Rha030 Workbook02 Student 6.1 2

Part Workbook 2.
Filesystem Basics
Table of Contents
1. Filesystem Navigation .................................................................................................... 5 Discussion .............................................................................................................. 5 The Filesystem Inverted Tree ............................................................................. 5 Exploring the Directory Tree using Nautilus .......................................................... 6 The Current Working Directory (cwd) .................................................................. 9 Where Am I? The pwd command ........................................................................ 9 Moving Around - The cd command ................................................................... 10 Absolute and Relative References ...................................................................... 11 Examples .............................................................................................................. 12 Example 1. Discovering your home directory ........................................................ 12 Example 2. Climbing up and down the tree. ......................................................... 12 Online Exercises .................................................................................................... 13 Online Exercise 1. Exploring the filesystem .......................................................... 13 Specification .......................................................................................... 13 Deliverables ........................................................................................... 13 Possible Solution .................................................................................... 13 Questions .............................................................................................................. 13 Navigating the Filesystem ................................................................................ 13 2. Important Directories .................................................................................................... 15 Discussion ............................................................................................................. 15 Standard Linux Directory Scheme ..................................................................... 15 The User's Home Directory .............................................................................. 15 The /tmp Temporary Directory .......................................................................... 16 The /etc Configuration Directory ....................................................................... 16 The /bin and /usr/bin Command Directories ......................................................... 16 The /sbin and /usr/sbin Command Directories ....................................................... 17 The /var "Variable" Directory ........................................................................... 17 root vs. /root vs. / (the filesystem root) ............................................................... 17 Examples .............................................................................................................. 17 Example 1. The /tmp Directory .......................................................................... 17 Example 2. Looking for commands in /bin and /usr/bin ........................................... 18 Online Exercises .................................................................................................... 18 Online Exercise 1. Exploring Important Directories ................................................ 18 Setup .................................................................................................... 19 Specification .......................................................................................... 19 Deliverables ........................................................................................... 19 Possible Solution .................................................................................... 19 Questions .............................................................................................................. 19 3. Managing Files ............................................................................................................ 22 Discussion ............................................................................................................. 22 Redirection .................................................................................................... 22 Copy files with cp .......................................................................................... 23 Examples: ............................................................................................. 23 Move/rename files with mv .............................................................................. 24 Examples: ............................................................................................. 24 Remove (delete) files with rm ........................................................................... 25 Examples: ............................................................................................. 25 Clobbering files (oops!) ................................................................................... 26 Examples .............................................................................................................. 27 Example 1. Moving/Archiving a Simple Website ................................................... 27 Example 2. Moving a Simple Website ................................................................. 28
Filesystem Basics
Example 3. Removing robots.txt ......................................................................... 28 Online Exercises .................................................................................................... 28 Online Exercise 1. Create a Website ................................................................... 28 Setup .................................................................................................... 28 Specification .......................................................................................... 28 Deliverables ........................................................................................... 29 Possible Solution .................................................................................... 29 Online Exercise 2. Investigate the "noclobber" option ............................................. 29 Specification .......................................................................................... 29 Deliverables ........................................................................................... 30 Possible Solution .................................................................................... 30 Questions .............................................................................................................. 30 4. Managing Directories .................................................................................................... 33 Discussion ............................................................................................................. 33 Creating (new, empty) directories: mkdir ............................................................ 33 Listing Directory Trees .................................................................................... 34 Removing (empty) directories: rmdir .................................................................. 34 Listing Directory Trees .................................................................................... 34 Copying directory trees: cp -r ........................................................................... 35 Removing directory trees: rm -r ........................................................................ 37 Examining Directories with tree ........................................................................ 38 Examples .............................................................................................................. 38 Example 1. Making a Backup of a Directory Tree .................................................. 38 Example 2. Creating a Local Copy of Example Files .............................................. 40 Online Exercises .................................................................................................... 40 Online Exercise 1. Managing Directories ............................................................. 40 Specification .......................................................................................... 40 Deliverables ........................................................................................... 41 Cleaning Up .......................................................................................... 41 Questions .............................................................................................................. 41 5. File Names and File Globbing ........................................................................................ 45 Discussion ............................................................................................................. 45 Filenames ...................................................................................................... 45 Hidden Files .................................................................................................. 46 "Globbing" .................................................................................................... 47 Examples .............................................................................................................. 48 Example 1. Finding That Configuration File ......................................................... 48 Example 2. Listing Static Libraries ..................................................................... 49 Example 3. Listing Man Pages ........................................................................... 50 Online Exercises .................................................................................................... 50 Online Exercise 1. Managing Files with File Globbing ............................................ 50 Specification .......................................................................................... 51 Deliverables ........................................................................................... 51 Suggestions ........................................................................................... 52 Questions .............................................................................................................. 52 6. Examining Files ........................................................................................................... 56 Discussion ............................................................................................................. 56 Viewing Files ................................................................................................ 56 The file Command .......................................................................................... 56 The cat Command .......................................................................................... 56 The more and less Pagers ................................................................................ 57 The head Command ........................................................................................ 58 The tail Command .......................................................................................... 59 Examples .............................................................................................................. 60
Filesystem Basics
Example 1. Identifying Documents with file ......................................................... 60 Example 2. Determining the Correct File Type with file .......................................... 60 Example 3. Skimming Files with head ................................................................. 61 Example 4. Combining Multiple Files into a Single File with head ............................ 61 Example 5. Monitoring Multiple Files with tail ..................................................... 62 Online Exercises .................................................................................................... 63 Online Exercise 1. Browsing Text Files ............................................................... 63 Specification .......................................................................................... 63 Deliverables ........................................................................................... 64 Cleaning Up .......................................................................................... 64 Questions .............................................................................................................. 64 7. Editing Files ............................................................................................................... 67 Discussion ............................................................................................................. 67 Text Editor vs. Word Processor ......................................................................... 67 Text Editors in Linux ...................................................................................... 67 nano ............................................................................................................. 67 gedit ............................................................................................................. 68 Online Exercises .................................................................................................... 69 Online Exercise 1. Text Editors .......................................................................... 69 Specification .......................................................................................... 69 Deliverables ........................................................................................... 70
Chapter 1. Filesystem Navigation

Key Concepts
The Linux filesystem is an "inverted tree" of directories and files, with a root directory called "/". Every process has a current working directory, often called it's "cwd". The pwd command reports the bash shell's cwd. The cd command changes the bash shell's cwd. Files can be referenced through either absolute or relative references.
Discussion
The Filesystem Inverted Tree
The concept of a filesystem organized into directories is common to many computer systems. An individual file is given a name (a filename), and filenames get organized into a directory (called a folder in some operating systems). But directories are themselves a kind of file, so they, too, can get collected into other directories. This process can continue, level after level, to create a highly structured environment. For example, a user alice might have several songs stored in files with names like song1.midi and song2.midi. These songs might be collected into a directory named songs. User alice might also have some photographs stored in the files picture1.png and picture2.png. These she might have placed into a directory called photos. These two directories, songs and photos might be organized into a directory called media. This new directory might be only one of several in a website directory. One way to visualize this is as a branching tree structure. The website directory may contain the subdirectories html, forms and media. Together with the directories and files named above, this arrangement might be diagrammed as shown in Figure 1-1.
Figure 1.1. Sample Directory Tree

-- html | website -- -- forms | | -| | | -- songs -| | | | | -| | -- media -| -| | -- photos-| --
song1.midi
song2.midi
picture1.png
picture2.png
It is the branching appearance of this diagram that gives the notion of a directory tree its name. This diagram is most often drawn with the branches hanging down, and with the root of the tree (the directory website in this case) at the top, and so is referred to as an inverted tree directory structure.
Filesystem Navigation
Linux uses a single directory tree for its entire collection of files and directories, all branching from a single root named "/" (read as "slash") and called the root directory. This directory tree can be visualized, in part, as shown in Figure 1-2.
Figure 1.2. Directory Tree Starting at /

-- bin... | / -- -- etc... | | -| | | -- alice -| | | | | -| | -- home -| | -| | | | -- elvis -| | | -...
mystuff...
website...
docs...
graphics...
The ellipses in the diagram are meant to suggest that there are many files and directories not shown here. You can see, for example, where the website directory tree drawn earlier could be attached to extend the diagram. The important feature of the diagram at this point is the tree structure growing out of /, not the specific contents. When naming a file or directory, start from the root of the Linux filesystem tree and list all of the directory branches out to the file or directory you want, separating each part with a slash (/). This is known as the item's fully qualified name, or FQN for short. For example, the FQN of the website directory above would be /home/alice/website. User alice's song1.midi file would be identified as /home/ alice/website/songs/song1.midi. The string of directory names leading up to the filename is the path to the file. This type of naming helps guarantee each file and directory has a unique fully qualified name. In actual practice, special shortcuts and default assumptions are used to simplify the typing of filenames
Exploring the Directory Tree using Nautilus

In the Red Hat Enterprise Linux X graphical environment, users can use Nautilus as a sophisticated tool for navigating the filesystem. In the Red Hat Enterprise Linux desktop, a Nautilus window can be opened by (left) double clicking on the home icon found in the upper left-hand corner.
Figure 1.3. Nautilus Window
As an alternative, you can use the Nautilus browser (which might be more familiar to people familiar with previous releases of Red Hat Enterprise Linux) by selecting the Applications : System Tools : File Browser menu item.
Figure 1.4. Nautilus's Browser View
If you would prefer to use the browser by default, you may choose "Alway open in Browser Windows" from the Behavior page of Nautilus's Preferences, found under the Edit menu.
Figure 1.5. Configuring Nautilus to use Browser View by Default
By default, the Nautilus browser's sidepane includes a "hotlist" of commonly used directories. The entire directory tree can be browsed, however, by choosing "tree" from within the side pane's popup menu.
Figure 1.6. Choosing the Nautilus tree sidepane
The Nautilus sidepane should now show list of directories (folders), arranged as two trees, one roooted at the root directory (labeled "File System"), and the other rooted at your home directory. A directory can be expanded by (left) clicking on the triangle icon next to the folder icon. When the folder icon, or folder name, is clicked, the contents of the folder will be displayed in Nautilus's main pane.
Figure 1.7. Exploring the directory tree with Nautilus
Spend some time exploring the filesystem with Nautilus, and viewing the contents of various directories. Try viewing the contents of the file /etc/sysconfig/network-scripts/ifcfg-lo, by first selecting the directory /etc/sysconfig/network-scripts in the sidepane, and then (left) double clicking on the ifcfg-lo icon in the main pane. Note that when you select a particular directory for viewing, the Location is represented as buttons above the main pain. Typing CTRL+L converts the buttons into a text entry, from which you can navigate by just typing a directory path. Notice the use of TAB completion, as well.
Figure 1.8. Typing a directory into Nautilus's Location Entry (using CTRL-L)
Now that you have spent some time exploring the concept of a directory tree, and how absolute references to directories and files are constructed, we will spend some time talking about Linux filesystem related concepts.
The Current Working Directory (cwd)

Each Linux process (i.e., program, command) is assigned a default directory when it runs. This default directory is used to complete the FQN for any file whose name is not specified as an FQN to start with. This has two effects. First, it allows the process to refer to files more simply, without using tiresomely long names. Second, it allows the process to operate more flexibly since its actions depend in part on the default directory. Change the default directory, and the action of the process is changed. This default directory is referred to as the current working directory, or cwd, for the process. One common approach to take with directories is to think of a directory as a place rather than a thing. From this point of view, then, the cwd for a process can be thought of as the process's current location. If a process changes its cwd, then it has moved from one directory to another. This view of a directory as a place is so common in the Linux world that the vocabulary of place is quite pervasive in the language. For instance, users talk about navigating the filesystem as they move from directory to directory. A process is first assigned a cwd by the parent process which starts it. However, a process is not locked into a particular cwd the whole time it runs. When needed, a process can change its cwd to another directory before continuing its work.
Where Am I? The pwd command

One of the processes found running on a Linux system is the command shell. If you login to the system through a virtual terminal, or start a terminal program in X, you see a command prompt where you can enter commands for the system to act on. This command prompt is produced by the command shell, the process which is responsible for reading and interpreting your commands. The default command shell for Red Hat Enterprise Linux systems is the bash (Bourne-again shell) command shell. Like any other process, the bash shell keeps track of its cwd while it runs. This current working directory affects the commands you type at the command prompt, and becomes their cwd as they run. In essence,
then, the cwd of the bash shell is your cwd, and can be thought of as where you are within the system. Obviously, keeping track of this cwd is important. Fortunately, there is help. First, the command prompt itself shows the last directory of the path to the cwd. For example, user alice, while working in her website directory, might see a command prompt like this:
[alice$station website]$
This prompt reminds her that she is logged in with username "alice" on the computer "station," and is in the directory website. But there can be other website directories elsewhere on the system. The complete (or absolute) path to the current working directory can be displayed with the pwd (print working directory) command. Usage: pwd User alice wants to verify she is in the correct directory. She would use pwd.
[alice@station student]$ pwd /home/alice/website
Moving Around - The cd command

As was noted earlier, processes are able to change their cwd as needed. This includes the bash command shell, which provides the cd (change directory) command for changing the current directory from the command prompt. Usage: cd [DIRECTORY] If not specified, DIRECTORY defaults to the user's home directory. Consider the following sequence of commands:
[alice@station website]$ pwd /home/alice/website [alice@station website]$ cd /home [alice@station home]$ pwd /home [alice@station home]$ cd /home/alice/website/songs [alice@station songs]$ pwd /home/alice/website/songs
Notice how the result of the pwd command and the last part of the command prompt both change to reflect the new cwd after each cd command. Because navigating through directories is so important, special abbreviations are available for referencing certain directories:
Table 1.1. Special Directory Names

