On Fri, 2006-21-04 at 14:48 -0400, Lewis, Cory (Genworth) wrote: > > > I have a cron job running perl and it is taking a very long time -- > > sometimes over 24 hours. > > > > How can I have cron schedule my job daily, or even hourly, > > and have the perl > > code exit if a previouse instance of the job is still running? > > > > Some have suggested creating a file in /tmp and checking it before > > proceding. But what if a cron job exits prematurely, perhaps > > because of a > > division by zero, and does not delete the /tmp file? > > > > How do I set up a signal handler to gaurentee that the > > /tmp/do_not_run_yet > > file gets deleted when the cron job exits? > > > > > I've not done it with perl, but we have several other scripts that have > this issue. We have the scripts output a file with it's PID. And anytime > the script starts it looks for that file and the PID. If it finds it, it > checks to see if the PID is still active - if it is, it exits, otherwise > it keeps running and updates the file with the new PID. > > Cory
If the PID is active and it matches a process of similar name already running, usually by checking /proc or /var/proc (see `man 5 proc). And then it sends a semaphore (see `man 5 ipc`) or a SIGUSR1 (see `man 7 signal`) to determine if it's not hung. Of course there other way is to stick in a /etc/rc?.d/* file and let init (see `man 8 init`) take care of everything. -- __END__ Just my 0.00000002 million dollars worth, --- Shawn "For the things we have to learn before we can do them, we learn by doing them." Aristotle * Perl tutorials at http://perlmonks.org/?node=Tutorials * A searchable perldoc is at http://perldoc.perl.org/ -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] <http://learn.perl.org/> <http://learn.perl.org/first-response>
