Skip to content
MikroTik RouterOS Docs

Ethernet Interface Configuration

For the impatient: check interface status, force speed if needed, verify link.

# Check interface status and negotiated speed
/interface ethernet monitor ether1 once
# Force 100Mbps (both ends must match)
/interface ethernet set ether1 auto-negotiation=no speed=100M-baseT-full
# Force 1G SFP (for SFP+ port running 1G module)
/interface ethernet set sfp-sfpplus1 auto-negotiation=no speed=1G-baseX
# View all interfaces and statistics
/interface ethernet print stats

MikroTik RouterOS supports a comprehensive range of Ethernet interfaces from 10Mbps copper to 400Gbps fiber optic connections. Ethernet interfaces form the physical foundation of your network, and proper configuration ensures optimal performance, reliable connectivity, and effective troubleshooting.

This guide covers:

  • Interface types: RJ45 copper, SFP/SFP+/SFP28 fiber, QSFP+ high-speed ports
  • Auto-negotiation: How RouterOS establishes link speed and duplex
  • Monitoring: Real-time diagnostics including SFP DDM (Digital Diagnostic Monitoring)
  • Troubleshooting: Cable testing and common issues

The following diagram shows common Ethernet interface types on a MikroTik device:

@startuml
skinparam backgroundColor white
skinparam componentStyle rectangle
node "MikroTik Router" as R1 {
[ether1-ether5\nRJ45 Copper\n10/100/1000M] as ETH
[sfp1\nSFP Cage\n1G Fiber/Copper] as SFP
[sfp-sfpplus1\nSFP+ Cage\n1G/10G] as SFPP
}
cloud "Network" as NET
ETH --> NET : Twisted Pair\n(Cat5e/6)
SFP --> NET : Fiber or\nRJ45 Module
SFPP --> NET : Fiber/DAC\nHigh Speed
note right of ETH
Auto-negotiation selects
best common speed
end note
@enduml
  • Access to a RouterOS device (v7.12+ recommended for new link mode syntax)
  • Physical connection to the interface you want to configure
  • For SFP interfaces: A compatible transceiver module

Auto-negotiation is enabled by default and handles most configurations automatically. However, behavior differs between copper and fiber interfaces.

On twisted-pair connections, devices exchange capability advertisements and select the highest common speed and duplex mode.

Community Tip: Gigabit Ethernet requires all 4 cable pairs. If your gigabit link negotiates down to 100Mbps, check for non-gigabit PoE injectors in the path - they pass auto-negotiation (which uses 2 pairs) but fail actual gigabit traffic.

On fiber/DAC connections, there is no capability exchange. Each side attempts its highest supported mode, and both ends must advertise the same highest mode for the link to establish.

Example failure scenario:

  • Device A advertises: 25G-baseCR, 10G-baseCR
  • Device B advertises: 10G-baseCR only
  • Result: No link (Device A tries 25G, Device B can’t match)

Solution: Configure both ends with matching highest advertise mode.

Check current interface configuration and link status.

/interface ethernet print

Expected Output:

Flags: R - RUNNING
Columns: NAME, MTU, MAC-ADDRESS, ARP
# NAME MTU MAC-ADDRESS ARP
0 R ether1 1500 AA:BB:CC:DD:EE:00 enabled
1 R ether2 1500 AA:BB:CC:DD:EE:01 enabled

For detailed real-time monitoring including negotiated speed:

/interface ethernet monitor ether1 once

Expected Output:

name: ether1
status: link-ok
rate: 1Gbps
full-duplex: yes
supported: 10M-baseT-half,10M-baseT-full,100M-baseT-half,
100M-baseT-full,1G-baseT-half,1G-baseT-full
advertising: 10M-baseT-half,10M-baseT-full,100M-baseT-half,
100M-baseT-full,1G-baseT-half,1G-baseT-full

Community Tip: The speed property in print/set is only used when auto-negotiation is disabled. Always use monitor to see the actual negotiated rate.

In some cases, you may need to force a specific link speed - typically when connecting to legacy equipment that doesn’t auto-negotiate properly.

Warning: Forcing speed requires matching settings on both ends. Mismatched duplex causes severe packet loss.

