summary.in   [plain text]


#! @PATH_PERL@ -w
# $Id$
# Perl version of (summary.sh, loop.awk, peer.awk):
# Create summaries from xntpd's loop and peer statistics.
#
# Copyright (c) 1997, 1999 by Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de>
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
# General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.

require 5.003; # "never tested with any other version of Perl"
use strict;

use Getopt::Long;

my $log_date_pattern = '[12]\d{3}[01]\d[0-3]\d';
my $statsdir = "/var/log/ntp";		# directory with input files
my $outputdir = "/tmp";			# directory for output files
my $skip_time_steps = 3600.0;		# ignore time offsets larger that this
my $startdate = "19700101";		# first data file to use (YYYYMMDD)
my $enddate=`date -u +%Y%m%d`; chomp $enddate; --$enddate;
my $peer_dist_limit = 400.0;

my %options = ("directory|input-directory=s" => \$statsdir,
	       "output-directory=s" => \$outputdir,
	       "skip-time-steps:f" => \$skip_time_steps,
	       "start-date=s" => \$startdate,
	       "end-date=s" => \$enddate,
	       "peer-dist-limit=f" => \$peer_dist_limit);

if ( !GetOptions(%options) )
{
    print STDERR "valid options for $0 are:\n";
    my $opt;
    foreach $opt (sort(keys %options)) {
	print STDERR "\t--$opt\t(default is ";
	if ( ref($options{$opt}) eq "ARRAY" ) {
	    print STDERR join(", ",  map { "'$_'" } @{$options{$opt}});
	} else {
	    print STDERR "'${$options{$opt}}'";
	}
	print STDERR ")\n";
    }
    print STDERR "\n";
    die;
}

# check possibly current values of options
die "$statsdir: no such directory" unless (-d $statsdir);
die "$outputdir: no such directory" unless (-d $outputdir);
die "$skip_time_steps: skip-time-steps must be positive"
    unless ($skip_time_steps >= 0.0);
die "$startdate: invalid start date|$`|$&|$'"
    unless ($startdate =~ m/.*$log_date_pattern$/);
die "$enddate: invalid end date"
    unless ($enddate =~ m/.*$log_date_pattern$/);

$skip_time_steps = 0.128 if ($skip_time_steps == 0);

sub min
{
    my ($result, @rest) = @_;
    map { $result = $_ if ($_ < $result) } @rest;
    return($result);
}

sub max
{
    my ($result, @rest) = @_;
    map { $result = $_ if ($_ > $result) } @rest;
    return($result);
}

# calculate mean, range, and standard deviation for offset and frequency
sub do_loop
{
    my ($directory, $fname, $out_file) = @_;
    print "$directory/$fname\n";
    open INPUT, "$directory/$fname" or warn "can't open $directory/$fname: $!";
    open OUTPUT, ">>$out_file" or die "can't open $out_file: $!";
    print OUTPUT "$fname\n";
    my ($loop_tmax, $loop_fmax) = (-1e9, -1e9);
    my ($loop_tmin, $loop_fmin) = (1e9, 1e9);
    my ($loop_time_rms, $loop_freq_rms) = (0, 0);
    my $loop_count = 0;
    my $loop_time = 0;
    my $loop_freq = 0;
    my ($freq, $offs);
    my @Fld;
    while (<INPUT>) {
	chop;	# strip record separator
	@Fld = split;
	next if ($#Fld < 4);
#NTPv3: 50529 74356.259 -0.000112 16.1230 8
#NTPv3: day, sec.msec, offset, drift_comp, sys_poll
#NTPv4: 51333 54734.582 0.000001648 16.981964 0.000001094 0.020938 6
#NTPv4: day, sec.msec, offset, drift_comp, sys_error, clock_stabil, sys_poll
	if ($Fld[2] > $skip_time_steps || $Fld[2] < -$skip_time_steps) {
	    warn "ignoring loop offset $Fld[2] (file $fname, line $.)\n";
	    next
	}
	$loop_count++;
	($offs, $freq) = ($Fld[2], $Fld[3]);
	$loop_tmax = max($loop_tmax, $offs);
	$loop_tmin = min($loop_tmin, $offs);
	$loop_fmax = max($loop_fmax, $freq);
	$loop_fmin = min($loop_fmin, $freq);
	$loop_time += $offs;
	$loop_time_rms += $offs * $offs;
	$loop_freq += $freq;
	$loop_freq_rms += $freq * $freq;
    }
    close INPUT;
    if ($loop_count > 1) {
	$loop_time /= $loop_count;
	$loop_time_rms = $loop_time_rms / $loop_count - $loop_time * $loop_time;
	if ($loop_time_rms < 0) {
	    warn "loop_time_rms: $loop_time_rms < 0";
	    $loop_time_rms = 0;
	}
	$loop_time_rms = sqrt($loop_time_rms);
	$loop_freq /= $loop_count;
	$loop_freq_rms = $loop_freq_rms / $loop_count - $loop_freq * $loop_freq;
	if ($loop_freq_rms < 0) {
	    warn "loop_freq_rms: $loop_freq_rms < 0";
	    $loop_freq_rms = 0;
	}
	$loop_freq_rms = sqrt($loop_freq_rms);
	printf OUTPUT
	    ("loop %d, %.0f+/-%.1f, rms %.1f, freq %.2f+/-%0.3f, var %.3f\n",
	     $loop_count, ($loop_tmax + $loop_tmin) / 2 * 1e6,
	     ($loop_tmax - $loop_tmin) / 2 * 1e6, $loop_time_rms * 1e6,
	     ($loop_fmax + $loop_fmin) / 2, ($loop_fmax - $loop_fmin) / 2,
	     $loop_freq_rms);
    }
    else {
	warn "no valid lines in $directory/$fname";
    }
    close OUTPUT
}