Symbol . .. ~ Meaning The current working directory The parent directory The user's home directory The previous working directory
10
The entries in this table require a little explanation. First, all but the last of these symbolic names are recognized by the bash shell in most contexts, not just in the cd command. Second, the directory tree is often described using a parent/child analogy. If dir1 contains dir2, then dir2 is said to be a child directory of dir1, and dir1 is the parent directory of dir2. Thus, the .. directory is one level closer to the root of the tree than is the cwd. Next, each user account is assigned a home directory, usually a subdirectory of /home that matches their username (more on this in the next chapter). The ~ directory represents that directory. Finally, the dash (-) is a special cd command option that refers to the previous working directory, making it possible to switch back-and-forth between directories easily. Again, consider the following sequence of commands:
[alice@station songs]$ pwd /home/alice/website/songs [alice@station songs]$ cd ~ [alice@station alice]$ pwd /home/alice [alice@station alice]$ cd [alice@station songs]$ pwd /home/alice/website/songs [alice@station songs]$ cd .. [alice@station website]$ pwd /home/alice/website [alice@station website]$ cd [alice@station alice]$ pwd /home/alice
The last set of commands in the example above shows that, by default, the cd command uses the ~ directory if no directory is given.
Absolute and Relative References

This section describes two ways of identifying the location of a file. In this section, as in almost every other place in these lessons, it is important to remember that in Linux a directory is a kind of file, so all the things said about naming files apply to directories as well as ordinary data files. To identify a file, enough information has to be provided to locate the file within the filesystem. This location can be given in two ways: as an absolute reference (or absolute pathname), or as a relative reference. Absolute references start with a slash (/) and give the FQN of the file. That is, the absolute reference names each branch of the filesystem directory tree, starting from /, that must be traversed to reach the file. Regardless of where you are in the filesystem (that is, regardless of the value of the cwd), an absolute reference unambiguously identifies a specific resource. Several examples of absolute references have already been discussed in this lesson. A relative reference does not describe a path to the file starting from /, but rather describes a path starting from the current directory. For example, if the cwd is /home/alice, then a relative reference to song1.midi might be website/songs/song1.midi. This is a relative reference because is does not start with /. This reference only names the directories that must be traversed starting from /home/ alice, not starting from /. In order to be valid, a relative reference must start by naming a directory (or file) in the cwd. All directories in Linux contain two special entries, the . and .. directories, that represent the current and parent directories, respectively. Thus, in the discussion earlier about the cd command, the example cd .. was really just a use of a relative reference. Table 1-2 contains some additional examples of relative references. Each of these is a reference to the file with FQN /home/alice/sample.txt. Some of these examples are intentionally "inefficient."
11
Table 1.2. Sample Relative References to /home/alice/sample.txt

cwd /home/alice /home/alice/website/songs /home/elvis/docs /home /home Relative Reference sample.txt or ./sample.txt ../../sample.txt ../../alice/sample.txt alice/sample.txt ../home/alice/website/../ sample.txt
Yes, that last example is pretty silly, but nevertheless is still perfectly valid.
Examples
Discovering your home directory
Hogan has just logged in and is unaware that, by default, his home directory is assigned as his cwd. Further, he is unaware that his home directory is /home/hogan. He sets out to discover what his home directory is. He also wants to see if, in fact, his default cwd is his home directory.
[hogan@station hogan]$ cd ~ [hogan@station hogan]$ pwd /home/hogan [hogan@station hogan]$ cd /home/hogan
Hogan knows the cwd ends in hogan (from the command prompt) but he doesn't know if this is home. He uses the cd command with the ~ directory symbol to change to his home directory. Next, Hogan uses the pwd to display his home directory's FQN. He now knows /home/hogan is home, but is that where he started from? To see where he started from, Hogan again uses the cd command, this time to go back to his previous working directory. The cd - command displays the cwd after making the change (why is this reasonable behavior?). Hogan now knows that his system logs him in with his home directory set as his initial working directory.
Climbing up and down the tree.

Alice needs to edit some of the html files in her website. After she logs in, she will need to change to the appropriate subdirectory under her home directory. When she is done, she wants to change to the / etc to examine a configuration file. Finally, she wants to change back to her home directory to begin working on another task.
[alice@station alice]$ cd website/html [alice@station html]$ ... (edits the html files) ... [alice@station html]$ cd /etc [alice@station etc]$ ... (examines the configuration file) ... [alice@station etc]$ cd [alice@station alice]$ ... (continues with other tasks) ...
12
Alice uses a relative path to change to the /home/alice/website/html directory because it is shorter than the FQN, but either would have worked. After finishing her work on the html files, Alice uses an absolute reference to change directories to /etc. In this case, the absolute path is much shorter and easier to understand than the equivalent relative reference, ../../../../etc. Note that the command prompt does not contain a / before the etc directory name. The default bash prompt displays only the final component of the cwd. Done with her work in /etc, Alice returns to her home directory using the default behavior of the cd command.
Online Exercises
Exploring the filesystem
Lab Exercise
Objective: Demonstrate the use of cd. Estimated Time: 10 mins.
Specification
This lab will have you set the current working directory of four simultaneously running bash shells. If you are using the X graphical environment, you may simply open four terminal windows. If you are using virtual consoles, you may use any four of the six provided virtual consoles. Open four terminals with bash shells (as described above), using your primary account. Using the cd command, set the current working directories of the shells to the following four directories: 1. ~/.. 2. /tmp 3. /etc/sysconfig 4. /usr/share/gnome
Deliverables
1. 1. Four concurrently running bash shells, with each shell's current working directory set to one of the four directories listed above.
Possible Solution
The following command would set a shell's current working directory to the first directory listed above.
[student@station student]$ cd ~/..
Questions
Navigating the Filesystem
All these questions assume you are logged in as user student on a typical classroom system.
13
1. a.
Which of the following is a relative directory reference? /home/student /etc .. ~
b. c. d. 2. a. b. c. d. 3. a. b. c. d.
Which of the following could have been displayed by pwd? home/student /etc .. ~
Following the command cd ~, which is the most likely result from pwd? /home/student /etc .. ~
Your screen shows the following:

[student@station html]$ cd ../bin [student@station bin]$ pwd /home/student/bin [student@station bin]$ cd -
4. a.
What would be the output of the command pwd? /home/student /home/student/html /home/html -
b. c. d. 5. a. b. c. d.
Which of the following pairs of commands always produce the same effect? cd .. and cd cd / and cd root cwd and pwd cd ~ and cd
14
Chapter 2. Important Directories

Key Concepts
Every user account is assigned a home directory. The /tmp directory is used for global scratch space. The /bin and /usr/bin directories contain commonly executed files. The /etc directory contains system configuration files. Root's home directory, /root, should not be confused with the filesystem root directory, /.
Discussion
Standard Linux Directory Scheme
Linux can be used to support many different types of computer systems: servers, development workstations, personal desktop systems, etc. In order to standardize the filesystem's directory structure across this varied range of systems, most Linux systems employ a common naming and utilization scheme that makes the systems much easier to use and maintain. Using the same type of tree diagram as in the last chapter, part of the first layer of the directory tree might look something like this:
Figure 2.1. Partial Directory Tree Starting at /

-- bin... | -- etc... | -- home... | -- root... / --| -- sbin... | -- tmp... | -- bin -- usr --| | -- sbin -- var...
This chapter presents a brief (sometimes very brief) description of the role played by each of these directories.
The User's Home Directory

Each user of a Linux system is assigned a special directory, called their home directory, that represents their "private" space on the system. Typically, this is a subdirectory under the /home directory whose name matches the user's login username. (Examples of this, such as /home/alice or /home/hogan were seen in the last chapter.) The one significant exception to this is the superuser (or root user) whose home directory is usually /root. For any user, the tilde character (~) represents the FQN of the user's home directory when used as the first character of a file or directory reference. The most obvious purpose for a user's home directory is as private data space, a place where they can store files away from the files of other users. Normally, users are free to create subdirectories under their home directory and organize their data as they see fit (subject to any quota restrictions that might
15
Important Directories
be in place). Two different users can even give files or directories the same name without conflict, since each is stored in their own space. For example, both Alice and Hogan can have public_html subdirectories of their own, /home/alice/public_html and /home/hogan/public_html respectively. Keeping each user's space separate from the others makes security much easier for the users, as well. Another important aspect of a user's home directory is that it provides a place to store user-specific configuration files. When Blondie logs in, for example, she may need a different environment than does Prince. Blondie may prefer different screen colors, different command shortcuts, perhaps even a completely different desktop environment than Prince. Local, user-specific configuration files make this possible. Many of these configuration files are created automatically in a default form when a user's account is created, or on first use of a particular system resource. Users normally are given relatively complete control over their own home directories, while access to the rest of the system is restricted. Alice, for example, most likely cannot modify or delete any of the configuration files in the /etc directory, even though she may be able to read them. And most likely she cannot read, modify or delete any of the files in the directories of other users of the system. When a user first logs in to the system, they are placed "in" their home directory; that is, the system sets their initial working directory as their home directory.
The /tmp Temporary Directory

In addition to their personal space in their home directories, users are given access to shared "scratch" space in the /tmp directory. A program to do data compression might store partial results in /tmp, only placing the final results in the user's home directory when the work was complete. Linux systems often implement quotas on user accounts to prevent any one user from consuming an unfair proportion of the available space. The /tmp directory gives all users access to additional space to meet short-term needs without charging the space against their quota. This is especially important since users may not always be aware of how much extra space a service needs, and some services (such as X) cannot run at all if there is no temporary storage space available. This "global" scratch space is available to all processes on the system as well as all users. Files placed in this directory are deleted automatically by the system after a few days.
The /etc Configuration Directory

One of the unique characteristics of a Linux system is its flexibility. Virtually every aspect of the system can be configured in some fashion by editing a configuration file. These configuration files are normally placed in /etc or a subdirectory of /etc. For example, the sendmail program, often used for running a mail server, uses configuration files placed in /etc/mail. System startup scripts are found in /etc/ rc.d, while network configuration files are found in /etc/sysconfig. Obviously, ordinary users cannot modify the files in /etc (or even read them, in some cases), but system administrators spend a lot of their time working with files stored here.
The /bin and /usr/bin Command Directories

Most system commands are stored as binary files in a machine-readable format. Commands appropriate for use by ordinary users are usually placed in the /bin or /usr/bin binary directories. Core utilities like ls, cd, cp, mv and the vi text editor, without which the system would not be usable, go in /bin. Supplementary utilities like compilers, your web browser and office suite go in /usr/bin, which can be made available to other systems over the network. Think of /bin and /usr/bin as unprivileged command directories, since no special privileges are required to use the commands found in them. Think of /bin and /usr/bin as unprivileged command directories, since no special privileges are required to use the commands found in them.
16
The /sbin and /usr/sbin Command Directories

Just as /bin and /usr/bin store command files for ordinary users, so also /sbin and /usr/sbin store command files for use by the superuser, root. These include commands for attaching and removing hardware, for starting and stopping the system, and for performing system maintenance. These privileged commands also are stored in two separate directories, for the same reasons as for /bin and /usr/bin.
The /var "Variable" Directory

The collections of commands and configuration files found in places like /bin or /etc change very little, if at all, from day to day. These files tend to be very stable. Some files, however, change frequently. These include incoming/outgoing email, system logs, news group collections, web sites, ftp file archives and the like. These variable content files and directories are usually collected in the /var directory. Placing such files here makes it easier to allocate space for them, and also makes it easier to protect the more stable files found elsewhere on the system.
root vs. /root vs. / (the filesystem root)

It is an unfortunate accident of history that the term root plays such an important and yet confusing role in Linux. The word "root" is the username of the superuser, that is, the user with supreme authority over the system. It is also the name of that user's home directory, /root. And it is also the term used for the base (top?) of the filesystem's directory tree, the / directory. Normally, the meaning of the term can be gleaned from context, but a phrase such as "root directory" can be ambiguous. Try to anticipate and avoid such confusion in your own communication, and seek clarification if you encounter a use of the word that cannot be resolved from context.
Examples
The /tmp Directory
Alice just learned that some processes automatically use /tmp for scratch space, and she wishes to see if anything she has done used that space. She also wants to test to see if she really can create files there. To create a file, Alice will use the command touch.
[alice@station alice]$ cd /tmp [alice@station tmp]$ ls orbit-alice ssh-XXDg4ke3 [alice@station tmp]$ ls -l total 8 drwx-----2 alice alice drwx-----2 alice alice [alice@station tmp]$ touch newfile [alice@station tmp]$ ls -l total 8 -rw-rw-r-1 alice alice drwx-----2 alice alice drwx-----2 alice alice
4096 Mar 16 08:04 orbit-alice 4096 Mar 16 07:07 ssh-XXDg4ke3
0 Mar 16 14:14 newfile 4096 Mar 16 08:04 orbit-alice 4096 Mar 16 07:07 ssh-XXDg4ke3
After changing to /tmp, Alice uses ls and ls -l to view the contents of the /tmp directory. While there is a lot about the long listing produced by ls -l she doesn't know, she sees her username enough to convince her that, as she had been told, some of the programs she runs had left files in /tmp. In fact, both orbit-alice and ssh-XXDg4ke3 were placed there when she started her desktop environment.
17
Alice uses the touch command to create a new, empty file named newfile in /tmp. Like many Linux commands, when touch succeeds, it succeeds silently (with no visual feedback on the screen). Alice verifies that touch worked with another ls -l.
Looking for commands in /bin and /usr/bin

