Dalin has asked for the wisdom of the Perl Monks concerning the following question:

Hey all, I'm writing a program to convert some data into a spreadsheet and then email the results back to the user. The machine doing the work is different than the one supplying the data. I need to run the program every few seconds to check for new data and the crontab is a little limiting in that it will only go to the minute. What is the best way to run the program persistently? I know that there are some "daemon" modules at there. Is this the right path? Do I just need some simple coding within the program? Any help would be appreciated. Thanks. Dalin 
I've often wondered what it would be like to have something to say.

[download]

Replies are listed 'Best First'.
(jeffa) Re: Persistence
by jeffa (Bishop) on Oct 08, 2001 at 20:23 UTC
    Personally, sounds to me like an event-driven model would be better - why roll your own if you have HTTPD at your disposal? But . . . if you _have_ to write a daemon, here are a couple of resources:

    jeffa

[reply]
Re: Persistence
by George_Sherston (Vicar) on Oct 08, 2001 at 21:01 UTC
    Glad you asked this question, as it's something I'm going to have to grapple with in a few days. I'll watch this thread with interest to see if the answer to my question comes up :)

    At the moment, how I'm planning to approach it (subject to revision by sibling monks) is 
    while (1) {
    open OFF, "off.txt";
    last if <OFF>;
    # all the code goes here
    sleep 5;
}
print "you turned me off!";

    [download]
    off.txt is a blank file. So then my script runs and runs and runs until I put something into off.txt, at which point the script breaks out of the loop.

    As I say, that's how I'd do it... but I'd also ask how to do it before I did it that way. As you have.

    § George Sherston
[reply]
[d/l]
      I prefer the file test operators to opening the file and reading from it to stop a daemon. It is probably faster and it is easyer to understand. 
      until (-e "off.txt") { #change -e to -s to behave like George's code
    #code here
    sleep 5;
}

      [download]
[reply]
[d/l]
Re: Best way to make a script run persistently (daemon)
by entropy (Sexton) on Oct 09, 2001 at 01:12 UTC
    The easiest way to make a script run in the background is just by putting an & after the command (if you are using a standard unix shell). If you want to make a program that backgrounds on its own, you should fork and then use the "setsid" command. 
    use POSIX qw(setsid);
exit if fork();  # parent dies
setsid;  # make this a new process group
... rest of your program ...

    [download]
[reply]
[d/l]
Re: Best way to make a script run persistently (daemon)
by zeitgheist (Novice) on Oct 09, 2001 at 07:11 UTC
    Dalin,
    There was a recent contest on Webreference about UNIX daemons in Perl. The contest is HERE. The tutorial is HERE. The sources were made available somewhere on those URLs.
    HTH

    cheers,
    zeitgheist
[reply]
Re: Best way to make a script run persistently (daemon)
by Starky (Chaplain) on Oct 09, 2001 at 00:32 UTC
    One additional mention that I don't often see in the literature in case you have some backend action going on: I have found that often after a fork I have to reconnect all my database handles. For some reason they frequently don't seem to survive the fork.

    Hope this helps :-)

[reply]
      If you read up on the InactiveDestroy database handle attribute, you'd know that forked processes should not share a database handle, and some databases won't even pass a db handle across a fork. If a child process needs a db handle, then you should connect in the child process.
[reply]

Back to Seekers of Perl Wisdom