Perform MTR/Traceroute Diagnosis - Documentation Infrawire

This guide will teach you how to use MTR (My Traceroute) and Traceroute to diagnose network connectivity issues, identify latency points, and analyze the quality of your network connection.

📋 Prerequisites

A VPS server with root or sudo access
An active SSH connection
Ubuntu/Debian (commands are adapted for these distributions)

🔧 Installation

Install Traceroute (basic tool)


Bash
# On Ubuntu/Debian
sudo apt update
sudo apt install traceroute -y

Install MTR (advanced tool)

MTR combines the functionalities of ping and traceroute into a single powerful tool.


Bash
1# On Ubuntu/Debian
2sudo apt update
3sudo apt install mtr-tiny -y
4
5# For the full version with graphical interface (optional)
6sudo apt install mtr -y

Verify Installation


Bash
1# Verify that traceroute is installed
2traceroute --version
3
4# Verify that MTR is installed
5mtr --version

🛠️ Using Traceroute

Basic Command


Bash
1# Traceroute to a destination
2traceroute google.com
3
4# Traceroute to an IP
5traceroute 8.8.8.8
6
7# Limit the number of hops (default 30)
8traceroute -m 20 google.com
9
10# Specify timeout in seconds
11traceroute -w 5 google.com

Useful Options


Bash
1# Use UDP (default)
2traceroute -U google.com
3
4# Use TCP (more reliable)
5traceroute -T google.com
6
7# Use ICMP (like ping)
8traceroute -I google.com
9
10# Don't resolve DNS names (faster)
11traceroute -n google.com

Example Output

traceroute to google.com (142.250.185.14), 30 hops max, 60 byte packets
 1  10.0.0.1 (10.0.0.1)  0.123 ms  0.089 ms  0.067 ms
 2  192.168.1.1 (192.168.1.1)  5.234 ms  5.189 ms  5.145 ms
 3  * * *
 4  8.8.8.8 (8.8.8.8)  12.456 ms  12.389 ms  12.334 ms
 ...

🚀 Using MTR (Recommended)

MTR is more powerful because it sends packets continuously and calculates statistics in real-time.

Basic Command


Bash
1# MTR in interactive mode (recommended)
2mtr google.com
3
4# MTR with report (non-interactive)
5mtr --report google.com
6
7# MTR with 10 test packets
8mtr --report --report-cycles 10 google.com
9
10# MTR to an IP
11mtr --report 8.8.8.8

Advanced Options


Bash
1# Specify the number of packets to send
2mtr --report --report-cycles 50 google.com
3
4# Use TCP instead of ICMP
5mtr --tcp --port 80 google.com
6
7# Use UDP
8mtr --udp --port 53 8.8.8.8
9
10# Don't resolve DNS names
11mtr --no-dns google.com
12
13# Interval between packets (in seconds)
14mtr --interval 2 google.com
15
16# Packet size (in bytes)
17mtr --psize 64 google.com

MTR Interactive Mode

When you launch mtr without the --report option, you enter interactive mode:

d : Show/hide detailed statistics
n : Enable/disable DNS mode (name resolution)
r : Reset statistics
s : Change packet size
j : Change interval between packets
q : Quit MTR

📊 Interpreting Results

Normal Latency

< 50 ms : Excellent, local network or very close
50-100 ms : Good, regional connection
100-200 ms : Acceptable, intercontinental connection
> 200 ms : High, may cause problems

Packet Loss

0% : Perfect
< 1% : Normal for an Internet connection
1-5% : Acceptable but may cause problems
> 5% : Problematic, investigation required

Meaning of Asterisks (*)

In traceroute, a * means:

A single * instead of a time : The router did not respond to that specific packet
*Three consecutive 's : The router did not respond at all (timeout or firewall filter)

🎯 Use Cases

Diagnosing High Latency


Bash
1# Test to Google DNS
2mtr --report --report-cycles 30 8.8.8.8
3
4# Test to Cloudflare DNS
5mtr --report --report-cycles 30 1.1.1.1
6
7# Identify the problematic hop
8mtr --report --report-cycles 50 your-domain.com

Diagnosing Packet Loss


Bash
1# Test with TCP to bypass ICMP filters
2mtr --tcp --port 443 --report --report-cycles 50 google.com
3
4# Test to a specific server
5mtr --report --report-cycles 100 your-server.com

Test from Your VPS to Your Location


Bash
1# Test to your public IP address
2# Get your public IP first
3curl ifconfig.me
4
5# Then test from your VPS
6mtr --report --report-cycles 30 YOUR_PUBLIC_IP

🔍 Troubleshooting Common Problems

Problem: High Latency on a Specific Hop

If you see high latency on an intermediate hop:


Bash
# Test with different protocols
mtr --tcp --port 443 --report google.com
mtr --udp --port 53 --report 8.8.8.8
mtr --report google.com

Important: High latency on an intermediate hop may be normal if that router doesn't respond quickly to diagnostic requests, but the final latency remains good.

Problem: Significant Packet Loss

If you see high packet loss:


Bash
1# Test with more packets to confirm
2mtr --report --report-cycles 100 destination.com
3
4# Test with TCP if ICMP is filtered
5mtr --tcp --port 80 --report --report-cycles 50 destination.com

Warning: If packet loss appears on all hops, the problem is likely on your VPS or your local connection. If it appears only on a specific hop, the problem is on the network between your VPS and the destination.

Problem: Repeated Timeouts


Bash
1# Increase timeout
2mtr --timeout 10 --report destination.com
3
4# Test with different protocols
5mtr --tcp --port 443 --timeout 10 --report destination.com

📝 Useful Commands

Complete Diagnostic Script

Create a script to test multiple destinations:


Bash
1#!/bin/bash
2echo "=== MTR Test to Google DNS ==="
3mtr --report --report-cycles 30 8.8.8.8
4
5echo -e "\n=== MTR Test to Cloudflare DNS ==="
6mtr --report --report-cycles 30 1.1.1.1
7
8echo -e "\n=== MTR Test to your domain ==="
9mtr --report --report-cycles 30 your-domain.com

Save Results


Bash
1# Save report to a file
2mtr --report --report-cycles 30 google.com > mtr-report-$(date +%Y%m%d).txt
3
4# With CSV format
5mtr --report --report-cycles 30 --csv google.com > mtr-report.csv

✅ Best Practices

Use MTR rather than traceroute : MTR provides more complete statistics
Test multiple destinations : Compare results to different servers
Run multiple tests : Networks can vary, run multiple tests at different times
Use TCP if ICMP is filtered : Some routers filter ICMP but not TCP
Interpret results correctly : High latency on an intermediate hop is not always a problem if the final latency is good

🆘 Troubleshooting

MTR Doesn't Display Correctly


Bash
1# If MTR doesn't work in interactive mode, use report mode
2mtr --report destination.com
3
4# Check MTR version
5mtr --version

Error "mtr: command not found"


Bash
1# Reinstall MTR
2sudo apt update
3sudo apt install mtr-tiny -y
4
5# Or install the full version
6sudo apt install mtr -y

All Results Show Asterisks

If all hops show *, it may mean:


Bash
1# Your firewall is blocking ICMP
2# Test with TCP instead
3mtr --tcp --port 443 --report destination.com
4
5# Check firewall rules
6sudo ufw status
7sudo iptables -L

📚 Additional Resources

❓ Frequently Asked Questions

Q: What is the difference between Traceroute and MTR?
A: Traceroute sends a few packets and displays results. MTR sends packets continuously and calculates statistics in real-time, which is more useful for diagnosis.

Q: Why do some hops show asterisks?
A: This means the router did not respond. This can be due to a firewall filter, router configuration, or simply that the router doesn't respond to diagnostic requests.

Q: Should I worry about high latency on an intermediate hop?
A: Not necessarily. If the final latency to the destination is good, high latency on an intermediate hop may be normal. It's the final latency that really matters.

Q: How do I interpret packet loss in MTR?
A: Packet loss < 1% is normal. Between 1-5%, it's acceptable but may cause problems. Beyond 5%, it's problematic and requires investigation.