Hogan wants to see if some of the common commands he uses are in either of the binary command directories /bin or /usr/bin. He decides to test cp, mozilla, cd and fdisk. Hogan could use ls to look for these commands, but he wants to try a new command instead, which. Simple usage: which PROGRAM Displays the FQN of the PROGRAM command file that will be used if the user runs the command.
[hogan@station hogan]$ which cp /bin/cp [hogan@station hogan]$ which mozilla /usr/bin/mozilla [hogan@station hogan]$ which cd /usr/bin/which: no cd in (/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/home/hoga n/bin:) [hogan@station hogan]$ which fdisk /usr/bin/which: no fdisk in (/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/home/h ogan/bin:) [hogan@station hogan]$ ls /sbin ... ether-wake ldconfig portmap usbmodules extendfs lilo poweroff vboxd fdisk logdump pppoe vgcfgbackup ...
Hogan sees that cp is found in the /bin directory, which seems right to him since files in /bin should be files that ordinary users would want that should always be present on a system. Hogan finds mozilla (his web browser of choice) in /usr/bin. This also seems right since files in /usr/bin should be files an ordinary user might want that might not always be available. Hogan is surprised at the results for cd, which seems to say there is no cd command on this system. He was expecting to see a response of /bin/cd. (Why?) What Hogan does not know is that cd is an example of an internal command, one that's provided by the bash shell itself. On the other hand, if he checked, he would find bash in /bin. Hogan is also surprised at the results for fdisk. He knows this is an important but dangerous command used to partition disk drives, and should only be available to the superuser, so he was expecting a response of /sbin. He can even see fdisk in /sbin when he looks using ls. Upon closer inspection, he discovers that which didn't even look for fdisk in /sbin. Later, Hogan will learn about the command path and discover why this makes perfect sense.
Online Exercises
Exploring Important Directories
Lab Exercise
Objective: Explore some of the characteristics of the important directories discussed in this chapter. Estimated Time: 15 mins.
18
Setup
In this exercise you will need to use the touch command to create files. This command was demonstrated in the examples for this chapter. You will need to use the which command to locate command files. This command was demonstrated in the examples as well.
Specification
1. Open a new terminal session or virtual console (this is necessary for the online grading mechanism to evaluate your work). You want to test whether or not you can use touch to create files in various directories. You expect this to work in your home directory and in /tmp, but not in other directories. Use the command touch newfile.username (where username is replaced by your username) in each of the following directories. a. ~ b. /tmp c. /bin d. /root Clearly, you should expect some of your efforts to fail. Return to your home directory when you are done. 2. Using which, determine the FQN of the binary command file for each of the following: pwd, ls and whoami. Run each of these commands in its "raw" form by using the FQN at the command prompt. Finally, exit the terminal session or virtual console. When you have finished, exit your shell, so that your command history is saved to the file ~/.bash_history.
Deliverables
1. 1. Two files called ~/newfile.username username is your primary account name. and /tmp/newfile.username, where
2. A ~/.bash_history file that contains a record of your "raw" commands.
Possible Solution
The following sequence of commands demonstrates the use of the which and touch commands, and running a command using its absolute reference.
[student@station student]$ touch /tmp/newfile.student [student@station student]$ which pwd /bin/pwd [student@station student]$ /bin/pwd /home/student
Questions
1. Which of these is most likely not someone's home directory?
19
a. b. c. d. 2. a. b. c. d. 3.
/home/student /root / None of these - all are likely to be home directories
The file named.conf is a system configuration file. This file belongs in /tmp /etc /bin /sbin
The file e2fsck is a privileged command that must always be available to the system. This file would be found in a. b. c. d. /tmp /etc /bin /sbin
4. a.
The /root directory is noteworthy because It is the root of the Linux filesystem It is the superuser's home directory It can always be abbreviated as ~ Its contents cannot be read by any user.
b. c. d. 5. a. b. c. d. 6. a. b. 7. a.
Which of the following commands could not be used to create a file in /tmp? touch /newfile touch /tmp/newfile touch ../newfile touch ../tmp/newfile
T/F: /bin and /usr/bin contain the same files. True False
T/F: Only the root user can read files in etc. True
20
b. 8. a. b. 9. a. b. 10. a. b.
False
T/F: Files for a website might often be found in /var. True False
T/F: Configuration files that customize each user's account are placed in /etc/users.d. True False
T/F: X uses /tmp. True False
21
Chapter 3. Managing Files

Key Concepts
Files can be easily created (or appended to) using shell redirection. The cp command copies files. The mv command moves files. The rm command removes files. Files can be "clobbered" as a result of redirection, moving, or copying.
Discussion
Redirection
Many commands produce visible output. Normally, output is displayed on the screen. Notice, for example, how the output from pwd is displayed for Julius:
[julius@station julius]$ pwd /home/julius [julius@station julius]$
Linux likes to think of everything as a file. In the example above, pwd sent the output to the standard output file, or STDOUT, which is by default the screen for users logged into the system. One of the features of the bash command shell is that output which would normally be directed at the screen through STDOUT can, instead, be redirected to some other file. This is done by using the special redirection symbol, >, as in this example:
[julius@station julius]$ ls [julius@station julius]$ pwd > results.txt [julius@station julius]$ ls results.txt
In this example, the usual results of the command were redirected to the file results.txt, creating the file in the process. A helpful command that can let us verify what has happened is the cat (concatenate) command. Usage: cat [OPTIONS] [FILE...] Concatenate FILE(s) to standard output. The cat command copies each of the files listed to standard output. If more than one file is listed, this effectively concatenates the files. As for any other command, this output will display on the screen by default, but could also be redirected to a file. We will use it to display a file by naming a single file and not redirecting the results.
[julius@station julius]$ pwd > results.txt [julius@station julius]$ ls results.txt [julius@station julius]$ cat results.txt /home/julius
22
Managing Files
This time, Julius captured the output of pwd by redirecting the results to results.txt, and then displayed results.txt with cat. If the file already exists, redirection will delete and re-create the file empty, then capture the new output. If, however, a double arrow (>>) is used, the new output is appended to the file. If >> is used and the target file does not exist, it is created, just as if > had been used. The > symbol is an example of a shell meta-character, that is, a symbol with special meaning which the bash shell sees and interprets before the rest of the command is acted upon. There are many of these as we shall learn in coming lessons, and all require careful attention. The simplest command of all that produces visible output is echo. The echo command takes whatever text is typed as part of the command and echoes it to STDOUT (usually the display). This simple command, together with redirection, can be used to create text files. Usage: echo [OPTIONS] [STRING...] Echoes STRING(s) to standard output. Example 1:
[julius@station julius]$ echo Hello > greetings.txt [julius@station julius]$ echo How are you >> greetings.txt [julius@station julius]$ cat greetings.txt Hello How are you
Example 2:
[julius@station [julius@station [julius@station /home/julius [julius@station [julius@station /home/julius results.txt julius]$ ls julius]$ pwd > results.txt julius]$ cat results.txt julius]$ ls >> results.txt julius]$ cat results.txt
Copy files with cp

Duplicate copies of files can be created with the cp (copy) command. Usage: cp [OPTIONS] {SOURCE} {TARGET} cp [OPTIONS] {SOURCE...} {DIRECTORY} In the first form, a copy of the SOURCE file is made as TARGET. In the second form, one or more files can be copied at one time to a directory. A copy of SOURCE, ..., is made in DIRECTORY and named DIRECTORY/SOURCE, .... With the appropriate options (not discussed here - try man cp) whole subdirectory trees can be copied at once.
Examples:
1. Make a copy of mysong.midi and name the copy backup.midi: cp mysong.midi backup.midi 2. Make a copy of mynovel.txt in /tmp:
23
Managing Files
cp mynovel.txt /tmp The resulting file will be named /tmp/mynovel.txt. 3. Copy the files songs.tar and novels.tgz into the directory /tmp: cp songs.tar novels.tgz /tmp The resulting files will be named /tmp/songs.tar and /tmp/novels.tgz. 4. Make a copy of webpage.html from your home directory to the current directory: cp ~/webpage.html . The resulting file will be named ./webpage.html.
Move/rename files with mv

Files can be moved from one directory to another, or from one name to another (renamed), with the mv (move) command. Usage: mv [OPTION...] {SOURCE} {TARGET} mv [OPTION...] {SOURCE...} {DIRECTORY} In the first form, the SOURCE file is renamed as TARGET. In the second form, one or more files can be moved at one time to a directory. The files SOURCE, ..., are moved to DIRECTORY and named DIRECTORY/SOURCE, .... SOURCE can be a directory, in which case the directory is moved/renamed. The mv command is particularly interesting because buried in the fabric of how it works is a critical fact about Linux filesystems: Linux treats the name of a file as being distinctly separate from the contents of the file. Even though the command mv stems from the word "move," mv rarely actually moves data around. Instead, the filesystem merely records a name change. If the name changes from /somedir/ somefile to /somedir/newname, we see this as "renaming" the file. If the name changes from / somedir/somefile to /newdir/somename, we see this as "moving" the file. If the name changes from /somedir/somefile to /newdir/newname, we see this as a double change, both moving and renaming. But to Linux, all of these are the same thing, a change in the FQN for the file.
Examples:
1. Rename mysong.midi as backup.midi: mv mysong.midi backup.midi 2. Move mynovel.txt to /tmp: mv mynovel.txt /tmp The resulting file will be named /tmp/mynovel.txt. 3. Move both songs.tar and novels.tgz to /tmp: mv songs.tar novels.tgz /tmp The resulting files will be named /tmp/songs.tar and /tmp/novels.tgz.
24
Managing Files
4. Move webpage.html from your home directory to the current working directory: mv ~/webpage.html . The resulting file will be named ./webpage.html. 5. Rename the html subdirectory of the current working directory to public_html: mv html public_html This renames ./html as ./public_html, assuming that ./public_html does not already exist - see the next example. 6. Your cwd is your home directory. You have two subdirectories, ~/images and ~/html. Move the images directory to the html directory: mv images html The resulting directory will be named ~/html/images. Note the similarity between this example and the previous one. The critical difference is that, in this case, the target directory ~/html already existed, so mv moved the source directory inside (underneath) the target. In the previous example, the target directory did not already exist, so mv renamed the source directory to the new name.
Remove (delete) files with rm

Files can be removed (erased, deleted) with the rm (remove) command. Usage: rm [OPTIONS] {FILE...} Removes the FILE(s) from the filesystem. Technically, it unlinks the FILE(s), a distinction which will be made more clear later. With the appropriate options (not discussed here - try man rm) whole subdirectory trees can be removed at once.
Note
rm cannot remove a directory unless special command options are used. There is a separate command, rmdir, for this purpose.
Warning
Yes, the rm command can remove entire directory trees at once. In the hands of the superuser it can delete the entire contents of the filesystem in a single stroke -- not usually what was intended.
Warning
The documentation for rm includes the following statement: "Note that if you use rm to remove a file, it is usually possible to recover the contents of that file." While this may be true, it requires expertise beyond the scope of this course, so, for all practical purposes, you should treat this command as non-reversible.
Examples:
1. Delete mysong.midi: rm mysong.midi
25
Managing Files
2. Remove both songs.tar and novels.tgz: rm songs.tar novels.tgz 3. Remove photos.html from your home directory: rm ~/photos.html
Clobbering files (oops!)

The warnings contained in the previous section hint at some of the dangerous potential commands like cp, mv and rm can have. Unfortunately, while we usually see the rapids, we often miss the rocks beneath the surface of the water, and Linux has its share of those. This is the price you pay for the power of an operating system like Linux. In this section we highlight one such danger: clobbering files. Command redirection with >, and the commands cp and mv all can name target files. Normally, these are new filenames, and the commands create the files. But if an existing file is named as a target for redirection, cp or mv, the existing file will be destroyed without warning. This is known as clobbering a file. Because the problem can be so surprisingly subtle to a new user, we'll work through several examples. Example 3:
[julius@station [julius@station /home/julius [julius@station [julius@station mycwd julius]$ pwd > file1 julius]$ cat file1 julius]$ ls > file1 julius]$ cat file1
No real surprises here if you've been paying attention -- simply note how the second redirection replaced the previous contents of file1 without any warning messages. The file file1 has been clobbered. Example 4:
[julius@station [julius@station file1 [julius@station [julius@station file1 file2 [julius@station /home/julius [julius@station [julius@station file1 [julius@station /home/julius julius]$ touch file1 julius]$ ls julius]$ pwd > file2 julius]$ ls julius]$ cat file2 julius]$ mv file2 file1 julius]$ ls julius]$ cat file1
No surprises here, either -- note how mv replaced the previous contents of file1 without any warning messages. Again, file1 has been clobbered. cp would have replaced file1 in the same way. In order to understand our final example fully, we need to learn just a little bit more about redirection. In particular, we need to learn when redirection happens. Example 5:
[julius@station julius]$ touch file1 [julius@station julius]$ ls file1 [julius@station julius]$ ls > file2
26
Managing Files
[julius@station julius]$ ls file1 file2 [julius@station julius]$ cat file2 file1 file2
Why does file2 list both file1 and file2, when the first ls only shows file1? The answer: because of the sequence in which the critical events took place. Event 1: bash saw the redirection to the file file2, and created it as an empty file (see the earlier discussion of redirection). Event 2: ls then generated a file listing, showing both the previously-created file1 file and the newly-created file2 file. Event 3: the output from ls was sent to file2. Example 6:
[julius@station julius]$ pwd > mycwd [julius@station julius]$ cat mycwd /home/julius [julius@station julius]$ cp mycwd mycwd cp: `mycwd' and `mycwd' are the same file [julius@station julius]$ cat mycwd /home/julius [julius@station julius]$ cat mycwd > mycwd cat: mycwd: input file is output file [julius@station julius]$ cat mycwd [julius@station julius]$
This takes some explanation. First Julius creates mycwd using pwd, and displays its contents with cat. Then he uses cp to try and copy the file onto itself, but cp detects the "error," displays a message, and does nothing (mv would have behaved the same way). The cat on the next line verifies that mycwd is unchanged. OK so far. The next sequence, however, produces a real surprise. Julius uses cat with redirection to try and copy a file onto itself. cat also complains about the source (input) and target (output) files being the same, and would seem to do nothing. But the final cat of mycwd produces no response -- the file's contents are lost! Why? Because of the sequence in which the critical events took place. Event 1: bash sees the redirection to an existing file, deletes the file and re-creates it as empty (see the earlier discussion of redirection) -- mycwd is clobbered. Event 2: it no longer matters what cat does, because the source file mycwd has already been destroyed and re-created. It is too late for the error to be prevented. There are a few command options and special command shell features that can be used to mitigate this behavior somewhat, but the objective in this chapter is simply to begin to learn how commands such as these work, and in this section, how they work on existing files. This is a feature, not a bug, that needs to be understood to be used correctly.
Examples
Moving/Archiving a Simple Website
Madonna attempted to set up a simple personal website, but after she was done her ISP told her she had placed all of the website's files in the wrong directory. She placed the files in ~/html, and they should have been in ~/public_html. She needs to move the files to the correct directory. While she's at it, she also wants to make a backup copy of the files in a directory called archive, just in case she accidentally damages or deletes one from ~/public_html. The archive directory already exists.
[madonna@station madonna]$ ls html archive [madonna@station madonna]$ mv html public_html [madonna@station madonna]$ cd public_html [madonna@station public_html]$ ls index.html photo.jpeg [madonna@station madonna]$ cp index.html photo.jpeg ../archive
27
Managing Files
To move the files, all Madonna had to do was rename the directory. To make backup copies, she used the ability of the cp command to copy more than one file at a time.
Moving a Simple Website

