parse_disk_log.py

   1 #!/bin/python
   2 # Copyright Arvid Norberg 2008. Use, modification and distribution is
   3 # subject to the Boost Software License, Version 1.0. (See accompanying
   4 # file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
   5
   6 import os, sys, time
   7
   8 lines = open(sys.argv[1], 'rb').readlines()
   9
  10 if len(sys.argv) < 2:
  11         print "usage: parse_disk_log.py logfile [seconds]"
  12         sys.exit(1)
  13
  14 keys = ['write', 'read', 'hash', 'move', 'release', 'idle']
  15 throughput_keys = ['write', 'read']
  16
  17 # logfile format:
  18 # <time(ms)> <state>
  19 # example:
  20 # 34523 idle
  21 # 34722 write
  22
  23 if len(sys.argv) > 2:
  24         quantization = long(sys.argv[2]) * 1000
  25 else:
  26         quantization = 5000
  27
  28 out = open('disk_io.dat', 'wb')
  29 out2 = open('disk_throughput.dat', 'wb')
  30 state = 'idle'
  31 time = 0
  32 i = 0
  33 state_timer = {}
  34 throughput = {}
  35 for k in keys: state_timer[k] = 0
  36 for k in throughput_keys: throughput[k] = 0
  37 for l in lines:
  38         l = l[:-1].split(' ')
  39         if len(l) < 2:
  40                 print l
  41                 continue
  42         try:
  43                 new_time = long(l[0])
  44                 while new_time > i + quantization:
  45                         i += quantization
  46                         state_timer[state] += i - time
  47                         time = i
  48                         for k in keys: print >>out, state_timer[k],
  49                         print >>out
  50                         for k in throughput_keys: print >>out2, throughput[k] / 1000.,
  51                         print >>out2
  52                         for k in keys: state_timer[k] = 0
  53                         for k in throughput_keys: throughput[k] = 0
  54                 state_timer[state] += new_time - time
  55                 time = new_time
  56                 state = l[1]
  57                 if state in throughput_keys:
  58                         throughput[state] += long(l[2])
  59         except:
  60                 print l
  61 out.close()
  62 out2.close()
  63
  64 out = open('disk_io.gnuplot', 'wb')
  65 print >>out, "set term png size 1200,700"
  66
  67 print >>out, 'set output "disk_throughput.png"'
  68 print >>out, 'set title "disk throughput per %s second(s)"' % (quantization / 1000)
  69 print >>out, 'set ylabel "throughput (kB)"'
  70 print >>out, 'plot',
  71 i = 0
  72 for k in throughput_keys:
  73         print >>out, ' "disk_throughput.dat" using %d title "%s" with lines,' % (i + 1, throughput_keys[i]),
  74         i = i + 1
  75 print >>out, 'x=0'
  76
  77 print >>out, 'set output "disk_io.png"'
  78 print >>out, 'set ylabel "time (ms)"'
  79 print >>out, 'set xrange [0:*]'
  80 print >>out, 'set title "disk io utilization per %s second(s)"' % (quantization / 1000)
  81 print >>out, "set key box"
  82 print >>out, "set style data histogram"
  83 print >>out, "set style histogram rowstacked"
  84 print >>out, "set style fill solid"
  85 print >>out, 'plot',
  86 i = 0
  87 for k in keys:
  88         if k != 'idle':
  89                 print >>out, ' "disk_io.dat" using %d title "%s",' % (i + 1, keys[i]),
  90         i = i + 1
  91 print >>out, 'x=0'
  92 out.close()
  93
  94 os.system('gnuplot disk_io.gnuplot');
  95