Lines Matching full:file

30 New file system organization
32 In the new file system organization (as in the
33 old file system organization),
34 each disk drive contains one or more file systems.
35 A file system is described by its super-block,
36 located at the beginning of the file system's disk partition.
39 This is done when the file system is created;
49 the minimum size of a file system block is 4096 bytes.
50 The size of file system blocks can be any power of two
52 The block size of a file system is recorded in the
53 file system's super-block
54 so it is possible for file systems with different block sizes
57 the file system is created;
58 it cannot be subsequently changed without rebuilding the file system.
60 The new file system organization divides a disk partition
72 the traditional file system's free list.
74 is allocated at file system creation time.
100 this would not work for file systems
104 the cylinder group information begin on a file system block boundary.
105 To start the cylinder group on a file system block boundary,
106 file systems with block sizes larger than 8 kilobytes
109 Without knowing the size of the file system blocks,
117 in a single disk transaction, greatly increasing file system throughput.
118 As an example, consider a file in the new file system
120 In the old file system this file would be composed of 1024 byte blocks.
121 By increasing the block size, disk accesses in the new file
130 file systems are composed of many small files.
132 Table 1 shows the effect of file system
133 block size on the amount of wasted space in the file system.
137 The measurements are based on the active user file systems containing
148 807.8 Mb	4.2	Data only, each file starts on 512 byte boundary
149 828.7 Mb	6.9	Data + inodes, 512 byte block UNIX file system
150 866.5 Mb	11.8	Data + inodes, 1024 byte block UNIX file system
151 948.5 Mb	22.4	Data + inodes, 2048 byte block UNIX file system
152 1128.3 Mb	45.6	Data + inodes, 4096 byte block UNIX file system
161 45.6% waste with 4096 byte file system blocks.
165 The new file system accomplishes this goal by allowing the division
166 of a single file system block into one or more
168 The file system fragment size is specified
169 at the time that the file system is created;
170 each file system block can optionally be broken into
179 Figure 1 shows a piece of a map from a 4096/1024 file system.
189 Figure 1 \- Example layout of blocks and fragments in a 4096/1024 file system.
204 On a file system with a block size of 4096 bytes
206 a file is represented by zero or more 4096 byte blocks of data,
208 If a file system block must be fragmented to obtain
212 As an example consider an 11000 byte file stored on
213 a 4096/1024 byte file system.
214 This file would uses two full size blocks and one
217 available at the time the file is created,
220 This remaining fragment can be allocated to another file as needed.
222 Space is allocated to a file when a program does a \fIwrite\fP
224 Each time data is written to a file, the system checks to see if
225 the size of the file has increased*.
227 * A program may be overwriting data in the middle of an existing file
230 If the file needs to be expanded to hold the new data,
237 The file contains no fragmented blocks (and the last
238 block in the file
253 The file contains one or more fragments (and the
270 The problem with expanding a file one fragment at a
275 except for a partial block at the end of the file.
276 Since file systems with different block sizes may reside on
278 the file system interface has been extended to provide
280 For files the optimal size is the block size of the file system
281 on which the file is being accessed.
290 and need the highest possible file system bandwidth.
292 The amount of wasted space in the 4096/1024 byte new file system
294 1024 byte old file system organization.
295 A file system with 4096 byte blocks and 512 byte fragments
297 block UNIX file system.
298 The new file system uses less space
300 file systems for indexing information for
306 when a new file system's fragment size
307 equals an old file system's block size.
310 a file system cannot be kept completely full.
311 For each file system there is a parameter, termed
313 gives the minimum acceptable percentage of file system
318 even when the file system is mounted and active.
319 The transfer rates that appear in section 4 were measured on file
322 the file system throughput tends to be cut in half,
323 because of the inability of the file system to localize
324 blocks in a file.
325 If a file system's performance degrades because
336 an old 1024 byte UNIX file system is roughly
337 comparable to a new 4096/512 byte file system
339 (Compare 11.8% wasted with the old file system
341 new file system.)
343 File system parameterization
346 the old file system ignores the parameters of the underlying hardware.
350 A goal of the new file system is to parameterize the
361 Each file system is parameterized so that it can be
366 the new file system tries to allocate new blocks
367 on the same cylinder as the previous block in the same file.
393 Given a block allocated to a file,
395 skip over so that the next block in the file will
429 even when the file system is mounted and active.
430 If a file system is parameterized to lay out blocks with
437 the file system can be parameterized for it
446 The file system layout policies are divided into two distinct parts.
447 At the top level are global policies that use file system
460 Two methods for improving file system performance are to increase
475 resembling the old file system.
484 the inode for each file in a directory.
504 In contrast, the old file system typically requires
505 one disk transfer to fetch the inode for each file in a directory.
508 Since data blocks for a file are typically accessed together,
510 blocks for a file in the same cylinder group,
517 causes future allocations for any file in the cylinder group
523 when a file exceeds 48 kilobytes,
527 at which a file on a 4096 byte block file system first
532 file has used about 25% of the data blocks in a cylinder group.
533 In observing the new file system in day to day use, the heuristics appear
581 File systems that are parameterized to maintain at least
583 File systems that are run without maintaining any free