Performance Statistics Collection for UNIX flavored OS
June 24, 2011 1 Comment
Introduction:
This post explains how to setup performance counter collection in Unix, Solaris, Linux or HP-UX environment and how to extract the log and present it in a Graphical format with simple ruby program which i have written.
Please download ruby from http://www.ruby-lang.org
The 3 major steps to follow to collect and report performance statistics are as follows:
1.Configure Performance counters
2.Conversion of Performance logs to csv
3.Reporting
1.Configure Performance counters:
90% of the time people are concerned with the memory and cpu statistics, so let’s see how to configure , collect and report memory and CPU statisctis in unix, solaris, linux or HP-UX environment.
Write .sh File:
1.memory statistics
Most of the time people are concerned about the total memory usage rather than breakdown of cache, swap, free, buffer memory etc. hence we can measure the overall memory usage by “free” command available in UNIX flavored OSes.Write a .sh file which will log the free and used memory as follows,
#!/bin/sh
free -m -s1 >> free_mem.log
The above command means the total free and used memory will be collected for every second and stored in the free_mem log file.
2.CPU and other statistics
sometimes people are concerned about knowing everything like cache, swap, free, buffer memory apart from CPU. we can measure all these by “vmstat” command available in UNIX flavored OSes.Write a .sh file which will log these information as follows,
#!/bin/sh
vmstat 1 >> vmstat.log
The above command means the statistics such as
Procs r: The number of processes waiting for run time. b: The number of processes in uninterruptible sleep.
Memory swpd: the amount of virtual memory used. free: the amount of idle memory. buff: the amount of memory used as buffers. cache: the amount of memory used as cache. inact: the amount of inactive memory. (-a option) active: the amount of active memory. (-a option)
Swap si: Amount of memory swapped in from disk (/s). so: Amount of memory swapped to disk (/s).
IO bi: Blocks received from a block device (blocks/s). bo: Blocks sent to a block device (blocks/s).
System in: The number of interrupts per second, including the clock. cs: The number of context switches per second.
CPU These are percentages of total CPU time. us: Time spent running non-kernel code. (user time, including nice time) sy: Time spent running kernel code. (system time) id: Time spent idle. wa: Time spent waiting for IO.
will be collected for every second and stored in the vmstat.log file.
Run the .sh Files during the tests and stop it once the test is done.collect the log files once the test is done for conversion.
2.Conversion of Performance logs to csv:
Convert memory log to CSV:
Execute the below ruby program to convert the log file to the csv file
infile = ARGV[0] || ‘C:\free_mem.log’
freq = ARGV[1] || 1
outfile = ‘C:\Results’
stat = File.stat(infile)
now = stat.mtime
t=Time.now
lines=[]
File.readlines(infile).each { |line|
line.chomp
if line.match(/^-/) then
data = line.split(/[^-\/\+\w+:$][\s\t]+/)
temp = ”
for i in 0 .. (data.length – 1)
temp = temp + data[i].to_s + ‘,’
end
temp[temp.length-1] = ”
lines << temp
end
}
lines.each do |row|
File.open(outfile+t.strftime(“%m%d%Y_%H%M%S”)+”.csv”,”w”) do |the_file|
the_file.puts “total,used,free”
the_file.puts lines
end
end
Convert vmstat log to CSV:
Execute the below ruby program to convert the log file to the csv file
infile = ARGV[0] || ‘C:\vmstat.log’
freq = ARGV[1] || 1
outfile = ‘C:\Results’
stat = File.stat(infile)
now = stat.mtime
t=Time.now
lines=[]
File.readlines(infile).each { |line|
line.chomp
if line.match(/^\s+\d+/m) then
now = now + freq
date = “#{now.day}/#{now.month}/#{now.year}”
time = “#{now.hour}:#{now.min}:#{now.sec}”
data = line.gsub(/^\s+/m,”").gsub(/[\s\t]+/m,”,”).gsub(/,$/m,”")
lines<<”#{date},#{time},#{data}”
end
}
lines.each do |row|
File.open(outfile+t.strftime(“%m%d%Y_%H%M%S”)+”.csv”,”w”) do |the_file|
the_file.puts “date,time,r,b,swpd,free,buff,cache,si,so,bi,bo,in,cs,us,sy,id,wa,st”
the_file.puts lines
end
end
3.Reporting:
Open the csv file in which the necessary performance counters data has been collected and generate graph using excel Chart Wizard feature as below CPU sample.
Thanks for Reading the blog. Hope it was useful for you. Please don’t hesitate to post comments if you got any clarification/feedback. Thanks a lot.
Recent Comments