Diagnosing connectivity issues

Reward Gateway monitors uptime from locations around the globe at frequent intervals.

Sometimes our clients may experience more localised connectivity issues than these checks can detect. In this case, we will need information from their IT team to diagnose and resolve these problems.

This article explains the information we will need from our clients in these circumstances.

Basic Connection Details

Clients can easily get details of their connection from ifconfig.io.

Clients should then send us screenshot this page and send it to us.

My Traceroute

My Traceroute, known as MTR, combines ping (used to test the reachability of a remote system) and traceroute (maps the other systems between the sender and receiver.)

Using MTR on Linux

For Red Hat based operating systems, our clients can download mtr by issuing the following command:

$ yum install mtr

For Debian based operating systems, download MTR by issuing the following command:

$ apt-get install mtr

After either of the above steps, issue the following command:

mtr --report --report-cycles 10 [destination_host]

Using MTR On Windows

Clients can download a Windows equivalent of mtr, WinMTR, from SourceForge

Open WinMTR and enter the destination host in the “Host” field.

Then, click start.

Using MTR on MacOS

Clients will need to use Homebrew to install MTR with the following command:

$ brew install mtr

After this, they should see the Linux command above.

Results

There are two important columns in the results:

1. Packet Loss
   Packet loss is an indicator of the quality of the link between source and destination. This may be caused by a device in between limiting or filtering traffic. A high level of packet loss indicates a problem that needs investigating.
   
   
2. Round-Trip Time (RTT)
   RTT measures the time it takes for messages to go from the source to destination and back. If the destination is a long way away, this number may be large. A high number may also suggest a problem with the quality of the link.
   
   

Sometimes devices that restrict or block traffic along the path from the source to the destination may prevent MTR from working properly. This does not mean there is no connectivity.

Clients need to provide us with screenshots of the MTR results.

A TCP traceroute

New versions of MTR also support conducting a TCP traceroute. This uses TCP SYN packets in place of ICMP pings to identify where traffic is being dropped instead. These tests are less useful for checking the end-to-end connectivity (many internet routers discard these packets) but can help determine if firewall rules are set properly.

TCP traceroute on Linux and MacOS

Clients should follow instructions above to install a new version of MTR.

Issue the following command:

mtr  --tcp --port [destination_port] --report --report-cycles 10 [destination_host]

Running MTR in TCP mode will require super-user privileges on most machines.

TCP traceroute on Windows

Before clients can do a TCP traceroute on Windows, they will need to download and install both the WinPcap library and tracetcp.

At a Command Prompt they will then be able to conduct a trace by running the following command:

> tracetcp [destination_host]:[destination_port]

Clients should provide us with screenshots of the MTR results.