# calculate mean, standard deviation, maximum offset, mean dispersion,
# and maximum distance for each peer
sub do_peer
{
    my ($directory, $fname, $out_file) = @_;
    print "$directory/$fname\n";
    open INPUT, "$directory/$fname" or warn "can't open $directory/$fname: $!";
    open OUTPUT, ">>$out_file" or die "can't open $out_file: $!";
    print OUTPUT "$fname\n";
# we toss out all distances greater than one second on the assumption the
# peer is in initial acquisition
    my ($n, $MAXDISTANCE) = (0, 1.0);
    my %peer_time;
    my %peer_time_rms;
    my %peer_count;
    my %peer_delay;
    my %peer_disp;
    my %peer_dist;
    my %peer_ident;
    my %peer_tmin;
    my %peer_tmax;
    my @Fld;
    my ($i, $j);
    my ($dist, $offs);
    while (<INPUT>) {
	chop;	# strip record separator
	@Fld = split;
	next if ($#Fld < 6);
#NTPv3: 50529 83316.249 127.127.8.1 9674 0.008628 0.00000 0.00700
#NTPv3: day, sec.msec, addr, status, offset, delay, dispersion
#NTPv4: 51333 56042.037 127.127.8.1 94f5 -0.000014657 0.000000000 0.000000000 0.000013214
#NTPv4: day, sec.msec, addr, status, offset, delay, dispersion, skew

	$dist = $Fld[6] + $Fld[5] / 2;
	next if ($dist > $MAXDISTANCE);
	$offs = $Fld[4];
	if ($offs > $skip_time_steps || $offs < -$skip_time_steps) {
	    warn "ignoring peer offset $offs (file $fname, line $.)\n";
	    next
	}
	$i = $n;
	for ($j = 0; $j < $n; $j++) {
	    if ($Fld[2] eq $peer_ident{$j}) {
		$i = $j;		# peer found
		last;
	    }
	}
	if ($i == $n) {		# add new peer
	    $peer_ident{$i} = $Fld[2];
	    $peer_tmax{$i} = $peer_dist{$i} = -1e9;
	    $peer_tmin{$i} = 1e9;
	    $peer_time{$i} = $peer_time_rms{$i} = 0;
	    $peer_delay{$i} = $peer_disp{$i} = 0;
	    $peer_count{$i} = 0;
	    $n++;
	}
	$peer_count{$i}++;
	$peer_tmax{$i} = max($peer_tmax{$i}, $offs);
	$peer_tmin{$i} = min($peer_tmin{$i}, $offs);
	$peer_dist{$i} = max($peer_dist{$i}, $dist);
	$peer_time{$i} += $offs;
	$peer_time_rms{$i} += $offs * $offs;
	$peer_delay{$i} += $Fld[5];
	$peer_disp{$i} += $Fld[6];
    }
    close INPUT;
    print OUTPUT
"       ident     cnt     mean     rms      max     delay     dist     disp\n";
    print OUTPUT
"==========================================================================\n";
    my @lines = ();
    for ($i = 0; $i < $n; $i++) {
	next if $peer_count{$i} < 2;
	$peer_time{$i} /= $peer_count{$i};
	eval { $peer_time_rms{$i} = sqrt($peer_time_rms{$i} / $peer_count{$i} -
					 $peer_time{$i} * $peer_time{$i}); };
	$peer_time_rms{$i} = 0, warn $@ if $@;
	$peer_delay{$i} /= $peer_count{$i};
	$peer_disp{$i} /= $peer_count{$i};
	$peer_tmax{$i} = $peer_tmax{$i} - $peer_time{$i};
	$peer_tmin{$i} = $peer_time{$i} - $peer_tmin{$i};
	if ($peer_tmin{$i} > $peer_tmax{$i}) {	# can this happen at all?
	    $peer_tmax{$i} = $peer_tmin{$i};
	}
	push @lines, sprintf
	    "%-15s %4d %8.3f %8.3f %8.3f %8.3f %8.3f %8.3f\n",
	    $peer_ident{$i}, $peer_count{$i}, $peer_time{$i} * 1e3,
	    $peer_time_rms{$i} * 1e3, $peer_tmax{$i} * 1e3,
	    $peer_delay{$i} * 1e3, $peer_dist{$i} * 1e3, $peer_disp{$i} * 1e3;
    }
    print OUTPUT sort @lines;
    close OUTPUT;
}

