Open jungin500 opened 1 year ago
There is something about the addressing which isn't matching what was programmed into Wireguard. [vbziu]
is the exit node.
2023-03-04 06:54:10.557418 +0000 UTC: peer keys: [vbziu]
2023-03-04 06:54:20.353862 +0000 UTC: wg: IPv4 packet with disallowed source address from [vbziu]
2023-03-04 06:54:20.354145 +0000 UTC: wg: [vbziu] - Received handshake initiation
2023-03-04 06:54:20.354175 +0000 UTC: wg: [vbziu] - Sending handshake response
2023-03-04 06:54:20.415299 +0000 UTC: wg: [vbziu] - Receiving keepalive packet
2023-03-04 06:54:22.425246 +0000 UTC: wg: IPv4 packet with disallowed source address from [vbziu]
2023-03-04 06:54:32.426778 +0000 UTC: wg: [vbziu] - Sending keepalive packet
2023-03-04 06:54:51.748462 +0000 UTC: wg: IPv4 packet with disallowed source address from [vbziu]
2023-03-04 06:55:01.749751 +0000 UTC: wg: [vbziu] - Sending keepalive packet
2023-03-04 06:55:37.773954 +0000 UTC: wg: IPv4 packet with disallowed source address from [vbziu]
2023-03-04 06:55:47.77561 +0000 UTC: wg: [vbziu] - Sending keepalive packet
2023-03-04 06:56:06.943972 +0000 UTC: wg: IPv4 packet with disallowed source address from [vbziu]
2023-03-04 06:56:16.945536 +0000 UTC: wg: [vbziu] - Sending keepalive packet
2023-03-04 06:56:20.682345 +0000 UTC: wg: IPv4 packet with disallowed source address from [vbziu]
...
2023-03-04 06:59:18.928263 +0000 UTC: open-conn-track: timeout opening (TCP [fd7a:115c:a1e0:ab12:4843:cd96:q:r]:50660 => [2606:x::x:x]:443) to node [vbziu]; online=yes, lastRecv=4s
2023-03-04 06:59:18.994624 +0000 UTC: open-conn-track: timeout opening (TCP [fd7a:115c:a1e0:ab12:4843:cd96:q:r]:50661 => [2606:x::x:x]:443) to node [vbziu]; online=yes, lastRecv=4s
2023-03-04 06:59:18.995222 +0000 UTC: open-conn-track: timeout opening (TCP [fd7a:115c:a1e0:ab12:4843:cd96:q:r]:50662 => [2606:x::x:x]:443) to node [vbziu]; online=yes, lastRecv=4s
2023-03-04 06:59:19.018358 +0000 UTC: open-conn-track: timeout opening (TCP [fd7a:115c:a1e0:ab12:4843:cd96:q:r]:50663 => [2606:x::x:x]:443) to node [vbziu]; online=yes, lastRecv=4s
2023-03-04 06:59:19.253875 +0000 UTC: open-conn-track: timeout opening (TCP [fd7a:115c:a1e0:ab12:4843:cd96:q:r]:50673 => [2606:x:x::x]:443) to node [vbziu]; online=yes, lastRecv=5s
2023-03-04 07:00:22.737974 +0000 UTC: allowing exit node access to local IPs: [127.0.0.0/8 fe80::/64]
https://github.com/tailscale/tailscale/blob/a9c17dbf939049ba7fc8f0be0bc1cb3510753287/ipn/ipnlocal/local.go#L3433 is trying to add the local interfaces used for VMs when --exit-node-allow-lan-access
is set, but it doesn't appear to be handling Parallels.
Likely duplicate of https://github.com/tailscale/tailscale/issues/2339
Is there any update on this or #2339 ?
What is the issue?
While using Tailscale along with Parallels Desktop, enabling exit node makes Parallels network interface (
bridge100
,bridge101
, ...) subnet is not reachable from Host machine. Guest VM still had internet connection but couldn't ping each other(Host->Guest, Guest->Host).Disabling exit node and restarting Parallels network interface through
ifconfig bridge101 down; ifconfig bridge101 up
resolves problem.Steps to reproduce
bridge101
succededtraceroute
at this point but traiffic falls through Host network interface's gateway (public network), not a Tailscale network norbridge101
interface subnet.Are there any recent changes that introduced the issue?
I haven't used this combination, so I can't conclude software update could be possible issue.
OS
macOS
OS version
macOS Ventura 13.2.1
Tailscale version
1.36.2 (App Store)
Other software
Firewalls
Ad-Blockers
Bug report
BUG-0a2c38da09ffdaf500b0f7174416f5772ab4fb18e7297f30e19c55beff1c6eb2-20230304065900Z-fae07145e379bdf9