This file is indexed.

/usr/share/munin/plugins/diskstat_ is in munin-node 1.4.6-3ubuntu3.

This file is owned by root:root, with mode 0o755.

The actual contents of the file can be viewed below.

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
#!/usr/bin/perl -w
# -*- perl -*-
# vim: sts=4 sw=4 ts=8

=head1 NAME

diskstat_ - Munin wildcard plugin to monitor various values provided
via C</proc/diskstats>

=head1 APPLICABLE SYSTEMS

Linux 2.6 systems with extended block device statistics enabled.


=head1 CONFIGURATION

None needed.

This plugin displays nicer device-mapper device names if it is run as
root, but it functions as needed without root privilege.  To configure
for running as root enter this in a plugin configuration file:

  [diskstat_*]
    user root

=head1 INTERPRETATION

Among the more self-describing or well-known values like C<throughput>
(Bytes per second) there are a few which might need further introduction.


=head2 Device Utilization

Linux provides a counter which increments in a millisecond-interval for as long
as there are outstanding I/O requests. If this counter is close to 1000msec
in a given 1 second timeframe the device is nearly 100% saturated. This plugin
provides values averaged over a 5 minute time frame per default, so it can't
catch short-lived saturations, but it'll give a nice trend for semi-uniform
load patterns as they're expected in most server or multi-user environments.


=head2 Device IO Time

The C<Device IO Time> takes the counter described under C<Device Utilization>
and divides it by the number of I/Os that happened in the given time frame,
resulting in an average time per I/O on the block-device level.

This value can give you a good comparison base amongst different controllers,
storage subsystems and disks for similiar workloads.


=head2 Syscall Wait Time

These values describe the average time it takes between an application issuing
a syscall resulting in a hit to a blockdevice to the syscall returning to the
application.

The values are bound to be higher (at least for read requests) than the time
it takes the device itself to fulfill the requests, since calling overhead,
queuing times and probably a dozen other things are included in those times.

These are the values to watch out for when an user complains that C<the disks
are too slow!>.


=head3 What causes a block device hit?

A non-exhaustive list:

=over

=item * Reads from files when the given range is not in the page cache or the O_DIRECT
flag is set.

=item * Writes to files if O_DIRECT or O_SYNC is set or sys.vm.dirty_(background_)ratio
is exceeded.

=item * Filesystem metadata operations (stat(2), getdents(2), file creation,
modification of any of the values returned by stat(2), etc.)

=item * The pdflush daemon writing out dirtied pages

=item * (f)sync

=item * Swapping

=item * raw device I/O (mkfs, dd, etc.)

=back

=head1 ACKNOWLEDGEMENTS

The core logic of this script is based on the B<iostat> tool of the
B<sysstat> package written and maintained by Sebastien Godard.

=head1 SEE ALSO

See C<Documentation/iostats.txt> in your Linux source tree for further
information about the C<numbers> involved in this module.

L<http://www.westnet.com/~gsmith/content/linux-pdflush.htm> has a nice
writeup about the pdflush daemon.

=head1 VERSION

  $Id: diskstat_.in 3050 2009-11-23 16:31:43Z feiner.tom $

=head1 MAGIC MARKERS

  #%# family=manual
  #%# capabilities=autoconf suggest

=head1 BUGS

Does not work correctly with multiple Munin masters as it calculates
averages between each time it is run.  In such a case it can be run
twice in the same second, this causes "division by zero" errors.  If
it is run two seconds apart the average it reports is over 2 seconds,
not 5 minutes.

=head1 AUTHOR

Michael Renner <michael.renner@amd.co.at>

=head1 LICENSE

GPLv2


=cut

use strict;

use File::Basename;
use Carp;
use POSIX;

# We load our own version of save/restore_state if Munin::Plugin is unavailable.
# Don't try this at home
eval { require Munin::Plugin; Munin::Plugin->import; };

if ($@) {
    die "Munin::Plugin is required for this plugin to run";
}

# Sanity check to ensure that the script is called the correct name.

if ( basename($0) !~ /^diskstat_/ ) {
    die
        qq(Please ensure that the name of the script and it's symlinks starts with "diskstat_"\n);
}

############
# autoconf #
############

if ( defined $ARGV[0] && $ARGV[0] eq 'autoconf' ) {
    my %stats;

    # Capture any croaks on the way
    eval { %stats = parse_diskstats() };

    if ( !$@ && keys %stats ) {

        print "yes\n";
        exit 0;
    }
    else {
        print "no\n";
        exit 1;
    }
}

###########
# suggest #
###########

if ( defined $ARGV[0] && $ARGV[0] eq 'suggest' ) {

    my %diskstats = parse_diskstats();

    my %suggested_devices;

