Troubleshooting and Operations
Troubleshooting questions reward a structured method: define the symptom, isolate the layer, verify with display commands, then change one variable. High-yield tools include ping/tracert, display interface (errors, CRC, duplex), display logbuffer, display diagnostic-information, and packet capture via port mirroring. Know how to read interface counters: CRC errors suggest cabling/duplex; runts and giants suggest MTU or L2 issues; discards suggest QoS or buffer limits.
Traps: assuming a link is up because it shows “up” without checking error counters; misreading display interface duplex mismatches; and forgetting that a routing loop shows as TTL expiry in tracert. Another trap is chasing a problem at L3 when the real cause is an L2 loop or STP topology change - check display stp and MAC flaps first. Also know how to capture with mirroring-group and analyze with a packet analyzer.
Worked example: intermittent connectivity to a server. Step 1: ping and tracert to localize. Step 2: display interface on each hop for CRC/discards. Step 3: display logbuffer for link flaps. Step 4: if L2 suspected, display mac-address for flapping and display stp brief for topology churn. Step 5: mirror the suspect port and capture. Change one thing at a time and re-verify.
Rule of thumb: isolate before you fix. A single verified display output beats several speculative configuration changes, and always save a baseline (display diagnostic-information) before altering a production device.