Use the Config module to find out whether you have reliable signals:
use Config;
print "Hurrah!\n" if $Config{d_sigaction};

RESULT: Hurrah!

To determine whether your interrupted system calls will automatically restart, look at your system's C signal.h include file: % egrep 'SAV_(RESTART|INTERRUPT)' /usr/include/*/signal.h

RESULT:
/usr/include/asm/signal.h: * SA_INTERRUPT is a no-op, but left due to historical reasons. Use the
/usr/include/asm/signal.h: * SA_RESTART flag to get restarting signals (which were the default long ago)
/usr/include/asm/signal.h:#define SA_RESTART 0x10000000
/usr/include/asm/signal.h:#define SA_INTERRUPT 0x20000000 /* dummy -- ignored */

Basically I do not know what to do. I've tried those examples that you gave me link to, but it still gives me the warning. However, when I used $SIG{INT} = 'IGNORE'; I managed to delay the interruption with 2 seconds.

UPDATE: I found out that if I'm running a single process/thread it works. It seems like the reasoning from crashing comes from the fact that I fork another process in the background. Do I have to put some handlers inside this fork aswell? Or is there any signal for "children" that I can use ?

SIGCHLD is the signal the parent gets from the kernel when the child terminates. Other than that, there are no special "children signals."

A forked child is an exact copy of the parent, until the child starts developing own ideas. Signal handlers should be the same in the child - if they are set up before forking, that is.

I guess at this point I need to see a bit more of the code involved to make an educated guess.