Good day! I posted this particular conundrum to the list a few days ago and have yet to hear any particular answer to the problem. Therefore, I am going to reiterate it and hopefully someone will have at least experienced this problem before, or can offer as to a solution to the problem.
We are running on a Solaris box, with 4 GB RAM and dual 1.2 GHz processors. We are using Ant 1.5.1 and JDK 1.3. We have one central build.xml that calls subsequent build projects in 14 other directories of our project. Each one of them read from a centralized setup.xml of initial properties. These are set the same for every subsequent project. The typical build process includes: setup & init, clean of old source, recompile of source, creation of jar/war files, EJB jar files created, and the moving of all created entities to a centralized location for development of an EAR file later to deploy to WebSphere Advanced Server 4.3. Now, the problem occurs during one particular combination of building procedures. If we attempt to run through the ENTIRE process above, it will proceed fine until the third subproject attempts to compile its source files, then we will get the following exception: Could not start java: Resource Temporarily Unavailable. If we remove the last step of MOVING/COPYING the created entities (jar/war) to a centralized location, then the process will succeed through every subproject to completion. It appears that the implementation of the move/copy on Unix causes the exec task to hang, and then is unavailable after a specific number of calls on the exec process itself. Has anyone else ever experienced this phenomenon before? If so, do you have any ideas as to a solution? Is there a bug on this particular error? Should one be created? Or is it simply an environment issue that I am missing? Any help would be greatly appreciated. Sincerely, M Damon Hill valtech technologies, inc. www.valtech.com Dallas, TX USA