Nero attempted to set up a simple personal website, but after he was done his ISP told him he had placed all of the website's files in the wrong directory. He placed the files in his home directory, and they should have been in ~/public_html. He needs to move the files to the correct directory. His problem is harder than Madonna's because he cannot simply rename a subdirectory. The public_html directory was created for him by his ISP.
[nero@station nero]$ ls index.html photo1.jpeg photo2.jpeg public_html [nero@station nero]$ mv index.html photo1.jpeg photo2.jpeg public_html [nero@station nero]$ ls public_html [nero@station nero]$ cd public_html [nero@station public_html]$ ls index.html photo1.jpeg photo2.jpeg
To move the files, Nero used the ability of the mv command to move more than one file at a time.
Removing robots.txt
Elvis notices a file in his ~/public_html directory that he doesn't recognize, a file named robots.txt. He decides to delete the file.
[elvis@station elvis]$ public_html [elvis@station elvis]$ index.html robots.txt [elvis@station elvis]$ [elvis@station elvis]$ index.html ls ls public_html rm public_html/robots.txt ls public_html
Online Exercises
Create a Website
Lab Exercise
Objective: Use redirection, cp and mv to create a simple website, and archive the website. Estimated Time: 15 mins.
Setup
In this exercise you will need to work with the directories ~/html and ~/archive. Login to your account, and use the following commands to create these two directories:
[student@station student]$ mkdir ~/html [student@station student]$ mkdir ~/archive
Specification
You wish to create a simple personal website using the techniques demonstrated in the Examples section of this chapter.
28
Managing Files
1. Create an "Under construction" webpage in the file ~/index.html. 2. Move this file to ~/html. 3. Rename the html directory as public_html. 4. Make a copy of your index.html file in ~/archive.
Note
The resulting webpage may or may not be visible with a browser, depending on additional features of your system's configuration. Your instructor will tell you if you can view the results of your work in this fashion.
Deliverables
1. 1. Two files, ~/public_html/index.html, and ~/archive/index.html, each of which contain identical content (The text "Under Construction", possibly with HTML markup.)
Possible Solution
The following sequence of commands provides one possible solution to this exercise.
[student@station [student@station [student@station [student@station [student@station [student@station [student@station [student@station index.html [student@station index.html student]$ student]$ student]$ student]$ student]$ student]$ student]$ student]$ mkdir html archive echo '<h1>' > index.html echo 'Under construction' >> index.html echo '</h1>' >> index.html mv index.html html mv html public_html cp public_html/index.html archive ls public_html
student]$ ls archive
Investigate the "noclobber" option

Lab Exercise
Objective: Investigate the consequences of the bash "noclobber" option. Estimated Time: 10 mins.
Specification
You wish to investigate the consequences of clobbering files, and the bash "noclobber" option. 1. Open a new terminal session or virtual console (this is necessary for the online grading mechanism to evaluate your work). 2. Create a file called source containing the word "source" and a file called target containing the word "target." 3. Use cp to clobber target with source. Verify the effect with cat. 4. Re-create target.
29
Managing Files
5. Clobber target again by catting source and redirecting the output. Verify the results. 6. Turn on the bash "noclobber" option with the following magic command:
[student@station student]$ set -o noclobber
7. Test both types of clobbering again, and note the results. 8. Turn off the bash "noclobber" option with the following magic command:
[student@station student]$ set +o noclobber
9. Exit the terminal session or virtual console.
Deliverables
1. 1. A ~/.bash_history file that contains a record of your enabling and disabling of the noclobber options.
Possible Solution
The following sequence of commands provides one possible solution to this exercise.
[student@station student]$ echo source > source [student@station student]$ echo target > target [student@station student]$ cp source target [student@station student]$ cat target source [student@station [student@station [student@station source [student@station student]$ echo target > target student]$ cat source > target student]$ cat target student]$ echo target > target
[student@station student]$ set -o noclobber [student@station student]$ cp source target [student@station student]$ cat target source [student@station student]$ echo target > target [student@station student]$ cat source > target -bash: target: cannot overwrite existing file [student@station student]$ set +o noclobber [student@station student]$ exit
Questions
1. a. b. c. d. The command to make a duplicate of a file is dupe rep copy cp
30
Managing Files
2. a.
To copy file1.txt to file2.txt, you could use cp file1.txt > file2.txt cp file1.txt file2.txt dupe file1.txt file2.txt mv file1.txt file2.txt
b. c. d. 3. a. b. c. d. 4. a. b. c. d. 5. a. b. c. d. 6. a. b. c. d. 7. a. b. c. d.
The successful command mv f1 f2 f3 f4 moves four items from the current directory moves four items to the current directory moves four items to the default directory moves three items from the current directory
To erase the file file1 from the current directory, you could use rm file1 erase file1 del file1 cut file1
The command cp file.a file.b will clobber file.b if file.a is an ordinary file file.a is non-empty file.b is an ordinary file file.b is a directory
If /home/student is a directory, then the command rm /home/student will remove all files in the directory delete the directory from the filesystem delete the student user account produce an error message
The command echo file.1 file.2 will copy file.1 to file.2 display the contents of file.1 and file.2 on STDOUT produce an error message because there is no redirection symbol display "file.1 file.2" on STDOUT
31
Managing Files
8. a.
The command ls > file.1 will produce an error message because > is an unprotected shell meta-character append the normal output of ls to file.1 replace file.1 if it already exists fail if file.1 already exists
b. c. d. 9. a. b. c. d. 10. a. b. c. d.
Assuming both exist, the command mv item.1 item.2 will move an entire subdirectory tree only if both item.1 and item.2 are existing subdirectories only if special command options are used only if item.2 is empty only if item.1 is empty
The command cp item.1 item.2 will copy an entire subdirectory tree only if both item.1 and item.2 are existing subdirectories only if special command options are used only if item.2 is empty only if item.1 is empty
32
Chapter 4. Managing Directories

Key Concepts
The contents of entire directory trees can be examined with the ls -R command. The mkdir command creates directories. The rmdir command removes (empty) directories. cp -r recursively copies directories rm -r recursively removes directories
Discussion
Creating (new, empty) directories: mkdir
Organizing files into directories requires the ability to create the directories you need. In Linux, the command for making a new directory is the mkdir (make directory) command. Usage: mkdir [OPTIONS] {DIRECTORY...} Creates the DIRECTORY(ies), if they do not already exist. Fails with an error message if one or more of the DIRECTORY(ies) do exist, but remaining new directories are still created. To create a ~/public_html subdirectory, Elvis could proceed as shown:
[elvis@station elvis]$ ls [elvis@station elvis]$ mkdir public_html [elvis@station elvis]$ ls public_html
Elvis can then create a subdirectory of public_html:

[elvis@station elvis] ls public_html [elvis@station elvis] mkdir public_html/images [elvis@station elvis] ls public_html images
Note, however, that by default mkdir will not create a subdirectory if the parent directory does not already exist:
[elvis@station elvis] ls work ls: work: No such file or directory [elvis@station elvis] mkdir work/spreadsheets mkdir: cannot create directory `work/spreadsheets': No such file or directory
Only if Elvis uses mkdir -p (as in 'parent') can he create an entire directory tree at once:
[elvis@station elvis] ls work ls: work: No such file or directory [elvis@station elvis] mkdir -p work/spreadsheets [elvis@station elvis] mkdir -p work/images/logos [elvis@station elvis] mkdir -p work/images/advertising [elvis@station elvis] ls work images spreadsheets
33
Managing Directories
[elvis@station elvis] ls work/images advertising logos
Listing Directory Trees

By now, you should be aware that the ls command, when given a directory as an argument, lists the contents of the directory. But what if the directory contains other directories? By default, the ls command will only show the contents of the top level directory. If you would like the ls command to list subdirectories as well, you can add the -R command line switch. Whenever a command iterates through a structure, operating on each element of the structure in the same manner as it operated on the topmost element, the command is said to act recursively. The -R command line switch specifies that the ls command should list with recursion. For example, hogan is exploring his machine's networking configuration, and suspects that the /etc/ NetworkManager directory is an important directory. In order discover what files it and its various subdirectories contain, hogan lists the contents of the directory recursively.
[hogan@station hogan]$ ls -R /etc/NetworkManager/ /etc/NetworkManager/: dispatcher.d NetworkManager.conf system-connections /etc/NetworkManager/dispatcher.d: 00-netreport 05-netfs 10-dhclient /etc/NetworkManager/system-connections: /etc/NetworkManager/VPN:
VPN
The ls -R can be combined with any of the previously learned command line switches, as well, such as ls -sR to include file sizes, or ls -lR to include file attributes.
Removing (empty) directories: rmdir

The command for removing a directory is rmdir (remove directory). Usage: rmdir [OPTIONS] {DIRECTORY...} Removes the DIRECTORY(ies), if they are empty. Fails with an error message if one or more of the DIRECTORY(ies) are not empty, but remaining (empty) directories are still removed. Note that a directory containing only a subdirectory is not empty. To remove his ~/public_html/images subdirectory, Elvis could proceed as shown:
[elvis@station public_html [elvis@station [elvis@station [elvis@station elvis]$ ls elvis]$ rmdir public_html/images elvis]$ ls elvis]$
As we have observed before, note again that commands such as mkdir and rmdir succeed silently. Silence is golden, as they say.
Listing Directory Trees

By now, you should be aware that the ls command, when given a directory as an argument, lists the contents of the directory. But what if the directory contains other directories? By default, the ls command will only
34
show the contents of the top level directory. If you would like the ls command to list subdirectories as well, you can add the -R command line switch. Whenever a command iterates through a structure, operating on each element of the structure in the same manner as it operated on the topmost element, the command is said to act recursively. The -R command line switch specifies that the ls command should list with recursion. For example, elvis would like a more comprehensive list of files and directories in his ~/work directory. Remember that previously elvis had to perform a separate ls command for each subdirectory of ~/work to see the subdirectory's contents. Now he can use ls -R instead.
[elvis@station elvis]$ ls -R work work/: images spreadsheets work/images: advertising logos work/images/advertising: work/images/logos: work/spreadsheets:
Note that the above command will display files as well as directories. In this example there just happen to be no files to display. The ls -R can be combined with any of the previously learned command line switches, as well, such as ls -sR to include file sizes, or ls -lR to include file attributes.
Copying directory trees: cp -r

In the previous chapter when cp was introduced, we noted that, with the correct options, the cp command could be used for copying entire directory trees. Here we discuss one such option: -r (for recursive). (Many commands have an option for acting recursively, that is, traversing all of the branches of a directory subtree, acting on each subdirectory in turn.) Consider the following directory structure in Hogan's home directory:
Figure 4.1. Directory structure before copying

-- html | hogan -- website -- -- forms | | -| | | -- songs -| | | | | -| | -- media -| -| | -- photos-| --
song1.midi
song2.midi
picture1.png
picture2.png
Even though this directory tree contains many branches and files, it can be copied with a single instruction:
35
[hogan@station hogan]$ cp -r website archive
Figure 4.2. Directory structure after copying

-- html | -- website -- -- forms | | -| | | -- songs -| | | | | -| | -- media -| -| | hogan --- photos-| | | -| | -- html | | -- archive -- -- forms | | -| | | -- songs -| | | | | -| | -- media -| -| | -- photos-| -| | | | | | | | | |
song1.midi
song2.midi
picture1.png
picture2.png
song1.midi
song2.midi
picture1.png
picture2.png
Remember the difference in the behavior of cp when the last argument is an existing directory: a copy of the source file(s) is made inside of the target directory. In the previous demonstration, the archive directory did not already exist and so was created. But if the archive directory had already existed, the effect would have been different.
Figure 4.3. Directory structure before copying

-- html | -- website -- -- forms | | | | -| | | | | -- songs -| | | | | | | -hogan -| | | -- media -| | -| | | | -- photos-| | | --- archive
song1.midi
song2.midi
picture1.png
picture2.png
Hogan uses exactly the same command as before to perform the copy:
[hogan@station hogan]$ cp -r website archive
36
Figure 4.4. Directory structure after copying with cp -r website archive

-- html | -- website -- -- forms | | -- song1.midi | | | -- songs -| | | | | -- song2.midi | | -- media -| -- picture1.png | | hogan --- photos-| | | -- picture2.png | | -- html | | -- archive -- website -- -- forms | | -- song1.midi | | | -- songs -| | | | | -- song2.midi | | -- media -| -- picture1.png | | -- photos-| -- picture2.png | | | | | | | | | |
In this second demonstration, the website directory is re-created inside the archive directory.
Removing directory trees: rm -r

In the previous chapter when rm was introduced, we noted that, with the correct options, the rm command could remove entire directory trees. Again, as for cp the option is -r (for recursive). Consider the following directory structure in Hogan's home directory:
Figure 4.5. Directory structure before removing

-- html | hogan -- website -- -- forms | | -| | | -- songs -| | | | | -| | -- media -| -| | -- photos-| --
song1.midi
song2.midi
picture1.png
picture2.png
If Hogan wished to remove the media subtree, he could use a single command:
37
[hogan@station hogan]$ rm -r website/media
Figure 4.6. Directory structure after removing

-- html | hogan -- website -| -- forms
It's worth observing one more time that, in general, removing files is non-reversible. Used carelessly, rm -r is a very dangerous command.
Examining Directories with tree

The tree command will recursively list the contents of the directory, much like the ls -R command. The output, however, is displayed more intuitively as an ASCII tree.
[elvis@station ~]$ tree /etc/NetworkManager/ /etc/NetworkManager/ ### dispatcher.d # ### 00-netreport # ### 05-netfs # ### 10-dhclient ### NetworkManager.conf ### system-connections ### VPN 3 directories, 4 files
Similarly, tree -d just shows directories.

