Why is TTL increasing? (and generally behaving erratically)

Question

Summary: TTL is not decremented when pinging hosts on a "network path"

My laptop (10.10.13.10) is connected via OpenVPN to a server (10.10.13.1) which hosts several network interfaces. One of them is 10.10.10.254, the gateway to the 10.10.10.0/24 network. 10.10.10.11 is one host on that network.

The path from the laptop to 10.10.10.11 is

>tracert -d 10.10.10.11

Tracing route to 10.10.10.11 over a maximum of 30 hops

  1    34 ms    32 ms    32 ms  10.10.13.1
  2    83 ms    34 ms    35 ms  10.10.10.11

I am now looking at TTLs when pinging the various interfaces, from my laptop up to the far remote host.

>ping 10.10.13.10
Reply from 10.10.13.10: bytes=32 time<1ms TTL=128

>ping 10.10.13.1
Reply from 10.10.13.1: bytes=32 time=37ms TTL=244

>ping 10.10.10.254
Reply from 10.10.10.254: bytes=32 time=30ms TTL=64

>ping 10.10.10.11
Reply from 10.10.10.11: bytes=32 time=29ms TTL=63

I thought that the TTL is expected to be decremented on each hop - which is not the case above. It starts with 128, jumps to 244, then 64 and 63. What can be the reason for this?

More detailed information about the setup: the test is done from Windows 10 (10.10.13.10, a hardware NIC), connected via OpenVPN to an Ubuntu 15.10 with several interfaces (10.10.13.1, a TUN and 10.10.10.254, a hardware NIC). The target device is a RPi running Raspbian (~Debian - 10.10.10.11 with a WiFi NIC).

Answer 1

The hop limit (or TTL) is decreased at each hop (unless something in your network is seriously broken). The values you are seeing are not from the packets you are sending, but rather from the replies you are receiving.

As you send packets onto the network, the hop limit will decrease at each hop until it reaches the destination. If the destination replies to your request the reply will be a new packet with a new hop limit to be decreasing.

The initial hop limit is chosen by the system generating the packet. Some of the common values for the initial value are 64 and 255. In some rare cases you will see systems copy the remaining hop limit from the request to the reply.

The different values you see are simply due to the systems you ping being configured differently.