Reset Search
 

 

Article

How to find the cause of high CPU utilization in Extreme Management on a linux server

« Go Back

Information

 
TitleHow to find the cause of high CPU utilization in Extreme Management on a linux server
Objective
How to find the cause of high CPU utilization in Extreme Management
Environment
  • Extreme Management Center
  • NetSight
  • Console
  • OneView
Procedure
The ideal way to pinpoint the cause of any CPU utilization is to narrow down the cause to a specific thread.
 
To get the PID of the netsight process
     
      cat /var/run/nsserver.pid
     
To get a java thread dump of the netsight process into a file jstack.out
     
      cd /usr/local/Enterasys_Networks/NetSight/java/bin
      ./jstack <PID> > jstack.out
     
To get information about CPU use in threads, but don’t show idle threads, and sort the output from highest CPU to lowest.
     
      ps -L -o pid,tid,%cpu <PID> | grep -v "0\.0" | sort -g -r -k 3,3
     
The second column of this ps command is the thread ID, in decimal. This corresponds to the “nid” (native id) value shown in the java thread dump, which is in hexadecimal. 
     
 For example:
     
 $ cat /var/run/nsserver.pid
 5552
     
 $ cd /usr/local/Enterasys_Networks/NetSight/java/bin
     
$ ./jstack <PID> > jstack.out
     
$ ps -L -o pid,tid,%cpu 5552 | grep -v "0\.0" | sort -g -r -k 3,3
5552  5717  9.8
5552  5718  1.3
5552  5570  1.2
     
$ printf "%x\n" 5717
1655

$grep 1655 jstack.out
 "Flow Engine - Active Flow Aging Timer" prio=10 tid=0x00007fba91ead800 nid=0x1655 in Object.wait() [0x00007fba49733000]
Additional notes
running a jstack against a NAC appliance will look a little different
The process is different, and the location is different
more /tmp/tag.pid
cd /usr/lib/jvm

then go to the latest jdkl directory 

./jstack <PID> > jstack.out
 

Feedback

 

Was this article helpful?


   

Feedback

Please tell us how we can make this article more useful.

Characters Remaining: 255