get system info using perl

[saurabh.hirani]

Hey all,

I am writing a daemon in perl which will check for load average, free disk space on a particular partition and free memory and send it as an updation query to another daemon.

I am supposed to check for these params at least twice in a minute. I was looking at perl modules which would give me this info - but after looking under their hood - most of them fire system command to launch a shell. Some of them do it via /proc which I would rather do myself. My queries are

1. What is faster - reading /proc/loadavg or uptime? I would say /proc/loadavg.

2. Are there any perl modules which are using XS to get these queries done faster without opening a shell?

3. Load average I can get through /proc/loadavg, free memory through /proc/meminfo. But how to obtain free disk space through proc? I am talking about 2.6.x kernels on RHEL4. I want to use /proc as much as I can because I would be reading from a file and not launching a shell.

Are there any other suggestions that you guys think would be useful in writing such daemons which do frequent updates? Thanks for going through my post.

regards,
Saurabh.

[moritz]

I think the speed differences can be neglected. When running time uptime and time cat /proc/uptime the variations are much larger than the difference between the respective execution times.

But why do you want to re-invent the wheel? There are tools like Nagios and collectd (which supports perl plugins) that can be used to obtain and collect such data.

[zentara]

If you want to save the spawning time of a new shell, you can keep a shell open in a pipe, using IPC::Open3, and send your commands to it. Put the command to-check in a while loop

#!/usr/bin/perl 
use warnings;
use strict;
use IPC::Open3;

$|=1;

#my $pid=open3(\*IN,\*OUT,\*ERR,'/bin/bash');
my $pid=open3(\*IN,\*OUT,0,'/bin/bash');
# set \*ERR to 0 to send STDERR to STDOUT

my $cmd = 'date';

#send cmd to bash
print IN "$cmd\n";

my $result = <OUT>;
print $result;

like you I thought about making it more efficient by reading the file driectly. What I did, was find an assembly version of "cat", and did something like

    sub refresh {
        my $pid = shift;
        #asmutils version of cat
    my @size = split "\n", `/home/zentara/perl5lib/cat /proc/$pid/stat
+us`;
    #my @size = split "\n", `cat /proc/$pid/status`;
    
        (my $vmsize) = grep { /VmSize/ } @size;
        my (undef, $size) = split '   ', $vmsize;
        $t->configure(-text => "PID: $pid ->  $size");
        if ($size eq '') { exit }
    }
}

I never really accurately tested which one is fastest, but the assembly cat version made me feel fast. :-)

[saurabh.hirani]

Thanks guys for your replies. I got a reply from comp.lang.perl.misc which pointed out the following 2 modules:
1.Filesys::Df
2. Linux::Sysinfo
They are pretty easy to use, fast, and provide many values. I know that my daemon is not going to bomb the other one once in a second. But I just wanted to know.


These modules really do pump up the speed from okay to burn-the-road. Just to compare system call and Filesys::Df, I wrote a small perl program which calls getting free disk using system call to df and using the module. I profiled them by running them around 1000 times and this was the output of dprofpp:


Forking df and getting the time:
Total Elapsed Time = 42.04899 Seconds
User+System Time = 5.668994 Seconds
Inclusive Times
%Time ExclSec CumulS #Calls sec/call Csec/c Name
96.7 5.485 5.485 1000 0.0055 0.0055 main::get_freequeue


Using df from the module
Total Elapsed Time = 0.618107 Seconds
User+System Time = 0.768107 Seconds
Inclusive Times
%Time ExclSec CumulS #Calls sec/call Csec/c Name
92.1 0.013 0.708 1000 0.0000 0.0007 main::get_freequeue
90.4 0.013 0.695 1000 0.0000 0.0007 Filesys::Df::df
88.7 0.682 0.682 1000 0.0007 0.0007
Filesys::Df::_df

The difference is staggering.

I will also play around with the IPC mod as pointed ou by zentara. Looks interesting

[slacker]

Regarding the third question.
I did some reading on this.
I do not believe there is a way, accessing only /proc, to obtain disk usage or space available.
I'd use system() with df -h to get a human readable output of all drive/partition information.

[zwon]

2. Are there any perl modules which are using XS to get these queries done faster without opening a shell?

There's no need to open shell to read /proc/loadavg. Just read it from perl:

while(1) {
    open my $la, '<', '/proc/loadavg';
    print <$la>;
    close $la;
    sleep 1;
}

Upd: script fixed

[mr_mischief]

Rather than a hard-wired system you write yourself, you might want to get familiar with Net-SNMP. Using NetSNMP you can monitor a the local host OS, remote OSes, routers, switches, firewalls, and much more. You can then wrap around that in Perl with, for example, Net::SNMP.

As for a place to monitor the monitoring, there's always Mon, nagios, or Big Brother. There's also Argus, which is written in Perl itself and has a plugin module system.

I used to monitor DNS, SMTP, POP3, IMAP, RADIUS, HTTP, MySQL, and more servers for their server protocols, memory usage, disk usage, processor load, temperature, and network throughput with this combination. I also used Argus to monitor dozens of routers, switches, firewalls, and ISDN/dialup access concentrators that had their own SNMP implementations that came with them.

Argus can be configured to email you when very particular circumstances happen (like after a router interface shows down once on its 5-minute test, it'll test every 30 seconds until it's back up and email you if it's not up within 7 minutes of showing down). It handles a dependency tree so that it doesn't test devices it can't reach because of a failure in the middle.

Doing the work in Perl is a good thing, because Perl is well-suited to this kind of project. Since there's already a good solution written in Perl, perhaps you should take a look at it before reinventing this particular wheel.