DEVICE:
    for my $devname ( sort keys %diskstats ) {

        # Skip devices without traffic
        next
            if ( $diskstats{$devname}->{'rd_ios'} == 0
            && $diskstats{$devname}->{'wr_ios'} == 0 );

        for my $existing_device ( @{ $suggested_devices{'iops'} } ) {

          # Filter out devices (partitions) which are matched by existing ones
          # e.g. sda1 -> sda, c0d0p1 -> c0d0
            next DEVICE if ( $devname =~ m/$existing_device/ );
        }

        push @{ $suggested_devices{'iops'} },       $devname;
        push @{ $suggested_devices{'throughput'} }, $devname;

        # Only suggest latency graphs if the device supports it
        if (   $diskstats{$devname}->{'rd_ticks'} > 0
            || $diskstats{$devname}->{'wr_ticks'} > 0 )
        {
            push @{ $suggested_devices{'latency'} }, $devname;
        }
    }

    for my $mode ( keys %suggested_devices ) {
        for my $device ( sort @{ $suggested_devices{$mode} } ) {

            my $printdev = translate_device_name( $device, 'TO_FS' );
            print "${mode}_$printdev\n";
        }
    }

    exit 0;
}

# Reading the scripts invocation name and setting some parameters,
# needed from here on

my $basename = basename($0);
my ( $mode, $device ) = $basename =~ m/diskstat_(\w+)_([-+:\w]+)$/;

if ( not defined $device ) {
    croak
        qq(Didn't get a device name. Please put this script into munin's plugin directory and run the commands suggested by "munin-node-configure --shell".\n);
}

$device = translate_device_name( $device, 'FROM_FS' );

##########
# config #
##########

if ( defined $ARGV[0] && $ARGV[0] eq 'config' ) {

    my $pretty_device = $device;

    if ( $device =~ /^dm-\d+$/ ) {
        $pretty_device = translate_devicemapper_name($device);
    }

    if ( $mode eq 'latency' ) {

        print <<EOF;
graph_title Disk latency for /dev/$pretty_device
graph_args --base 1000
graph_category disk

util.label Device utilization (percent)
util.type GAUGE
util.info Utilization of the device. If the time spent for I/O is close to 1000msec for a given second, the device is nearly 100% saturated.
util.min 0
svctm.label Average device IO time (ms)
svctm.type GAUGE
svctm.info Average time an I/O takes on the block device
svctm.min 0
avgwait.label Average IO Wait time (ms)
avgwait.type GAUGE
avgwait.info Average wait time for an I/O from request start to finish (includes queue times et al)
avgwait.min 0
avgrdwait.label Average Read IO Wait time (ms)
avgrdwait.type GAUGE
avgrdwait.info Average wait time for a read I/O from request start to finish (includes queue times et al)
avgrdwait.min 0
avgwrwait.label Average Write IO Wait time (ms)
avgwrwait.type GAUGE
avgwrwait.info Average wait time for a write I/O from request start to finish (includes queue times et al)
avgwrwait.min 0

EOF

    }
    elsif ( $mode eq 'throughput' ) {

        print <<EOF;
graph_title Disk throughput for /dev/$pretty_device
graph_args --base 1024
graph_vlabel Bytes/second
graph_category disk

rdbytes.label Read Bytes
rdbytes.type GAUGE
rdbytes.min 0
wrbytes.label Write Bytes
wrbytes.type GAUGE
wrbytes.min 0

EOF
    }
    elsif ( $mode eq 'iops' ) {

        print <<EOF;
graph_title Disk IOs for /dev/$pretty_device
graph_args --base 1000
graph_vlabel Units/second
graph_category disk

rdio.label Read IO/sec
rdio.type GAUGE
rdio.min 0
wrio.label Write IO/sec
wrio.type GAUGE
wrio.min 0
avgrqsz.label Average Request Size (KiB)
avgrqsz.type GAUGE
avgrqsz.min 0
avgrdrqsz.label Average Read Request Size (KiB)
avgrdrqsz.type GAUGE
avgrdrqsz.min 0
avgwrrqsz.label Average Write Request Size (KiB)
avgwrrqsz.type GAUGE
avgwrrqsz.min 0

EOF

    }
    else {
        croak "Unknown mode $mode\n";
    }
    exit 0;
}

########
# MAIN #
########

my %cur_diskstat = fetch_device_counters($device);

my ( $prev_time, %prev_diskstat ) = restore_state();

save_state( time(), %cur_diskstat );

