#!/bin/sh
#
# This script takes vxstat output, converts it to gnuplot-friendly
# data, and then generates the graphs.
#
# A standard way to get the vxstat output is to run:
#
#   # vxstat -g <diskgroup> -i 120 >vxstat.<diskgroup> &
#
# Then, once you've collected a bunch of data, run it through
# vxplot:
#
#   (uses default png options in script)
#   # vxplot -t png vxstat.*
#
#   (uses monochrome png settings for older png)
#   # vxplot -t "png monochrome; set size 1.5" vxstat.*
#
# Author: Todd Stansell <todd@stansell.org>
#
# $Id: vxplot,v 1.13 2004-04-13 00:23:47-07 todd Exp $
#

BASE=`basename $0`
PATH=.:$PATH:/usr/local/bin	# most likely gnuplot is in /usr/local/bin
outputdir=.			# dir where output files should go
no=false			# run the command?
DATA=4				# default data to gather (write ops/sec)
AVE=5				# num of datapoints to average
LINES=1440			# graph this number of data points
				  # (48 hours with 120 sec data interval)
ext=gif				# default graph type
complete=true			# recreate datafile?
tmpfile=/tmp/$$.tmpfile		# general use tempfile

USAGE="usage: $BASE [-n] [-q] [-h] [-d num] [-a num] [-o dir] file ...
	-n	don't run commands
	-Q	quick run (don't recreate datafile)
	-h	help
	-t type	change the graph type that gnuplot produces [$ext]
		(use 'png-color' if your gnuplot supports color with png)
	-l num	number of data points to graph [$LINES]
	-a num	number of data points to average (smoothing) [$AVE]
	-d num	column number of vxstat data to graph [$DATA]
	-o dir	directory to put graph [$outputdir]
	file	vxstat data files"

i=`echo -n a`
if [ "X$i" = "Xa" ] ; then
	N='-n'
	X=''
else
	N=''
	X='\c'
fi
echon () {
	echo $N "${@}$X"
}

cleanup() {
	if [ -n "$tmpfile" ] ; then
		rm -f $tmpfile
	fi
}

#
# return true if $1 is a number greater than 0
#
isnum(){
	expr $1 : '^[0-9]*$' >/dev/null ; status=$?
	if [ "$status" -eq 0 ] ; then
		if [ "$1" -gt 0 ] ; then
			return 0
		else
			echo "$1 must be greater than 0" >&2
			return 1
		fi
	else
		return $status
	fi
}

#
# Try to figure out what the interval between datapoints is
# in a given datafile.  We only look at the hh:mm:ss entry,
# which should almost always give us the right value.
#
getinterval() {
	# Sat Aug 11 10:41:33 2001
	cat $1 | \
	nawk 'BEGIN {n=0}
	    /^[A-Z]/ && NF==5 {
		n=n+1
		Time[n]=$4
		if (n == 2) {
		    split(Time[1],a,/:/)
		    split(Time[2],b,/:/)
		    seca=a[3]+(a[2]*60)+(a[1]*60*24)
		    secb=b[3]+(b[2]*60)+(b[1]*60*24)
		    printf "%d\n",secb-seca
		    exit
		}
	    }'
	    
}

#
# Convert vxstat datafile to a gnuplot format datafile
#   usage: vxstat2gnuplot datafile
#
# Variables used:
#   DATA = column of vxstat data to extract
#   AVE  = number of datapoints to average (for smoothing)
#
# Notes: Currently extracts only volume data.
#
vxstat2gnuplot() {
    cat $1 | \
    nawk '
	BEGIN { mon["Jan"]=1 ; mon["Feb"]=2 ; mon["Mar"]=3 ; mon["Apr"]=4 ;
		mon["May"]=5 ; mon["Jun"]=6 ; mon["Jul"]=7 ; mon["Aug"]=8 ;
		mon["Sep"]=9 ; mon["Oct"]=10; mon["Nov"]=11; mon["Dec"]=12;
		test=2; data='$DATA'}
	/NAME/ {
	    test=2
	}			# reset when we see headers
	NF==5 && test>0 {
	    test=test-1
	    next
	}			# skip timestamps
	NF==5 && test<1 {
	    b=$2
	    split($4,a,/:/) ;
	    printf "%04s/%02s/%02s %2s:%2s:%2s",$5,mon[b],$3,a[1],a[2],a[3]
	}			# output date/time in gnuplot-happy format
	/^vol/ && test<1 {
	    printf " %-7s",$data
	}			# get stats for each vol
	/^$/ && test<1 {
	    printf "\n"
	}			# new set of data, so new line
	END {
	    print "END"
	}			# print known ending line
	# nuke last line, which may be incomplete
     ' | grep -v "END" | \

