Identify and Resolve Link Failures

Resolve Physical Link Degragation

If the link does not experience a hard signal loss, configure mechanisms to isolate a faulty link before it degrades traffic. Use this approach when isolating the link is preferable to operating in a degraded state and when sufficient capacity or alternate paths are available for rerouted traffic.

Configure optical thresholds to isolate the link when the signal crosses the low warning level. To view optical thresholds, use the following command:

To trigger a link-down when the optical level crossed the low threshold, apply this configuration:

Use link‑degrade monitoring to track link quality without shutting down the interface when it crosses a warning threshold. Use this feature when you want to monitor degradation in real time and take action only after it reaches an unacceptable level.

link-degrade-monitor tracks physical link quality on high‑speed Ethernet interfaces by monitoring error rates such as FEC, CRC, and symbol errors. When error rates exceed configured thresholds, the system can generate alarms, syslog messages, or shut down the interface to prevent traffic loss.

Packet loss can occur at relatively low bit error rates (BER), especially for real‑time applications such as video. Traditional protocols (BFD, OAM, LACP) and Ethernet Local Fault Signaling (LFS) might not detect these conditions early enough. Use link‑degrade monitoring to detect and address these issues sooner.

The system monitors physical‑layer errors by using a sliding window algorithm that tracks:

• PCS block errors and BER
• MAC FCS errors
• FEC errors
• BIP‑8 or B1, B2, B3 counters

When BER crosses configured thresholds, the system generates notifications and can shut down the local and remote link by using LFS.

Configuration options include:

• link-degrade-monitor-enable—Enable monitoring
• no-link-degrade-monitor-enable—Disable monitoring
• actions media-based—Apply media-based actions (bring down the link locally and remotely)
• recovery { auto | manual }—Configure automatic or manual recovery
• recovery timer <seconds>—Specify the recovery interval
• thresholds { set | clear | warning-set | warning-clear } <value>—Define BER thresholds
• thresholds interval <seconds>—Specify the monitoring interval

The following example shows a basic configuration without automatic recovery:

The configuration implements the following behavior:

• Normal: BER < 1e‑9—No action
• Warning: BER > 1e‑6—Generate syslog messages and alarms only
• Critical: BER > 1e‑3—Shut down the interface after 100 consecutive link‑degrade events (adjust this value as needed)
• Clear: Require manual clearing of the link‑degrade condition

To restore the interface, manually issue the appropriate request command.

Note:

The MX304 router does not support link-degrade-monitor due to the PHY hardware limitation.

Resolve Unidirectional Logical Degradation

A failure can be asymmetric, occurring in only one direction. Use BER monitoring to detect Layer 1 degradation when supported. If the hardware or peer device does not support this feature, use sub‑second detection protocols to respond quickly.

Do not rely on IGP adjacency timeouts because they use longer detection intervals. LACP provides a minimum detection time of 3 seconds in fast mode, which might not be sufficient.

Use the following point‑to‑point protocols for sub‑second detection of unidirectional link failures:

• BFD and micro‑BFD (for LAG)
• OAM LFM

These protocols operate in a distributed manner. Line card CPUs handle packet transmission and reception, while the Routing Engine processes only state changes.

BFD is protocol‑dependent and supports single‑hop and multi‑hop operation. For link degradation detection, configure single‑hop BFD and associate it with the IGP (IS‑IS or OSPF). The IGP maintains standard keepalive timers, while BFD uses sub‑second intervals for faster detection.

The following example shows BFD configuration for IS‑IS (family inet) with 100‑ms intervals:

Note:

BFD uses UDP port 3784.

For LAG interfaces, micro-BFD is used. It automatically creates a single-hop BFD session for each member link and operates in parallel with LACP PDUs:

Note:

micro-BFD uses UDP port 6784.

You can check the state of BFD and micro-BFD sessions with the following command:

BFD is a routed protocol that uses UDP transport and relies on IPv4 or IPv6. It is typically tied to a routing protocol. If you need a protocol-agnostic solution to detect non-clean failures or degradation, you can use OAM LFM. This standardized protocol operates directly at the Ethernet layer (agnotics to any families configured) and can be enabled on all Ethernet interfaces, including LAG.

To configure the routing protocol parameters, apply the settings from the following template. Detailed parameter explanations follow the configuration snippet:

OAM LFM configuration includes two components. The first component defines an action profile (for example, OAM_DOWN) that specifies tracked events, such as loss of OAM adjacency, and actions, such as generating a syslog message or setting the interface operational state to down.

The second component applies the configuration to interfaces. Configure the keepalive interval, threshold (multiplier), and reference the action profile. When you apply this configuration to a LAG interface, OAM LFM automatically creates a session for each member link.

To monitor OAM LFM session state, use the following command: