Hi,

We are trying to use JNA to get some of the platform-specific functions done in our java-based project.

In one such situation, one of my threads forks and execs a script (using JNA), say forked_executable to be run on the local machine. The behaviour we are observing is as follows:

Machine : i686 OS : Linux (Ubuntu)

Summary: Sometimes we see the forked_executable running successfully and everything is fine. Occassionally (3times out of 5) we see two instances of our original program running, which leads us to conclude that execv failed or wasn't executed for some reason.

-- Relevant Code --

//linux is the libc.INSTANCE

if((childPid = linux.fork()) == 0){ try{ System.out.println(Thread.currentThread ().getName() + " - execing right now"); Thread.currentThread().setName("New_Execed Thread"); System.out.println(Thread.currentThread ().getName() + " - just before exec"); System.out.println("return value" + linux.execv ("/bin/sleep", testArray)); }catch(Exception e){ System.out.println("exception in abs"); e.printStackTrace(); } } else{ System.out.println(Thread.currentThread().getName() + " - I am the parent"); }

-- Relevant Code --

Have you guys encountered anything related to this? Is there a possibility of a race condition somewhere? Any help is appreciated!

Thanks, -Jags