N10-009 ยท Troubleshooting methodology, cabling and interfaces
Link down, flapping and speed/duplex problems
Applying the methodology to link-down, flapping and speed/duplex scenarios.
Key points
- Link down: check LEDs, patch cable, port status and the far end, then swap known-good parts one at a time.
- Flapping: look for a failing cable, connector, NIC or optic and for interference.
- Speed/duplex: if one side is hard-set and the other auto-negotiates, the auto side may fall back to half duplex, which shows as late collisions and CRC errors; set both sides the same or use auto on both.
Practice questions
When a port has no link light on either end, which OSI layer should you check first?
- Layer 1, physical
- Layer 3, network
- Layer 4, transport
- Layer 7, application
Show answer
Answer: A. Layer 1, physical
No link at both ends is a Layer 1 symptom: cable, connector, port or optic. IP and application settings cannot cause a dead link light.
Which are plausible causes of a link that keeps flapping?
- A failing patch cable or loose connector
- A dying transceiver or dirty fiber end
- An aggressive power-saving setting on the NIC
- A wrong DNS server address
- A deny entry in an ACL
Show answer
Answer: A. A failing patch cable or loose connector, B. A dying transceiver or dirty fiber end, C. An aggressive power-saving setting on the NIC
A failing patch cable, a worn connector or optic and a NIC power-saving feature can all drop the link repeatedly. A wrong DNS server and ACL entries do not toggle the physical link.
If one end of a link is hard-set to 100/full and the other end is on auto, the auto end will correctly detect full duplex.
Show answer
Answer: False
A fixed side stops sending negotiation signals. The auto side can usually sense speed but defaults to half duplex, causing a mismatch.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