For 100Mbps full-duplex:

/interface ethernet set ether1 auto-negotiation=no speed=100M-baseT-full

For 1Gbps on SFP interface:

/interface ethernet set sfp1 auto-negotiation=no speed=1G-baseX

Common Mistake: Some MikroTik devices (hEX, hAP ac2, RB750Gr3) require auto-negotiation enabled for gigabit links. If forcing 1G doesn’t work, leave auto-negotiation enabled.

SFP+ ports often support multiple speeds. To use a 1G SFP module in an SFP+ cage:

/interface ethernet set sfp-sfpplus1 auto-negotiation=no speed=1G-baseX

For 2.5G operation (RouterOS 7.3+):

/interface ethernet set sfp-sfpplus1 auto-negotiation=no speed=2.5G-baseX

Forward Error Correction (FEC) is required for some high-speed links (50G+) and recommended for 25G connections between CRS switches.

/interface ethernet set sfp28-1 fec-mode=fec91

FEC Mode Reference:

SpeedSupported FECNotes
25Gfec74, fec91Recommended for CRS3xx/5xx links
40Gfec74Optional
50G+fec91Required

Community Tip: Some optical modules rely on FEC functionality in the MAC. If you have unexplained packet loss on 25G+ links, enable FEC even if not strictly required.

If you see this log message:

module auto-initialization failed, try forced-mode

The SFP module has a corrupted or unrecognized EEPROM. Try forcing the speed manually:

For DAC cables:

/interface ethernet set sfp-sfpplus1 auto-negotiation=no speed=10G-baseCR

For optical modules:

/interface ethernet set sfp-sfpplus1 auto-negotiation=no speed=10G-baseSR-LR

Combo ports have both RJ45 and SFP cages but only one can be active.

# Auto mode - first connected port wins
/interface ethernet set combo1 combo-mode=auto
# Force copper (RJ45)
/interface ethernet set combo1 combo-mode=copper
# Force SFP
/interface ethernet set combo1 combo-mode=sfp

Common Mistake: Combo ports cannot use both media types simultaneously. “Combo is one or the other, not both.”

Confirm the interface is up and running at expected speed.

/interface ethernet monitor ether1 once

Expected Output:

status: link-ok
rate: 1Gbps
full-duplex: yes

For SFP interfaces, verify the transceiver is recognized.

/interface ethernet monitor sfp-sfpplus1 once

Expected Output:

sfp-module-present: yes
sfp-rx-loss: no
sfp-tx-fault: no
sfp-type: SFP/SFP+/SFP28/SFP56
sfp-vendor-name: Mikrotik
sfp-temperature: 33C
sfp-tx-power: -5.8dBm
sfp-rx-power: -5.2dBm

Community Tip: Third-party SFP modules showing “sfp-type: (unknown)” often work fine. If the link is up and traffic flows, the module is functional.

Diagnose cable problems when link won’t establish. The interface must be in “no-link” state.

/interface ethernet cable-test ether2

Expected Output:

name: ether2
status: no-link
cable-pairs: open:4,open:4,open:4,open:4

Interpreting Results:

  • open:X - Cable disconnected or broken at X meters
  • short:X - Cable pairs shorted at X meters
  • Four values represent the four twisted pairs

Check traffic counters and error rates:

/interface ethernet print stats

Key Statistics to Monitor:

CounterMeaningAction if High
rx-fcs-errorChecksum errorsCheck cable/module quality
rx-align-errorAlignment errorsCheck duplex match
rx-overflowBuffer overflowCheck for congestion
tx-collisionCollisionsCheck for duplex mismatch

Clear counters for clean troubleshooting baseline:

/interface ethernet reset-counters ether1
  1. Mismatched duplex settings - If one end is full-duplex and the other half-duplex, you’ll see massive packet loss. Always ensure both ends match.

  2. Gigabit over 2-pair cable - Gigabit requires all 4 pairs. Cat3 or damaged Cat5 with only 2 working pairs will negotiate 100Mbps (or fail intermittently at gigabit).

  3. Non-gigabit PoE injectors - These pass auto-negotiation (2 pairs) but break gigabit (4 pairs). A common hidden cause of speed issues.

