Messages

Anyone know how to debug this not starting?

the error is

Code Select Expand

Notice root /usr/local/etc/rc.d/os-udpbroadcastrelay: WARNING: failed to start osudpbroadcastrelay
I've found a github thread saying wire guard was an issue but I don't use it. Other than that error I cant work out whats wrong.

Running     Opnsense 24.7.10_2 at present

I had the same problem, in my case I wanted to support mDNS on port 5353, but it was already occupied by Unbound because I run AdGuard on port 53.

I had to move Unbound to port 5354 and then I could start the relay

I'm starting to believe I cannot have it working with one of my ISP handing out a dynamic prefix.


I've setup a gateway group for IPv6, then I've enabled NTPv6 for my static prefix on my dynamic one if the interface is starlink.

With this setup, if my primary ISP is online, I was routed trough it, but when it went offline, the gateway group forced my gateway to starlink and the NTP engaged.

The problem is that I had to hardcode my current starlink delegation in NTP, and this will stop working as soon as the prefix rotate.

It is also a lot to configure, as I have 6 VLAN in my network

I spend more time on this over the weekend and I managed to get to a better situation.

First of all, I was wrong about virtual IP, I had to restart radvd, but it actually announce all the prefixes in that way.


Current status:

- I have a working IPv4 load balancing between the two ISP
- My local interfaces are tracking IPv6 from the dynamic ISP (starlink) and have a virtual IP on the static prefix from my other provider (that does not support DHCPv6-PD)
- For IPv6 I don't have a gateway group (should I have one?) and I've put two rules on each interface to select the correct ipv6 gateway based on the source ip. Because only one is static, I had to use reverse logic on the only prefix I know
- With this, if both ISP are online, clients works fine

Now the problems:

- If my static ISP goes down, connections start failing randomly, some goes trough the starlink and work, others don't
- I was using DNS over TLS with Unbound with only IPv6 servers, for some reasons it fails with SERVFAIL if the starlink connection is down

For the DNS stuff, I've added also IPv4 resolvers and now it's working, but the other issue is still there and makes the hole setup unusable.

Hello, I've installed an opnsense box. I'm coming from an openwrt background.

I have 2 ISP with "funny" ipv6 support:

one fiber provider gives me a static ipv4/32 and a static ipv6/61. IPv6 is not doing PD, I have to manually set the next hop on their appliance web interface.

The other one hands out a /56 with PD but it's dynamic and IPv4 is GCNAT.

In my openwrt setup, I have the first provider set as static with link-local addresses and the /61 set as static PD. With this setup, each internal network gets an ipv6 on both providers and can do failover and route selection at the device level.

With opnsense I was not able to replicate this setup.
It seems I cannot set a static PD and use track interface, but I have to set static ipv6 manaully on each subnet.

Moreover it seems that I cannot add the second ISP prefix.

I've tried setting up my networks using "track interface" from isp2 (that has dynamic PD) and then add the well-known static as a virtual IP, but in that case I don't have RA on the additional network.

How can I get radvd to advertise prefixes from all the upstream interfaces?