cs 471 - lecture 8 file systems ch. 10,11 george mason university fall 2009
TRANSCRIPT
10.2GMU – CS 571
File-System Interface
File Concept File Operations Access Methods Directory Structure Access control
10.3GMU – CS 571
Files
A file is a named collection of related information that is recorded on secondary storage
Several information storage media (magnetic/optical disks)
The operating system provides a uniform logical view of information storage
10.4GMU – CS 571
Files Files
• are mapped onto physical storage devices.
• represent programs (both source and object forms) and data.
• have a certain structure that may be considered as sequence of bits, bytes, lines, records…
• meaning defined by file’s creator
• have attributes that are recorded by the O.S. (name, size, type, location, protection info, time info, etc.)
• logically contiguous
Information about files are kept in the directory structure, which is also maintained on the secondary storage.
10.5GMU – CS 571
Basic File Operations
Create Write Read Delete Others
• reposition within the file, append, rename, truncate, ...
For write/read operations, the operating system needs to keep a file position pointer for each process• Need to update it dynamically and properly
10.6GMU – CS 571
File Operations
To avoid searching the directory entries repeatedly, many systems require that an open() system call be issued before that file is first used actively.
Operating System keeps • a system-wide open-file table containing
information about all open files
• per-process open-file tables containing information about all open files of each process
The open operation takes a file name and searches the directory, copying the directory entry into the open-file table. It returns a pointer to the entry in the open file table.
10.7GMU – CS 571
File Operations
The per-process open table contains info about• Position pointer (current location within file)
• Access rights
• Accounting
• Pointer to the system-wide open-file table entry
The system-wide open table includes info about• File location on the disk
• File size
• File open count (the number of processes using this file)
A process that completes its operations on a given file will issue a close() system call.
10.8GMU – CS 571
File Operations (Cont.)Process A’s Open-File
Table
.
.
. ..
Process B’s Open-File
Table
.
.
.. ..
.
.
.. ...
.
.
System-Wide Open-File
Table
10.9GMU – CS 571
An Example Program Using File System Calls (1/3)
/* File copy program. Error checking and reporting is minimal. */
/* “myfilecopy oldfile newfile” will copy the contents of “oldfile” to “newfile” */
/* The program will read blocks of 4K from the “oldfile” to a buffer, and store them to “newfile” sequentially */
#include <sys/types.h> /* include necessary header files */
#include <fcntl.h>
#include <stdlib.h>
#include <unistd.h>
int main(int argc, char *argv[]); /* ANSI prototype */
#define BUF_SIZE 4096 /* use a buffer size of 4096 bytes */
#define OUTPUT_MODE 0700 /* protection bits for output file */
10.10GMU – CS 571
An Example Program Using File System Calls (2/3)
int main(int argc, char *argv[])
{
int in_fd, out_fd, rd_count, wt_count;
char buffer[BUF_SIZE];
if (argc != 3) exit(1); /* error if argc is not 3 */
/* Open the input file and create the output file */
in_fd = open(argv[1], O_RDONLY); /* open the source file */
if (in_fd < 0) exit(2); /* if it cannot be opened, exit */
out_fd = creat(argv[2], OUTPUT_MODE); /* create the destination file */
if (out_fd < 0) exit(3); /* if it cannot be created, exit */
10.11GMU – CS 571
An Example Program Using File System Calls (3/3)/* Copy loop */
while (TRUE) {
rd_count = read(in_fd, buffer, BUF_SIZE); /* read a block of data */
if (rd_count <= 0) break; /* if end of file or error, exit loop */
wt_count = write(out _fd, buffer, rd_count); /* write data */
if (wt_count <= 0) exit(4); /* wt_count <= 0 is an error */
}
/* Close the files */
close(in_fd);
close(out_fd);
if (rd_count == 0) /* no error on last read */
exit(0);
else
exit(5); /* error on last read */
}
10.12GMU – CS 571
File Types
Most operating systems associate a type with a file
File type can be used to operate on files in reasonable ways • ex: Windows – file type (i.e. suffix) used to
determine what program to open a file with
• ex: Unix – info stored in file (‘magic number’) can be used for differentiation – suffix not always used
10.14GMU – CS 571
File Structure
None - sequence of words, bytes Simple record structure
• Lines • Fixed length• Variable length
Complex Structures• Formatted document• Relocatable load file
Can simulate last two methods with first method by inserting appropriate control characters
Who decides:• Operating system• Program
10.15GMU – CS 571
Internal File Structure
Disk systems have a well-defined block size determined by the size of a sector.