[elvis@station ~]$ tree -d /etc/sysconfig/ /etc/sysconfig/ ### cbq ### console ### modules ### networking # ### devices # ### profiles # ### default ### network-scripts ### rhn ### allowed-actions # ### configfiles # ### script ### clientCaps.d 13 directories
The output of tree --help should entice you to explore more features of tree.
Examples
Making a Backup of a Directory Tree
The user hogan has been working hard on a report, which several files which he has organized into subdirectories under a directory called report.
report/ |-- chapter01/
38
| | | | | `--
|-- figures/ | |-- image01.png | `-- image02.png |-- section01.html `-- section02.html chapter02/ |-- figures/ | |-- image01.png | `-- image02.png |-- section01.html `-- section02.html
4 directories, 8 files
He is about to use an automated spell checker to check and replace misspelled words. Because this is the first time he has used the spell checker, he wants to make backup of his work before continuing. He first tries to make a backup with the cp command.
[hogan@station hogan]$ cp report report.bak cp: omitting directory `report'
Realizing his mistake, he adds the -r command line switch, to tell the cp command to copy recursively.
[hogan@station hogan]$ cp -r report report.bak
Now, the contents of the directory report are recursively copied to the directory report.bak.
. |-| | | | | | | | | | | | `-report/ |-- chapter01/ | |-- figures/ | | |-- image01.png | | `-- image02.png | |-- section01.html | `-- section02.html `-- chapter02/ |-- figures/ | |-- image01.png | `-- image02.png |-- section01.html `-- section02.html report.bak/ |-- chapter01/ | |-- figures/ | | |-- image01.png | | `-- image02.png | |-- section01.html | `-- section02.html `-- chapter02/ |-- figures/ | |-- image01.png | `-- image02.png |-- section01.html `-- section02.html
10 directories, 16 files
Once hogan has run the spell checker, and is happy with the results, he wants to remove the backup directory report.bak. He first uses the rmdir command.
[hogan@station hogan]$ rmdir report.bak/ rmdir: `report.bak/': Directory not empty
Remembering that rmdir only works on empty directories, he uses rm -r instead.

[hogan@station hogan]$ rm -r report.bak/
39
[hogan@station hogan]$ ls report
Creating a Local Copy of Example Files

The user ventura has taken an interest in programming with the python programming language. He has discovered sample scripts included in the /usr/share/doc/pygtk2-2.16.0/examples/ directory. He would like to be able to fiddle with the scripts, in order to try out new ideas, but doesn't have writable permissions to the files. He decides to make a local copy of the scripts.
[ventura@station ventura]$ cp -r /usr/share/doc/pygtk2-2.16.0/examples/ . [ventura@station ventura]$ ls -R examples/ examples/: atk gobject ide Makefile.am pango simple glade gtk Makefile Makefile.in pygtk-demo examples/atk: atk-demo.py examples/glade: autoconnect.py glade-demo.py examples/gobject: editable-interface.py examples/gtk: application.py bin.py ...
README
test2.glade
test.glade
properties.py
signal.py
customtreemodel.py filechooser.py
scrollable.py sizegroup.py
testprint.py uimanager.py
widget.py
Note the use of the directory name . to mean "the local directory". When cp receives a directory as its destination (last argument), it copies into that directory, preserving original filenames. In this case, the directory examples was copied into the directory .. The ls -R command generates a recursive listing of every file in the newly created examples directory. Now ventura has a local copy of the python scripts, which he can modify at will.
Online Exercises
Lab Exercise
Objective: List, copy, move, and remove directories efficiently. Estimated Time: 20 mins.
Specification
You have become interested in gedit, a simple editor, and want to explore its design. 1. Use the ls command (with the appropriate command line switches) to obtain a recursive listing, including sizes (-s), of the contents of the /usr/share/gedit-2 directory. Redirect the command's output to a file in your home directory called lsgedit.txt. (For grading purposes, it is important that you specify the directory as an absolute reference, i.e., not share/gedit-2).
40
2. Make a (recursive) copy of the directory /usr/share/gedit-2 in your home directory. The copied directory should also be called gedit-2. 3. Make a second copy of the newly created directory gedit-2, called gedit-2.bak, again in your home directory. 4. From your backup copy, remove (recursively) the gedit-2.bak/plugins subdirectory. 5. Again from your backup copy, move the subdirectory gedit-2.bak/icons to your home directory. (The directory should still be named icons).
Deliverables
1. 1. A file in your home directory called lsgedit.txt which contains the redirected output of a recursive listing, including sizes, of the /usr/share/gedit-2 directory. 2. A directory in your home directory called gedit-2, which is an exact copy of the directory /usr/share/gedit-2. 3. A directory in your home directory called gedit-2.bak, which starts as an exact copy of the directory gedit-2, with the subdirectory gedit-2/plugins removed, and the subdirectory gedit-2.bak/icons moved to your home directory. 4. A directory in your home directory called icons, which is a result of the previous deliverable.
Cleaning Up
The directories you have created are fairly significant in size, so you may want to conserve disk space by removing them when you are done.
Questions
1. a. b. c. d. e. Which command will list the contents of the directory report recursively? lsdir report ls --recur report ls -r report ls -R report ls report \r
The user hogan is executing the following command.

[hogan@station hogan]$ cp rep1.txt rep2.txt rep3.txt reps
2. a.
For the command to succeed, which of the following statements must be true? The file rep1.txt must be a text file. The file rep1.txt must be an existing directory. The file reps must be an existing directory.
b. c.
41
d. e.
The files rep1.txt, rep2.txt, and rep3.txt must be existing directories. The command cannot succeed, because the -r command line switch must be specified when using multiple arguments.
3. a.
Which command creates a new, empty directory? md nd mkdir newdir touch -d
b. c. d. e.
In his home directory, hogan has a directory called website, which contains various files and subdirectories, and a currently empty directory called archive, as illustrated in the following graph.
. |-- archive/ `-- website/ |-- forms/ |-- html/ | |-- index.html | |-- page1.html | |-- page2.html | `-- page3.html `-- media/ |-- photos/ | |-- picture1.png | `-- picture2.png `-- songs/ |-- song1.midi `-- song2.midi
For the following questions, assume that the current working directory is hogan's home directory, /home/ hogan. 4. a. What would be the effect of the following command? cp -r website archive The directory archive would be clobbered, and replaced with a copy of the website directory. The directory archive would contain a new subdirectory called website. An error, because the cp command does not work on directories. An error, because -r is not a legitimate switch for the cp command. None of the above What would be the effect of the following command? cp website archive/website.bak a. b. c. The directory archive would contain a new subdirectory called website.bak. An error, because the directory archive/website.bak does not exist. An error, because the -r command line switch must be included when copying directories.
b. c. d. e. 5.
42
d. e. 6. a. b. c. d. e. 7.
An error, because the correct command to use is cpdir. None of the above
What would be the effect of the following command? mv -r website archive The directory archive would be clobbered, and replaced with the website directory. The directory archive would contain a new subdirectory called website. An error, because the mv command does not work on directories. An error, because the mv command does not recognize the -r command line switch. None of the above Which of the following commands would move the website/media directory and its subdirectories to archive/media a. b. c. d. e. mv -r website/media archive/media mv website/media archive/media mv website/media archive Both B and C All of the above Which of the following commands would remove the website/media directory and its subdirectories? a. b. c. d. e. rmdir website/media rm website/media rm -r website/media Both A and C All of the above Which of the following commands would copy the website directory, and all of its subdirectories except website/media/photos? a. b. c. d. e. cp -x photos website archive cp -d2 website archive cp -x photos website archive/website cp -d2 website archive/website None of the above Why would someone be tempted to use rmdir instead of rm -r? a. To ensure that overlooked files in the directory are not unintentionally erased.
8.
9.
10.
43
b.
Because the rmdir command works recursively, without needing to specify a command line switch. The rmdir command creates an automatic backup of everything it removes. Both B and C
c. d.
44
Chapter 5. File Names and File Globbing

Key Concepts
Filenames can contain almost any character except /. Although filenames can contain almost any character, that doesn't mean they should. Files that start . are "hidden" files. The *, ?, [...], and [^...] characters can be used to match filenames, through a process commonly called "file globbing".
Discussion
Filenames
Many operating systems restrict the choice and number of characters that can be used in naming files. In Linux, virtually any printable character can be used in a filename, and filenames can be of almost any length. Linux filenames can be up to 255 characters long, excluding any directory components. When used in a command, an absolute or relative filename, including directory components, may be up to 4095 characters long! This allows names of files and directories to be very descriptive. Linux filenames can contain any printable character (and some unprintable ones!) except for / (slash). Slash can't be part of a filename since it's the character used to separate directory name components in a relative or fully-qualified name. Since many of the more "unusual" characters are shell metacharacters, they must be protected in quotes to be used in a filename:
[julius@station julius]$ touch [julius@station julius]$ touch [julius@station julius]$ touch [julius@station julius]$ touch [julius@station julius]$ ls !@#$%^&*()_+=-|\}]{[:;?>.<,~` [julius@station julius]$ 'a and b' '"' "'" '!@#$%^&*()_+=-|\}]{[:;?>.<,~`' ' " a and b
As you can see, four files were created with rather strange names.
Warning
The preceding example is for illustrative purposes only. Many of the characters used in the example are shell meta-characters whose significance we have not discussed. You should avoid using punctuation in filenames since the uninformed use of shell meta-characters can produce disastrous results. The warning given above should not be taken lightly. In general, filenames should be composed of alphabetic and numeric characters (A-Z, a-z, 0-9) and the punctuation symbols ._-+~ (dot, underscore, dash, plus, tilde). Normally, filenames should start with an alphanumeric character or a dot (see the section
45
File Names and File Globbing
below on Hidden Files). While you will encounter them with some frequency, filenames with embedded spaces are discouraged.
Figure 5.1. "Safe" Filename Characters

A-Z a-z 0-9 . _ - + ~
Here are two examples to show what the ill-advised use of otherwise-meaningful characters can do. We'll abuse two shell meta-characters: the space and greater-than characters. Example 1: Watch those spaces!
[hogan@station [hogan@station [hogan@station [hogan@station dir hogan]$ hogan]$ hogan]$ hogan]$ mkdir bad dir touch this and that mv this and that bad dir ls
Hogan attempts to create a directory called 'bad dir' but unwittingly creates two directories, one named 'bad' and the other 'dir'. Next, Hogan tries to create a file named 'this and that' with touch, but creates three files instead, named 'this', 'and', and 'that'. Now Hogan attempts to move the file he thinks is called 'this and that' to the directory he thinks is called 'bad dir' but instead moves all four of 'this', 'and', 'that', and 'bad' (three ordinary files and one directory) to the directory dir. Because all of the above commands "succeeded," only now does Hogan realize something has gone wrong. Of course, if Hogan had used quotes around the names as was done in the explanatory notes, things would have worked as he intended. This is worth remembering since you will encounter filenames with embedded spaces. On a related note, a filename with leading or trailing spaces is legal, but very tricky to identify in a file listing. Example 2: A little misdirection.
[hogan@station [hogan@station [hogan@station [hogan@station bad- dir hogan]$ hogan]$ hogan]$ hogan]$ mkdir bad->dir touch this mv this bad->dir ls
Hogan agrees to avoid spaces. This time he attempts to create a directory called 'bad->dir' but unwittingly creates a directory named 'bad-' and a file named 'dir' (why?). Next, Hogan creates a file named 'this' with touch. This step he gets right! Now Hogan attempts to move the file called 'this' to the directory he thinks is called 'bad->dir' but instead moves 'this' to the directory bad-, redirecting the output (there is none) to the file 'dir'. Again only now does Hogan realize something has gone wrong. As before, if Hogan had used quotes around the names things would have worked as he intended. Quotes mask a multitude of sins, but don't use this as an excuse to create poor filenames. Many of the tasks we have yet to learn are made easier if the files involved have "normal" names.
Hidden Files
File and directory names (remember, a directory is a type of file) that start with a dot are "hidden" files. Hidden files do not appear in directory listings produced by ls unless the special command option -a (all)
46
is used, or unless you specify the leading dot as part of the name. This makes it possible to reduce clutter and confusion by keeping certain files "out of sight, out of mind." In the next section we will learn how to name multiple files through a technique known as "globbing." In general, hidden files also do not appear in a "glob" unless the leading dot was specifically listed. Except for hiding the file, the leading dot has no other significance, and hidden files and directories can be used just like any others.
[alice@station [alice@station [alice@station sample [alice@station . .. sample [alice@station [alice@station [alice@station sample [alice@station [alice@station sample [alice@station alice]$ mkdir .secret alice]$ touch sample alice]$ ls alice]$ .secret alice]$ alice]$ alice]$ ls -a mv sample .secret ls ls .secret
alice]$ cd .secret .secret]$ ls .secret]$
"Globbing"
Frequently it is necessary to issue a command that acts on more than one file. Commands like cp -r and rm -r work on entire directory sub-trees, but Linux has a more flexible way of identifying sets of files. The bash command shell treats some of its special meta-characters as wildcard characters. When bash reads the command line, the shell performs a process called meta-character expansion or wildcard expansion, generates a list of filenames that match the pattern described by the wildcard expression, and then passes the generated list on to the command. This is commonly referred to as "filename globbing."
Table 5.1. Wildcard Characters

Character * ? [...] [^...] Effect matches zero or more characters (except leading dot) matches exactly one character (except leading dot) matches exactly one character from the list or range matches exactly one character not from the list or range
The use of * and ? is fairly straightforward. The list wildcards are a bit more complicated, but also quite powerful and useful. The bracket wildcards represent a list of single characters. Thus, [aeiou] matches any one vowel. A contiguous range of characters may be represented using a dash, as in [a-z] for the lowercase alphabet. A leading carat (^) negates the list, so that [âeiou] is any character except a vowel. A "real" dash or carat can be represented by escaping the characters using a backslash (\) (more on this later). Thus, the expression [a\-z] matches a, z or a dash only. Ranges and single characters can be mixed together. The pattern [A-Za-z0-9._\-+~] matches any one of the "safe" filename characters mentioned above in Figure 5-1. Suppose the current directory contains the following files:
image1.jpg image2.jpg image10.jpeg image11.jpeg page1.html page2.html page3.htm page40.htm script1.pl
47
The following table illustrates the use of these wildcard characters.
Table 5.2. Wildcard Examples

Pattern * *.html page*.htm* image?.* [ps]* [^ps]* Generates the list all the files in the list page1.html page2.html page1.html page2.html page3.htm page40.htm image1.jpeg image2.jpeg page1.html page2.html page3.htm page40.htm script1.pl image1.jpeg image2.jpeg image10.jpeg image11.jpeg
Wildcards can be used with any command that expects filenames. For example, the command
[bob@station bob] mv *.htm public_html
is the same as the command