    # Do a floating average thing to smooth out the lines
    # vxstat can produce a *very* jaggy graph without this.
    # (for each data point, average over the past 5 data points...)
    nawk '
	BEGIN   { i=0 ; ave='$AVE' }
	i<ave	{
		  # for each column of data ... 
		      # rotate our $ave saved values that we will average
		      #   null <- val(1) <- val(2) <- val(3) <- val(4)
		      # get the new value
		      #   val(4) <- new value
		      # increment the number of values we have (we need
		      #   a minimum of $ave values)
		  for (x=3; x<=NF ; x++) {
		      for (y=1; y<ave; y++)
			val[x,y]=val[x,y+1]
		      val[x,ave]=$x
		      i++
		  }
		}
	i>=ave	{
		  printf "%s %s",$1,$2
		  for (x=3; x<=NF ; x++) {
		      sum=0
		      for (y=1; y<ave; y++) {
			val[x,y]=val[x,y+1]
			sum=sum+val[x,y]
		      }
		      val[x,ave]=$x
		      sum=sum+val[x,y]
		      printf " %7.2f",sum/ave
		  }
		  printf "\n"
		}' | \
    
    awk 'NR%1=="0" {print}'

}

#
# usage: customplot datafile data [...]
#
customplot() {
	infile=$1
	[ ! -f $infile ] && return 1

	vols=`awk '
		BEGIN	{ a=2 }
		NF==5	{ a=a-1 }
		 a==0	{ exit }
		 a==1 && /^vol/ { print $2 }' $file`
	#
	# set up the gnuplot information to get better looking graphs.
	#
	start="`head -1 $infile | cut -c1-14`00:00"   # start at hour boundary
	end=`tail -1 $infile | cut -c1-19`	      # end wherever data ends
	title="$fname -- `date +\"%m/%d/%y %H:%M:%S\"`"
	case $DATA in
		3)	ylabel='read (operations/sec)' ;;
		4)	ylabel='write (operations/sec)' ;;
		5)	ylabel='read (blocks/sec)' ;;
		6)	ylabel='write (blocks/sec)' ;;
		7)	ylabel='read ave time(ms)' ;;
		8)	ylabel='write ave time(ms)' ;;
	esac
	ntics=16			# show this many x-labels
	nsecs="`expr $LINES \* $INTERVAL / $ntics`"	# guess at # of sec between x-labels
	mxtics="`expr $nsecs / 3600 + 1`"
	nsecs="`expr \( $mxtics - 1 \) \* 3600`"
	case $ext in
		png)	terminal="png monochrome; set size 1.5,1"
			;;
		png-color)
			terminal="png small size 960,480 xffffff x000000 xe2e2e2 \
			xdd0000 x0000dd xff9900 x33cc00 xc0c0c0"
			;;
		gif)	terminal="gif small size 960,480 xffffff x000000 xe2e2e2 \
			xff0000 xbf00bf xa0a0a0 xc0c0c0 xe0e0e0"
			;;
		jpg|jpeg)
			ext=jpg
			terminal="jpeg small size 960,480 xffffff x000000 xe2e2e2 \
			xff0000 xbf00bf xa0a0a0 xc0c0c0 xe0e0e0"
			;;
		*)	terminal="$ext"
			;;
	esac
	fext=`echo $ext | awk '{print $1}'`

