Linux Certif

Toute la documentation sur la certification Linux LPI

Unix::Statgrab.3pm

NAME

Unix::Statgrab - Perl extension for collecting information about the machine

SYNOPSIS

     use Unix::Statgrab;
 
     local $, = "\n";
     
     my $host = get_host_info or 
         die get_error;
         
     print $host->os_name, 
           $host->os_release,
           $host->os_version,
           ...;
 
     my $disks = get_disk_io_stats or
         die get_error;
         
     for (0 .. $disks->num_disks - 1) {
         print $disks->disk_name($_),
               $disks->read_bytes($_),
               ...;
     }
 
 

DESCRIPTION

Unix::Statgrab is a wrapper for libstatgrab as available from http://www.i-scream.org/libstatgrab/ <http://www.i-scream.org/libstatgrab/>. It is a reasonably portable attempt to query interesting stats about your computer. It covers information on the operating system, CPU, memory usage, network interfaces, hard-disks etc.

Each of the provided functions follow a simple rule: It never takes any argument and returns either an object (in case of success) or "undef". In case "undef" was returned, check the return value of "get_error". Also see ``ERROR HANDLING'' further below.

FUNCTIONS

drop_privileges()

Unix::Statgrab can be told to discard setuid and setgid privileges which is usually a good thing. If your program doesn't need the elevated privileges somewhere else, call it right after "use"ing the module.

get_host_info()

Returns generic information about this machine. The object it returns supports the following methods:

*

os_name

*

os_release

*

os_version

*

platform

*

hostname

*

uptime

get_cpu_stats

Returns information about this machine's usage of the CPU. The object it returns supports the following methods, all of which return the number of ticks the processor has spent in the respective states:

*

user

*

kernel

*

idle

*

iowait

*

swap

*

nice

*

total

*

systime

The system time in seconds.

get_cpu_stats_diff

Returns the differences in ticks for each of the states since last time "get_cpu_stats" or "get_cpu_stats_diff" was called. If "cpu_get_stats_diff" is called for the first time (and "get_cpu_stats" wasn't called before) its return values will be the same as "get_cpu_stats".

Its return value supports the same methods as "get_cpu_stats". "systime" then will be the seconds since the last call of this function.

get_cpu_percents

Calls "get_cpu_stats_diff" under the hood but instead of returning ticks, it returns percentages. Its return value provides the same methods as "get_cpu_stats" and "get_cpu_stats_diff".

get_disk_io_stats

Returns the disk IO per disk stored in the kernel which holds the amount of data transferred since bootup. Unlike most other methods presented in this manpage, the methods you can call on its return value take an additional optional parameter which specifies which disk you want information about. If you do not provide this parameter, 0 (= first disk) is assumed.

*

num_disks

The number of disks that were found on this machine.

*

disk_name($disk)

*

read_bytes($disk)

*

write_bytes($disk)

*

systime($disk)

The system time in seconds over which "read_bytes" and "write_bytes" were transferred.

get_disk_io_stats_diff

The same as "get_disk_io_stats" except that it will report the difference to the last call of either "get_disk_io_stats" or "get_disk_io_stats_diff". Provides the same methods as "get_disk_io_stats".

get_fs_stats

Returns statistics about the mounted filesystems, including free space and inode usage. The provided methods again take one optional argument which specifies which partition you want information about. If you do not provide this parameter, 0 (= first mounted filesystem) is assumed:

*

num_fs

The number of mounted filesystems that were found on this machine.

*

device_name($fs)

*

fs_type($fs)

*

mnt_point($fs)

*

size($fs)

Size in bytes.

*

used($fs)

*

avail($fs)

*

total_inodes($fs)

*

used_inodes($fs)

*

free_inodes($fs)

*

avail_inodes($fs)

*

io_size($fs)

The recommended size in bytes when doing I/O operations on this device.

*

block_size($fs)

*

total_blocks($fs)

*

free_blocks($fs)

*

used_blocks($fs)

*

avail_blocks($fs)

get_load_stats()

Returns the load average over various span of times. The following methods are provided:

*

min1

Load average over 1 minute.

*

min5

*

min15

get_mem_stats()

Returns statistics about memory usage. The following methods exist:

*

total

Total memory in bytes.

*

free

*

used

*

cache

Amount of cache used in bytes.

get_swap_stats()

Returns statistics about swap usage. The following methods exist:

*

total

Total swap memory in bytes.

