High CPU usage consumed by U2000 server

Since we upgraded our U2000-Server to V200R016C60CP2023 from version V200R016C60SPC200 our server is consuming much more cpu-power. There is a system task, which uses 50% of all 16 vcores.

The name of the process is BulkCollectorDM.I think it’s not normal, because it’s running for days now. Since our upgrade we didn’t change anything in our network and we didn’t add any new device.

[Root Cause]

[Solution]

·       U2000 V200R016C60 CP2025 (for U2000 V200R016C60CP2023)

Step 2      Delete the core file core.BulkCollectorDm* generated by the BulkCollectorDm_1 process. The asterisk (*) indicates that the file contains a file name extension and is uniformly identified. This operation frees up disk space.

On Solaris, assume that the installation directory is /opt/oss. Run the following commands as the ossuser user to delete the core file:

cd /opt/oss/server/var/logs

rm -rf core.BulkCollectorDm*

On Windows, manually delete the core.BulkCollectorDm* file. If the installation directory is D:\oss, delete the file in D:\oss\server\var\logs.

[Confirmation]

The BulkCollectorDm_1 process is running properly and does not automatically restart. Historical performance data of the bulk instance can be collected and viewed.

Step 1      On the Service Monitor tab page of the System Monitor client, check whether the BulkCollectorDm_1 process is running properly. Check the start time of the process to determine whether the process is running properly. If the start time does not change, the process does not automatically restart.

Step 2      After two collection periods, verify that the historical performance data of PTN V8 NEs (such as PTN7900/PTN990/PTN970/PTN905E) or OSN9800/OSN9600 NEs is displayed.

Step 3      After two collection periods, verify that the historical performance data of the instance created by using the default bulk template under the Router/Switch/MA/ME NE node is displayed.