JBoss crashes with no errors on Seam application initializat

[dolanp]

I am having an issue with JBoss application server and a Seam application being developed. I have developed the application in JBoss Developer Studio and it runs fine inside that JBoss server (which is EAP CP01).

When I deploy the EAR to the server (which is EAP CP02) it deploys without error but when I navigate to the main page it just crashes the server. The server is explicitly killed and there are no error messages or anything printed to the log files or STDOUT. A little background: I initially had this problem with the latest GA version of JBoss, so I switched over to EAP and it worked, then somewhere between builds something happened/changed and now it does not deploy on EAP either. I am suspecting a possible library conflict or something like that but without any errors it's hard to troubleshoot.

Is there any way to put the server in some kind of debug mode or somehow get more information from it on why it could be crashing?

FWIW we have another app that's JSP/Struts and that works fine still.

[peterj]

Operating system? I ask because if the java process is going away, you will need to use operating-system specific tools to debug this.

Also, are you running JBossEAP from a command line or as a service?

[dolanp]

It's RHEL 4. I tried attaching strace to the process but this is all I get:

futex(0x8145b94, FUTEX_WAIT, 1, NULL
 = -1 EINTR (Interrupted system call)
PANIC: attached pid 21048 exited
Process 21048 detached

[dolanp]

I get a few of these messages, maybe this is related...

DEBUG [org.jboss.mx.loading.RepositoryClassLoader] setRepository, repository=org.jboss.mx.loading.HeirarchicalLoaderRepository3@53f89f, cl=org.jboss.mx.loading.HeirarchicalLoaderRepository3$CacheClassLoader@1e81ab3{ url=null ,addedOrder=0}

[peterj]

I am not that familiar with process debugging tools on Linux, maybe someone else can help with that. I would suspect that there is some setting that will cause the OS to take a dump of the process when it crashes, and some tool to read that dump.

One thing you can try is to run the same app on another box. Often, a particular box will have a problem - bad hardware, a broken device driver, or a problematic system library. Any of those can cause a crash. Trying on another box will at least eliminate application issues.

[dolanp]

Oddly enough this turned out to be an issue with X forwarding through SSH. I had gotten this message 'X connection to localhost:10.0 broken (explicit kill or server shutdown)' whenever it crashed but I figured it was incidental to the process being killed. Turns out it is actually the cause. Once I disabled X forwarding on my SSH client the application works as expected...

:)