  4. Forcing 1G on incompatible hardware - Some devices (hEX, hAP ac2) won’t link at forced 1G. Leave auto-negotiation enabled.

  5. FEC mismatch - Both ends must use the same FEC mode. Auto vs disabled may work; fec74 vs fec91 will not.

  6. Mixing bridge and switch VLANs - Don’t configure bridge VLANs and switch chip VLANs simultaneously for the same task.

  7. Duplicate MACs after restore - Using backup/restore on different hardware copies MAC addresses. Use export compact instead, or run /interface ethernet reset-mac-address.

Hardware Offload and VLANs: Enabling vlan-filtering=yes on a bridge disables hardware offload except on CRS3xx series. For other devices, configure VLANs through the switch chip directly to maintain wire-speed performance. Users report CPU dropping from high load to 0-3% after switching approaches.

Link Flapping on RouterOS 7.x: Some users report Ethernet links going down/up repeatedly on v7.x that didn’t occur on v6.49. Workarounds include placing a switch between WAN and router, or generating supout.rif during failures for MikroTik support.

RB5009 Multi-Chip VLAN: Traffic between switch chips must travel via CPU ports. Include the CPU port in switch VLAN configuration, or “sacrifice two switch ports to link them together externally” for wire-speed inter-chip switching.

SFP Cold Start Issues: Some SFP+ ports won’t establish link after cold restart, showing “no link, not running, slave” until manually disabled/enabled. Script a workaround if this affects your deployment.

  1. Check physical connection - Verify cable is seated, SFP module clicked in
  2. Check auto-negotiation - Try matching settings on both ends
  3. For SFP: Check sfp-module-present, sfp-rx-loss, sfp-tx-fault
  4. For high-speed: Verify FEC mode compatibility
  5. Run cable-test on copper interfaces
  1. Check link-partner-advertising for remote capabilities
  2. Verify cable category (Cat5e minimum for gigabit)
  3. Check for non-gigabit devices in the path (PoE injectors!)
  4. Verify SFP module specifications match expected speed
  1. Check error counters: rx-fcs-error, rx-align-error
  2. Monitor SFP temperature (overheating causes issues)
  3. Check cable quality with cable-test
  4. Look for rx-overflow indicating congestion
  1. Check eeprom-checksum - “bad” means corrupted module EEPROM
  2. Try forced-mode configuration
  3. Verify module compatibility with your device
  4. Reseat the module
CommandDescription
/interface ethernet printList all Ethernet interfaces
/interface ethernet print detailShow full configuration including MAC
/interface ethernet print statsDisplay traffic statistics
/interface ethernet monitorReal-time link status and SFP DDM
/interface ethernet setConfigure interface settings
/interface ethernet cable-testDiagnose cable problems (when no-link)
/interface ethernet reset-countersClear interface statistics
/interface ethernet reset-mac-addressRestore factory MAC address
/interface ethernet blinkFlash interface LEDs for identification
TaskCommand
View all interfaces/interface ethernet print
Monitor link status/interface ethernet monitor ether1
Force 100Mset ether1 auto-negotiation=no speed=100M-baseT-full
Force 1G SFPset sfp1 auto-negotiation=no speed=1G-baseX
Enable FECset sfp28-1 fec-mode=fec91
Cable test/interface ethernet cable-test ether1
Reset counters/interface ethernet reset-counters ether1
Reset MAC/interface ethernet reset-mac-address ether1

Ethernet interface configuration in RouterOS:

  1. Auto-negotiation is enabled by default and handles most configurations
  2. Copper interfaces exchange capabilities; both ends select the best common speed
  3. SFP/QSFP interfaces require matching highest advertise mode on both ends
  4. Force speed only when required (legacy equipment, troubleshooting)
  5. FEC is required for 50G+ links and recommended for 25G

Key points:

  • Use /interface ethernet monitor to see actual negotiated speed (not print)
  • Gigabit requires all 4 cable pairs - check for non-gigabit PoE injectors
  • Some devices (hEX, hAP ac2) require auto-negotiation for gigabit links
  • Match FEC mode on both ends of high-speed connections
  • Use cable-test to diagnose copper wiring issues
  • Third-party SFPs showing “unknown” type often work fine