Constant stream of handshake for peer did not complete

[Michael Brookes]

Hi

I have one wg endpoint which is CentOS 7.9.2009
(3.10.0-1160.71.1.el7.x86_64) with kmod-wireguard
9:1.0.20220627-1.el7_9.elrepo installed. Has about 300ish peers.

We have one, there could be more, cases of wireguard logging on the
other peer (typically laptops, I've seen this happen on Mac using the
app and Ubuntu)

Handshake for peer 6 ... did not complete after 5 seconds, retrying (try 4)

The handshake on the CentOS end never updates and of course the tunnel
doesn't work on the Mac/Ubuntu laptop.

To get the tunnel working again all I do is on the CentOS end remove
and add the peer, and the tunnel suddenly springs to life and traffic
flows normally.

I've since enabled debug on the CentOS end and see a constant stream of:

Handshake for peer 6 ... did not complete after 5 seconds, retrying (try 4)

Since I enabled debugging ~200 different peers have logged that.
Also see this - invalid address - a few times too for different peers:

Handshake for peer 1809 ((invalid address)) did not complete after 5
seconds, retrying (try 5)

Is this expected debug output for a setup where you have a CentOS
server with peers which are all laptops? i.e. the CentOS end is trying
to initiate the handshake and the laptop is off/wireguard app turned
off.

Is it down to some UDP traffic potentially being dropped? Some network
card tweaks needed?

Any help greatly appreciated