File system implementation

Unit – 4

File System Implementation

Overview

Free Space Management
Efficiency and Performance
Recovery
Log-Structured File Systems
NFS

Free-Space Management
 How should unallocated blocks be managed?
 Need a data structure to keep track of them
Index block
 same as for conventional file, but no info in any block
 initially very large; makes it impractical: index table too large
Linked list
 Very large
 Hard to manage spatial locality
 want to allocate closely located blocks to a file
 this minimizes seek time
 hard to do with a linked list; have to traverse the list to find appropriate
blocks

Linked Free Space List on Disk

 Need a data structure to keep track of them
Grouping
 Modification of free-list (linked-list) approach
 Store the addresses of n free blocks in the first free block
 n - 1 of these are actually free; last block contains pointers to the next n
free block. Etc.
 Can now find a large number of free blocks quickly.
Counting
 Don’t keep the address of every free block.
 Free blocks are often consecutive
 So keep the address of the first free block and the number of following
consecutive blocks that are free.
 Each entry in the free-space list consists of a disk address and a count.

Bit vector (block status map or “disk map”) (n blocks)

0 1 2 n-1
…


0 block[i] free
bit[i] =
1 block[i] occupied
Block number calculation

(number of bits per word) *
(number of 0-value words) +
offset of first 1 bit

Most CPUs have special bit-manipulation instructions.
Example: Intel and PowerPC have instructions that return the offset in a
word of the first bit with the value 1

Free-Space Management (Cont.)

 Bit map requires extra space
Example:
block size = 212 bytes
disk size = 230 bytes (1 gigabyte)
n = 230/212 = 218 bits (or 32K bytes)
 Easy to get contiguous files
 Counting

Free-Space Management (Cont.)

Need to protect:
 Pointer to free list
 Bit map
 Must be kept on disk
 Copy in memory and disk may differ
 Cannot allow for block[i] to have a situation where
bit[i] = 1 in memory and bit[i] = 0 on disk
Solution:
Set bit[i] = 1 in disk
Allocate block[i]
Set bit[i] = 1 in memory

Efficiency and Performance

 Efficiency dependent on:
 disk allocation and directory algorithms
 types of data kept in file’s directory entry
Example: keep a “last write date” and a “last access date”
 Thus whenever a file is read, a field in the directory structure must be
written to.
 This means that the block must be read into memory, a section changed,
and the block written back out to disk.

Efficiency and Performance

Performance
 disk cache – separate section of RAM in the disk controller for
frequently used blocks
 Some systems maintain a separate section of main memory for a
buffer cache.
 Keep blocks read from disk on assumption that they’ll be used again
 free-behind and read-ahead – techniques to optimize sequential
access
 improve PC performance by dedicating section of memory as virtual
disk, or RAM disk

Recovery
Consistency checking
Problem:
1. Directory information kept in main memory
2. Main memory more up to date than info on disk
What happens when computer crashes?
 Changes to directories may be lost.
 Actual state is not reflected in the directory structure.
 compares data in directory structure with data blocks on
disk
 tries to fix inconsistencies

Recovery
Fixable problems
 Depend on free-space-management algorithms
 If linked allocation used and each block links to next block
 Then directory structure can be reconstructed from data
block
Example: UNIX
 Indexed system, blocks have no knowledge of one another
 Impossible to recreate files
 This is why UNIX caches read
 But any action that results in metadata changes is done
synchronously, before data blocks are written

Recovery
 Use system programs to back up data from disk to another
storage device (floppy disk, magnetic tape, other magnetic disk,
optical)
 Recover lost file or disk by restoring data from backup
Backup programs take advantage of metadata about file
updates:
Day 1 Copy all files from disk to backup drive (full backup)
Day 2 Copy changes since day 1 (incremental backup)
Day 3 Copy changes since day 2
…
Day n Copy changes since day n-1.
Start over again.
 Allows restoration by starting with full backup, then applying all
incremental changes.

Log Structured File Systems
 Log structured (or journaling) file systems record each update to
the file system as a transaction
 All metadata transactions are written to a log
 Each set of operations for performing a specific task is a
transaction
 A transaction is considered committed once it is written to the log
(even though the operations are not yet performed)
 The transactions in the log are asynchronously written to the file
system
 As the operations are completed, a pointer is updated to indicate
which actions have completed and which are still incomplete
 When all operations are complete, the transaction is removed from
the log file
 If the file system crashes, all remaining transactions in the log
must still be performed

The Sun Network File System (NFS)

 An implementation and a specification of a software system for
accessing remote files across LANs (or WANs)
 The implementation is part of the Solaris and SunOS operating
systems running on Sun workstations using an unreliable
datagram protocol (UDP/IP protocol) and Ethernet
 Interconnected workstations viewed as a set of independent
machines with independent file systems, which allows sharing
among these file systems in a transparent manner

NFS
 Accesing file over a network
 A remote directory is mounted over a local file system
directory
 The mounted directory looks like an integral subtree of the
local file system, replacing the subtree descending from the
local directory
 Specification of the remote directory for the mount operation
is nontransparent; the host name of the remote directory has
to be provided
 Files in the remote directory can then be accessed in a
transparent manner
 Subject to access-rights accreditation, potentially any file
system (or directory within a file system), can be mounted
remotely on top of any local directory

NFS Protocol
 Provides a set of remote procedure calls for remote file
operations.
 The procedures support the following operations:
1. searching for a file within a directory
2. reading a set of directory entries
3. manipulating links and directories
4. accessing file attributes
5. reading and writing files
 NFS servers are stateless; each request has to provide a full set
of arguments
 (NFS V4 is just coming available – very different, stateful)
 The NFS protocol does not provide concurrency-control
mechanisms

Three Major Layers of NFS
Architecture
UNIX file-system layer
 interface (based on the open, read, write, and close calls, and
file descriptors)
Virtual File System (VFS) layer
 Distinguishes local files from remote ones, and local files are
further distinguished according to their file-system types
 The VFS activates file-system-specific operations to handle
local requests according to their file-system types
 Calls the NFS protocol procedures for remote requests
NFS service layer
 bottom layer of the architecture
 Implements the NFS protocol

Hope u have learnt something !!

Thank you

File system implementation

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to File system implementation (20)

More from Rotract CLUB of BSAU (7)

Recently uploaded (20)

File system implementation

Editor's Notes