java 多个虚拟机,java虚拟机在执行多个java应用程序时将执行的操作

By reading this article, I know that each java application will run in a specific Java Virtual Machine Instance. So if I execute the following commands("Java -jar test1.jar","Java -jar test2.jar", I will get two processes in the system. And If each command used the default heap size, for example, 256M. The total memory cost is 512M, is that right?

Also I have other questions:

Is the Java virtual Machine a daemon process, start up with the system?

When I execute "java -jar test1.jar", it will create an instance of Java Virtual Machine, then execute the main function. Does it mean every running java application is a sub thread or process of Java Virtual Machine?

Is each running java application individual, other application can not get variable, method, constant, etc, from this running java application?

If one running java application is crashed, will it affect other running java application?

PS: I googled and got lots of different answers, I was totally confused. Anyone who can help me on this kind of questions or even more depth of Java virtual Machine. For example, How it works.

解决方案

The JVM is a standard process, just like any other. As such there's no implicit communication or state sharing between the two. Each will have their own heap, threads etc. If you kill one it won't affect the other.

What will get shared are the code pages of the JVM itself. The kernel is intelligent enough to identify the same binary (any binary -not just the JVM) running twice and reuse the image. This only applies to the actual binary code - not its state. See here for more info re. Linux.

The JVM isn't a daemon process, but could be started upon system startup as a Windows service or Unix/Linux process (via /etc/init.d scripts). This is how you'd (say) run a web service written in Java when a machine is booted up.