QOS_processor consumes high CPU

Document ID : KB000047903
Last Modified Date : 14/02/2018
Show Technical Document Details

Symptoms


QOS_processor is using up to 100% CPU





Solution


There may be an issue with a corrupt or missing QOS table in the database. You will see the following in the qos_processor log:

May 20 10:00:15:984 [Qos Monitor Database Updater #1, qos_processor] Did not update qos [S_QOS_DATA { qos_name: QOS_NET_CONNECT, source: cbdnmshubw1, target: 192.168.2.105:ping, nim_origin: CBDNMSHUBW1hub, origin: CBDNMSHUBW1hub, modifier: nimsoft, host: 192.168.2.105, robot: cbdnmshubw1, probe: net_connect, table_id: null, qos_def_id: null }] with update s_qos_data set origin = 'CBDNMSHUBW1hub', host = '192.168.2.105', robot = 'cbdnmshubw1', probe = 'net_connect', nim_origin = 'CBDNMSHUBW1hub', modifier = 'nimsoft', samplerate = 120 where checksum = 'D514787B69B9D6C6F9F70546C9D00F82460172FC'

This entry will fill up the log as it is constantly trying to update the table.

In this case we ran the following query:

SELECT *
FROM S_QOS_DATA
WHERE qos = 'QOS_NET_CONNECT' AND
source = 'cbdnmshubw1' AND
target = '192.168.2.105:ping';

And the query returned 0 rows.

A deactivate and activate of the probe in question did not republish the QOS. We had to deactivate, clear niscache, and activate the probe. This published the QOS and cause the QOS_processor to run correctly.