heap memory of vmware probe suddenly rises

book

Article ID: 7419

calendar_today

Updated On:

Products

DX Infrastructure Management NIMSOFT PROBES

Issue/Introduction

vmware probe heap memory increases rapidly when polling cycle is not enough. 

Cause

When polling cycle is not enough and vmware needs more time than polling cycle, it will poll as soon as last poll ends. This causes memory increase because vmware probe does not have enough time to perform garbage collection.

Environment

Release: CNMSPP99000-8.4-Unified Infrastructure Mgmt-Server Pack-- On Prem
Component:

Resolution

Please set loglevel=3 and wait for 1 polling cycle. 

Then please check for following vmware.log. 

 

Apr 28 23:08:59:344 [pool-186059-thread-1, vmware] PERF: START: ----- Pass-7876 --- Data Collector Check Interval for: <monitor name> 

Apr 28 23:08:59:921 [pool-186059-thread-1, vmware] PERF: START: Check for quick stat updates: https://<monitor name>:443/sdk 

Apr 28 23:14:07:222 [pool-186059-thread-1, vmware] PERF: DONE: Check for quick stat updates: https://<monitor name>:443/sdk {Seconds=307.301} 

Apr 28 23:14:30:006 [pool-186059-thread-1, vmware] Polling interval was less than the collection time (300 <= 330) for system DataCollector for <monitor name> 

Apr 28 23:14:30:006 [pool-186059-thread-1, vmware] PERF: DONE: ----- Pass-7876 --- Data Collector Check Interval for: <monitor name> {Seconds=330.662} 

 

If  "Polling interval was less than the collection time" log generated, we recommend to set more polling interval. 

Also, following log is telling, time took for vmware api call for monitoring. 

Check for quick stat updates: https://<monitor name>:443/sdk {Seconds=307.301} 

If this is taking unreasonably, we recommend to check for vCenter.