ECC errors indicate faulty memory on a specific hardware component.
ECC errors on the Routing Engine (RE):
user@lab-router> show log chassisd | match "ECC" Apr 1 01:41:15 CHASSISD_SBE_DETECTED: Too many single bit ECC errors in routing engine DRAM Apr 1 01:41:15 send: red alarm set, device Routing Engine 0, reason Host 0 memory ECC S
ECC errors on the Compact Forwarding Engine Board (CFEB):
user@router> show log messages | match "ECC" Jun 6 11:14:56 router cfeb BCHIP 1: multiple correctable ECC errors Jun 6 11:14:56 router cfeb BCHIP 1: ECC from SDRAM bank 1, at bit 39 was corrected Jun 6 11:14:56 router cfeb CM: Slot 1: Recoverable error detected; multiple ECC errors Jun 6 11:15:02 router cfeb BCHIP 1: multiple correctable ECC errors Jun 6 11:15:02 router cfeb BCHIP 1: ECC from SDRAM bank 1, at bit 39 was corrected
ECC errors on the Flexible PIC Concentrator (FPC) in slot 2, as reported by the System Control Board (SCB):
user@router> show log messages | match "ECC" Feb 28 11:30:49 router scb BCHIP 2: correctable ECC error Feb 28 11:30:49 router scb BCHIP 2: ECC from SDRAM bank 0, at bit 62 was corrected Feb 28 11:30:49 router scb CM: Slot 2: Recoverable error detected; ECC error Feb 28 12:58:52 router scb BCHIP 2: correctable ECC error Feb 28 12:58:52 router scb BCHIP 2: ECC from SDRAM bank 0, at bit 62 was corrected Feb 28 12:58:52 router scb CM: Slot 2: Recoverable error detected; ECC error
There is Dynamic Random Access Memory (DRAM) on the RE, in addition to Synchronous Dynamic Random Access Memory (SDRAM) on various PFE components. All of this memory uses ECC, which allows for the detection (and sometimes correction) of bit errors.
ECC errors can be either correctable or uncorrectable. Any hardware that logs an ECC error should be replaced, unless they occur infrequently.