Use the sleep function.

sleep 300;
[download]

"Perl makes the fun jobs fun
and the boring jobs bearable" - me

Or you could use select():
select(undef, undef, undef, 298.6);    # Be precise That way, your program will sleep for exactly 298.6 seconds. sleep() will actually sleep more or less than you want depending on what architecture you are on.

To most programmers, "pause" means to relinquish the CPU for some amount of time. This is typically done by a sleep() call, which lets your program relinquish the CPU for multiples of 1 second, or by a select() call, which relinquishes the CPU for multiples of a millisecond.

By "relinquish" we mean "not run". The operating system won't schedule your program for execution until at least the specified period of time has elapsed. None of your program's instructions are executed. The CPU is then available for other processes.

Rather than relinquishing the processor, the script above runs in place, in a fairly tight loop, until it sees that the clock has reached the desired time. As you've noted, this isn't processor friendly, and probably isn't what you intended.

sleep 300;
[download]

I've thought about making it a cron job that is called every five minutes, but that would add some complexity, since the process has to know how often it failed when it is checking out some stats every five minutes.

Everyone was quick (and correct) to point out the merits of sleep(), but to answer this part of your question - persist your data out to a file after you run, and persist it back in when you get restarted from cron.

Depending on the details of what you're trying to do, this might be the better soloution.

For anything beyond very simple persistance, may I heartily reccomend getting cozy with:

use Data::Dumper;

which has been part of the standard Perl distribution for some time. Great stuff.

Peace,
-McD

ChOas