[bob@station bob] mv page3.htm page40.htm public_html
assuming Bob's home directory contains the files listed above. [Technical note: Many commands behave differently when there are no files listed as opposed to when one or more files are given as part of the command. So what happens in such a command if a wildcard pattern is used, and the pattern does not match anything so that the list of files is empty? In the special case where the bash shell attempts to expand a wildcard expression and there is no match, it leaves the original expression as part of the command. This is illustrated in the following:
[bob@station bob]$ rm rm: too few arguments Try `rm --help' for more information. [bob@station bob]$ rm z* rm: cannot lstat `z*': No such file or directory
In the first rm, the command saw zero filenames and failed with an error message. In the second version, the shell could not generate a list of names from the pattern z*, and so the string z* was sent to rm as the filename itself. Since Bob doesn't have a file named z* in his directory, the command failed with a different message.]
Examples
Finding That Configuration File
Alice was recently examining the DNS (Domain Name Service) configuration file, but has since forgotten its name. She remembers that the file ended .conf. In order to reduce the number of files she has to consider, she used file globbing to list all files in the /etc directory that end .conf.
[alice@station images]$ ls /etc/*.conf /etc/aep.conf /etc/lftp.conf /etc/aeplog.conf /etc/libuser.conf /etc/cdrecord.conf /etc/logrotate.conf /etc/esd.conf /etc/lpd.conf /etc/fam.conf /etc/ltrace.conf /etc/gpm-root.conf /etc/modprobe.conf /etc/pnm2ppa.conf /etc/pwdb.conf /etc/resolv.conf /etc/rndc.conf /etc/scrollkeeper.conf /etc/sysctl.conf
48
/etc/grub.conf /etc/host.conf /etc/initlog.conf /etc/jwhois.conf /etc/krb5.conf /etc/krb.conf /etc/ldap.conf /etc/ld.so.conf
/etc/mtools.conf /etc/named.conf /etc/nscd.conf /etc/nsswitch.conf /etc/ntp.conf /etc/pam_smb.conf /etc/pbm2ppa.conf /etc/pine.conf
/etc/syslog.conf /etc/updatedb.conf /etc/updfstab.conf /etc/warnquota.conf /etc/webalizer.conf /etc/wvdial.conf /etc/xinetd.conf /etc/yp.conf
With a little more looking, she recognizes the resolv.conf configuration file.
Listing Static Libraries

Bob has been learning the difference between statically and dynamically linked libraries, that are used by programs. He knows that statically linked libraries conventionally have the name libsomething.a, where something is the name of the library. He also knows that libraries are generally found in the / usr/lib directory. When he goes to see what static libraries are installed on his system, he finds nearly 1000 files in the / usr/lib directory. In order to pair down the number of files to examine, he tries listing all files that end in an a.
[bob@station bob]$ ls /usr/lib/*a /usr/lib/liba2ps.a /usr/lib/libimlib-ps.a /usr/lib/libacl.a /usr/lib/libimlib-tiff.a /usr/lib/libanl.a /usr/lib/libimlib-xpm.a ... /usr/lib/mozilla: plugins /usr/lib/samba: vfs
Many lines in this long listing have been deleted, but enough have been left to illustrate the problem, bob's "glob" was not specific enough, and caught some directories that ended a as well. With a little more refinement, he produces the list he wants.
[bob@station bob]$ ls /usr/lib/lib*.a /usr/lib/liba2ps.a /usr/lib/libimlib-ps.a /usr/lib/libacl.a /usr/lib/libimlib-tiff.a /usr/lib/libanl.a /usr/lib/libimlib-xpm.a ... /usr/lib/libimlib-jpeg.a /usr/lib/libimlib-png.a /usr/lib/libimlib-ppm.a /usr/lib/libz.a /usr/lib/libzvt.a
Next, bob suspects that the most commonly used libraries would have short, efficient (if not descriptive) names. He proceeds to list all of the libraries that have 1, 2, and 3 letter names.
[bob@station bob]$ ls /usr/lib/lib?.a /usr/lib/libc.a /usr/lib/libl.a /usr/lib/libz.a /usr/lib/libg.a /usr/lib/libm.a [bob@station bob]$ ls /usr/lib/lib??.a /usr/lib/libdl.a /usr/lib/libgd.a /usr/lib/librt.a /usr/lib/libfl.a /usr/lib/libmp.a [bob@station bob]$ ls /usr/lib/lib???.a /usr/lib/libacl.a /usr/lib/libgdk.a /usr/lib/libosp.a /usr/lib/libanl.a /usr/lib/libgif.a /usr/lib/libpam.a /usr/lib/libapm.a /usr/lib/libgmp.a /usr/lib/libpbm.a
/usr/lib/librle.a /usr/lib/librpm.a /usr/lib/libSDL.a
49
/usr/lib/libbfd.a /usr/lib/libbsd.a /usr/lib/libefs.a /usr/lib/libesd.a /usr/lib/libgal.a
/usr/lib/libgpm.a /usr/lib/libgtk.a /usr/lib/libIDL.a /usr/lib/libnsl.a /usr/lib/libogg.a
/usr/lib/libpci.a /usr/lib/libpgm.a /usr/lib/libpng.a /usr/lib/libpnm.a /usr/lib/libppm.a
/usr/lib/libssl.a /usr/lib/libttf.a /usr/lib/libusb.a /usr/lib/libxml.a /usr/lib/libzvt.a
Is there any easy way, using file globbing, that bob could have listed libraries with 1, 2, or 3 character names with one expression? No.
Listing Man Pages

The user alice has discovered that the files that contain the man pages are installed into the directory / usr/share/man, in subdirectories that relate to the man page chapters.
[alice@station man]$ ls /usr/share/man/ cs de_DE fr hu it man1 man3 man5 de es fr_FR id ja man2 man4 man6 man7 man8 man9 mann pl pt_BR ru sk sv
As a standard user of the system, she knows that she is most interested in chapters 1, 5, and 7. She uses the file globbing match expression to just list the contents of those directories.
[alice@station man]$ ls /usr/share/man/man[157] /usr/share/man/man1: :.1.gz pbmtolj.1.gz [.1.gz pbmtoln03.1.gz 411toppm.1.gz pbmtolps.1.gz a2p.1.gz pbmtomacp.1.gz ... groff.7.gz groff_char.7.gz groff_diff.7.gz wireless.7.gz x25.7.gz
If alice had been interested in all of the man pages in numbered chapters, but not the oddly named chapter "n", she could have used the file globbing range expression instead.
[alice@station man]$ ls /usr/share/man/man[1-9] /usr/share/man/man1: :.1.gz pbmtolj.1.gz [.1.gz pbmtoln03.1.gz 411toppm.1.gz pbmtolps.1.gz a2p.1.gz pbmtomacp.1.gz ... named-checkconf.8.gz named-checkzone.8.gz /usr/share/man/man9: zic.8.gz
Online Exercises
Managing Files with File Globbing
Lab Exercise
Objective: Manage files efficiently with file globbing. Estimated Time: 20 mins.
50
Specification
This lab will use file globbing to manage a large number of files that have loosely structured names. All of the files are found in the /usr/share/tcl8.5/encoding directory. To make your life easier, try to perform each of the following steps with a single command. 1. Create the following directories in your home directory: cp_even, cp_mid, cp_thousand, iso_mid, mac, and xxxn. 2. Examine the files in the /usr/share/tcl8.5/encoding directory. Copy all files that begin mac into the mac directory. 3. Copy all files which start cp and contain an even number into the directory cp_even. 4. Copy all files which start cp and contain a number greater than 1000 into the cp_thousand directory. 5. Copy all files which start cp and contain numbers greater than 800 but less than 1000 into the cp_mid directory. 6. Copy all files which start iso8859- with a number following the hyphen between 3 and 8 (inclusive) into the iso_mid directory. 7. Copy all files whose first four characters are composed of exactly three letters followed by a number into the directory xxxn.
Deliverables
1. If the exercise has been performed correctly, you should have the following six directories, with the following files.
[bob@station bob]$ ls * cp_even: cp1250.enc cp1256.enc cp852.enc cp1252.enc cp1258.enc cp860.enc cp1254.enc cp850.enc cp862.enc cp_mid: cp850.enc cp852.enc cp855.enc
cp864.enc cp866.enc cp874.enc
cp949.enc cp950.enc
cp_thousand: cp1250.enc cp1252.enc cp1251.enc cp1253.enc iso_mid: iso8859-3.enc iso8859-4.enc mac: macCentEuro.enc macCroatian.enc macCyrillic.enc xxxn: big5.enc iso2022.enc iso2022-jp.enc iso2022-kr.enc iso8859-10.enc
cp1254.enc cp1255.enc
cp1256.enc cp1257.enc
cp1258.enc
iso8859-5.enc iso8859-6.enc
iso8859-7.enc iso8859-8.enc
macDingbats.enc macGreek.enc macIceland.enc
macJapan.enc macRoman.enc macRomania.enc
macThai.enc macTurkish.enc macUkraine.enc
iso8859-13.enc iso8859-14.enc iso8859-15.enc iso8859-16.enc iso8859-1.enc
iso8859-2.enc iso8859-3.enc iso8859-4.enc iso8859-5.enc iso8859-6.enc
iso8859-7.enc iso8859-8.enc iso8859-9.enc jis0201.enc jis0208.enc
jis0212.enc koi8-r.enc koi8-u.enc ksc5601.enc
51
Suggestions
The following command would copy the files that meet the first criteria (assuming the shell's current working directory is bob's home directory).
[bob@station bob]$ cp /usr/share/tcl8.5/encoding/mac* mac/
Questions
Note
The following questions all end in an appropriate question mark, which should not be considered part of any specified commands. Use the output from the following command to answer the next 5 questions.
[elvis@station files]$ ls ebbs eras lakes lit echo erect lamed lives elope ergo lash loads enact erupt lead loath end evens leaks lobby enjoy evoke leaps locus entry ewes learn logic envoy lace leech login epic laced lick loin loop lorry lost loves lows loyal lucks oaks obese olden one ounce oust ovary overt racks radon raged rank rapid rasp rays razor read reaps recta recur renew reply retry rho rigid rigs rime riots risks robe robed rock rocky rooms roses rouge rove row run wagon wares ways weak weep wench west wet whirr whorl wile windy witty wools wound wrens writ
1. a.
Which of the following files would be included in the output of the command ls r*s ? eras raged rigs retry racks wrens Which of the following files would be included in the output of the command ls r?? ? a. b. c. d. e. f. eras row riots rho wet run Which of the following files would be included in the output of the command ls [aeiou]??? ?
b. c. d. e. f. 2.
3.
52
a. b. c. d. e. f. 4. a. b. c. d. e. f. 5. a. b. c. d. e. f.
ebbs lead oft whip oaks oust Which of the following files would be included in the output of the command ls *[âeiou] ? ebbs rove one rasp ways era Which of the following files would be included in the output of the command ls [aeiou]?e* ? ebbs rove one rasp ways rouge
Use the following recursive listing of the /etc/X11/gdm directory to answer the following questions.
[student@station student]$ ls -R /etc/X11/gdm/ /etc/X11/gdm/: factory-gdm.conf gnomerc locale.alias PreSession gdm.conf Init PostSession Sessions /etc/X11/gdm/Init: Default /etc/X11/gdm/PostSession: Default /etc/X11/gdm/PreSession: Default /etc/X11/gdm/Sessions: default Default Failsafe
XKeepsCrashing
GNOME
KDE
6.
Which of the following files would be included in the output of the command ls /etc/X11/ gdm/S*/*a* ?
53
a. b. c. d. e. f. 7.
/etc/X11/gdm/Init/Default /etc/X11/gdm/locale.alias /etc/X11/gdm/factory-gdm.conf /etc/X11/gdm/Sessions/Default /etc/X11/gdm/Sessions/Failsafe /etc/X11/gdm/Sessions/default Which of the following files would be included in the output of the command ls /etc/X11/ gdm/*/*[Ef] ?
a. b. c. d. e. f. 8. a. b. c. d. e. f. 9.
/etc/X11/gdm/gdm.conf /etc/X11/gdm/factory-gdm.conf /etc/X11/gdm/Init/Default /etc/X11/gdm/Sessions/GNOME /etc/X11/gdm/Sessions/KDE /etc/X11/gdm/Sessions/Failsafe Which of the following files could be included in the output of the command ls /etc/*.?? ? /etc/issue.net /etc/mail.rc /etc/php.ini /etc/huh.?? /etc/xpdfrc.ja /etc/gss-1.0 Which of the following files could be included in the output of the command ls /etc/[akp]?? [0-9]* ?
a. b. c. d. e. f. 10.
/etc/krb5.conf /etc/as5-0 /etc/pbm2ppa.conf /etc/a2ps.cfg /etc/php.ini /etc/krc8 Which of the following files would be included in the output of the command ls /etc/[aeiou] [âeiou]??[aeiou]* ?
54
a. b. c. d. e. f.
/etc/adjtime /etc/issue /etc/aeplog.conf /etc/at.deny /etc/at.allow /etc/auto.tab
55
Chapter 6. Examining Files

Key Concepts
The file command displays a file's type The cat command displays a file's contents The more and less commands "page" through a file, one screen full at a time. The head and tail commands display the first or last few lines of a file. In the X graphical environment, nautilus can be used to examine text files.
Discussion
Viewing Files
Linux offers several commands for examining files and file contents. This Lesson looks at some of the most commonly used of these tools.
The file Command

The contents of any given file might be ASCII (plain text, HTML, shell script, program source code, etc.) or binary (compiled executable, compressed archive, audio, etc.). The commands discussed in this chapter are intended for use on ASCII files only, and attempting to use them with binary files can lead to problems ranging from the mildly irritating (screens full of strange character flashing by) to the more significant (locking the terminal display). This is because binary files can contain arbitrary binary codes, and some of these codes have special meaning when being (mis)interpreted as ASCII text. Therefore, it is a good idea to check the file type of any file you do not recognize using file before attempting to view the file with one of the other tools shown here. file [OPTIONS] [FILE...] Tests FILE(s) to determine the file type, and displays results to standard output. Example of use with ASCII files:
[nero@station nero]$ file index.html novel.ps mailer index.html: HTML document text novel.ps: PostScript document text conforming at level 3.0 mailer: perl script text executable
Example of use with binary files:

