Wireguard 1to1 crashing after long period of time

Routers
1

I was running a wireguard 1 to 1 connection between a Flint (wg server) and Opal (wg client) for about 18 days with no issue. Today, I couldnt reach the lan anymore from Flint to Opal, yet the wireguard connection was still connected. I can reach the Opal but no lan behind it anymore, Opal web interface showing this where the wg connection is connected though:

image

Log is spamming this:

Sun Feb 4 12:04:52 2024 kern.info kernel: [1633232.084828] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:07:41 2024 kern.info kernel: [1633400.424249] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:09:46 2024 kern.info kernel: [1633525.666319] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:11:47 2024 kern.info kernel: [1633646.325443] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:13:49 2024 kern.info kernel: [1633768.232739] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:15:53 2024 kern.info kernel: [1633892.498445] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:17:53 2024 kern.info kernel: [1634012.936047] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:19:53 2024 kern.info kernel: [1634133.020117] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:21:53 2024 kern.info kernel: [1634253.152861] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:23:55 2024 kern.info kernel: [1634375.012528] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED
Sun Feb 4 12:25:55 2024 kern.info kernel: [1634495.068701] wireguard: wireguard-hotplug IFNAME=wgclient ACTION=KEYPAIR-CREATED

connection is on though:

image

Connecting via SSH on the Opal and I can ping all clients on the subnet with no problem.

Seems some service crashed on the Opal mostly GLInet logic scripts? Anyone seen this happening before? Workaround solutions except rebooting the Opal?

Doing this remotely on the Opal via ssh had no effect:

/etc/init.d/firewall reload

But doing this fixed the issue:

ifup wgclient

Now web interface shows properly connected status again and reaching Opal lan also works:

image

Seems there is some flaw in the Glinet watchdog script. any way to write a simple workaround script for this? how would you check if the connection is broken though, if you can reach the wgserver and other way around.

2

Do you connect Opal by cable or repeater? I’ll do some tests.

3

Opal is connected to the Internet via USB tether with a USB 4g stick. Like seen above, the wg connection was still up, I could ssh remotely into the Opal via WG link, yet the Opal web interface was showing the above issue, and I couldnt reach the LAN anymore from wg server Flint to Opal lan. SSH into Opal over wg link and doing “ifup wgclient” fixed the issue. I noticed the issue, because I couldnt reach LAN of Opal anymore from other side, where it was working for 18 days.

1 Like
4

Thanks. I’ve set up a wireguard network with Opal and Flint.
Could you please export the log, when issue happens again? I’d like to see the debug info included.
image

5

Have you encountered this issue for opal firmware 4.3.11?

6

Might be, the issue has vanished since Opal fw 4.3.11, I cant say with 100 % chance, but I have not encountered the issue anymore since some weeks. It was not reproducable before though and happened randomly.

1 Like