On 2013-05-30 16:15, Edward Ned Harvey (openindiana) wrote:
This problem has two parts. Atomicity of signaling operations (acquiring / releasing mutex, etc), and inter-process signaling. (Let the later instance signal the earlier instance that it should die.) It seems easy enough, as long as you have a good atomic operation for locking, the process that acquired lock can write its PID into a file, and later instances can "kill" that PID.
Well, you can use signalling ;) You can use "trap" to set up a signal handler; this code would be executed instantly if a signal is caught. You could invoke some clean_exit procedure (made by you) to clean up working data and exit with a non-error exit code (0), or you could set a variable value instantly (i.e. to break out of a loop when its cycle is completed). Then your other invokation has to find the first one somehow (i.e. "ps -ef|grep ..." or "svcs -p" parsing) and send the signal. Usually you'd use SIGUSR1 and such for custom signals, and handle the typical exit signals as exit ones (0 1 2 3 15). Note that the signal 0 is a placeholder for shell exit when the script ends. Also note that if you trap the exit signals, then you program does not end instantly, but tries to end gracefully by its own definition. You can end a script quickly with an uncatchable SIGKILL (9). For example: END_LOOP=0 trap "END_LOOP=1" 1 2 3 15 while [ "$END_LOOP" = 0 ]; do ... done trap "" 1 2 3 15 In your loop you might also decide to set END_LOOP=nonzero to end it. The last line clears the handler and reverts to default ones. HTH, //Jim Klimov _______________________________________________ OpenIndiana-discuss mailing list [email protected] http://openindiana.org/mailman/listinfo/openindiana-discuss
