While running a tomcat application process, tomcat dies with no errors in catalina.out or localhost. So we enabled trace on the process
strace -f -p 23752 > cst.trace.txt 2>&1

and we found the following errors

pid 23857] gettimeofday({1259793502, 901659}, NULL) = 0
[pid 23857] rt_sigprocmask(SIG_BLOCK, NULL, [QUIT], 8) = 0
[pid 23857] gettimeofday({1259793502, 901737}, NULL) = 0
[pid 23857] write(4, "# [ 1800883104 ] Wed Dec 2 17:3"..., 101) = 76
[pid 23857] write(4, "PI> get,s0,q0,a_webc_url\n", 25) = -1 EFBIG (File too large)
...
...
[pid 23838] <... futex resumed> ) = -1 EINTR (Interrupted system call)
[pid 23839] <... futex resumed> ) = -1 EINTR (Interrupted system call)
[pid 23820] <... accept resumed> 0xffffa934, [760209211420]) = ? ERESTARTSYS (To be restarted)
[pid 23855] <... futex resumed> ) = -1 EINTR (Interrupted system call)
[pid 23837] <... accept resumed> 0x6c9b7fb0, [760209211420]) = ? ERESTARTSYS (To be restarted)
...
...
[pid 23824] +++ killed by SIGXFSZ +++
[pid 23843] +++ killed by SIGXFSZ +++
[pid 23845] +++ killed by SIGXFSZ +++
[pid 23825] +++ killed by SIGXFSZ +++
[pid 23840] +++ killed by SIGXFSZ +++

Looks like some file size exceeded limits. But how do we determine which file size is exceeding the limit.
Also what should be the limit threshold. This was a new server built recently

Any suggestions/help is greatly appreciated.
Thanks in advance

Ok....I was able to locate the culprit file by using the command

ls -l /proc/<pid>/fd
The file was a trace enabled file through the application and was becoming huge in seconds. After disabling the trace, looks good. But I have a feeling there is a file size limit somewhere in linux. As I recall, we had enabled tracing on another server and we did not see any issues there, though the trace file had grown as big.