Keywords

SNCP switching event, slow response, core dump, MML process

Version Mapping

NMS	OS	Database	NE
All	All	-	-

Symptom

After the U2000 works for a period of time, the following symptoms occur:

l   It is slow to load data after the U2000 client is restarted. The progress is always 99%. This symptom can be solved by restarting NE management processes.

l   The U2000 responds slowly. NE-level and E2E operations are often suspended.

l   Core dumps or other exceptions occur on the SDH NE management process, and the CPU usage of this process is high.

Problem Identification

This is a U2000 bug. Specifically, the U2000 server throws exceptions when processing numerous SNCP switching events.

Cause Analysis

1.         Analyze logs of NE management processes in the following directory:

%imap_root%\var\logs\Develop\nemgr_sdh_*\*

The logs of NE management processes are wrapped at a fast pace, one package in about 10 minutes.

Over 30 switching events are reported every minute, including MSP and SNCP switching events.

Qx5 header file: NSEVT_CFG_MSSPR_SWITCH_EX = 28595 (0x6FB3)  //Switching event for MSP squelch services

Qx5 header file: NSEVT_CFG_SNCP_SW = 28567 (0x6F97)  //SNCP switching event

2.         Check whether the MML process (EML_MML) has been deployed on the U2000.

If the process has been deployed, frequent reporting of SNCP switching events is the root cause of this problem.

Solution

Preventive measure:

1.         Restart the SDH NE management process.

2.         Undeploy the MML process.

3.         Disable the reporting of switching events.

Solution:

Patches are provided for different U2000 versions, which is traced by DTS2013052307892 and DTS2013052407837.

T2000 V200R007C03SPC113

U2000 V100R002C01CP5036

U2000 V100R005C00CP6035

U2000 V100R006C02SPC303 (SPC306 is recommended.)

U2000 V100R008C00CP3111 (SPC302 is recommended.)

Suggestion and Summary

Deploy the MML process only when necessary.