Skip to content

Latest commit

 

History

History
580 lines (397 loc) · 20.4 KB

shell_tools_101.rst

File metadata and controls

580 lines (397 loc) · 20.4 KB

Useful shell tools

When you work in a unix environment, you will need to make frequent use of the command line tools available to you in order to complete the tasks you have.

.. todo:: Mention that the philosophy here is to have many tools which perform a few discrete tasks, and run them together to get what you want.

Working with your system

ps

ps shows the processes currently running on the system. ps takes many arguments but some good recipies are ps aux to see all processes from a user standpoint, whether they have a tty or not. Also good is ps -lfU <username> to see all processes owned by <username> in long format. ps has output formatting with the -o switch and works well with grep. pgrep combines ps and grep into one command.

$ ps aux | grep vim
opsschool      5424  1.1  0.2 247360 10448 pts/0    Sl+  17:01   0:00 vim shell_tools_101.rst
opsschool      5499  0.0  0.0  13584   936 pts/5    S+   17:01   0:00 grep --color=auto vim

Note that the grep command also shows up in the proccess table.

top

top shows the top most cpu intensive processes in a table form with system stats summarized at the top. It refreshes every 1 second. By pressing m while running, top will sort not by processor use but by memory use.

df

df looks at all your mounted filesystems and reports on their status. This includes the filesystem type, total size, available space, percent used, and mountpoint. df -h will show the sizes in human readable form. df -h . will show only the filesystem the current working directory is on, very useful when debugging nfs and autofs mounts.

$ df -h .
Filesystem      Size  Used Avail Use% Mounted on
/dev/sda5        43G   31G  9.6G  77% /

du

du estimates the size on disk of a file or files. du -h returns the usage information in human readable format. If the argument to du is a directory, du will run and return a new value for each file in that directory, recursing into subdirectories as well. du -sh can be run on a directory to prevent this behavior and return the sum of the space on disk of all files under that directory.

$ du -sh drawing.svg
24K drawing.svg
4.0K    style.css
20K sitemeter.js

find

find is for finding files on the system. find . recursively walks the entire filesystem tree below the current working directory and prints out each file by its relative path. find . -name 'opsschool' again recursively walks the entire filesystem tree, but only prints out files named opsschool. find . -name 'opschool' -type d prints out only directories and -type f prints out only regular files. The -name filter also accepts wildcards. find . -name '*.html' will print out all files ending in .html. As you become more skilled at Unix/Linux, use the find command often, as it becomes more and more useful with experience.

kill

kill is used to send a signal to a process, which is a way of interrupting it. The most common use of this is to stop, or kill, processes. For example, to stop a process with the PID '123':

$ kill 123

Once the signal is sent the process can decide what to do with it. For instance the above kill 123 sends a TERM signal to the process with PID 123. TERM (also known as SIGTERM) means "Software Termination Signal". Some processes will notice that signal, finish what they are doing, and shut down. Others will just shut down immediately or ignore it entirely.

There are many more types of signals. (Check out man signal for a good list.) For instance, if the above kill command did not succeed in terminating your process, a more heavy-handed option is to send a KILL signal to the process:

$ kill -KILL 123

Every signal has a name and a number. You can reference them by either one. Another way of running kill -KILL is:

$ kill -9 123

Be careful when using the KILL signal as it is the one signal that cannot be caught by the process. It will not have a chance to gracefully shut down. This can lead to temporary files not being removed, open files not being closed, or even corruption of database files.

Signals can be used in a wide variety of ways, not just for terminating processes. One interesting use: if your system is running Unicorn processes, you can send a TTIN signal to the master process and it will spawn an additional worker. Likewise, you can send a TTOU signal and it will remove one of the workers. Another example is Apache HTTPD Web Server which accepts USR1, which causes it to close and re-open log files, which is useful when you need to rotate your log files.

