[SNMP4J] high CPU usage with SNMP4J 1.11 and DefaultTcpTransportMapping

[Brice Fines]

Hi Frank,

I updated my environment, I am now running our app in WAS 6.1.0.29 (fix 
pack 29, the last one available).
I also updated the associated JVM (Service Release 11, the one for fix 
pack 29):

java version "1.5.0"
Java(TM) 2 Runtime Environment, Standard Edition (build pwi32dev-20091106a 
(SR11
 ))
IBM J9 VM (build 2.3, J2RE 1.5.0 IBM J9 2.3 Windows XP x86-32 
j9vmwi3223-2009110
4 (JIT enabled)
J9VM - 20091103_45935_lHdSMr
JIT  - 20091016_1845_r8
GC   - 20091026_AA)
JCL  - 20091106

Unfortunately, it does not solve the high CPU usage when sending SNMP 
commands.
When opening the TCP connection to send and receive the command, CPU goes 
up between 55%-100%; with 
org.snmp4j.transport.DefaultTcpTransportMapping$SocketTimeout, it falls 
again down to 2%)
I hope the next IBM update will help more.

Is there a simple way to cause/check a high CPU usage boost related to 
java NIO select?
I am trying the Ping example (which use a NIO selector) from 
http://java.sun.com/j2se/1.4.2/docs/guide/nio/example/index.html , but the 
cpu usage is 0% while running it.
But if I execute the following from a command window, I get a CPU usage 
peak while running:
C:\>C:\RAD7\SDP70\runtimes\base_v61\java\bin\java -cp snmp4j-1.11.jar 
org.snmp4j
.tools.console.SnmpRequest  -c private -v 2c -p SET udp:localhost/161 
"1.3.6.1.2
.1.1.5.0={s}SNMP4J"
Received response after 2016 millis
Request timed out.

Regards