Transcript What Is A File?

Operating Systems
File Management
What Is A File?




Named collection of data stored on hard disk,
CD-ROM etc.
Can be programs (both source and object
forms) or data.
The UNIX/Linux and Windows OS’s do not
know anything about the file structure.
The application program associated with the
file knows how to interpret the file.
–
Interpretation logged by OS
What Is File Management?





The creation and deletion of files.
The creation and deletion of directories/
folders.
The manipulation of files and directories
The mapping of files onto disk storage.
Backup of files on stable (non volatile)
storage.
What is a Directory / Folder?

A logical grouping of files.
–

Directories can contain other directories
known as sub-directories, creating a tree
structure.
–

Actually a special type of file.
Hierarchical
UNIX/Linux uses the term directory, Windows
uses the term folder.
Example of a Directory Structure

Windows Explorer
–
–
The yellow icons are
directories (folders).
The plus (+) sign next to
the directory indicates
that it contains subdirectories.
Pathnames




In order to uniquely specify a file in a tree-structured
directory, a pathname must be used.
A pathname prefixes the filename with a list of the
directories from the base directory which have to be
traversed in order to reach the file.
The pathname separator is the forward-slash (/) in
UNIX/Linux and the backslash (\) in Windows.
At any instant, a process has a current directory,
which is where the operating system will look for a
file whose pathname is not specified.
Pathname Examples

UNIX/Linux
–
–

Full pathname - /cprogs/work/file1.txt
Relative pathname - work/file1.txt
Windows
–
–
Full pathname - C:\CPROGS\WORK\FILE1.TXT
Relative pathname – WORK\FILE1.TXT
Common Directory Commands
FUNCTION
Change Directory
UNIX/Linux Windows/
DOS
cd
CD
CHDIR
Make Directory
mkdir
MD
MKDIR
Remove Directory
rmdir
RD
RMDIR
ls
DIR
pwd
CD
CHDIR
List Directory Contents
Display Current Directory
File Naming

UNIX/Linux
–
–
–
–
–
File names are case-sensitive.
File name can be up to 255 characters in length.
File name can contain any character except /.
Characters that have special meaning to the shell
(* > < ? \ ; : | [ ] ( )) should not be included in a file
name.
Linux does allow spaces in the file name, other
flavours of UNIX may not.
File Naming

Windows
–
–
–
–
–
–
File names are not case sensitive, but the case will be
preserved.
File name can be up to 255 characters in length.
File name can contain any character except / \ “ ; * ? < > |.
The portion of the file name after the last dot (.) is known as
the extension.
The extension is used to specify the nature of the file.
Windows also maintains a short file name using the first 8
characters of the file name and the first three characters of
the extension.
File Naming



Both UNIX/Linux and Windows allow a file to be
referenced using wildcard characters.
In both OS’s - * can be replaced by zero or more
characters; ? can be replaced by one character.
UNIX/Linux also supports more complex wildcarding.
–
–
–
[a,b,c] - character can be either a, b or c.
[a-d] - character can be either a,b,c or d
[a-d,f] – character can be either a,b,c,d or f
Hard Disk Partitioning







Partitioning is a means of sub-dividing a hard disk.
A hard disk must contain at least one partition.
There are two types of partitions – primary and extended.
A disk can contain up to 4 primary partitions or up to 3 primary
and 1 extended partition.
One primary partition may be designated as active, this will be
used for booting the computer.
Different operating systems can be installed on different
primary partitions – allowing multi-booting.
An extended partition cannot be accessed directly by the
operating system but is instead sub-divided into logical drives.
Hard Disk Partitioning


In UNIX/Linux
the fdisk
command is
used to
manage hard
disk partitions.
In Windows
2000+ use the
DISKPART
command or
the Disk
Management
tool.
Hard Disk Initialization




In order to be usable by the OS a partition must be
initialized.
In UNIX/Linux the mkfs command is used, in
Windows the command is FORMAT.
The initialisation process sets up the necessary
structures on a disk partition.
In UNIX/Linux an initialized partition is referred to as
a file system on Windows it is called a drive.
File Space Allocation




