RRDp - Attach rrdtool from within a perl script via a set of pipes;
use RRDp
RRDp::start path to rrdtool executable
RRDp::cmd rrdtool commandline
$answer
= RRD::read
$status
= RRD::end
$RRDp::user, $RRDp::sys, $RRDp::real
With this module you can safely communicate with the rrdtool.
After every RRDp::cmd you have to issue an RRDp::read command to get rrdtools answer to your command. The answer is returned as a pointer, in order to speed things up. If the last command did not return any data, RRDp::read will return an undefined variable.
If you import the PERFORMANCE variables into your namespace, you can access rrdtools internal performance measurements.
Load the RRDp::pipe module.
start rrdtool. The argument must be the path to the rrdtool executable
pass commands on to rrdtool. check the rrdtool documentation for more info on the rrdtool commands.
read rrdtools response to your command. Note that the $answer
variable will only contain a pointer to the returned data. The reason for
this is, that rrdtool can potentially return quite excessive amounts of
data and we don't want to copy this around in memory. So when you want to
access the contents of $answer
you have to use
$$answer
which dereferences the variable.
terminates rrdtool and returns rrdtools status ...
these variables will contain totals of the user time, system time and real time as seen by rrdtool. User time is the time rrdtool is running, System time is the time spend in system calls and real time is the total time rrdtool has been running.
The difference between user + system and real is the time spent waiting for things like the hard disk and new input from the perl script.
use RRDp; RRDp::start "/usr/local/bin/rrdtool"; RRDp::cmd qw(create demo.rrd --step 100 DS:in:GAUGE:100:U:U RRA:AVERAGE:0.5:1:10); $answer = RRDp::read; print $$answer; ($usertime,$systemtime,$realtime) = ($RRDp::user,$RRDp::sys,$RRDp::real);
For more information on how to use rrdtool, check the manpages.
Tobias Oetiker <oetiker@ee.ethz.ch>