain the issue, are you using the bridge on eth1? If not disable it.

Ok, I set eth1 to no bridging and eth2 back to none for ip assignment. The “two NICs on one network” warning has disappeared, however the system still is not able to access the internet. 
 

doing this also removed an entry from the routing table. Attached photos and an updated set of diagnostics 

nancenas-diagnostics-20230808-1449.zip

Edited August 8, 20232 yr by rnance
Added photo of routing table

Like mentioned before the docker dedicated NIC should not interfere with internet access, for that the server uses eth0, I would suggest renaming /boot/config/network.cfg and rebooting, that will restore the network settings to default DHCP, confirm that you have internet access and if yes then follow the guide to create the dedicated docker NIC, if not post new diags with the default settings.

So I did this and it worked for a brief period (several minutes) and then stopped. Same issue. I am now getting a slightly different issue. At first CA was not getting through to the internet. FCP was not giving a failure to access github warning, and unraid connect was not throwing an error. After following This post about manually setting DNS if CA does not load, I was able to get CA working again. However, now docker will not update. All of my apps have "unavailable" listed for the version number and when I click check for updates it shows the blue "checking" symbol for a bit and then just stops. 

Attached are new logs. 

Thanks

 

nancenas-diagnostics-20230811-1354.zip