All disk I/O is performed in units of one block (physical record).• Each block is one or more sectors• A sector can hold 32 – 4096 bytes
Files are made of logical records. Often, a number of logical records
will be packed into physical records. Operating System will perform
translation from logical records to physical records.
Internal fragmentation
10.16GMU – CS 571
File Access Methods
Sequential Access• Information is processed in order, one record after the other
(tape model)• Example: editors and compilers
read nextwrite next reset (rewind)
10.17GMU – CS 571
File Access Methods Direct Access
• The file is made up fixed-length logical records that allow programs to read and write records rapidly in any order
read nwrite n
or alternatively:position to nread nextwrite next
n = relative block number request to read block N translated into physical
address B*N + start (for block size B) ex: database
Other access methods often built on top of direct access
10.18GMU – CS 571
Directory Structure
The directory acts as a symbol table that translates file names into their directory entries.
Operations on a directory• Search for a file
• Create a file
• Delete a file
• List a directory
• Rename a file
• …
10.19GMU – CS 571
Organize the Directory (Logically) to Obtain
Efficiency – locating a file quickly Naming – convenient to users
• Two users can have same name for different files
• The same file can have several different names
Grouping – logical grouping of files by properties, (e.g., all Java programs, all games, …)
10.20GMU – CS 571
Single-Level Directory
A single directory for all users
Naming problem
Grouping problem
10.21GMU – CS 571
Two-Level Directory
Separate directory for each user
Path name
Can have the same file name for different user
Efficient searching
No grouping capability
10.22GMU – CS 571
Tree Directory Structure
Tree-structured directories extend the structure to a tree of arbitrary height• User-imposed structure• Relative paths vs. absolute paths• Directory deletion policy• Concept of a ‘current directory’
10.23GMU – CS 571
Acyclic-Graph Directories
Allows shared subdirectories and files. A shared file will “exist” in multiple directories
at once.
10.24GMU – CS 571
Achieving File Sharing
Option 1: Duplicate all information about the shared file in both directories (Problem?)
Option 2: Create a new directory entry called link• The link is effectively a pointer to another file or
directory
• When the directory entry of a referred file is a link, we resolve the link by using the path name (symbolic link in Unix)
• “ln –s reports/report1.txt myreport”
10.25GMU – CS 571
Achieving File Sharing (Cont.) Option 3: Each entry in a directory can point to a
little data structure (File Control Block [FCB], or “i-node”) that keeps information about the file • The directory entries corresponding to a shared
file will all point to the same file control block
• Non-symbolic or “hard” links in Unix
• “ln reports/report1.txt myreport”
FCB of the file
“reports” Directory report1.txt
“root“ Directory myreport
10.26GMU – CS 571
Achieving File Sharing (Cont.) What to do when a shared file is deleted by a user?
• The deletion of a link should not affect the original file
• If the original file is deleted, we may be left with dangling pointers.
Solutions• Using backpointers, delete also all links. The search
may be expensive.
• Alternatively, leave the links intact until an attempt is made to use them (Unix symbolic links). May lead to infrequent but subtle problems.
• In case of non-symbolic (or in Unix, “hard”) links: Preserve the file until all references are deleted. Keep the count of the number of the references, delete the file when the count reaches zero.
