Interface status and error troubleshooting
Read interface status codes and counters, and isolate collisions, CRC errors, duplex and speed mismatches, and cabling faults.
Reading interface status
Interface status combinations
| Line protocol | Status | Meaning | First suspects |
|---|---|---|---|
| up / up | Working | L1 and L2 operational | Look elsewhere — issue is higher in the stack |
| up / down | L2 problem | Physical link OK, protocol failing | Encapsulation mismatch, keepalive issues |
| down / down | L1 problem | No carrier detected | Cable, optic, far-end shutdown, bad port |
| administratively down / down | Shutdown | shutdown configured | Someone disabled it — no shutdown |
| down / down (err-disabled) | Disabled by feature | Port security, BPDU guard, link-flap, etc. | show interface status err-disabled for reason |
GigabitEthernet0/1 is up, line protocol is upL1 and L2 both operational. Hardware is iGbE, address is 00aa.bb01.0201 (bia 00aa.bb01.0201)Burned-in vs configured MAC. Full-duplex, 1000Mb/s, media type is 10/100/1000BaseTXNegotiated speed and duplex — compare both ends. 0 runts, 0 giants, 0 throttlesClean physical layer so far. 0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignoredAny growth here points at media/duplex. 0 output errors, 0 collisions, 1 interface resetsCollisions should be zero on full-duplex.Duplex mismatches and error counters
Duplex mismatch happens when one end runs full-duplex and the other half-duplex — usually because one side was hard-coded while the other still autonegotiates (autonegotiation then fails and falls back to half-duplex). The full-duplex side transmits whenever it likes; the half-duplex side treats those simultaneous transmissions as collisions and backs off, producing CRC errors, runts and late collisions and terrible throughput that still 'passes ping'. The fix is symmetric configuration: autonegotiation on both ends, or identical hard-coded speed/duplex on both ends.
Error counters and what they indicate
| Counter | Meaning | Likely cause |
|---|---|---|
| CRC / FCS errors | Frame checksum failed on receipt | Physical-layer suspect — compare counter deltas over time with cable/optic condition |
| Runts | Frames shorter than 64 bytes | Collisions (half-duplex), duplex mismatch, faulty NIC |
| Giants | Frames over the interface MTU | MTU mismatch, baby giants/jumbo misconfig |
| Collisions | Contention on shared media | Half-duplex link, duplex mismatch |
| Late collisions | Collision detected after 512 bits | Duplex mismatch is the classic cause |
| Input/output errors | General receive/transmit failures | Buffer overruns, physical faults |
- Symptom
- Server link 'up/up' but backups crawl; collisions and late collisions increment
- Config
- Switch hard-coded 1000/full; NIC set to autonegotiate
- With switch autonegotiation disabled, the NIC cannot negotiate and falls back to half-duplex at 100 Mb while the switch stays hard-coded 100/full.
- The switch transmits full-duplex frames whenever queued; the NIC's half-duplex CSMA/CD logic calls them collisions.
- Result: link is up, pings pass, throughput collapses and late collisions accumulate — the classic signature.
Set the NIC to 1000/full explicitly (or re-enable autonegotiation on the switch) so both ends agree.
Verification workflow
Baseline interface state and error counters — Privileged EXEC
Cisco IOS / IOS XE
show ip interface brief show interfaces status show interfaces gi0/1 show interfaces gi0/1 counters errors
show ip interface briefreveals status/protocol pairs at a glance.show interfaces statusexposes negotiated speed/duplex and err-disabled.- Error counters near zero and non-incrementing indicate a healthy physical link.
- Counter snapshots are meaningless alone — clear and re-check to see increments.
Read-only verification commands.
Common mistakes
- Declaring 'cable fault' on an up/down status — up/down means L1 is fine; look at L2 causes first.
- Reading one snapshot of error counters — only increments over time are diagnostic.
- Hard-coding speed/duplex on one end only — the autonegotiating side can fall back to half-duplex, creating a mismatch.
- Expecting collisions on a healthy full-duplex link — any late collisions are abnormal.
- Treating err-disabled as a hardware failure — it is a feature action; find and fix the trigger first.
Observation: Users report brief drops; interface shows up/up with incrementing CRC errors.
- Damaged cable or connector
- Duplex mismatch
- Failing NIC or transceiver
| Check | Indicates |
|---|---|
show interfaces over time | Growing CRC with clean duplex counters points at media, not negotiation. |
show interfaces status speed/duplex vs far end | Mismatched duplex explains collisions; CRC alone favors a physical fault. |
| Swap patch lead / move to a known-good port | Fault following the cable or the port localizes the component. |
Correction: Replace the faulty media or set matching speed/duplex on both ends.
Counters stay flat and user drops stop; no new CRC increments after a busy period.
Related resources
Sources
- Cisco CCNA 200-301 v1.1 Exam Topics (official PDF) — reviewed 2026-11-02
- Cisco — Interface and Hardware Component Command Reference — reviewed 2026-11-02
Targets CCNA 200-301 v1.1 objectives. Supplemental blocks are labeled beyond exam scope.