I noticed Community Apps couldn't connect so I cranked up a Terminal window and noticed I couldn't ping out.  Went and looked at the network settings page and it looks like it seems to think the primary default gateway is 192.168.1.1...  

 

The 192 address goes to a router that has no Internet access.  That is used for my security camera system.  The correct gateway is the 10.10 address.  I tried changing the metric on the 192 address and can't, nor can I delete it.

 

How can I get rid of that default address?

Did some digging (here, here) and playing around:

 

 

...and I still can't get it to work.  It seems that you can't change a route once it's there.  You can only delete it and re-add it.  Cool.  No problem.  

When I try to do that, though, I get the "RENETLINK answers: No such process" error.  Not sure why.

Ideas, anyone?  I just need the 10.10.1.1 GW to be default!

Edited August 23, 20205 yr by rmp5s

Could you static the 192 IP NIC without gateway (blank) and no bridge on it.

Edited August 23, 20205 yr by Benson

37 minutes ago, Benson said:

Could you static the 192 IP NIC without gateway (blank) and no bridge on it.

I can't do anything with it. Supposedly you can't modify an entry...only add or delete. Which is fine...I'd delete the 192 and add it back with a metric of 1000 so it's not used...but I can't get it to delete.

K.  So.  What I ended up doing was stopping the VM and Docker services, going into the Network Settings and manually entering a metric of 1000 for that interface.  Cranked everything back up and we're back in business.  Simple enough.

you made my day 🙂 thanks for the solution

 

```

route add default gw 192.168.20.1 br0

```

On 8/23/2020 at 8:28 PM, rmp5s said:

Did some digging (here, here) and playing around:

 

 

...and I still can't get it to work.  It seems that you can't change a route once it's there.  You can only delete it and re-add it.  Cool.  No problem.  

When I try to do that, though, I get the "RENETLINK answers: No such process" error.  Not sure why.

Ideas, anyone?  I just need the 10.10.1.1 GW to be default!

You missed a /0 after 0.0.0.0, I tried and the route gets deleted

 

ip route del 0.0.0.0/0 via 192.168.1.0/24

 

Just want to say I solved this by delete the route I didn't want.

In my case I wanted to delete the default route pointing to a 192.168.100.1 gateway

I used the following command to do it: 
 

ip route del default via 192.168.100.1

And it's gone! Hope this helps others!

How to make these changes persistent? After reboot everything goes back to the same before changing default route.

 

The command "ip route change default via [IP]" works but does not stick.

I've been waiting for an answer to this question for a very long time. I'd like to know how the OS determines which interface to use as the default route. As an example, I have a 10GBE PCI-E card, that has two interfaces: eth0 and eth1. Eth1 is always disconnected, but always gets picked for the default GW for the network. Every time unraid starts, I have to go into it, and manually reconfigure the network.

 

So, is there a way to change this? Even if it's going into the console and changing some files?

 

Because I'm pretty sure that if we could have this knowledge, we'd solve the problems discussed here.

This is annoying, just added a 10Gbe card to one of my unraid servers (the only one that didn't start with a 10g card) and I see all sorts of connection issues.  It's like Unraid just WANTS to use the original 1G nic no matter what I do.  Server is up but I get things like community apps and Unraid Connect connection errors.

@MiguelBazil  I am no expert but I had thought that "Interface Rules" would determine that.  Whatever NIC you assign to eth0: would be the default route.  But despite putting in my 10Gbe card as eth0 Unraid insists on sending the default route through br1 (bridge interface).  >brctl show command tells me that eth1 is attached to br1.  I've tried adding routes, messing with the metric to no avail.

 

@rmp5s I don't know if this will be any help but I have noticed that if you have more then one NIC (typically your eth0) with bridging enabled that I would get issues with various things not connecting (Unraid Connect and Community Apps being some of them).  Which makes sense since (I THINK) bridging is what allows apps/docker to communicate to that interface so having more than one can be an issue?).

Edited January 24, 20251 yr by Nnyan

18 hours ago, Nnyan said:

@MiguelBazil  I am no expert but I had thought that "Interface Rules" would determine that.  Whatever NIC you assign to eth0: would be the default route.  But despite putting in my 10Gbe card as eth0 Unraid insists on sending the default route through br1 (bridge interface).  >brctl show command tells me that eth1 is attached to br1.  I've tried adding routes, messing with the metric to no avail.

 

I thought so too, but that's not the case. It goes even further: if I have an interface assigned to eth0 and another to eth1, and I switch assignments, the default route moves with the assignment. In a sense, it's similar to rmp5s issue. That's why I'd like to understand exactly how it's determining these things. There are a few things I'd like to also be able to do, like configuring an interface to ALWAYS start down, for example. Maybe that could solve both problems. The reality is that I don't know enough of the base OS to even try to do it through a script that at that level.

 

Up until now, I've connected to unraid after each restart, and manually recreate the routes. With the latest version (7.x.x), I can't even do it through the network, as the interface is not responding properly until I set the correct default route... That means I have to physically go to the server, and add it there, otherwise I don't have network connectivity. Not an option if I'm away and there's a power outage.

Edited January 25, 20251 yr by MiguelBazil

@MiguelBazil have you ever found a solution?

I wish, but no. Eventually I got a fiber optics upgrade in my house because I started playing with homelabbing, which led to have a single fiber-optics card with 2 ports, and the issue got solved that way. Port 2 of that card doesn't have an SFP, and as such, never gets configured, which means, single port. I still would like to know those things because I do think it'd help to create some more specific configurations for users, but no dice. If anyone ever finds the info, feel free to tag me.

