|
Re: Help with Perl
by i5513 (Pilgrim) on Jan 23, 2012 at 23:03 UTC
|
Hi Ashok.g,
Ok, your task can be programmed dynamicaly, but .. why don't you want to use cron (if you are in Unix wolrd)?
Two approaches:
- Regenerate /var/spool/cron/user when you need, cron will reload in the next minute
- Regenerate /etc/cron.d/yourtask when you need it, cron will reload it, and surely will be more readable and maybe easy to implement.
Of course where you write "a.pl 25" will be "*/25 * * * * /path/to/a.pl"
So, if you are in Unix environment, why not cron ?
Thanks
| [reply] |
|
|
I gotta chime in on this one ... agreeing wholeheartedly. Always look for the simplest available solution that leverages existing facilities. If cron can reasonably do the job, without you creating something that will be impossible to understand three months from now, then that is probably the way that you want to do it.
Another notion would be that you could use cron simply to execute some script at the smallest possible interval ... then maintain a list or database (SQLite?) of “what needs to be done” and “the last time it was done.” The script simply executes each command that has come due, then exits until next time. But, I am very reluctant to build complicated solutions to what ought to be simple, routine problems.
| |
|
|
As I said in my post, the scripts and interval are always dynamic..... So I didn't want to always add and delete CRON entries... may be this will be my last option....
| [reply] |
|
|
| [reply] |
Re: Help with Perl
by choroba (Cardinal) on Jan 23, 2012 at 17:36 UTC
|
I do not get it. What separator is used, space or colon? How do you get an interval from a single number? | [reply] |
|
|
Here Space or tab is used and the interval is in seconds so it might be a single digit or any.
Please suggest.
| [reply] |
|
|
Oh, do you mean each script should be called repeatedly after the given interval?
| [reply] |
|
|
|
|
Re: Help with Perl
by choroba (Cardinal) on Jan 24, 2012 at 09:02 UTC
|
I still do net get it. If the file contains the line a.pl 30
b.pl 35
does it mean that in 30 secs, the script a.pl should be run? And after the 30 secs, another run of a.pl should be scheduled based on the file (which might have changed in the meantime, hence "dynamic")? But you do not want to start a new b.pl that is scheduled to happen in 5 seconds - but what if there is a new line in the file saying c.pl 15
Please, be much more specific. | [reply]
[d/l]
[select] |
|
|
What I want is, for example the script invoker.pl will read the file conf.txt:
a.pl 30
b.pl 20
Split each line and validate *.pl and a digit and run a.pl at every 30 minutes and b.pl for every 20 minutes.
If invoker.pl dies then when again I will run the script, the contents of conf.txt may get changed. Hope this will make more understanding.
The challenge here is to run each script for the specified interval continuously.
Can we do that?
Thanks in advance.
Thanks,
Ashok | [reply]
[d/l] |
Re: Help with Perl
by choroba (Cardinal) on Jan 24, 2012 at 18:03 UTC
|
Searching for "timer" on CPAN can give you some solutions, e.g. IO::Async::Timer::Periodic (not tested myself). Or, you can write your own:
#!/usr/bin/perl
use warnings;
use strict;
my $start_time = time;
my %dispatch;
my %interval_of;
while (<>) {
my ($script, $interval) = split;
$interval *= 60;
push @{ $dispatch{ $start_time + $interval } }, $script;
$interval_of{$script} = $interval;
}
while (1) {
foreach my $time (keys %dispatch) {
if (time >= $time) {
for (@{ $dispatch{$time} }) {
system "$_ &";
push @{ $dispatch{ $time + $interval_of{$_} } }, $_;
}
delete $dispatch{$time};
}
}
sleep 1;
}
| [reply]
[d/l] |