Cisco interfaces: down/down versus administratively down

Distinguish a disabled interface from a missing signal

An administratively disabled port may never have been enabled or may have been shut down intentionally for maintenance or security. An example from show ip interface brief reads: GigabitEthernet0/1 unassigned YES unset administratively down down. Confirm the intended state before enabling it.

A different example reads: GigabitEthernet0/1 is down, line protocol is down. An enabled port without a physical signal may have an unplugged or damaged cable, an offline peer, the wrong port or module, or negotiation trouble. No shutdown addresses administrative disabling, not these underlying physical causes.

• Inspect: show ip interface brief
• Inspect the detailed port: show interface gigabitEthernet0/1
• When authorised and intended: interface gigabitEthernet0/1 → no shutdown

Two status values describe different levels

Down/down typically points to cable, peer, transceiver, speed or hardware. Up/down indicates a physical carrier with a line protocol that is not operational, potentially due to encapsulation, keepalives or peer settings depending on the interface.

Interpretation depends on platform and interface type. Read show interfaces, show interfaces status, configuration and logs together. A premature shutdown/no shutdown can change symptoms before the cause is recorded.

Counters describe quality as well as state

CRC errors often indicate damaged frames from media, connectors, optics or interference. Input Errors is an aggregate; drops can signal overload or buffer limits. Duplex mismatches can cause classic performance problems, but verify negotiation instead of assuming a mismatch.

Optical links require compatible wavelength, fibre type and supported transceivers. DOM readings can show transmitted and received power. For copper, include cable testing, pair state and the remote port.

Make changes in a controlled way

Record the initial state, counters and time, then test one hypothesis at a time: another cable, a known-good port, a compatible transceiver or a corrected setting. Clear counters only after documenting the baseline and agreeing an observation window.

For production uplinks, account for impact, redundancy and rollback. Restoring a local link may still leave VLAN, port-channel or Spanning Tree problems. Multiple affected interfaces, repeated flapping or production outages call for structured diagnosis rather than repeated resets.