# Probably the first run for the given device, we need state to do our job,
# so let's wait for the next run.
exit if ( not defined $prev_time or not %prev_diskstat );

calculate_and_print_values( $prev_time, \%prev_diskstat, \%cur_diskstat );

########
# SUBS #
########

sub calculate_and_print_values {
    my ( $prev_time, $prev_stats, $cur_stats ) = @_;

    my $bytes_per_sector = 512;

    my $interval = time() - $prev_time;

    my $read_ios  = $cur_stats->{'rd_ios'} - $prev_stats->{'rd_ios'};
    my $write_ios = $cur_stats->{'wr_ios'} - $prev_stats->{'wr_ios'};

    my $rd_ticks = $cur_stats->{'rd_ticks'} - $prev_stats->{'rd_ticks'};
    my $wr_ticks = $cur_stats->{'wr_ticks'} - $prev_stats->{'wr_ticks'};

    my $rd_sectors = $cur_stats->{'rd_sectors'} - $prev_stats->{'rd_sectors'};
    my $wr_sectors = $cur_stats->{'wr_sectors'} - $prev_stats->{'wr_sectors'};

    my $tot_ticks = $cur_stats->{'tot_ticks'} - $prev_stats->{'tot_ticks'};

    my $read_io_per_sec  = $read_ios / $interval;
    my $write_io_per_sec = $write_ios / $interval;

    my $read_bytes_per_sec  = $rd_sectors / $interval * $bytes_per_sector;
    my $write_bytes_per_sec = $wr_sectors / $interval * $bytes_per_sector;

    my $total_ios         = $read_ios + $write_ios;
    my $total_ios_per_sec = $total_ios / $interval;

    # Utilization - or "how busy is the device"?
    # If the time spent for I/O was close to 1000msec for
    # a given second, the device is nearly 100% saturated.
    my $utilization = $tot_ticks / $interval;

    # Average time an I/O takes on the block device
    my $servicetime
        = $total_ios_per_sec ? $utilization / $total_ios_per_sec : 0;

    # Average wait time for an I/O from start to finish
    # (includes queue times et al)
    my $average_wait
        = $total_ios ? ( $rd_ticks + $wr_ticks ) / $total_ios : 0;
    my $average_rd_wait = $read_ios  ? $rd_ticks / $read_ios  : 0;
    my $average_wr_wait = $write_ios ? $wr_ticks / $write_ios : 0;

    my $average_rq_size_in_kb
        = $total_ios
        ? ( $rd_sectors + $wr_sectors ) 
        * $bytes_per_sector / 1024
        / $total_ios
        : 0;
    my $average_rd_rq_size_in_kb
        = $read_ios ? $rd_sectors * $bytes_per_sector / 1024 / $read_ios : 0;
    my $average_wr_rq_size_in_kb
        = $write_ios
        ? $wr_sectors * $bytes_per_sector / 1024 / $write_ios
        : 0;

    my $util_print = $utilization / 10;

    if ( $mode eq 'latency' ) {
        print <<EOF;

util.value $util_print
svctm.value $servicetime
avgwait.value $average_wait
avgrdwait.value $average_rd_wait
avgwrwait.value $average_wr_wait

EOF
    }
    elsif ( $mode eq 'throughput' ) {

        print <<EOF;

rdbytes.value $read_bytes_per_sec
wrbytes.value $write_bytes_per_sec

EOF
    }
    elsif ( $mode eq 'iops' ) {

        print <<EOF;

rdio.value $read_io_per_sec
wrio.value $write_io_per_sec
avgrqsz.value $average_rq_size_in_kb
avgrdrqsz.value $average_rd_rq_size_in_kb
avgwrrqsz.value $average_wr_rq_size_in_kb

EOF

    }
    else {
        croak "Unknown mode $mode\n";
    }

}

sub read_diskstats {

    open STAT, '< /proc/diskstats'
        or croak "Failed to open '/proc/diskstats': $!\n";

    my @lines;

    for my $line (<STAT>) {

        # Strip trailing newline and leading whitespace
        chomp $line;
        $line =~ s/^\s+//;

        my @elems = split /\s+/, $line;

        # We explicitly don't support old-style diskstats
        # There are situations where only _some_ lines (e.g.
        # partitions on older 2.6 kernels) have fewer stats
        # numbers, therefore we'll skip them silently
        if ( @elems != 14 ) {
            next;
        }
        push @lines, \@elems;
    }

    close STAT or croak "Failed to close '/proc/diskstats': $!";
    return @lines;
}

sub read_sysfs {

    my ($want_device) = @_;

    my @devices;
    my @lines;

    if ( defined $want_device ) {

        # sysfs uses '!' as replacement for '/', e.g. cciss!c0d0
        $want_device =~ tr#/#!#;
        @devices = $want_device;
    }
    else {
        @devices = glob "/sys/block/*/stat";
        @devices = map {m!/sys/block/([^/]+)/stat!} @devices;
    }

    for my $cur_device (@devices) {
        my $stats_file = "/sys/block/$cur_device/stat";

        open STAT, "< $stats_file"
            or croak "Failed to open '$stats_file': $!\n";

        my $line = <STAT>;