10.27GMU – CS 571
File Protection
File owner/creator should be able to control:• what can be done
• by whom
Types of access• Read
• Write
• Execute
• Append
• Delete
• List
10.28GMU – CS 571
Access Lists and Groups Mode of access: read, write, execute Three classes of users
RWXa) owner access 7 1 1 1RWXb) group access 6 1 1 0RWXc) public access 1 0 0 1
Ask manager to create a group (unique name), say G, and add some users to the group.
For a particular file (say game) or subdirectory, define an appropriate access.
10.31GMU – CS 571
File System Implementation
File System Structure File System Implementation Allocation Methods File System Performance
10.32GMU – CS 571
File System Structure
An operating system may allow multiple file systems.
Once the user interface is determined, the file system must be implemented to map the logical file system to the physical secondary-storage devices.
File control block – storage structure that keeps information about a given file (Unix “i-nodes”).• Ownership, size, permissions, access date info,
location of data blocks
11.33 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
Schematic View of Virtual File SystemSchematic View of Virtual File System
10.34GMU – CS 571
Layered File System File system is organized into layers
Logical File System Layer manages the file-system structure (through directories and FCBs).
File-Organization Module performs mapping between logical blocks and physical blocks. It also includes free-space manager and block allocation manager.
Basic File System Layer issues generic commands to the appropriate device driver (I/O Control Layer) to read and write physical blocks on the disk
10.35GMU – CS 571
Storage Structure
A disk is a physical memory storage device that can be used for:• a single file system (in its
entirety)• multiple file systems• in part for file systems, in
part for other purposes (e.g. for swap space or unformatted (raw) disk space)
These parts are known as partitions, slices or minidisks.
10.36GMU – CS 571
Storage Structure (Cont.)
Each partition can be either “raw” (containing no file system), or “cooked” (with a file system)
Raw disk • contains a large sequential array of logical blocks,
without any file-system data
• can be used as swap space
• can be used for special (e.g. database) applications
10.37GMU – CS 571
Storage Structure (Cont.)
Each partition that contains a file system has a device directory
The device directory keeps information (name, location, size, type, owner) for files on that partition.
10.38GMU – CS 571
Accessing Disk Sub-system
Disks allow direct access to stored data
Disk access time has two components• Random access time
(positioning) that includes seek time and rotational latency (5-10 ms)
• Transfer time (10 MB/s)
Compare to the memory access time of 10-100 nanoseconds
10.39GMU – CS 571
Accessing Disk Sub-system
When a process needs I/O, it issues a system call to the OS• input or output
• from what disk address
• to what memory address
• how many sectors
At any point in time, the disk may have several pending requests that must be scheduled:• FCFS
• SSTF (shortest seek time first)
• SCAN
• …
10.40GMU – CS 571
Implementation of “Open” and “Read”
Figure (a) refers to opening a file. Figure (b) refers to reading a file.
10.41GMU – CS 571
Allocation Methods
The allocation method refers to how disk blocks are allocated for files:
• Contiguous allocation
• Linked allocation
• Indexed allocation
10.42GMU – CS 571
Contiguous Allocation Each file occupies a set of contiguous blocks on the disk. Simple – only starting location (block #) and length (number of blocks) are
required.
10.43GMU – CS 571
Contiguous Allocation Efficient access to multiple blocks of a file Both sequential and direct access can be supported.
A major problem is determining how much space is needed for a new file.
How to let files grow?
Finding space for a new file: First-fit and best-fit …
These algorithms suffer from external fragmentation:free space is broken into multiple chunks.
10.44GMU – CS 571
Extent-Based Systems
Many newer file systems (I.e. Veritas File System) use a modified contiguous allocation scheme
Extent-based file systems allocate disk blocks in extents
An extent is a contiguous block of disks• Extents are allocated for file allocation
• A file consists of one or more extents.
10.45GMU – CS 571
Linked Allocation
Each file is a linked list of disk blocks: blocks may be scattered anywhere on the disk.
pointerblock =
10.46GMU – CS 571
Linked Allocation Each file is a linked list of disk blocks: blocks may be
scattered anywhere on the disk. Each block contains a pointer to the next block. Each directory entry has a pointer to the first and last
disk blocks of the file.