For more on signals see :doc:`unix_signals`.

ls

ls is used to list the contents of a directory. It's most basic usage would be to list the contents of your shell's current working directory:

$ ls
bar  foo

You can also pass a directory name to the ls command and it will list the contents of that directory:

$ ls /usr/local
bin  etc  games  include  lib  libexec  sbin  share  src

There are a number of options that can be passed to the ls command to control both what is output and how it's formatted. Files and directories that begin with a '.' are refered to as hidden files. Two of the more useful options are: -a and -l:

  • ls -a will list these hidden files and directories.
  • ls -l outputs what's called a long listing, where various attributes are given in addition to the filenames.

Emaple of using both:

$ ls -la
total 26636
drwx-----x. 39 luser luser   4096 Jun 28 01:56 .
drwxr-xr-x.  4 root root     4096 Dec 11  2012 ..
drwxrwxr-x   7 luser luser   4096 May 23 00:40 Applications
-rw-------.  1 luser luser  16902 Jun 28 01:33 .bash_history
-rw-r--r--.  1 luser luser     18 May 10  2012 .bash_logout
-rw-r--r--.  1 luser luser    176 May 10  2012 .bash_profile

In a long listing the first field lists the file type, its permissions, and also any special attributes it might have. The very first letter in the first field indicates the file type. Notice directories are indicated by a "d" and regular files are indicated by a "-". After the first field, from left to right the fields are filetypeattributespermissions, links, owner, group, file size, modification date, and file name.

Notice also the files named "." and "..". These are the current directory and the directory up one level respectively.

lsof

lsof lists open files. This command can be very useful in examining what a particular process or user happens to be doing on a system. For each open file information is listed such as the process id that holds the file open, the command that started the process, and the name of the user running the process.

lsof doesn't just list regular files. Of particular use is examing what network activity is currently going on. This can be viewed with by issuing lsof -i.

man

The man command is used to access the man pages. A man page, short for manual page, is documentation on a particular aspect of your operating system, be it a command, a configuration file, or even functions from a library for a programming language. To access a man page, simply type the man command followed by the name of the command, file, etc. that you want to view documentation on. In the old days these manuals were hardcopy and on some systems (e.g. Solaris) you will still see evidence of their page layout. There are man pages for most, if not all, programs on your system. If you install new software from a package manager, usually new man pages will be installed. When man is invoked, it will display the man page to you, when you press 'q', the page will disappear.

The man pages are split up into different sections based on their types. For example if you access the bash man page, at the very top you will see "BASH(1)", indicating that the bash manual is in section 1: general commands. Depending on what you're trying to access, you may have to include a section number when you run man. For example man printf will show you the printf commands man page. But if instead you were wanting to view documentation on the C printf function you would type man 3 printf as section 3 contains documentation on library functions.

The man page sections are as follows:

  • Section 1: General commands
  • Section 2: System calls (functions provided by the kernel)
  • Section 3: Library functions (functions within program libraries)
  • Section 4: Special files (usually found in /dev)
  • Section 5: File formats and conventions (eg /etc/passwd)
  • Section 6: Games and screensavers
  • Section 7: Miscellaneous
  • Section 8: System administration commands

To search through the man pages run either man -k or apropos followed by your search term. This will return a list of man pages who's descriptions match your search term.

The info command is another way to find out information about the system and its utilities. Most system administrators are comfortable with the 'basics' of their core command set, but are frequently checking the man pages to look up odd flags and functionality.

mount

The mount command is used to mount filesystems. For example, mounting an ext4 filesystem that resides on the :file:`/dev/sda2` partiton could be done as follows: mount -t ext4 /dev/sda2 /mnt In this example the "-t" switch tells the mount command that the filesystem type is ext4.

When passed no arguments the mount command lists the filesystems that are currently mounted:

$ mount
/dev/sda2 on / type ext4 (rw)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
tmpfs on /dev/shm type tmpfs (rw)
/dev/sda1 on /boot type ext4 (rw)

The mount command will also consult :file:`/etc/fstab` and if it's able to and use the entries and options it finds there. If an entry for /home exists in /etc/fstab one would be able to simply issue the command mount /home. This command would mount whatever parition is found that is associated with the :file:`/home` entry in /etc/fstab, and use any options that happen to be present.

