Connection problems with macOS machine on tailscale network

Hi, I have four machines that or on my Tailscale network and which are currently online:

  • two Synology NASes running Tailscale 1.16.2 [named “orb” and “gem”]
  • one Linux machine running Tailscale 1.22.2 [named “wok”]
  • one macOS 12.3.1 machine running Tailscale 1.20.1 [named “ink”]

The NASes and macOS machine are on the same subnet. The Linux machine is on a different network, over the Internet.

The macOS machine can’t connect to any other machine on the Tailscale network, nor can the others connect to it. The others can connect to each other fine.

tailscale status on the macOS machine says:

# Health check:
#     - haven't heard from home DERP region 5 in 13m40s

100.68.121.31   ink                  heycam@      macOS   -
100.113.228.4   gem                  heycam@      linux   -
100.69.101.96   orb                  heycam@      linux   -
100.85.134.13   wok                  heycam@      linux   -

Earlier today (before a tailscale down / tailscale up) it was saying this:

# Health check:
#     - ReceiveDERP is not running
#     - ReceiveIPv4 is not running
#     - ReceiveIPv6 is not running

100.68.121.31   ink                  heycam@      macOS   -
100.113.228.4   gem                  heycam@      linux   -
100.69.101.96   orb                  heycam@      linux   -
100.85.134.13   wok                  heycam@      linux   -

tailscale netcheck on the macOS machine:

cam@ink:~ $ /Applications/Tailscale.app/Contents/MacOS/Tailscale netcheck

Report:
	* UDP: true
	* IPv4: yes, XXX
	* IPv6: no
	* MappingVariesByDestIP: false
	* HairPinning: false
	* PortMapping:
	* Nearest DERP: Sydney
	* DERP latency:
		- syd: 20.9ms  (Sydney)
		- sin: 90ms    (Singapore)
		- blr: 123.2ms (Bangalore)
		- tok: 148.5ms (Tokyo)
		- sfo: 167.1ms (San Francisco)
		- sea: 183.6ms (Seattle)
		- dfw: 194.3ms (Dallas)
		- ord: 211.4ms (Chicago)
		- nyc: 215.2ms (New York City)
		- fra: 243.1ms (Frankfurt)
		- lhr: 246.5ms (London)
		- sao: 349ms   (São Paulo)

What might be causing this lack of connectivity?

Thanks,

Cameron

Holding option while clicking on the Tailscale icon will add a Debug selection with a “Bug report”, which would provide more information for troubleshooting.

Thanks, filed BUG-a3a46d5d48d1d4fbd2ba9a5f95f7fb826df14830ae60e3d38fcae6a4346bad3a-20220410025519Z-326c6a56644229f3.

Hello.

the first interesting thing I see around that bug report is that your computer time is jumping every couple of minutes. Do you have a very short delay before your computer goes to sleep?

I do have “Turn display off after” on battery set to “2 mins”. Not sure if that also sleeps the machine.

That alone shouldn’t. But the logs show it is falling asleep very quickly. I’d try disabling that to test, and run it for a bit to see if the unwanted behaviour resurfaces.

I’m still getting this connection problem when I am on mains power, where the sleep time is much longer.

Another data point: I just tried running /Applications/Tailscale.app/Contents/MacOS/Tailscale down then Tailscale up, and the command is hung on the Tailscale up.

I relented and restarted my Mac machine, and the connection is now working. So, not sure what the problem was. Will report back if it happens again.

The macOS app is constructed in two pieces:

  • a Network Extension, which is the macOS technology that can manipulate routing tables and provide the VPN functionality but has to run in a sandbox with little access to the rest of the OS
  • the GUI, which communicated with the Network Extension

It sounds like the Network Extension was not running any more. The next time it happens if you open /Applications/Utilities/Console.app and search for “tailscale” and “IPN”, it might turn up information about some kind of crash.