        # Trimming whitespace
        $line =~ s/^\s+//;
        chomp $line;

        my @elems = split /\s+/, $line;

        croak "'$stats_file' doesn't contain exactly 11 values. Aborting"
            if ( @elems != 11 );

        # Translate the devicename back before storing the information
        $cur_device =~ tr#!#/#;

        # Faking missing diskstats values
        unshift @elems, ( '', '', $cur_device );

        push @lines, \@elems;

        close STAT or croak "Failed to close '$stats_file': $!\n";
    }

    return @lines;
}

sub parse_diskstats {

    my ($want_device) = @_;

    my @stats;

    if ( glob "/sys/block/*/stat" ) {

        @stats = read_sysfs($want_device);
    }
    else {
        @stats = read_diskstats();
    }

    my %diskstats;

    for my $entry (@stats) {

        my %devstat;

        # Hash-Slicing for fun and profit
        @devstat{
            qw(major minor devname
                rd_ios rd_merges rd_sectors rd_ticks
                wr_ios wr_merges wr_sectors wr_ticks
                ios_in_prog tot_ticks rq_ticks)
            }
            = @{$entry};

        $diskstats{ $devstat{'devname'} } = \%devstat;
    }

    return %diskstats;
}

sub fetch_device_counters {

    my ($want_device) = @_;

    my %diskstats = parse_diskstats($want_device);

    for my $devname ( keys %diskstats ) {

        if ( $want_device eq $devname ) {
            return %{ $diskstats{$devname} };
        }
    }
    return undef;
}

# We use ':' (and formerly '+' and '-') as placeholder for '/' in device-names
# used as calling name for the script.
sub translate_device_name {

    my ( $device, $mode ) = @_;

    if ( $mode eq 'FROM_FS' ) {

        # Hackaround to mitigate issues with unwisely chosen former separator
        if ( not( $device =~ m/dm-\d+/ ) ) {
            $device =~ tr#-+:#///#;
        }

    }
    elsif ( $mode eq 'TO_FS' ) {

        $device =~ tr#/#:#;

    }
    else {
        croak "translate_device_name: Unknown mode\n";
    }

    return $device;
}

sub translate_devicemapper_name {
    my ($device) = @_;

    my ($want_minor) = $device =~ m/^dm-(\d+)$/;

    croak "Failed to extract devicemapper id" unless defined($want_minor);

    my $dm_major = find_devicemapper_major();
    croak "Failed to get device-mapper major number\n"
        unless defined $dm_major;

    for my $entry ( glob "/dev/mapper/\*" ) {

        my $rdev  = ( stat($entry) )[6];
        my $major = floor( $rdev / 256 );
        my $minor = $rdev % 256;

        if ( $major == $dm_major && $minor == $want_minor ) {

            my $pretty_name = translate_lvm_name($entry);

	    $entry =~ s|/dev/||;

            return defined $pretty_name ? $pretty_name : $entry;

        }
    }

    # Return original string if the device can't be found.
    return $device;
}

sub translate_lvm_name {

    my ($entry) = @_;

    my $device_name = basename($entry);

# Check for single-dash-occurence to see if this could be a lvm devicemapper device.
    if ( $device_name =~ m/(?<!-)-(?!-)/ ) {

        # split device name into vg and lv parts
        my ( $vg, $lv ) = split /(?<!-)-(?!-)/, $device_name, 2;
        return undef unless ( defined($vg) && defined($lv) );

        # remove extraneous dashes from vg and lv names
        $vg =~ s/--/-/g;
        $lv =~ s/--/-/g;

        $device_name = "$vg/$lv";

        # Sanity check - does the constructed device name exist?
        # Breaks unless we are root.
        if ( stat("/dev/$device_name") ) {
            return "$device_name";
        }

    }
    return undef;
}

sub find_devicemapper_major {

    open( FH, '< /proc/devices' )
        or croak "Failed to open '/proc/devices': $!";

    my $dm_major;

    for my $line (<FH>) {
        chomp $line;

        my ( $major, $name ) = split /\s+/, $line, 2;

        next unless defined $name;

        if ( $name eq 'device-mapper' ) {
            $dm_major = $major;
            last;
        }
    }
    close(FH);

    return $dm_major;
}