Disk space is dynamically allocated to files.
Space is allocated in units of fixed size called
allocation units or clusters.
When a new file is created, it is allocated one
cluster.
As the file is extended the cluster eventually
becomes filled and more clusters are
assigned to the file.
File Space Allocation



Ideally, successive file clusters should occupy contiguous
space on the disk.
In general, the disk will contain many different files whose
clusters are intermingled with each other.
Clusters are released when a file is deleted or shrunk, so free
clusters may be intermingled with allocated clusters.
A
A
B
A
End of
file A
B
C
C
B
Free
cluster
B
End of
file B
UNIX/Linux File System

4 classes of files:
–
–
–
–
regular file.
directory.
character or block special file: device such as a
disk, printer, terminal.
named pipe: queuing buffer which passes the
output of one process to the input of another.
UNIX/Linux File System




A file is described by an information block
called an inode.
Every file on disk has an associated inode.
Kernel memory contains a copy of the inode
for every open file.
A directory entry holds the filename and a
reference to its associated inode or its inode
number.
UNIX/Linux File System

The inode contains all information about a file,
except its name. This information includes
–
–
–
–
–
File access and type information, collectively known as the
mode.
File ownership information.
Time stamps for last modification, last access and last mode
modification.
File size in bytes.
Addresses of physical blocks.
UNIX/Linux File System - Inodes





The inode records which disk blocks/clusters have
been allocated to a file.
This information is stored in 13 “pointers” in the
inode structure.
The first 10 are direct pointers – they map directly to
the physical address of a block of a file.
The other 3 pointers point to index blocks which
themselves contain lists of pointers.
The pointers in the index blocks reference either
data blocks or further index blocks.
UNIX/Linux File System – Volume
Structure

When the mkfs command is run to initialize a
partition the following layout is created:
–
–
–
–
Boot block: first stage boot program.
Super block: contains details of disk size and
allocation of inodes and data blocks.
Inodes: storage space for inodes.
Data Blocks: rest of disk available for directories
and files.
UNIX/Linux File System – Free Space
Control



When the mkfs command is run it creates a
chain of linked blocks.
The super block contains the start block in
the chain.
Each block contains an array of free data
block numbers and a pointer to the next
block in the chain.
Windows File System

Windows 2000+ supports 3 different file
systems:
–
–
–

FAT also called FAT16
FAT32
NTFS
FAT and FAT32 are implemented similarly.
FAT16/FAT32 File System




The FAT system maintains a table of entries known
as the File Allocation Table (hence the name FAT).
The table consists of an array of 16-bit (FAT16) or
32-bit (FAT32) values.
The first value in the array identifies the disk system
type and the second entry is set to all 1’s.
Each of the remaining entries maps to a cluster on
the disk.
FAT16/FAT32 File System

Each entry records one of five things:
–
–
–
–
–
the address of the next cluster in a chain
a special end of file (EOF) character that indicates
the end of a chain
a special character to mark a bad cluster
a special character to mark a reserved cluster
a zero to note that that cluster is unused
FAT16/FAT32 File System



Prior to any data being written to the disk, the
FAT entries are all set to zero, indicating a
free cluster.
As file space is allocated, the FAT entries
contain pointers to the next entry in the
chain.
When more file space is requested, the OS
searches the FAT for an entry with a value of
zero (unused) and allocates it to the file.
FAT16/FAT32 File System – Volume
Structure

When the FORMAT command is run on a partition to
create a FAT file system, the following layout is
created:
–
–
–
–
–
Boot sector: details about the disk’s characteristics and a
bootstrap loader.
FAT: principal file allocation table.
Additional FAT(s): Duplicate table(s) which are updated at
the same time as the principal FAT.
Root directory
File space: The rest of the disk.
Disk Fragmentation




Over time, allocation of clusters becomes less
contiguous.
Deletion and resizing of files may result in free
clusters located in random positions on the disk.
Such a situation is called fragmentation and can
result in reduced performance of the computer.
To correct this situation a defragmenting tool must be
used which can put all the clusters of a file in
contiguous locations and coalesce the free clusters.
Disk Fragmentation

Windows comes with a disk defragmenting tool.