10.47GMU – CS 571
Linked Allocation External fragmentation is eliminated. The size of a file does not need to be declared at
the time of creation. However, it can be used effectively only for
sequential access files. Inefficient for direct-access files.
Another disadvantage is the space required for the pointers.
One solution is to collect blocks into multiples (clusters) and to allocate the clusters rather than blocks.
Another problem of linked allocation is reliability: what will happen if a pointer is lost or damaged?
10.48GMU – CS 571
File-Allocation Table (FAT)
A variation of the linked allocation method
A section of the disk at the beginning of each partition is used as the File Allocation Table.
The table entries give the block number of the next block in the file.
The scheme can result in a significant number of disk head seeks, unless the FAT is cached.
10.49GMU – CS 571
Indexed Allocation Indexed allocation supports direct access,
without suffering from external fragmentation or size-declaration problems.
However, wasted space may be a problem. How large the index block should be?
• To reduce the wasted space, we want to keep the index block small
• If the index block is too small, it will not be able to hold pointers for a large file.
• Linked scheme• Multilevel scheme• Combined scheme
index table
10.51GMU – CS 571
Combined Scheme (Unix)
Keep the first N pointers of the index block in the file’s i-node (FCB).
The first 12 of these pointers point to direct blocks The next three pointers point to indirect blocks
10.52GMU – CS 571
File System Performance
Disk access is the bottleneck for the file system performance
Caching • Most disk controllers have an on-board cache that
can store entire tracks at a time• Subsequent requests can be served through the on-
board cache
Most systems maintain a separate section of main memory for a disk cache (block cache, or buffer cache), where blocks are kept under the assumption that they will be re-used in near future
10.53GMU – CS 571
Caching
A page cache caches pages rather than disk blocks using virtual memory techniques
Memory-mapped I/O uses a page cache
Routine I/O through the file system uses the buffer (disk) cache
10.54
Memory-mapped I/O
Memory-mapped I/O uses the same address bus to address both memory and I/O devices, and the CPU instructions used to access the memory are also used for accessing devices.
Port-mapped I/O uses a special class of CPU instructions specifically for performing I/O.
A device's direct memory access (DMA) is a memory-to-device communication method, that bypasses the CPU.
GMU – CS 571
10.55GMU – CS 571
Unified Buffer Cache
A unified buffer cache uses the same page cache to cache both memory-mapped pages and ordinary file system I/O
10.56GMU – CS 571
File System Performance (Cont.) LRU is a reasonable block replacement policy
BUT: if a critical block (such as File Control Block, or i-node) is read into the cache and modified, but not re-written to the disk, a crash will leave the file system in an inconsistent state.
Critical blocks must be written immediately.
Avoiding inconsistency• Write through-cache: write every modified block to disk
as soon as it has been written
UNIX solution• The system call sync forces all the modified blocks out
onto the disk immediately.• A program, usually called update, is invoked in the
background to call sync every 30 seconds.