Items to you wish to mount don't necessarily have to be a partition on a disk to be mounted either. Mounting an ISO file, an image of a optical disk, is especially handy: mount -o loop -t iso9660 /home/luser/installer.iso /mnt/cdrom

mount can also operate on currently mounted filesystems. Of particular use is switching a currently mounted filesystem from read-write to read-only so that a filesystem check can be performed: mount -o remount,ro / This particular command tells mount to remount the currently mounted / filesystem as read-only.

There are quite a number of options that can be passed to the mount command's "-o" switch. Some are filesystem independent, while others depend entirely on the type of filesystem that's being mounted. Further documentation on either can be found in the man pages.

stat

.. todo:: stat command

vmstat

.. todo:: vmstat command

strace

.. todo:: strace command

ulimit

.. todo:: ulimit command


Extracting and manipulating data

A very common pattern in unix is to take some data (a text file, a directory listing, the output from a command) and either extract specific data from it, change some of the data, or both. These tools help you when you do this.

cat

cat outputs the contents of a file either to the shell, another file that already exists, or a file that does not yet exist.

Perhaps most frequently, cat is used to print the contents of a file to the shell. For example, if file :file:`foo.txt` contains the word 'foo':

$ cat /tmp/foo.txt
foo

When cat is called on multiple files, the output is in the same order as the files. If we have another file :file:`bar.txt` that contains 'bar' and run:

$ cat /tmp/foo.txt /home/jdoe/bar.txt
foo bar

If you want to combine the contents of the two files:

$ cat /tmp/foo.txt /home/jdoe/bar.txt > /home/jdoe/foobar.txt
$ cat /home/jdoe/foobar.txt
foo
bar

It is important to note that :file:`foobar.txt` did not exist before running this command. For this particular usage, cat can create a file "on the fly".

cat can also be used to output the contents of one file to another file.

Warning

You should be careful when using cat this way since it will overwrite the contents of the receiving file.

$ cat /tmp/foo.txt > /home/jdoe/bar.txt
$ cat /home/jdoe/bar.txt
foo

There are many tools that you may use to parse the output of fiels, and since cat can provide a comfortable input method for other tools, it is not always necessary. Read more on Useless Use of cat.

cut

The cut utility cuts out selected portions of each line and writes them to the standard output.

As an example, let's take a file students.txt that stores a list of student names, ages and email addresses in columns separated by a tab:

$ cat students.txt
John  Doe     25      [email protected]
Jack  Smith   26      [email protected]
Jane  Doe     24      [email protected]

Here, you can see that the first two columns contain the student's name, the third has an age and the fourth, an email address. You can use cut to extract just the student's first name and email like this:

$ cut -f1,4 students.txt
John [email protected]
Jack [email protected]
Jane [email protected]

The flag, -f is used to select which fields we want to output.

cut, by default, uses tab as a delimiter, but we can change that by using the -d flag.

Suppose the students.txt instead stores data like this:

$ cat students.txt
John Doe|25|[email protected]
Jack Smith|26|[email protected]
Jane Doe|24|[email protected]

Now, if the | character is used as a delimiter, the first column would be the student's full name:

$ cut -f1 -d| students.txt
John Doe
Jack Smith
Jane Doe

If you want to use the space to delimit strings, you would do it like so:

$ cut -f1 -d' ' students.txt
John
Jack
Jane

cut also has some other options. If you have some input with fixed width columns, you can use -c to break them apart. For example, to show the login names and times of the currently logged in users:

$ who | cut -c 1-9,19-30
mike     Aug  1 23:42
mike     Aug  5 20:58
mike     Aug 22 10:34
mike     Aug  6 19:18

You might have to change some of the character positions to make it work on your system.

grep

.. todo:: Talk a LOT about grep

awk

awk is a very powerful utility that lets you extract and manipulate data from files.

For example, if you had a file students.txt similar to the one above, but with the fields separated by a space:

$ cat students.txt
John Doe 25 [email protected]
Jack Smith 26 [email protected]
Jane Doe 24 [email protected]

You can use awk to extract just the student's first name and email like this:

$ awk '{print $1, $4}' students.txt
John [email protected]
Jack [email protected]
Jane [email protected]

By default, awk uses the space character to differentiate between columns. Using this, $1 and $4 told awk to only show the 1st and 4th columns of the file.