	cat <<-'EOT' | sed -e "s|OUTPUT|$outputdir/$fname.$fext|" \
		  -e "s|YLABEL|$ylabel|" \
		  -e "s|START|$start|" \
		  -e "s|END|$end|" \
		  -e "s|NTICS|$nsecs|" \
		  -e "s|MXTICS|$mxtics|" \
		  -e "s|TITLE|$title|" \
		  -e "s|TERMINAL|$terminal|"
	set terminal TERMINAL 
	set key outside Left reverse
	set title "TITLE"
	set output 'OUTPUT'
	set mxtics MXTICS
	set mytics 2
	set tics out
	set grid xtics ytics mxtics mytics 
	set xlabel "Time"
	set timefmt "%Y/%m/%d %H:%M:%S"
	set xdata time
	set ylabel 'YLABEL'
	set ytics nomirror
	set format x "%Y/%m/%d\n%k:%M:%S"
	set xtics rotate NTICS
	set xrange ["START":"END"]
	plot \
	EOT

	#
	# Now add custom data information.
	#
	i=3	# cols 1 and 2 are timestamps, 3 begins data
	t=`echo $vols | wc -w | awk '{printf "%d\n",$1+2}'`	# total cols
	for v in $vols ; do
		echon "'$gp' using 1:$i title '$v' with lines"
		if [ $i -lt $t ] ; then
			echo ', \'
		else
			echo
		fi
		i=`expr $i + 1`
	done
}

run() {
	cmd="$@"
	echo "  # $cmd" >&2
	$no || $cmd
}


#############################################
#         BEGIN MAIN PROGRAM
#############################################

while getopts nQht:l:d:a:d:o: c; do
	case $c in
	l|d|a)    if isnum $OPTARG ; then
			case $c in
				l) LINES=$OPTARG ;;
				d) DATA=$OPTARG ;;
				a) AVE=$OPTARG ;;
			esac
		    else
			echo "$OPTARG is not a valid number" >&2
			exit 64
		    fi
		    ;;
	n)	no=true ;;
	Q)	complete=false ;;
	t)	ext=$OPTARG ;;
	o)	outputdir="$OPTARG" ;;
	h)	echo "$USAGE"; exit ;;
	esac
done
shift `expr $OPTIND - 1`

if [ $# -eq 0 ] ; then
	echo "Please specify a datafile."
	echo "$USAGE"
	exit 1
fi
if [ -z "$outputdir" -o ! -d $outputdir ] ; then
	echo "\"$outputdir\" is an invalid output directory."
	exit 1
else
	outputdir=`cd $outputdir && /bin/pwd`
fi

trap cleanup 1 2 3 5 15

for file in $@ ; do
    if [ ! -f $file ] ; then
	echo "$file does not exist."
	continue
    fi
    fname=`basename $file | sed -e 's/vxstat\.//'`
    gp=$outputdir/$fname.gp
    INTERVAL=`getinterval $file`

    #
    # Extract the data for the disk group
    #
    echo "Working on $file..."
    if $complete || [ ! -s $gp ] ; then
	echo "  Converting datafile and using $LINES data points..."
	run vxstat2gnuplot $file | tail -${LINES}l >$tmpfile

	if [ ! -s $tmpfile ] ; then
	    echo "  no data...quitting."
	    rm $tmpfile
	    continue
	fi
	mv $tmpfile $gp
	echo
    fi
    LINES=`grep -c . $gp`

    echo "  building plot.$fname..."
    run customplot $gp >$tmpfile
    cat $tmpfile | sed 's/^/  /' >&2	# show the plot file
    echo

    echo "  creating graph..."
    gnuplot $tmpfile
    err=$?
    if [ $err -eq 0 ] ; then
	echo "graph created."
    else
	echo "gnuplot exited with error $err."
    fi
    rm -f $tmpfile $gp
done