10.57GMU – CS 571
File System Performance
Block-read-ahead: When reading block k to the cache in memory, read also block k+1
Reduce disk arm motion through• Putting blocks that are likely to be accessed in
sequence close to each other
• Disk scheduling algorithms that serve pending disk access requests in an order that reduces the delay
10.58GMU – CS 571
Distributed File Sharing
Sharing of files on multi-user systems is desirable
On distributed systems, files may be shared across a network• Manually via programs like FTP• Automatically, seamlessly using distributed file
systems• Semi automatically via the world wide web
Network File System (NFS) is a common distributed file-sharing method
10.59GMU – CS 571
File Sharing – Remote File Systems Client-server model allows clients to mount remote file
systems from servers• Server can serve multiple clients• Client and user-on-client identification is insecure or
complicated• NFS is standard UNIX client-server file sharing protocol• CIFS is standard Windows protocol• Standard operating system file calls are translated into
remote calls
Distributed Information Systems (distributed naming services) such as LDAP, DNS, NIS, Active Directory implement unified access to information needed for remote computing
10.60GMU – CS 571
File Sharing – Failure Modes
Remote file systems add new failure modes, due to network failure, server failure
Recovery from failure can involve state information about status of each remote request
Stateless protocols such as NFS include all information in each request, allowing easy recovery but less security
10.61GMU – CS 571
File Sharing – Consistency Semantics
Consistency semantics specify how multiple users are to access a shared file simultaneously• Similar to process synchronization algorithms
Tend to be less complex due to disk I/O and network latency (for remote file systems)
• Andrew File System (AFS) implemented complex remote file sharing semantics
• Unix file system (UFS) implements: Writes to an open file visible immediately to other users
of the same open file Sharing file pointer to allow multiple users to read and
write concurrently
• AFS has session semantics Writes only visible to sessions starting after the file is
closed
11.62 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
The Sun Network File System (NFS)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)
11.63 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
NFS (Cont.)NFS (Cont.)
Interconnected workstations viewed as a set of independent machines with independent file systems, which allows sharing among these file systems in a transparent manner 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
11.64 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
NFS (Cont.)NFS (Cont.)
NFS is designed to operate in a heterogeneous environment of different machines, operating systems, and network architectures; the NFS specifications independent of these media
This independence is achieved through the use of RPC primitives built on top of an External Data Representation (XDR) protocol used between two implementation-independent interfaces
The NFS specification distinguishes between the services provided by a mount mechanism and the actual remote-file-access services
11.65 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
Three Independent File SystemsThree Independent File Systems
11.66 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
Mounting in NFS Mounting in NFS
Mounts - S1:/usr/shared
Over U:/usr/local/
Cascading mounts - S2:/usr/dir2
Over U:/usr/local/dir1
11.67 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
NFS Mount ProtocolNFS Mount Protocol
Establishes initial logical connection between server and client
Mount operation includes name of remote directory to be mounted and name of server machine storing it
Mount request is mapped to corresponding RPC and forwarded to mount server running on server machine
Export list – specifies local file systems that server exports for mounting, along with names of machines that are permitted to mount them
Following a mount request that conforms to its export list, the server returns a file handle—a key for further accesses
File handle – a file-system identifier, and an inode number to identify the mounted directory within the exported file system
The mount operation changes only the user’s view and does not affect the server side
11.68 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
NFS ProtocolNFS Protocol
Provides a set of remote procedure calls for remote file operations. The procedures support the following operations: searching for a file within a directory reading a set of directory entries manipulating links and directories accessing file attributes 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) Modified data must be committed to the server’s disk before results
are returned to the client (lose advantages of caching) The NFS protocol does not provide concurrency-control
mechanisms
11.69 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
Three Major Layers of NFS Architecture Three Major Layers of NFS Architecture
UNIX file-system 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
11.70 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
Schematic View of NFS Architecture Schematic View of NFS Architecture
11.71 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
NFS Path-Name TranslationNFS Path-Name Translation
Performed by breaking the path into component names and performing a separate NFS lookup call for every pair of component name and directory vnode
To make lookup faster, a directory name lookup cache on the client’s side holds the vnodes for remote directory names
11.72 Silberschatz, Galvin and Gagne ©2005Operating System Concepts – 7th Edition, Jan 1, 2005
NFS Remote OperationsNFS Remote Operations
Nearly one-to-one correspondence between regular UNIX system calls and the NFS protocol RPCs (except opening and closing files)
NFS adheres to the remote-service paradigm, but employs buffering and caching techniques for the sake of performance
File-blocks cache – when a file is opened, the kernel checks with the remote server whether to fetch or revalidate the cached attributes
Cached file blocks are used only if the corresponding cached attributes are up to date
File-attribute cache – the attribute cache is updated whenever new attributes arrive from the server
Clients do not free delayed-write blocks until the server confirms that the data have been written to disk