Thanks everyone (@MiguelBazil@rmp5s and others I won't bug with tagging) I was helped by this thread... eventually.  I ran into this issue wile switching my home network to a DIY opnsense router, WAP, vlans, etc.  I also added a 10G SFP+ card to my unraid server.  So many variables... 

 

I couldn't connect to my unraid server across VLANs.  I thought it was a network vlan issue at first.

I had a really tough time figuring out where my problem was because I had to reasonably assume that my lack of understanding of my brand-new opnsense and vlan tagging on switches adventure was my problem.  I was wondering if I had collision issues, firewall, all sorts of stuff I knew little about.

 

Eventually, after a ton of self-doubt and trying to fix firewall issues, I narrowed down that it was not an inter-vlan issue.  I could iperf3 across vlans between other devices, and I could even ping unraid from other vlans, but the iperf3 would only send a packet or so then nothing when trying to connect to unraid from specific vlans.

 

I noticed that the only VLAN connections to the server having issues were those VLANs I had created connections for to Unraid on my other 2.5G NIC (I was already having issues, probably the same reason, when I had all the vlans running through the same 10G nic).

 

After a LOT more time and frustration, and thanks to this thread, I was able to successfully 'fix' the issue by deleting some routing table entries.  However, @MiguelBazil now has me worried that this is something I will need to do every reboot.

Right now, I have a fix (maybe) and more questions:

1. Deleting some of the ipv4 entries related to the gateways for the vlans (which I shouldn't need), appears to have allowed traffic to my primary bridge as expected. 

2. I had to use the CLI to delete the 'default' one, which did not immediately fix traffic.  I don't know for sure if this was required.

3. I don't know enough about routing tables to know what should be there and what shouldn't.  Or if there's any other stuff there I don't need.

4. Will I need to re-do this after every reboot?

5. How will I ever know that my routing table is correct?

13 hours ago, QMoney said:

2. I had to use the CLI to delete the 'default' one, which did not immediately fix traffic.  I don't know for sure if this was required.

In essence, the default route is the IP that is selected for traffic forwarding when no next hop, or other rules, matches. Usually you assign it to the router you're connected to, because that's the point on your network that decides where traffic goes. You can either get that info from DHCP if you receive IP configuration from the network, or you'll have to configure it on the card manually. Last resort, it's what I used to do, where Unraid selected the wrong default route interface, and I had to go to Unraid's console, delete the route from 1 interface, and manually create it for the other.

 

If the default route is pointing to the right router IP, using the right network interface, you won't have to delete it. If the IP/interface combo is not right, then yes, you need to recreate it like I did, and probably figure out why the system is not pointing to the right place.

 

13 hours ago, QMoney said:

4. Will I need to re-do this after every reboot?

It really depends on your scenario. I'd say, write down the current configuration you have, restart unraid, and observe. If you have to reconfigure it after the restart, then it's possible that you will have to continue doing it every restart until something changes. In my case, it was essentially making sure that I had a single network interface available on Unraid. That way, he had to use that one. Once that happened, Unraid automatically selected the right default route.

 

13 hours ago, QMoney said:

5. How will I ever know that my routing table is correct?

If you can access everything you want, then it should be. 95% of the times, you shouldn't need to care about that, as long as the network allows you to do what you need. The other times are when you want to screw around and find out 😁

On 5/14/2025 at 4:42 AM, MiguelBazil said:

In essence, the default route is the IP that is selected for traffic forwarding when no next hop, or other rules, matches. Usually you assign it to the router you're connected to, because that's the point on your network that decides where traffic goes. You can either get that info from DHCP if you receive IP configuration from the network, or you'll have to configure it on the card manually. Last resort, it's what I used to do, where Unraid selected the wrong default route interface, and I had to go to Unraid's console, delete the route from 1 interface, and manually create it for the other.

 

If the default route is pointing to the right router IP, using the right network interface, you won't have to delete it. If the IP/interface combo is not right, then yes, you need to recreate it like I did, and probably figure out why the system is not pointing to the right place.

Thanks! It has been working like a champ since deleting the 3 lines. I want to keep the multiple NICs so I can isolate tagged vlan traffic from the untagged port and offload a bit of noise from my main line.

I know that the last 2 entries were required because each fixed a problem immediately, it's just the first of the three I'm not sure about. It's the 1st one below. I don't think it was needed vlan if the VLAN is only for container access:

On 5/14/2025 at 4:42 AM, MiguelBazil said:

It really depends on your scenario. I'd say, write down the current configuration you have, restart unraid, and observe. If you have to reconfigure it after the restart, then it's possible that you will have to continue doing it every restart until something changes. In my case, it was essentially making sure that I had a single network interface available on Unraid. That way, he had to use that one. Once that happened, Unraid automatically selected the right default route.

 

If you can access everything you want, then it should be. 95% of the times, you shouldn't need to care about that, as long as the network allows you to do what you need. The other times are when you want to screw around and find out 😁

I have not had an opportunity to do a restart and see if it needs fixing. I took a copy of the interface and console versions of the routing tables for when I do.

I'm trying to do some research on linux routing tables. I don't think it needs to be very complicated for unraid but some of the automated entries make it look a bit more complicated than it needs to be. It is not intuitive though.