sub do_clock
{
    my ($directory, $fname, $out_file) = @_;
    print "$directory/$fname\n";
    open INPUT, "$directory/$fname";
    open OUTPUT, ">>$out_file" or die "can't open $out_file: $!";
    print OUTPUT "$fname\n";
    close INPUT;
    close OUTPUT;
}

sub peer_summary
{
    my $in_file = shift;
    my ($i, $j, $n);
    my (%peer_ident, %peer_count, %peer_mean, %peer_var, %peer_max);
    my (%peer_1, %peer_2, %peer_3, %peer_4);
    my $dist;
    my $max;
    open INPUT, "<$in_file" or die "can't open $in_file: $!";
    my @Fld;
    $n = 0;
    while (<INPUT>) {
	chop;	# strip record separator
	@Fld = split;
	next if ($#Fld < 7 || $Fld[0] eq 'ident');
	$i = $n;
	for ($j = 0; $j < $n; $j++) {
	    if ($Fld[0] eq $peer_ident{$j}) {
		$i = $j;
		last;			# peer found
	    }
	}
	if ($i == $n) {			# add new peer
	    $peer_count{$i} = $peer_mean{$i} = $peer_var{$i} = 0;
	    $peer_max{$i} = 0;
 	    $peer_1{$i} = $peer_2{$i} = $peer_3{$i} = $peer_4{$i} = 0;
	    $peer_ident{$i} = $Fld[0];
	    ++$n;
	}
	$dist = $Fld[6] - $Fld[5] / 2;
	if ($dist < $peer_dist_limit) {
	    $peer_count{$i}++;
	    $peer_mean{$i} += $Fld[2];
	    $peer_var{$i} += $Fld[3] * $Fld[3];
	    $max = $Fld[4];
	    $peer_max{$i} = max($peer_max{$i}, $max);
	    if ($max > 1) {
		$peer_1{$i}++;
		if ($max > 5) {
		    $peer_2{$i}++;
		    if ($max > 10) {
			$peer_3{$i}++;
			if ($max > 50) {
			    $peer_4{$i}++;
			}
		    }
		}
	    }
	}
	else {
	    warn "dist exceeds limit: $dist (file $in_file, line $.)\n";
	}
    }
    close INPUT;
    my @lines = ();
    print
	"       host     days    mean       rms       max   >1  >5 >10 >50\n";
    print
	"==================================================================\n";
    for ($i = 0; $i < $n; $i++) {
	next if ($peer_count{$i} < 2);
	$peer_mean{$i} /= $peer_count{$i};
	eval { $peer_var{$i} = sqrt($peer_var{$i} / $peer_count{$i} -
				    $peer_mean{$i} * $peer_mean{$i}); };
	$peer_var{$i} = 0, warn $@ if $@;
	push @lines, sprintf
	    "%-15s %3d %9.3f% 9.3f %9.3f %3d %3d %3d %3d\n",
	    $peer_ident{$i}, $peer_count{$i}, $peer_mean{$i}, $peer_var{$i},
	    $peer_max{$i}, $peer_1{$i}, $peer_2{$i}, $peer_3{$i}, $peer_4{$i};
    }
    print sort @lines;
}

my $loop_summary="$outputdir/loop_summary";
my $peer_summary="$outputdir/peer_summary";
my $clock_summary="$outputdir/clock_summary";
my (@loopfiles, @peerfiles, @clockfiles);

print STDERR "Creating summaries from $statsdir ($startdate to $enddate)\n";

opendir SDIR, $statsdir or die "directory ${statsdir}: $!";
rewinddir SDIR;
@loopfiles=sort grep /loop.*$log_date_pattern/, readdir SDIR;
rewinddir SDIR;
@peerfiles=sort grep /peer.*$log_date_pattern/, readdir SDIR;
rewinddir SDIR;
@clockfiles=sort grep /clock.*$log_date_pattern/, readdir SDIR;
closedir SDIR;

# remove old summary files
map { unlink $_ if -f $_ } ($loop_summary, $peer_summary, $clock_summary);

my $date;
map {
    $date = $_; $date =~ s/.*($log_date_pattern)$/$1/;
    if ($date ge $startdate && $date le $enddate) {
	do_loop $statsdir, $_, $loop_summary;
    }
} @loopfiles;

map {
    $date = $_; $date =~ s/.*($log_date_pattern)$/$1/;
    if ($date ge $startdate && $date le $enddate) {
	do_peer $statsdir, $_, $peer_summary;
    }
} @peerfiles;

map {
    $date = $_; $date =~ s/.*($log_date_pattern)$/$1/;
    if ($date ge $startdate && $date le $enddate) {
	do_clock $statsdir, $_, $clock_summary;
    }
} @clockfiles;

print STDERR "Creating peer summary with limit $peer_dist_limit\n";
peer_summary $peer_summary if (-f $peer_summary);