*

used

*

free

get_network_io_stats()

Returns statistics about the network traffic per network interface as stored in the kernel. Again, the provided methods support one optional parameter specifiying which network interface to query. If the parameter is missing, 0 (= first interface) is assumed.

*

num_ifaces

The number of network interfaces found on your machine.

*

interface_name($if)

*

tx($if)

The number of bytes transmitted.

*

rx($if)

The number of bytes received.

*

ipackets($if)

The number of packets received.

*

opackets($if)

The number of bytes transmitted.

*

ierrors($if)

The number of receive errors

*

oerrors($if)

The number of transmit errors

*

collisions($if)

*

systime

The time period over which "tx" and "rx" were transferred.

get_network_io_stats_diff()

The same as "get_network_io_stats" except that it will report on the difference to the last time "get_network_io_stats" or "get_network_io_stats_diff" was called. It supports the same methods as "get_network_io_stats".

get_network_iface_stats()

Returns statistics about each of the found network interfaces in your computer. The provided methods take one optional argument being the interface to query. If this parameter is missing, 0 (= first interface) is assumed.

*

num_ifaces

The number of interfaces found.

*

interface_name($if)

*

speed($if)

The speed of the interface, in megabits/sec

*

dup($if)

One of "SG_IFACE_DUPLEX_FULL", "SG_IFACE_DUPLEX_HALF" and "SG_IFACE_DUPLEX_UNKNOWN". Unknown could mean that duplex hasn't been negotiated yet.

*

up($if)

Whether the interface is up.

get_page_stats()

Returns the number of pages the system has paged in and out since bootup. It supports the following methods:

*

pages_pagein

*

pages_pageout

*

systime

The time period over which pages_pagein and pages_pageout were transferred, in seconds.

get_page_stats_diff()

The same as "get_page_stats" except that it will report the difference to the last time "get_page_stats" or "get_page_stats_diff" was called. Supports the same methods as "get_page_stats".

get_user_stats()

Returns information about the currently logged in users. It supports the following methods:

*

num_entries

The number of currently logged in users.

*

name_list

A list of the users currently logged in.

get_process_stats()

Returns loads of information about the current processes. This function only returns a container. If you want to look at the processes returned, call "all_procs" on its return value.

The processes can also be sorted by various criteria by using the "sort_by" method. This will change the internal order of the container. This method returns the container object so you can do some method chaining:

     my $procs = get_process_stats;
     $procs->sort_by("name");
     print $_->proc_name, "\n" foreach $procs->all_procs;
 
     # syntactically sweeter
 
     print $_->proc_name, "\n" 
         foreach get_process_stats->sort_by("name")->all_procs;
 
 

Available sorting methods are ``name'', ``pid'', ``uid'', ``gid'', ``size'', ``res'', ``cpu'' and ``time''.

You can also sort the list returned by "all_procs". For that you can use one of the eight sorting routines thusly:

     my $p = get_process_stats;
     
     my @by_name = sort sort_procs_by_name $p->all_procs;
     my @by_pid  = sort sort_procs_by_pid  $p->all_procs;
     my @by_uid  = sort sort_procs_by_uid  $p->all_procs;
     # etc.
 
 

Each object returned by "all_procs" supports the following methods:

*

proc_name

*

proc_title

The full command line with which the process was started.

*

pid

*

parent_pid

*

pgid

Process ID of process group leader.

*

uid

*

euid

Effective user ID.

*

gid

*

egid

Effective group ID.

*

proc_size

In bytes.

*

proc_resident

In bytes.

*

time_spent

Time running in seconds.

*

cpu_percent

*

nice

*

state

One of "SG_PROCESS_STATE_RUNNING", "SG_PROCESS_STATE_SLEEPING", "SG_PROCESS_STATE_STOPPED", "SG_PROCESS_STATE_ZOMBIE" and "SG_PROCESS_STATE_UNKNOWN".