[nero@station nero]$ file logo.jpeg symphony.mp3 archive.tgz logo.png: PNG image data, 387 x 69, 8-bit grayscale, non-interlaced symphony.mp3: MP3, 128 bits, 44.1 kHz, JStereo archive.tgz: gzip compressed data, deflated, last modified: Thu Mar 20 21:27:27 2003, os: Unix
The cat Command

This command was first introduced in Chapter 3 of this workbook. Here we show some of its options. cat [OPTIONS] [FILE...]
56
Examining Files
Concatenate FILE(s) to standard output. Switch -A -s -n Effect Show all characters, including control characters and non-printing characters " Squeeze" multiple adjacent blank lines into a single blank line Number lines of output
The cat command, when used for viewing files, simply displays the contents all at once. Large files scroll by too quickly to read, making cat most suitable for files with less than a screen of text.
The more and less Pagers

Both more and less are designed for viewing text files on screen. Their use is quite similar, except that the more modern less command has a few extra features, such as responding correctly to [PgUp], [PgDn] and navigation arrow keystrokes correctly. After all, less is more. It is especially important to be familiar with less because other tools (such as man) use it behind the scenes to provide paging capabilities. more [OPTIONS] [FILE...] Displays FILE(s) to standard output one screen full at a time under keyboard control. Option -c -s Action Clear screen and redraw, instead of scrolling "Squeeze" multiple adjacent blank lines into a single blank line
less [OPTIONS] [FILE...] Displays FILE(s) to standard output one screen full at a time under keyboard control. Option -c -r -s Action Clear screen and redraw, instead of scrolling Display raw control characters "Squeeze" multiple adjacent blank lines into a single blank line
Table 6.1. less (more) navigation commands

Command <space> [PgDn] b [PgUp] [DnArrow] [UpArrow] /text ?text more? Yes No Yes No No No Yes No Action move ahead one full screen move ahead one full screen move back one full screen move back one full screen move ahead one line move back one line search forward for text search backward for text
57
Examining Files
Command n N q h
more? Yes No Yes Yes
Action repeat last search repeat last search, but in opposite direction quit help (which introduces many new commands)
One feature of the less pager relies on a standard Unix concept called pipes. While pipes are officially discussed later, this feature of less is too useful and simple enough to use that it merits introducing now. A pipe acts like redirection, in that the output of a command is redirected somewhere other than a terminal. With redirection (using >), the output gets redirected to a specified file. When piping, the output of one command gets redirected into the input of another command. The bash shell uses the | character (usually found above the RETURN key) to construct pipes. When running a command that produces a lot of output, the output of the command can be piped into less. Rather than getting dumped to a terminal, the output can be browsed like a file, paging down, paging up, and searching. When finished, quit less, and the output disappears. The user elvis is using the ps aux command to view a table of all processes running on the machine. Because he expects the output of the command to be lengthy, he chooses to pipe the command into the less browser.
[elvis@station elvis]$ ps aux USER PID %CPU %MEM VSZ RSS TTY root 1 0.0 0.0 1392 92 ? root 2 0.0 0.0 0 0 ? root 3 0.0 0.0 0 0 ? root 4 0.0 0.0 0 0 ? root 9 0.0 0.0 0 0 ? ... [elvis@station elvis]$ ps aux | less STAT S SW SW SWN SW START 17:27 17:27 17:27 17:27 17:27 TIME 0:04 0:00 0:00 0:00 0:00 COMMAND init [ [keventd] [kapmd] [ksoftirqd_CPU0] [bdflush]
Now he can use the same key sequences he uses when browsing a file (or man page) to browse the ps commands output: SPACE to page forward, b to page back, /textRETURN to search for the text text, and q to quit.
The head Command

Sometimes all that is needed in examining files is to see the first few lines of the file. The head command provides this capability. head [OPTIONS] [FILE...] Print the first 10 lines of each FILE to standard output. Switch -num, -n num -q Effect Print first num lines (default is 10). Suppress headers giving file names
In the following example, elvis will examine the first 10, and then the first 5, lines of the /etc/group file.
[elvis@station elvis]$ head /etc/group root:x:0:root bin:x:1:root,bin,daemon daemon:x:2:root,bin,daemon sys:x:3:root,bin,adm
58
Examining Files
adm:x:4:root,adm,daemon tty:x:5: disk:x:6:root lp:x:7:daemon,lp mem:x:8: kmem:x:9: [elvis@station elvis]$ head -n 5 /etc/group root:x:0:root bin:x:1:root,bin,daemon daemon:x:2:root,bin,daemon sys:x:3:root,bin,adm adm:x:4:root,adm,daemon
If given more than one filename as arguments, head displays the first few lines of each file individually, separated by a header displaying the file name. In the following, elvis lists the first few line of all of the Red Hat service scripts, found in the /etc/rc.d/init.d directory.
[elvis@station elvis]$ head -5 /etc/rc.d/init.d/* ==> /etc/rc.d/init.d/aep1000 <== #!/bin/sh # # Tags # chkconfig: - 30 70 # description: load and unload AEP1000/AEP2000 coprocessor driver ==> /etc/rc.d/init.d/anacron <== #!/bin/sh # Startup script for anacron # # chkconfig: 2345 95 05 # description: Run cron jobs that were left out due to downtime ==> /etc/rc.d/init.d/apmd <== #!/bin/sh # # chkconfig: 2345 26 74 # description: apmd is used for monitoring battery status and logging it via \ # syslog(8). It can also be used for shutting down the machine when \ ==> /etc/rc.d/init.d/atd <== #!/bin/bash # # /etc/rc.d/init.d/atd # # Starts the at daemon ...
As with the less pager, above, the output from long commands can be truncated by piping it into the head command.
[student@station student]$ ps aux | USER PID %CPU %MEM VSZ RSS root 1 0.0 0.1 2500 416 root 2 0.0 0.0 0 0 root 3 0.0 0.0 0 0 root 4 0.0 0.0 0 0 [student@station student]$ head -5 TTY ? ? ? ? STAT S SW SW SWN START 10:51 10:51 10:51 10:51 TIME 0:03 0:00 0:00 0:00 COMMAND init [5] [keventd] [kapmd] [ksoftirqd/0]
The tail Command

tail [OPTIONS] [FILE...] Complementing head, the tail command prints the last 10 lines of each FILE to standard output.
59
Examining Files
Switch -num, -n num -q -f
Effect Print last num lines (default is 10). Suppress headers giving file names Keep file open, and print new lines as appended
In the following, elvis prints the last 10, and then the last 5, lines of the /etc/group file.
[elvis@station elvis]$ tail /etc/group named:x:25: gdm:x:42: postgres:x:26: apache:x:48: postdrop:x:90: postfix:x:89: squid:x:23: webalizer:x:67: ldap:x:55: elvis:x:500: [elvis@station elvis]$ tail -n 5 /etc/group postfix:x:89: squid:x:23: webalizer:x:67: ldap:x:55: elvis:x:500:
The tail command has another, very useful, option: the -f (follow) option. With this option, tail will display the last lines of the file, and then "waits" and continues to display any new lines as they are added to the file. This is often used to monitor system log files in real time. Once started with this option, the command prompt does not return. Instead the tail -f stays active and will continue to display new lines as they occur until CTRL+C is pressed.
Examples
Identifying Documents with file
While exploring the cups printing system's configuration, prince came across the following files in the / usr/share/cups/banners directory.
[prince@station banners]$ ls /usr/share/cups/banners/ classified confidential secret standard topsecret unclassified
Unsure what types of files these were, he tries to identify them with the file command.
[prince@station banners]$ cd /usr/share/cups/banners/ [prince@station banners]$ file * classified: PostScript document text conforming at level confidential: PostScript document text conforming at level secret: PostScript document text conforming at level standard: PostScript document text conforming at level topsecret: PostScript document text conforming at level unclassified: PostScript document text conforming at level
3.0 3.0 3.0 3.0 3.0 3.0
As the file command reveals, these are PostScript documents, intended to be sent to the printer as banner pages.
Determining the Correct File Type with file

The user elvis has just downloaded some images from his friend's photo album website, and stored them in a directory called pics.
60
Examining Files
[elvis@station images]$ ls pics/ dscn0936.png dscn0942.png dscn0953.jpg dscn0938.png dscn0952.png dscn0959.jpg
He notices from the file extensions that 4 of the files seem to be in PNG format, while 2 of them seem to be JPEG format. Thinking this odd, he confirms the file types by running the file command.
[elvis@station images]$ file pics/dscn0936.png: PNG image pics/dscn0938.png: PNG image pics/dscn0942.png: PNG image pics/dscn0952.png: PNG image pics/dscn0953.jpg: PNG image pics/dscn0959.jpg: PNG image pics/* data, 360 data, 640 data, 360 data, 360 data, 640 data, 640 x x x x x x 480, 480, 480, 480, 480, 480, 8-bit/color 8-bit/color 8-bit/color 8-bit/color 8-bit/color 8-bit/color RGB, RGB, RGB, RGB, RGB, RGB, non-interlaced non-interlaced non-interlaced non-interlaced non-interlaced non-interlaced
The last two files are in the PNG format, like the other four. The last two files were apparently misnamed with the wrong extension.
Skimming Files with head

The user blondie has recently learned that all of the files in the /etc/profile.d directory that end .sh are included in her bash shell's configuration every time she logs on. Curious about what the files contain, she skims the first few lines of each file with the following single command.
[blondie@station blondie]$ head /etc/profile.d/*.sh ==> /etc/profile.d/glib2.sh <== ## This caused GLib2 applications to convert filenames from ## locale encoding to UTF-8. If the locale encoding is already ## UTF-8 then it makes no difference. export G_BROKEN_FILENAMES=1 ==> /etc/profile.d/gnome-ssh-askpass.sh <== SSH_ASKPASS=/usr/libexec/openssh/gnome-ssh-askpass export SSH_ASKPASS ==> /etc/profile.d/jdk.sh <== export PATH=/usr/local/j2sdk1.3.1/bin:$PATH ==> /etc/profile.d/lang.sh <== # /etc/profile.d/lang.sh - set i18n stuff sourced=0 for langfile in /etc/sysconfig/i18n $HOME/.i18n ; do [ -f $langfile ] && . $langfile && sourced=1 done if [ -n "$GDM_LANG" ]; then sourced=1 LANG="$GDM_LANG" ...
Combining Multiple Files into a Single File with head

The user madonna is learning more about PAM, Linux's Pluggable Authentication Modules. She has discovered the /usr/share/doc/pam-0.75/txts directory, which contains a large number of README files.
[madonna@station madonna]$ cd /usr/share/doc/pam-0.75/txts/ [madonna@station txts]$ ls pam_appl.txt README.pam_ftp README.pam_shells
61
Examining Files
pam_modules.txt pam.txt README README.pam_access README.pam_chroot README.pam_console README.pam_cracklib README.pam_deny README.pam_env README.pam_filter
README.pam_limits README.pam_listfile README.pam_localuser README.pam_mail README.pam_nologin README.pam_permit README.pam_pwdb README.pam_rhosts README.pam_rootok README.pam_securetty
README.pam_stack README.pam_stress README.pam_tally README.pam_time README.pam_timestamp README.pam_unix README.pam_userdb README.pam_warn README.pam_wheel README.pam_xauth
Rather than deal with each file individually, she would like to combine them into a single file which she can browse easily. She knows she could use the cat command, and redirect the output to a file, but the she would no longer know the filenames associated with each of the files. Instead, she decides to use the head command, giving a ludicrously large number of lines to select. That way, she will be able to browse the file contents, with an identifying banner giving the file name.
[madonna@station txts]$ head -99999 README* > /tmp/pam_READMEs.txt [madonna@station txts]$ less /tmp/pam_READMEs.txt ==> README <== $Id: 010_text.dbk,v 1.1.1.1 2003/08/11 13:18:31 bowe Exp $ This is a directory for text versions of the pam documentation ==> README.pam_access <== # Description of its configuration file # # (The default config file is "/etc/security/access.conf". This # default can be overridden with a module config argument ...
Monitoring Multiple Files with tail

