SA

Codes for the system administration class.

p1.sh

Requirement

• Inspect the current directory(“.”) and all sub-directory.
• Calculate the number of directories. (Do not include ‘.’ and ‘..’)
• Calculate the number of files.
• Calculate the sum of all file size.
• List the top 5 biggest files.
• Only consider the regular file. Do not count in the link, FIFO, block device... etc.

p1.awk

command's awk part's code.

cpuplot.sh

Requirement

• Use gnuplot to draw the CPU usage.
• CPU usage is logged in a log file. You only need to read and plot it.
• cpuplot [-h] [-o out_file_name] [-t type] [-c color] -n <60-600>
• -h print the help.
• -o set the output file name. (default: out.png)
• -t set the graph type. (one of ‘filledcurve’, ‘lines’, default: ‘filledcurve’)
• -c set graph color. (in hexadecimal form, default: #1E90FF)
• -n set the number of point should use. (Must be set. Should be in range [60-600])
• Read LOGFILE environment variable. If it is not set, use “/tmp/sysmonitor”
• If type is not one of ‘filledcurve’, ‘lines’, you should print error message and help.

type should be one of 'filledcurve' and 'lines'.

• If color is not in hexadecimal form (a leading sharp ‘#’ and 6 hex digits.), you should print error message and help.

color format error.

• If the number specified by –n is not in [60-600], you should print error message and help.

num should be in range 60 - 600.

• It doesn’t matter which is checked first.
• You can generate temporary files, but you need to clean them when exit. -Image output format

The title should be “CPU Usage”. The y-title should be “CPU Usage (%)”. The x-title should be ”time from now (sec)”. The y-axis range should be 0-100. The x-axis means “time from now”. You can’t predict future, so the value should be negative. The CPU usage is appended to the log file every second. If you get “-n 60” it means you should draw the last 60 lines in the log file.

gnuplot.cmd

• Using gnuplot this tool to draw picture, and this file contains the command test.

sysmonitor.sh

Requirement

• Your script should be in /usr/local/bin/.
• sysmonitor --logfile filename --email yourname@server
• Log the usage per second to the file specify by --logfile
• Do not need to implement fool-proof. You can add additional arguments if you need.
• If the user is not root, you should print some error message and exit.

Part 1: Get and calculate the CPU usage and log it to a file

• Get CPU usage of each process from ps(1).
• (sum of CPU usage exclude [idle] process) / (sum of CPU usage)
• Assume the PID of idle process is 11.
• You can only call ps once per second.
• If the CPU usage is higher than 90% over 5 seconds, you should send a email out to notify you.
• When you send a mail, you are allowed to call ps at the second time.
• Only send one mail if the loading is continuously high.
• Sample mail format

You should exclude the [idle] process. List top 5 processes of CPU usage. You should sort the result by CPU usage. You should display the same columns as below.

Part 2: Write a RC script to manage it

• Your RC script should be in /usr/local/etc/rc.d/.
• Start it when system start.
• Can use service(8) to start or stop it.
• You should log the running pid in "/var/run/sysmonitord.pid"