ERROR HANDLING

One function "get_error" exists that will return the error encountered during the last operation, if any. Its return value is dual-typed. In string context, it returns a text representation of the error. In numeric context it returns one of the following values:

     SG_ERROR_ASPRINTF
     SG_ERROR_DEVSTAT_GETDEVS
     SG_ERROR_DEVSTAT_SELECTDEVS
     SG_ERROR_ENOENT
     SG_ERROR_GETIFADDRS
     SG_ERROR_GETMNTINFO
     SG_ERROR_GETPAGESIZE
     SG_ERROR_KSTAT_DATA_LOOKUP
     SG_ERROR_KSTAT_LOOKUP
     SG_ERROR_KSTAT_OPEN
     SG_ERROR_KSTAT_READ
     SG_ERROR_KVM_GETSWAPINFO
     SG_ERROR_KVM_OPENFILES
     SG_ERROR_MALLOC
     SG_ERROR_NONE
     SG_ERROR_OPEN
     SG_ERROR_OPENDIR
     SG_ERROR_PARSE
     SG_ERROR_SETEGID
     SG_ERROR_SETEUID
     SG_ERROR_SETMNTENT
     SG_ERROR_SOCKET
     SG_ERROR_SWAPCTL
     SG_ERROR_SYSCONF
     SG_ERROR_SYSCTL
     SG_ERROR_SYSCTLBYNAME
     SG_ERROR_SYSCTLNAMETOMIB
     SG_ERROR_UNAME
     SG_ERROR_UNSUPPORTED
     SG_ERROR_XSW_VER_MISMATCH
 
 

Based on the above, you have finer control over the error handling:

     my $disks = get_disk_io_stats;
     
     if (! $disks) {
         if (get_error == SG_ERROR_PARSE) {
             ...
         } else if (get_error == SG_ERROR_OPEN) {
             ...
         } 
         etc. 
     }
 
 

EXPORT

All by default. This means all of the above functions plus the following constants:

   SG_ERROR_ASPRINTF
   SG_ERROR_DEVSTAT_GETDEVS
   SG_ERROR_DEVSTAT_SELECTDEVS
   SG_ERROR_ENOENT
   SG_ERROR_GETIFADDRS
   SG_ERROR_GETMNTINFO
   SG_ERROR_GETPAGESIZE
   SG_ERROR_KSTAT_DATA_LOOKUP
   SG_ERROR_KSTAT_LOOKUP
   SG_ERROR_KSTAT_OPEN
   SG_ERROR_KSTAT_READ
   SG_ERROR_KVM_GETSWAPINFO
   SG_ERROR_KVM_OPENFILES
   SG_ERROR_MALLOC
   SG_ERROR_NONE
   SG_ERROR_OPEN
   SG_ERROR_OPENDIR
   SG_ERROR_PARSE
   SG_ERROR_SETEGID
   SG_ERROR_SETEUID
   SG_ERROR_SETMNTENT
   SG_ERROR_SOCKET
   SG_ERROR_SWAPCTL
   SG_ERROR_SYSCONF
   SG_ERROR_SYSCTL
   SG_ERROR_SYSCTLBYNAME
   SG_ERROR_SYSCTLNAMETOMIB
   SG_ERROR_UNAME
   SG_ERROR_UNSUPPORTED
   SG_ERROR_XSW_VER_MISMATCH
   SG_IFACE_DUPLEX_FULL
   SG_IFACE_DUPLEX_HALF
   SG_IFACE_DUPLEX_UNKNOWN
   SG_PROCESS_STATE_RUNNING
   SG_PROCESS_STATE_SLEEPING
   SG_PROCESS_STATE_STOPPED
   SG_PROCESS_STATE_UNKNOWN
   SG_PROCESS_STATE_ZOMBIE
 
 

If you don't want that, use the module thusly:

     use Unix::Statgrab ();
 
 

or provide a list of those symbols you want:

     use Unix::Statgrab qw/get_network_iface_stats 
                           SG_IFACE_DUPLEX_FULL
                           SG_IFACE_DUPLEX_HALF
                           SG_IFACE_DUPLEX_UNKNOWN/;
 
 

SEE ALSO

The excellent and very complete manpage of statgrab(3). You can get additional information for each of the above functions by prefixing the function name with ``sg_'' and feed it to "man":

     man sg_get_network_iface_stats
 
 

libstatgrab's home is at http://www.i-scream.org/libstatgrab/ <http://www.i-scream.org/libstatgrab/>

AUTHOR

Tassilo von Parseval, <tassilo.von.parseval@rwth-aachen.de>

COPYRIGHT AND LICENSE

Copyright (C) 2004-2005 by Tassilo von Parseval

This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version.