ETH-CFM grace is an indication that MEPs on a node undergoing a maintenance operation may be expected to be unable to transmit or receive ETH-CC PDUs, failing to satisfy the peers requirements. Without the use of a supporting grace function, CCM-enabled MEPs time out after an interval of 3.5 ✕ ccm-interval. During planned maintenance operations, the use of grace can extend the timeout condition to a longer interval.
The Ethernet CFM system-wide configuration eth-cfm>system>[no] grace-tx-enable command controls the transmission of ETH-CFM grace. The ETH-CFM grace function is enabled by the Soft Reset notification by default. The ETH-CFM grace function determines the individual MEP actions based on their configured parameters.
To transmit a grace PDU, the MEP must be administratively enabled and ETH-CC must also be enabled. The ETH-CC interval is ignored. Grace transmission uses the class 1 DA, with the last nibble (4 bits) indicating the domain level, for all grace-enabled MEPs. When a grace event occurs, all MEPs on a node that are configured for grace actively participate in the grace function until the grace event has completed. When a soft reset occurs, ETH-CFM does not determine which peers are directly affected by a soft reset of a specific IOM or line card. This means that all MEPs enter a grace state, regardless of their location on the local node.
The grace process prevents the local MEP from presenting a new timeout condition, and prevents its peer, also supporting a complementary grace process, from declaring a new timeout defect (DefRemoteCCM). Other defects, unrelated to timeout conditions, are processed as during normal operation. This includes the setting, transmission, and reception processing of the RDI flag in the CCM PDU. Because the timeout condition has been prevented, it can be assumed that the RDI is caused by some other unrelated CCM defect condition. Entering the grace period does not clear existing defect conditions, and any defect condition that exists at the start of the grace period is maintained and cleared using normal operation.
Two approaches are supported for ETH-CFM grace:
Both approaches use the same triggering infrastructure but have unique PDU formats and processing behaviors. Only one grace transmission function can be active under an individual MEP. MEPs can be configured to receive and process both grace PDU formats. If a MEP receives both types of grace PDUs, the last grace PDU received becomes the authority for the grace period, using its procedures. If the operator needs to clear a grace window or expected defect window on a receiving peer, the appropriate authoritative reception function can be disabled.
Active AIS server transmissions include a vendor-specific TLV that instructs the client to extend the timeout of AIS during times of grace. When the grace period is completed, the server MEP removes the TLV and the client reverts to standard timeout processing based on the interval in the AIS PDU.