Routing Loop Causing TTL Exceeded Errors

Step 1: Run traceroute to Identify the Loop

# Linux / macOS
traceroute -m 30 203.0.113.1

# Windows
tracert -h 30 203.0.113.1

# Expected loop output:
#  8  10.1.0.1  2.341 ms
#  9  10.1.0.2  3.112 ms
# 10  10.1.0.1  2.890 ms   <-- same as hop 8, loop confirmed
# 11  10.1.0.2  3.067 ms
# ...
# 30  * * *                <-- TTL exhausted

Step 2: Check Routing Tables on Suspected Routers

# On router-A (Cisco IOS style)
show ip route 203.0.113.0
# Expected loop: "via 10.1.0.2, GigabitEthernet0/1"

# On router-B
show ip route 203.0.113.0
# "via 10.1.0.1, GigabitEthernet0/0"  <- sends back to router-A

# Linux router
ip route show 203.0.113.0/24
# If both point to each other, the loop is confirmed

Step 3: Look for Conflicting Static and Dynamic Routes

# Linux: check all routes including policy tables
ip route show table all | grep 203.0.113

# Cisco: compare static routes against OSPF/BGP learned routes
show ip route static
show ip route ospf
show ip route bgp

# A static route with lower administrative distance can override the correct dynamic route,
# pointing traffic back to a peer that in turn has no route (and uses its own default)

Step 4: Check for Null Route / Black Hole vs Loop Distinction

# A black hole drops silently; a loop cycles with TTL exceeded messages
# If traceroute shows repeating hops = loop; if it shows * * * = black hole

# Check for explicit null routes
ip route show type blackhole
ip route show type unreachable

Step 5: Examine Recent Configuration Changes

# Linux: check recently modified files
ls -lt /etc/network/ /etc/netplan/ /etc/iproute2/rt_tables
# Identify any new static route entries

# Router: check running config change log
show archive log config  # Cisco IOS-XE

Fix 1: Remove the Offending Static Route

# Linux: remove the conflicting static route
sudo ip route del 203.0.113.0/24 via 10.1.0.2

# Verify the loop is broken
traceroute 203.0.113.1

# Make permanent by removing the route from /etc/network/interfaces, /etc/netplan/*.yaml,
# or /etc/iproute2/rt_tables as appropriate

Fix 2: Correct the Next-Hop

If the route is needed but points to the wrong next-hop:

# Replace the incorrect route
sudo ip route replace 203.0.113.0/24 via 198.51.100.1 dev eth0

# For Cisco IOS:
no ip route 203.0.113.0 255.255.255.0 10.1.0.2
ip route 203.0.113.0 255.255.255.0 198.51.100.1

Fix 3: Add a Null Route as Loop Prevention

Install a less-specific null route so that if the dynamic route disappears, traffic falls to the null route rather than the default gateway loop:

# Linux
sudo ip route add blackhole 203.0.113.0/16

# Cisco
ip route 203.0.113.0 255.255.0.0 Null0 254  # AD 254 = only used as last resort

Fix 4: Fix OSPF Redistribute Issue

# Cisco: avoid redistributing connected/static into OSPF if it creates feedback
router ospf 1
  no redistribute static subnets
  # Or use a route-map to filter:
  redistribute static subnets route-map FILTER_STATIC

route-map FILTER_STATIC permit 10
  match ip address prefix-list ALLOWED_STATICS

• Use a route policy that requires explicit next-hop verification before adding static routes affecting the default path
• Enable OSPF loop prevention features (stub areas, route filtering) and validate with show ip ospf database after changes
• Add null routes (administrative distance 254) for all major summary prefixes as a loop-of-last-resort safeguard
• Implement change management that requires traceroute validation of affected prefixes before and after routing changes
• Monitor ICMP TTL Exceeded rates via SNMP or syslog; a spike indicates a routing loop in progress