Problem Description

The SSN2EFS4 single board at OSN3500 site repeatedly reports COMMUN_FAIL alarms, the service is normal, the host version: 5.21.17.31, EFS4 version: 5.30. The EFS4 single board is configured for the first-level aggregation service, and the external ports are not used.

Alarm message

commun_fail

Processing

1. De-enable the corresponding port of the EFT that reported the ETH_LOS alarm. After the ETH_LOS alarm disappears, observe that the EFS board is no longer reset and the fault is eliminated.

2、Upgrade the software to V1R8C01B01c or later versions can also solve this problem.

Root cause

1, collect the single board black box bb5.log, found that there are a lot of application memory failure records, ERROR CODE: 70001 indicates that the module ID for 0x7E ( 126 ) module frequently apply for 0x400 (i.e., 1K ) size of memory, due to the exhaustion of the memory failed to apply for, resulting in a soft reset of the single board.

2. After analyzing the reason why the memory was not released by R&D, it was confirmed that it was caused by N2EFS4 not being able to handle the unknown GFP management frame.

3, finally located in the docking M500 device is the source of the unknown GFP management frame. EFS single board docked M500 EFT single board port ETH_LOS alarm, constantly sending GFP customer signal loss frame to the EFS single board, resulting in the EFS single board GFP management frame processing task to apply for memory after the application of memory is not processed to release, which ultimately led to the single board memory depletion reset, reported COMMUN! _FAIL alarm

Recommendation and Summary

None