Introduction, UNIX Fundamentals

• HH 310 (Linux/Mac Lab), HH 305 (Mac Lab)
• Remote access via Internet or LAN
  • UNIX servers (Linux)
    • rockhopper.monmouth.edu
    • plato.monmouth.edu
  • UNIX lab machines (Linux)
  • Terminal access via SSH
    • PuTTY telnet/ssh client for Windows
  • Graphical desktop access
    • See MUCSremote for remote Linux desktop access

(Also see http://www.levenez.com/unix/ for a constantly updated UNIX history banner.)

• AT&T's Bell Labs creates UNIX
  • Evolves into “System V” UNIX
• Univ. of California-Berkeley takes AT&T's UNIX and develops BSD (Berkeley Standard Distribution)
• Today, proprietary versions of UNIX and Linux contain some of both UNIX “standards”
• Some differences in utilities
  • lpr (BSD) and lp (SysV) for printing
  • different options for utilities like ps
  • Some UNIX variants include both kinds of utilities

• Multi-user
  • Allows multiple users to access a system at the same time

• Allows processes and peripherals to talk to each other, even on different machines
• pipes and sockets
  • work between processes
  • work between machines on a network
    • standard output from a process on machine A is piped to standard input of a process on machine B
  • Also see http://en.wikipedia.org/wiki/Pipes_and_filters

• Large number of standard command-line utilities
  • for text editing and text processing, compiling, file processing
  • the focus of much of this course because using these utilities together is the core of the UNIX philosophy
    • even in these days of advanced or dumbed-down GUIs and GUI applications

• An “open” system
  • allows programmers to easily access OS features via system calls

• A portable OS
  • Relatively easy to compile and run on a wide variety of platforms because it's mostly written in C
    • A contributing reason for so many proprietary variants of UNIX that were developed: Solaris, Irix, HP-UX, AIX, Tru64, Mac OS X…
    • A contributing reason for free UNIX variants that have been available on so many different platforms: Intel, PowerPC, Alpha, MIPS, ARM (Android), Playstation …

• Everything is (or should be) a file, for example …
  • System processes are listed as pseudo files in /proc
  • System devices are accessed as special files in /dev

• Everything is (or should be) text.
  • The core UNIX utilities were written to handle text streams allowing them to easily pass data between them.

• No drive letters, e.g. C:, D:, etc.

• File systems on hard drive partitions are mounted and appear as directories under the root “/” directory
  • Also goes for remote file systems…
  • … and removable media drives, e.g. floppy drives, CD-ROM, USB pen drives

            $ ls /

            bin/        df.txt   initrd/        media/  root/     sys/       var/
            boot/       etc/     lib/           mnt/    sbin/     tftpboot/
            denyhosts/  export/  linux-images/  opt/    selinux/  tmp/
            dev/        home/    lost+found/    proc/   srv/      usr/



            $ df
            Filesystem           1K-blocks      Used Available Use% Mounted on
            /dev/sda1               522028    404048    117980  78% /
            none                   1557628        48   1557580   1% /dev/shm
            /dev/sda2             10241088   8681048   1560040  85% /usr
            /dev/sda3               819232    639592    179640  79% /var
            /dev/sda5             21823584  20355644   1467940  94% /linux-images
            ...
            ...
            netapp:/home/cslab   286703264 129809216 156894048  46% /export/home

• varies between different variants of UNIX, but usually…
• Files and directories associated with applications and programs are usually under usr
  • bin directories contain executable programs and scripts
  • lib directories contain libraries that programs link with
  • include directories contain program header files that correspond with the libraries in lib
  • man directories contain manual pages used with the man command
  • doc directories contain package documentation
    • /usr/share/doc is the usual location for documentation
  • share directories contain other application data such as fonts, pixmaps
  • var directories contain application generated data such as database files, log files
  • tmp directories contain temporary, transient directories and files
    • /tmp (tmp in the root directory) is usually readable/writeable for every user and process
  • etc directories contain application and system configuration files
    • /etc (etc in the root directory) contains most system and application configuration files.
      • the configuration files, in the UNIX tradition, are almost always plain text files
        • …and many have names that end with .conf

                   $ find /etc -name '*.conf' | less

• Another set of bin, lib, include, man, doc, share, var, etc … directories is under /usr/local

• Nice picture and explanation at https://www.blackmoreops.com/2015/02/14/linux-file-system-hierarchy

• A large enough program is split into the various directories mentioned above
  • whereis will usually tell you how the program is split up

         $ whereis geany
         geany: /usr/bin/geany /usr/include/geany /usr/share/geany /usr/share/man/man1/geany.1.gz

(See also http://en.wikipedia.org/wiki/Unix_philosophy)

(See also http://www.faqs.org/docs/artu/ch01s06.html)

• A program should do one thing and do it well
  • Small programs should be combined to accomplish more complex tasks
  • Use “pipes,” which allow output of one process to be used as input (“piped”) to another process
    • …without the need for intermediate temporary files