In Red Hat Enterprise Linux, system events get logged into text files found in the /var/log directory, such as /var/log/messages (for general system events) and /var/log/secure (for events which entail sensitive information). As events occur, new lines get appended to these files in real time. Suppose a system administrator would like to monitor both of these log files, and see new information as soon as it arrives. She uses the tail -f command to "follow" the ends of both files. The tail command immediately displays the last 10 lines of both files, decorated by filename.
[root@station root]# cd /var/log [root@station log]# tail -f messages secure ==> Jun Jun Jun Jun Jun Jun Jun Jun Jun Jun messages <== 25 11:20:59 station 25 11:20:59 station 25 11:20:59 station 25 11:20:59 station 25 11:20:59 station 25 11:22:06 station 25 11:22:10 station 25 11:23:06 station 25 11:23:55 station 25 11:24:00 station
kernel: 11: @cf13e4c0 length 800000de status 800100de kernel: 12: @cf13e500 length 800000de status 000100de kernel: 13: @cf13e540 length 80000042 status 00010042 kernel: 14: @cf13e580 length 800005ea status 000105ea kernel: 15: @cf13e5c0 length 8000031e status 0001031e su(pam_unix)[2229]: session closed for user madonna su(pam_unix)[2352]: session opened for user root by elvis(uid=2291) su(pam_unix)[2399]: session opened for user root by elvis(uid=2291) su(pam_unix)[2399]: session closed for user root su(pam_unix)[2441]: session opened for user elvis by elvis(uid=2291)
==> secure <== Jun 24 15:32:15 station sshd[11535]: Accepted publickey for root from 127.0.0.1 port 36253 ssh2 Jun 24 15:32:15 station sshd[11545]: Accepted publickey for root from 127.0.0.1 port 36254 ssh2 Jun 24 15:32:16 station sshd[11555]: Accepted publickey for root from 127.0.0.1 port 36256 ssh2
62
Examining Files
Jun Jun Jun Jun Jun
24 24 24 25 25
17:00:22 17:29:05 17:30:49 10:06:28 10:09:01
station sshd[819]: Received signal 15; terminating. localhost sshd[792]: Server listening on 0.0.0.0 port 22. localhost xinetd[803]: START: sgi_fam pid=1322 from=<no address>Jun 24 19:24:47 localho station sshd[834]: Server listening on 0.0.0.0 port 22. station xinetd[845]: START: sgi_fam pid=1134 from=<no address>
At this point, the tail command is still running, but the output pauses until a new system event occurs which gets logged to one of the files. The tail command then shows the newly appended lines, again decorated with the filename of the relevant file.
==> messages <== Jun 25 11:24:54 station su(pam_unix)[2481]: session opened for user root by elvis (uid=501) Jun 25 11:25:02 station su(pam_unix)[2481]: session closed for user root
Finally, when the administrator is finished monitoring the files, she cancels the tail command with the CTRL+C control sequence.
Online Exercises
Browsing Text Files
Lab Exercise
Objective: Examine files using several browsing utilities. Estimated Time: 15 mins.
Specification
This Lab Exercise will require three terminals (either virtual consoles, or terminal windows within the X graphical environment), each using your primary account. 1. In the first terminal, cat the contents of the file /etc/services, including the correct switch to number all lines. After you have viewed the results, repeat the command (with the same switch), redirecting the output into the file services.cat.txt in your home directory. 2. While still in the first terminal, take a recursive listing of all files with the ls -R / command, and "catch" the output (using a pipe) with the less pager. Use the basic pager commands, such as SPACE and b, to browse the output. Explore other pager commands by using the h command to open a help screen. When you are finished, DO NOT exit the less pager, but move to a new terminal for the next step. 3. In a second terminal, Use the head command to collect the first 5 lines of all files in the /etc/ sysconfig directory whose filename starts system- into the file confheaders.txt in your home directory. Use the appropriate switch to suppress the filename headers that head prints by default. 4. While still in the second terminal, run the command vmstat 1. The command tracks memory and CPU statistics about your machine, generating a new line once per second. When you have seen enough, use the CTRL+C control sequence to kill the command. Run the command again, this time redirecting output to the file vmstat.out in your home directory, as in the following.
[elvis@station elvis]$ vmstat 1 > vmstat.out
63
Examining Files
DO NOT stop this command, but leave it running. Move to a new terminal for the next step. 5. In a third terminal, follow the growing vmstat.out file with the tail -f command. DO NOT stop the tail command until you have confirmed your exercise below.
Deliverables
1. 1. A file in your home directory called services.cat.txt, which contains the contents of the /etc/services file, with line numbers as produced by the cat command. 2. A running less pager which is using a pipe to browse the output of the ls -R / command. 3. The file confheaders.txt in your home directory, that contains first five lines of every file in the /etc/sysconfig that begins with redhat-, as produced by the head command, suppressing filenames (with the appropriate command line switch(es)). 4. A running tail command, which is following the growing file vmstat.out in your home directory.
Cleaning Up
After completing this exercise, quit the less pager with the q command, and terminate the vmstat and tail commands with a CTRL+C control sequence.
Questions
Use the file command to help answer the following 2 questions. 1. a. b. c. d. e. 2. a. b. c. d. e. 3. What type of file is /usr/bin/run-parts? A compiled ELF Executable An Awk script A Bash (Bourne-Again) shell script A Symbolic Link A /usr/bin/perl script What type of file is /dev/log? A character special file A block special file A socket A named pipe A compiled ELF executable Which of the following commands would act the same as cp a.txt b.txt, assuming the file b.txt does not exist?
64
Examining Files
a. b. c. d. e. 4. a. b. c. d. e. 5.
cat -5 a.txt b.txt cat a.txt b.txt cat -A a.txt > b.txt cat a.txt > b.txt None of the above. Which of the following key sequences can be used to quit the less pager? q x CTRL+C Both A and C All of the above While reading the file /etc/services, which of the following is an advantage of the less pager over the more pager?
a. b. c.
less allows the user to browse output one page at a time, while more does not. less recognizes familiar navigation keys, such as UpArrow and PgDown. less allows users to page backwards through the file, while more only allows users to page forwards. less allows the user to quit with a simple q, while more requires a CTRL+C All of the above Which of the following commands would display the first 5 lines of the file /etc/passwd?
d. e. 6. a. b. c. d. e. 7.
head -5 /etc/passwd head -n /etc/passwd head --five /etc/passwd head /etc/passwd > 5 head /5 /etc/passwd Why would you expect trouble from the following command?
[bob@station bob]$ head /bin/ls
a. b.
The user bob does not have permissions to read the file /bin/ls. The head command must be called with the -n command line switch, to specify how many lines to show. The file /bin/ls is too small for the head command
c.
65
Examining Files
d. e. 8. a. b. c. d. e.
The file /bin/ls is a binary file, and the head command works primarily on text files. None of the above.
How can a user stop the tail -f command? There is no need to, as the command times out after 10 seconds. By typing the q key. With the CTRL+Q control sequence. By switching virtual consoles. None of the above.
Use the following output to help in answering the next two questions.
[alice@station alice]$ cat -An /etc/fstab 1 LABEL=/ / 2 LABEL=/boot /boot 3 none /dev/pts 4 none /proc 5 none /dev/shm 6 /dev/hda2 swap ext3 ext3 devpts proc tmpfs swap defaults 1 1$ defaults 1 2$ gid=5,mode=620 0 0$ defaults 0 0$ defaults 0 0$ defaults 0 0$
9. a.
Which of the following is NOT true? The /etc/fstab file is 6 lines long. The columns in the /etc/fstab are separated by tabs instead of spaces. The /etc/fstab file is a text, instead of binary, file. The commands head -8 /etc/fstab and tail -6 /etc/fstab would produce identical output. None of the Above. How would the output of the commands cat /etc/fstab /etc/fstab and head /etc/ fstab /etc/fstab differ? a. b. c. d. e. The first command would only list the contents of the /etc/fstab file once. The first command would display tabs as visible characters. The second command would only show the first 10 lines of the duplicated /etc/fstab file. The second command would include banners identifying the filename in the output. None of the Above.
b. c. d. e. 10.
66
Chapter 7. Editing Files

Key Concepts
A text editor is not a word processor. Linux ships with many text editors nano is a commonly used command line text editor. gedit is a commonly used command graphical text editor.
Discussion
Text Editor vs. Word Processor
One of the ways Linux achieves its high degree of flexibility is through the robustness of its configuration. Configuration information typically is stored in plain ASCII text files. System administration often involves updating or correcting these configuration files. The tool for doing this work is a text editor. A text editor is not the same as a word processor. In addition to the words of a document, word processors allow the user to specify fonts, tab settings, margins, justification, and many other features designed to affect the way a document appears on paper. Documents generated by word processors contain embedded binary codes that specify the settings for these extra features. A plain text file such as a configuration file does not contain such codes, and is not designed for printing on paper. A text editor is designed just for modifying the text content of a plain ASCII file, and does not contain any embedded binary codes. A word processor would be an appropriate choice for writing a letter, a memo, a bulletin, an advertising flier, or any document where the visual presentation is controlled by the editing software. A text editor is the appropriate choice for creating or modifying configuration files, program source code, system documentation, or any document (such as an HTML document) where the visual presentation is controlled by the viewing software (such as a web browser) and not the editing software. Since this course focuses on Linux system administration, text editors will be our tool of choice for most situations.
Text Editors in Linux

Linux comes complete with a full suite of text editors, each with specific strengths and weaknesses. Linux also includes word processors for document creation as part of products such as OpenOffice. Since all text editors work on simple text files, you may choose any one you like - there are no problems of compatibility between one product and another. Normally, you make your choice on the basis of availability, ease of use, familiarity, and special features. Some editors, for example, make it easy to generate correctly-formatted program code, making programming easier and faster. This chapter presents two easy-to-use text editors, nano and gedit. The most popular text editor on Linux systems, vi, is also the most robust and the most complicated. The complexity of this editor makes it difficult to address in an introductory course.
nano
nano is a simple, display-oriented (full screen) text editor. Commands are displayed at the bottom of the screen, and context-sensitive help is provided. As characters are typed they are immediately inserted into the text.
67
Editing Files
Usage: nano [OPTIONS] [FILE] Opens FILE in insert mode for editing. As mentioned above, nano commands are displayed at the bottom of the screen throughout the edit session, making recalling editor commands very easy. The primary nano commands are summarized below. In this summary, the carat symbol followed by a letter is used to represent Ctrl-key combinations, so that ^K represents the CTRL+K key combination.
Table 7.1. nano commands

Command ^G ^X Ô ^J ^R ^W ^Y ^V ^K Û ^C ^T Function Get Help - opens a small help screen Exit - exits (and optionally saves file) Write Out - saves file Justify Read File - and insert at cursor position Where is - search for text Previous Page - move up one page Next Page - move down one page Cut Text - cuts one line at a time UnCut Text - inserts last cut block of text Cursor Position - identifies line and column number To Spell - starts spell check
Table 7.2. Selected nano Command Line Switches

Switch -w -v Function disables word wrap "view" file in read-only mode
System configuration files can be damaged by inappropriate line breaks, so nano is frequently started with the -w option when used to edit these files.
gedit
gedit is a graphical text editor available for the X-window environment. It can be accessed by entering gedit at a command prompt in a terminal window, or by selecting Accessories -> Text Editor from the GNOME Menu. It provides a familiar document window interface for creating and editing text files, and toolbar icons for accessing such features as search and replace, cut-copy-paste, and printing. Its tabbed window format allows for more than one document to be open at a time. It also supports a standard clickto-browse interface for interacting with the filesystem.
68
Editing Files
Figure 7.1. The gedit Graphical Text Editor
Online Exercises
Text Editors
Lab Exercise
Objective: Text Editors Estimated Time: 10 mins.
Specification
The /etc/services file is an example of a Linux ASCII text configuration file. This lab will have you make two copies of this file, and edit one using the nano text editor, and edit the other using the gedit graphical text editor. If your system does not have the X server installed, use the nano editor for both files. The nano Editor 1. In your home directory, create a copy of the file /etc/services, called services.nano 2. Open the file services.nano using the nano text editor. 3. Using the control sequences specified at the bottom of the screen, search for, and remove, all occurrences of the + character. Preserve the spacing of the columns as you do so. Save your file as services.noplus. 4. Continuing in the same editing session, move your cursor to line number 145. The CTRL+C key combination will report your current line number. (Note that CTRL+C, which usually terminates a process, has been overridden by the nano editor, and has new behavior). Delete that line, and the following line. Save your file as services.nonext. 5. Close the nano editor. The gedit Editor
69
Editing Files
1. In your home directory, create a copy of the file /etc/services, called services.gedit 2. Open the file services.gedit using the gedit text editor. 3. Move your cursor to line number 104. (A little exploration through gedit's menus may simplify this task). Delete that line, and the following 2. Save your file as services.nosmux. 4. Replace every occurrence of the letter tcp with udp. Make your search and replace case sensitive (don't replace TCP, for instance). Save your file as services.notcp. 5. Close the gedit editor.
Deliverables
1. All files are to be created in your home directory. 1. A file named services.noplus, which is a duplicate of the file /etc/services, with + characters removed. 2. A file named services.nonext, which, in addition to the previous edit, has lines number 145 and 146 removed. 3. A file named services.nosmux, which is a duplicate of the file /etc/services, with lines numbered 104, 105, and 106 removed. 4. A file named services.notcp, which in addition to the previous edit, has all occurrences of the characters tcp replaced with udp.
70

Rha030 Workbook02 Student 6.1 2

Uploaded by

Copyright:

Available Formats

Rha030 Workbook02 Student 6.1 2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rha030 Workbook02 Student 6.1 2

Uploaded by

Copyright:

Available Formats

Part Workbook 2.

Chapter 1. Filesystem Navigation

Figure 1.1. Sample Directory Tree

Figure 1.2. Directory Tree Starting at /

Exploring the Directory Tree using Nautilus

Figure 1.3. Nautilus Window

Figure 1.4. Nautilus's Browser View

Figure 1.5. Configuring Nautilus to use Browser View by Default

Figure 1.6. Choosing the Nautilus tree sidepane

Figure 1.7. Exploring the directory tree with Nautilus

The Current Working Directory (cwd)

Where Am I? The pwd command

Moving Around - The cd command

Table 1.1. Special Directory Names

Absolute and Relative References

Table 1.2. Sample Relative References to /home/alice/sample.txt

Climbing up and down the tree.

Which of the following is a relative directory reference? /home/student /etc .. ~

Your screen shows the following:

Chapter 2. Important Directories

Figure 2.1. Partial Directory Tree Starting at /

The User's Home Directory

The /tmp Temporary Directory

The /etc Configuration Directory

The /bin and /usr/bin Command Directories

The /sbin and /usr/sbin Command Directories

The /var "Variable" Directory

root vs. /root vs. / (the filesystem root)

4096 Mar 16 08:04 orbit-alice 4096 Mar 16 07:07 ssh-XXDg4ke3

Looking for commands in /bin and /usr/bin

2. A ~/.bash_history file that contains a record of your "raw" commands.

/home/student /root / None of these - all are likely to be home directories

T/F: X uses /tmp. True False

Chapter 3. Managing Files

Copy files with cp

Move/rename files with mv

Remove (delete) files with rm

Clobbering files (oops!)

Moving a Simple Website

Investigate the "noclobber" option

9. Exit the terminal session or virtual console.

Chapter 4. Managing Directories

Elvis can then create a subdirectory of public_html:

[elvis@station elvis] ls work/images advertising logos

Listing Directory Trees

Removing (empty) directories: rmdir

Listing Directory Trees

Copying directory trees: cp -r

Figure 4.1. Directory structure before copying

[hogan@station hogan]$ cp -r website archive

Figure 4.2. Directory structure after copying

Figure 4.3. Directory structure before copying

Figure 4.4. Directory structure after copying with cp -r website archive

Removing directory trees: rm -r

Figure 4.5. Directory structure before removing

[hogan@station hogan]$ rm -r website/media

Figure 4.6. Directory structure after removing

Examining Directories with tree

Similarly, tree -d just shows directories.

Remembering that rmdir only works on empty directories, he uses rm -r instead.