The order in which the columns is specified is important, because awk will print them out to the screen in exactly that order. So if you wanted to print the email column before the first name, here's how you would do it:

$ awk '{print $4, $1}' students.txt
[email protected] John
[email protected] Jack
[email protected] Jane

You can also specify a custom delimiter for awk and override the default one (the space character) by using the -F option. Suppose the students.txt instead stored data like this:

$ cat students.txt
John Doe - 25 - [email protected]
Jack Smith - 26 - [email protected]
Jane Doe - 24 - [email protected]

Now, if the - character is used as a delimiter, the first column would be the student's full name:

$ awk -F '-' '{print $1}' students.txt
John Doe
Jack Smith
Jane Doe

Using this same logic, the second column would be the student's age, and the third their email address.

$ awk -F '-' '{print $1, $3}' students.txt
John Doe [email protected]
Jack Smith [email protected]
Jane Doe [email protected]

awk functionality doesn't stop at printing specific columns out to the screen; you can use it to:

  • extract a specific row from the file using the NR command
$ awk 'NR==2' students.txt
Jack Smith - 26 - [email protected]

Note

The -F option was not used here since rows are being manipulated, and the -F option specifies a delimiter for column manipulation

  • extract lines longer than a specific length by using the length($0) command
$ awk 'length($0) > 30' students.txt
John Doe - 25 - [email protected]
Jack Smith - 26 - [email protected]
Jane Doe - 24 - [email protected]

$ awk 'length($0) > 32' students.txt
Jack Smith - 26 - [email protected]
  • find the average of numbers in a column
$ awk -F '-' '{sum+=$2} END {print "Average age = ",sum/NR}' students.txt
Average age =  25

In the last example, with the average age, {sum+=$2} tells awk to take each value in the second column and add it to the existing value of the variable sum. It's important to note here that the variable sum didn't have to be declared or initialised anywhere, awk creates it on-the-fly. The END pattern tells awk what to do after all lines in the file have been processed. In our case, that involves printing out the average age of all students. To get the average age, the sum of all ages (stored in variable sum) was divided by the total number of lines in the file, represented by NR.

In addition to the END pattern, awk also provides a BEGIN pattern, which describes an action that needs to be taken before a the first line of the file is processed.

For example:

$ awk 'BEGIN {print "This is the second line of the file"} NR==2' students.txt
This is the second line of the file
Jack Smith - 26 - [email protected]

sed

.. todo:: Only talk about replacing text for now? It's the most common / needed piece of sed at this level.

sort

sort can be used to sort lines of text.

For example, if you had a file coffee.txt that listed different types of coffee drinks:

$ cat coffee.txt
Mocha
Cappuccino
Espresso
Americano

Running sort would sort these in alphabetical order:

$ sort coffee.txt
Americano
Cappuccino
Espresso
Mocha

You can also reverse the order by passing in -r to the command:

$ sort -r coffee.txt
Mocha
Espresso
Cappuccino
Americano

All very easy so far. But, say we have another file orders.txt that is a list of how many of each drink has been bought in a day:

$ cat orders.txt
100 Mocha
25 Cappuccino
63 Espresso
1003 Americano

What happens when we run sort on this file?

$ sort orders.txt
1003 Americano
100 Mocha
25 Cappuccino
63 Espresso

This isn't what we want at all. Luckily, sort has some more flags, -n is what we want here:

$ sort -n orders.txt
25 Cappuccino
63 Espresso
100 Mocha
1003 Americano

What if we want to sort the new list by name? We will have to sort by the second column, not the first one. Sort assume that columns are space separated by default. sort has the flag -k that let us specify what key we want to use.

$ sort -k2 orders.txt
1003 Americano
25 Cappuccino
63 Espresso
100 Mocha

There are man more flags available, man sort will show you them all. There is probably already something there for whatever you can throw at it.