I am testing my application with the latest JDK using the -server switch 
and I am seeing some strange behavior with the Swift side of the app. The 
setup is NT4+SP6, Swift 2.1.0 and JDK 1.3.1. If I start the Swift router 
first (as is the norm), then start my app, the cpu usage of  the app stays 
at 100%. If I start the app before the router the cpu behavior is as 
expected. The difference in my code being. that if the router is present at 
app startup all the normal connections are made, ready for sending 
messages, but if the router is not present a timer is started to check for 
the router every so often. Running the profiler with the -server switch 
shows that the cpu is spending the vast majority of it's time in a socket 
read of a particular thread as shown below:

        java.net.SocketInputStream.socketRead(SocketInputStream.java:Native method)
        java.net.SocketInputStream.read(SocketInputStream.java:86)
        java.io.BufferedInputStream.fill(BufferedInputStream.java:186)
        java.io.BufferedInputStream.read(BufferedInputStream.java:204)
        java.io.DataInputStream.readInt(DataInputStream.java:338)
        com.swiftmq.tools.dump.Dumpalizer.construct(Dumpalizer.java:50)
        com.swiftmq.jms.ConnectionImpl.readObject(ConnectionImpl.java:339)
        com.swiftmq.jms.ConnectionImpl.run(ConnectionImpl.java:424)
        java.lang.Thread.run(Thread.java:484)

Note that this problem does not occur with the -hotspot switch so it is 
probably a Sun problem, but I thought I would ask the list first to see if 
anyone can shed some light on it.

TIA

Robert

PS It makes no difference if I start the router with -server or -hotspot, 
only my application.


------------------------------------------------------
SwiftMQ developers mailing list * http://www.swiftmq.com
To unsubscribe from this list, send an eMail to 
[EMAIL PROTECTED] and write in the body of your message:
UNSUBSCRIBE developers <your-email-address>
Archive: http://www.mail-archive.com/developers@mail.iit.de/




Reply via email to