• Examples

# upside down cal
cal | tac
  
# number of processes that are associated with me
ps aux | grep jchung | wc -l
  
# ??????
wget -q -O - https://www.gutenberg.org/cache/epub/11231/pg11231.txt | sed 's/ /\n/g' | sed 's/[^A-Za-z0-9]*$//g' | sed 's/^[^A-Za-z0-9]*//g' | tr A-Z a-z | sort | uniq -c | sort -rn

• The UNIX approach to problem solving
  • If you can solve the problem by combining multiple existing utilities using pipes, do it; otherwise
  • …ask people on the network if they know how to solve it. If they do, great; otherwise
  • …if you could solve the problem with the aid of some other hand-written utilities, write the utilities yourself and add them to the UNIX repertoire.
    • Each utility you write should do one thing and do it well so that it can be reused to solve other problems.
  • If more utilities won't do the trick, then write your own program to solve the problem.

• The GNU project's list of core Unix utilities
• IEEE Single UNIX Specification list of standard Unix utilities
• moreutils

• The “command line”

• Relationship between the shell and “Terminals”
  • The Terminal is not the shell.
    • The shell runs in the Terminal to interpret and execute commands and display output.
    • The shell can run invisibly outside of any Terminal.

• Environment variables are everywhere in UNIX.
  • Used by running processes and in shells
  • Run env | less # (or set | less)
    • Look for value of the SHELL environment variable
      • env | grep SHELL
      • echo $SHELL
        • In the shell, all variables are accessed by prepending the $.

• system-wide shell config = /etc/profile
• user-specific shell config
  • ~/.profile (for ksh and bash)
  • ~/.bashrc, ~/.bash_profile (for bash)
    • ~ is shorthand for $HOME
      • env | grep HOME or echo $HOME
        • … so ~/.bash_profile for me is the same as /export/home/hawkdom2/jchung/.bash_profile
        • ls ~ or ls $HOME are equivalent

• When you start an interactive shell, your shell reads its shell config file.
• Shell configs set env vars, define functions, aliases, etc. and run programs.
• The Shell prompt
  • PS1 shell variable

• scrolling
  • Use up and down arrow keys to scroll the command history.

• searching
  • history command lists previously issued commands
  • history | grep cat searches for cat in the command history

• shell history files
  • (Bash) Command history written to ~/.bash_history when user exits
    • Number of commands kept in command history and command history file set w/ HISTSIZE and HISTFILESIZE env vars, respectively
      • Default value of HISTFILESIZE is 500
      • HISTSIZE and HISTFILESIZE set in shell configuration file
      • (Do in class) Change HISTSIZE and HISTFILESIZE to 5000 to increase command line history.

• Reissuing previous commands
  • previous commands can be issued again with “!” directive

# Display your command history
history
   
...
   
548  cat .bash_history
549  man bash
550  nano .bashrc
551  source .bashrc
   
!549       # issues 'man bash' command again
!man       # reissues last command starting w/ 'man'
!nan       # reissues the 'nano .bashrc' command

• Type first few letters of a command or file name and hit Tab for completion.

• typically set in shell profile so they are available for every interactive shell session
  • type alias to list aliases
    • setting an alias
      • alias commandalias='command(s) plus options'

• (Do in class) Add some useful aliases to bash shell profile (~/.bashrc).

• Before proceeding, see the collection setup notes.
  • Create an examples/cat directory under your course directory (cs370 or se370).

• Searching for apropriate man pages
  • man -k (or the equivalent apropos command)

man -k vim   # Search for all man pages containing keyword 'vim'
            
eview (1)            - easy Vim, edit a file with Vim and setup for modeless editing
evim (1)             - easy Vim, edit a file with Vim and setup for modeless editing
gvim (1)             - Vi IMproved, a programmers text editor
gvimdiff (1)         - edit two, three or four versions of a file with Vim and show differences
gvimtutor (1)        - the Vim tutor
rgvim (1)            - Vi IMproved, a programmers text editor
rvim (1)             - Vi IMproved, a programmers text editor
vim (1)              - Vi IMproved, a programmers text editor
vimdiff (1)          - edit two, three or four versions of a file with Vim and show differences
vimtutor (1)         - the Vim tutor

• Navigating in man pages
  • 1G - move to top of man page, G - move to bottom of man page
  • u, b or <Page Up> to page up
  • d, space or <Page Down> to page down
  • q to quit

• Searching in man pages
  • Type '/' then the word or phrase you are searching for.
    • 'n' and 'p' to search for next and previous instances of search string, respectively

