PlantUML DiagramUnidirectional Link Failure
A structured lesson workspace with readable content, hands-on examples, and a clean path to completion.
Every few hours, users on one floor experience a 30-second outage. The pattern is irregular but happens multiple times per day. No config changes, no log errors you can see without digging. This is the shape of an STP problem — a periodic reconvergence you don't know about.
STP failures rarely announce themselves. They manifest as intermittent slowness, brief outages, or traffic that flows through the wrong switch. Diagnosis requires reading counters and logs that most engineers ignore until something breaks.
A new switch joins the network with priority 4096 (default for some vendors) and becomes root. Every flow in the campus now traverses it. If it's a small access switch with 1 Gbps uplinks, performance tanks. This is why you configure explicit root priorities on your intended root switches — set them lower than any accidental candidate can achieve.
Fiber pairs sometimes fail on just one strand. If BPDUs keep flowing out but not in, a switch thinks the topology is healthy when it isn't. STP can move a blocked port to Forwarding because it hasn't received a BPDU in Max Age seconds, causing a loop.
UDLD (UniDirectional Link Detection) catches this. It sends periodic echoes and shuts down ports where the echo doesn't return. Configure `udld aggressive` on all fiber uplinks. It's the cheapest insurance against a nasty failure mode.
Every time a port transitions to Forwarding or a link goes down, the switch sends a Topology Change Notification (TCN) up to the root. The root then floods TCNs to every switch, and every switch shortens its MAC address aging time from 300 seconds to 15 seconds for the next 35 seconds. This flushes MAC tables and causes temporary flooding while the tables rebuild.
Frequent TCNs mean frequent MAC table flushes. If a port flaps every 30 seconds, you're in a constant state of MAC table rebuild — massive unnecessary flooding. The fix is stabilizing the flapping port, not tuning STP.
47 topology changes in the lifetime of the switch, with the last one 2 minutes ago. That's a lot. For a stable network, you'd expect single-digit changes over months. The pattern — high count, recent — is a clue that something is flapping.
Gi0/17 is generating TCNs. That port is probably flapping — either a cable issue, a NIC cycling, or a device on the far end resetting. Check the interface counters and the connected device.
STP problems usually surface as 'the network is slow sometimes.' The path to diagnosis is almost always the same: check for topology changes, find the source, fix the flapping port, and stop tuning STP while you're at it. Add UDLD on fiber uplinks, verify BPDU Guard and PortFast are set correctly on access ports, and the vast majority of STP failures simply won't happen.
Finish the lesson once you have worked through the material. This awards ★ 30 XP.
Select each vulnerable line, then classify the finding.
Findings
Review target: 16 lines · 2 expected findings