hole punching fails about nebula HOT 4 CLOSED

yes, 192.168.122.1 is interface virbr0

nodes 192.168.111.2 and 192.168.111.4 are on different routers, but has same ip segment 192.168.1.1/24
When I restarted many times nebula, it can work ok

from nebula.

Are 192.168.111.2 and 192.168.111.4 on the same local network? It seems surprising that the local packets are also failing, if so. IE:

 level=info msg="Handshake message sent" handshake="map[stage:1 style:ix_psk0]" initiatorIndex=1352070048 remoteIndex=0 udpAddr="192.168.1.13:46888" vpnIp=192.168.111.4
 level=info msg="Handshake message sent" handshake="map[stage:1 style:ix_psk0]" initiatorIndex=1352070048 remoteIndex=0 udpAddr="192.168.122.1:46888" vpnIp=192.168.111.4

which appears to show handshake attempts to 192.168.1.13 and 192.168.122.1. (I'm guessing 192.168.122.1 indicates libvirt is in use?)

from nebula.

Not out to necro this issue too much, but I also tried to set this up with a public lighthouse and a single node behind NAT tonight. I was unable to ping the lighthouse until I opened a node firewall port for 4242, even with reverse punching enabled. Wasn’t able to determine exactly why, but just wanted to mention it

from nebula.

I seem to be seeing a similar issue. One lighthouse in digital ocean, and a node behind NAT at home. ping seems to be sent to lighthouse, inbound/outbound ACL set to all tcp/udp, yet ping blackhole'd.

from nebula.