• cat (short for catenate) takes input from standard input or from a list of files and sends them to standard output
  • Access contents of a file w/ cat

cat filename

• Creating a simple text file w/ cat

cat > fileFromCat.txt # The '>' means that keyboard input is 'redirected' to fileFromCat.txt
The first line
second line
third line
EOF          # EOF is usually entered using Control-D
   
cat fileFromCat.txt
   
The first line
second line
third line

• Concatenating a series of files w/ cat

# output contents of four files to screen
$ cat file1 file2 file3 file4

# output to a file, overwriting contents of allFilesTogether   
$ cat file1 file2 file3 file4 > allFilesTogether

# output to a file, appending to contents of allFilesTogether
$ cat file1 file2 file3 file4 >> allFilesTogether

# using a shell wildcard (*),
# cat the contents of all files whose names begin w/ 'file'; files are
# concatenated in sorted order, i.e. file1 file2 file3 file4
$ cat file*

• Shell facilities for job control
  • suspend, bring job to foreground, put job in background
  • suspend with Control-Z
    • Start the nano editor and then Control-Z
    • Control-Z suspends nano
      • The nano process is suspended
    • Type the fg command (fg = foreground).
      • nano is brought to the foreground
    • Control-Z again
    • Type the bg command (bg = background).
      • The nano process resumes running in the background.
    • Start man nano and Control-Z to suspend it.
    • Type the jobs command to list the two jobs currently running.
    • Resume nano with fg %1

• ps command and jobs
  • The ps command displays process IDs (PIDs)
    • The PID and jobs number are not the same
  • ps aux or ps fax show all processes (Linux)
    • ps options depend on SysV or BSD implementation of ps
      • On SysV-type systems, ps -ef displays all processes
      • On BSD-type systems, ps -aux displays all processes

• The top command shows top processes on system in “real time”

• The ./.. shortcuts
  • . means current directory
  • .. means parent directory
• mkdir -p dirname
  • “-p” option of mkdir creates entire subdirectory tree
  • Example: mkdir -p cs370/examples/cat
• rmdir vs. rm -r dirname
  • rmdir is rarely used because it cannot remove non-empty directories
  • The more commonly used rm -r dirname must be used with care.
  • Example: rm -rf a_directory_name
    • Again, use rm -rf with great care.
• simple shell file name globbing
  • * - wildcard
    • ls abc*: list files and directories that begin with “abc”
    • ls *abc*: list files and directories that contain “abc”
  • [ ] - subset
    • ls [a-z]*: list files and directories that begin with a lower case letter
• file managers
  • mc: a 2-pane console file manager for power users
  • various GUI file managers

• ls -l to view permissions
• chmod to change permissions of files or directories
  • directories must be executable to be accessible
• chown to change ownership
  • Only useful to root (system admin)
• groups to see what groups you are in
• chgrp to change group ownership of files or directories
• setfacl, getfacl to use access control lists
  • much more flexibility than standard chmod/chown/chgrp commands
• Certain GUI file managers provide frontends to chmod/chown/chgrp/setfacl/getfacl

• lpr and lp to print from commandline
• lpq and lpstat to view queued print jobs
• lprm to remove your own print jobs from print queues
• enscript to convert text to Postscript for prettier output

• If using a GUI desktop environment in UNIX, copying and pasting of text can be done between Terminals and other applications using common clipboard-like functionality.
  • Right-clicking in a GUI Terminal may bring up a context menu which includes “copy” and “paste” options.
  • Keyboard shortcuts to copy and paste in Terminals are usually Control-Shift-C and Control-Shift-V, respectively.
• UNIX desktops have an older copy-and-paste functionality that is commonly used with the mouse.
  • Select text by swiping with the left mouse button …
    • or double click a word to select it …
    • or triple click a line to select the entire line.
    • The selected text is automatically copied to a different clipboard area in memory.
  • Paste text by placing the mouse cursor at the insertion point and clicking the middle mouse button.

• Ssh is the de facto method and protocol for logging into a console (terminal) session on a remote UNIX system.
• The most common way to use ssh is to run

ssh some_host_name  # Example: ssh plato, ssh plato.monmouth.edu

• Your user account must exist on the remote host or you must specify another existing account when ssh'ing, e.g.,

ssh s1100841@plato  # or ssh s1100841@plato.monmouth.edu (if off-campus)

• Set up no-password, key authentication for ssh
  • This method uses public key encryption instead of passwords
  • The set up of ssh public key authentication requires the following:
    • Generate a public/private key pair if you don't already have one.
      • command: ssh-keygen
      • local files: ~/.ssh/id_rsa, ~/.ssh/id_rsa.pub
    • Copy the public key contents to authorized_keys on the remote UNIX system that you want to ssh to.
      • command: ssh-copy-id
      • remote file: ~/.ssh/authorized_keys

• For maximum convenience, most users generate ssh keys without a passphrase.
  • But the best practice for ssh key authentication is to generate the key pair using a strong passphrase, and then use ssh-agent to type in the passphrase for us when ssh'ing to remote systems.

• Remote file transfer programs that also use the ssh protocol are scp (secure copy), sftp (secure ftp) and rsync.