Hey dear Unraid community,

I've been having problems using Wireguard with ipv6 since 7.1.0.

I've created an IPv6-only configuration and whenever I try to enable it, the switch immediately disables itself again.

To be more precise, the tunnel can be switched on if no peer is present. As soon as a peer is present, the tunnel cannot be activated.

In the logs of Unraid I see again and again that Unraid tries to retrieve the WAN IPv6 address via https://wanip6.unraid.net/.

Unfortunately, the query times out after a while.

I did check where the error is using “curl -v https://wanip6.unraid.net/”. The address is resolved correctly but I get no response. Briding is enabled in my network configuration.

When I try to retrieve the WAN IPv6 address over br0 using “curl -v --interface br0 https://wanip6.unraid.net”, everything works and I get my correct address.

I've already been able to rule out that this is due to the rest of my network hardware.

I'm using a Fritzbox as a router and a RasPi as an Adguard Home server. I've activated the RasPi for PING6 on the Fritzbox as a test.

The RasPi's IPv6 GUA address is publicly accessible and also provides a response when pinged. Likewise, “curl -v https://wanip6.unraid.net/” also works on the RasPi.

I enabled the Unraid for PING6 via the Fritzbox, but unfortunately I get no response.

I also used “trace route / tracert” to check where exactly the traffic is discarded.

The RasPi responds normally and shows my Fritzbox in the penultimate hop and the RasPi itself as the last hop.

Unraid, on the other hand, also shows the Fritzbox in the penultimate hop, but after that all connections go nowhere.

I have currently been on Unraid 7.1.4 for a week. Unfortunately, the problems are also present in the latest version.

IPv6 worked under Unraid 7.0.1.

Edited July 7, 20251 yr by purplehexagon
Add image

Please post a screenshot of the Wireguard settings you are using to create the tunnel, without the keys naturally, and also the diagnostics.

23 minutes ago, JorgeB said:

Please post a screenshot of the Wireguard settings you are using to create the tunnel, without the keys naturally, and also the diagnostics.

Edited July 8, 2025Jul 8 by purplehexagon
removed diagnostics for privacy reasons

43 minutes ago, purplehexagon said:

Please post a screenshot of the Wireguard settings

Sorry if I wasn't clear, a screenshot from the WireGuard page when creating the tunnel, with the options used, to try and duplicate

33 minutes ago, JorgeB said:

Sorry if I wasn't clear, a screenshot from the WireGuard page when creating the tunnel, with the options used, to try and duplicate

Sorry for the missunderstanding.

Edited July 7, 20251 yr by purplehexagon
Hide endpoint address

Thanks, I'll see if I can duplicate as soon as I have some time

Thanks in advance. If you need something else f.e. my routing table or network settings don't hesitate to ask.

Edited July 7, 20251 yr by purplehexagon

Looks like it may be a bug similar to one before for another config, at least I see the same error, please post the output from cat /var/log/wg-quick.log after trying to start the tunnel.

30 minutes ago, JorgeB said:

please post the output from cat /var/log/wg-quick.log after trying to start the tunnel.

Yep, same error I see: Error: inet address is expected rather than "dev".`

I'll report it to LT, seems similar to this one https://forums.unraid.net/bug-reports/prereleases/wireguard-broken-on-710-rc1-r3757/

17 minutes ago, JorgeB said:

Yep, same error I see: Error: inet address is expected rather than "dev".`

Thank you. Yes, it seems like Unraid is trying to create an ipv4 route, but this can't work because there is no ipv4 address for an ipv6-only tunnel.

However, this is only one half of the errors mentioned.
I still don't understand why I can't ping my server from outside my internal network via ipv6 and that Unraid keeps throwing timeouts in the logs when trying to resolve the WAN IPv6 address.

Inside my network everything seems normal my global unique address ist reachable from inside but from the outside it isn't and infact for unraid only.

I run the connectivity checks via a smartphone hotspot paired to my laptop without connecting to my internal network, if that matters

Edited July 8, 20251 yr by purplehexagon
typo

35 minutes ago, purplehexagon said:

and that Unraid keeps throwing timeouts in the logs when trying to resolve the WAN IPv6 address.

I also see this, part of my report.

36 minutes ago, purplehexagon said:

I still don't understand why I can't ping my server from outside my internal network via ipv6

I would expect that to be normal without the tunnel working, or do you mean a different thing?

48 minutes ago, JorgeB said:

I also see this, part of my report.

Yes, I see it as part of your report too, but I suspect it's not the final cause. In the end though, I'm no expert.

49 minutes ago, JorgeB said:

I would expect that to be normal without the tunnel working, or do you mean a different thing?

Yes, I think it is something else. If I set up a tunnel where both ipv4 and ipv6 are used (screenshots attached), the tunnel can be activated, but the WAN IPv6 request to https://wanip6.unraid.net/ is still not answered.

As I described at the beginning, the request can be resolved on the DNS side, but it ends in a timeout.

To me it looks like the IPv6 traffic can leave Unraid, but receiving is impossible if the traffic comes from outside my local network.

I have now tested again and made both my smartphone and my Raspi reachable via an ipv6 address, and I was able to ping both devices from outside my local network.

So it must be my Unraid server blocking incoming IPv6 traffic, or am I on the wrong track here?

That means even if the WireGuard tunnel could be switched on, it would be of no use as Unraid remains unreachable via IPv6.

However, I have no idea what the problem is either.

Sceenshots for wg3 where using IPv4 and IPv6:

wg-quick of wg3 after enabling the tunnel:

Timeout error:

Edited July 8, 20251 yr by purplehexagon
typo

This should be fixed in 7.2.0-beta.1, please retest with that release, or wait for the rc/stable if you don't want to use a beta.

On 8/2/2025 at 9:47 AM, JorgeB said:

This should be fixed in 7.2.0-beta.1, please retest with that release, or wait for the rc/stable if you don't want to use a beta.

I waited for the official release of 7.2.0 so that I could continue testing.

Now that 7.2.0 has been released, I tried again to create a Wireguard tunnel with IPv6 only and a tunnel with IPv6+IPv4 combined.

And indeed: a connection is possible, everything seems to be working again.

The only thing that still appears in the logs is the following line:

Nov 4 09:25:32 PurpleGravity http_get_contents: Curl error 28: Connection timed out after 15002 milliseconds. Requested url: 'https://wanip6.unraid.net'

Before 7.2.0 i was able to curl -v --interface br0 https://wanip6.unraid.net/ and get a corret response. Now with 7.2.0 i can't get a response with any interface.

The exact answer of my above command is:

Host wanip6.unraid.net:443 was resolved.

* IPv6: 2600:1f14:b8f:3b02:2f81:a0d1:f4d9:9e12

* IPv4: (none)

* Trying [2600:1f14:b8f:3b02:2f81:a0d1:f4d9:9e12]:443...

* socket successfully bound to interface 'br0'

* connect to 2600:1f14:b8f:3b02:2f81:a0d1:f4d9:9e12 port 443 from <redacted> port 43296 failed: Connection timed out

* Failed to connect to wanip6.unraid.net port 443 after 133307 ms: Could not connect to server

* closing connection #0

curl: (28) Failed to connect to wanip6.unraid.net port 443 after 133307 ms: Could not connect to server

But other than the message, everything appears to be working?

5 minutes ago, JorgeB said:

But other than the message, everything appears to be working?

Correct. Anything but the message is working! Thanks for the amazing work done here.