Contents of /xml/htdocs/doc/en/handbook/hb-install-filesystems.xml

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1.15 - (hide annotations) (download) (as text)
Thu Jan 23 19:16:03 2014 UTC (4 years, 5 months ago) by swift
Branch: MAIN
Changes since 1.14: +4 -4 lines
File MIME type: application/xml
Fix bug #498580 - inode count needs to be higher on at least 5Gb as well, move it to 8

1 neysx 1.1 <?xml version='1.0' encoding="UTF-8"?>
2 swift 1.15 <!-- $Header: /var/cvsroot/gentoo/xml/htdocs/doc/en/handbook/hb-install-filesystems.xml,v 1.14 2013/12/17 10:15:36 swift Exp $ -->
3 neysx 1.1 <!DOCTYPE included SYSTEM "/dtd/guide.dtd">
5     <included>
7 swift 1.15 <version>13</version>
8     <date>2014-01-23</date>
9 neysx 1.1
10     <section id="filesystemsdesc">
11     <title>Filesystems</title>
12     <body>
14     <p test="contains('x86 Alpha',func:keyval('arch'))">
15 swift 1.7 The Linux kernel supports various filesystems. We'll explain ext2, ext3, ext4,
16 neysx 1.1 ReiserFS, XFS and JFS as these are the most commonly used filesystems on Linux
17     systems.
18     </p>
20     <p test="func:keyval('arch')='IA64'">
21     The Linux kernel supports various filesystems. We'll explain vfat, ext2, ext3,
22 swift 1.8 ext4, ReiserFS, XFS and JFS as these are the most commonly used filesystems on
23     Linux systems.
24 neysx 1.1 </p>
26     <p test="func:keyval('arch')='AMD64'">
27     Several filesystems are available. Some of them are found stable on the amd64
28     architecture, others aren't. The following filesystems are found to be stable:
29 swift 1.7 ext2, ext3, ext4 and XFS. JFS and ReiserFS may work but need more testing. If
30     you're really adventurous you can try the other filesystems.
31 neysx 1.1 </p>
33     <p test="func:keyval('arch')='arm'">
34     Several filesystems are available. Some of them are found stable on the arm
35 swift 1.8 architecture, others aren't. ext2, ext3 and ext4 are found to be stable. JFS,
36     XFS and ReiserFS may work but need more testing. If you're really adventurous
37     you can try the other filesystems.
38 neysx 1.1 </p>
40     <p test="func:keyval('arch')='HPPA'">
41 swift 1.9 Several filesystems are available. Ext2, ext3, ext4, XFS and reiserfs are found
42 neysx 1.1 stable on the HPPA architecture. The others are very experimental.
43     </p>
45     <p test="func:keyval('arch')='MIPS'">
46 swift 1.7 Several filesystems are available. ReiserFS, EXT2, EXT3 and EXT4 are found
47     stable on the MIPS architectures, others are experimental.
48 neysx 1.1 </p>
50     <p test="func:keyval('arch')='PPC'">
51     Several filesystems are available for use on the PowerPC architecture including
52 swift 1.12 ext2, ext3, ext4, ReiserFS and XFS, each with their strengths and faults.
53 neysx 1.1 </p>
55     <note test="func:keyval('arch')='PPC64'">
56 swift 1.12 Several filesystems are available. ext2, ext3, ext4 and ReiserFS support is built in
57 neysx 1.1 the Installation CD kernels. JFS and XFS support is available through kernel
58     modules.
59     </note>
61     <p test="func:keyval('arch')='SPARC'">
62     Several filesystems are available, some are known to be stable on the
63 swift 1.8 SPARC architecture. Ext2, ext3 and ext4, for example, are known to work well.
64 neysx 1.1 Alternate filesystems may not function correctly.
65     </p>
67     <note test="func:keyval('arch')='Alpha'">
68     <c>aboot</c> only supports booting from <b>ext2</b> and <b>ext3</b>
69     partitions.
70     </note>
72     </body>
73     <body>
75     <p test="func:keyval('arch')='IA64'">
76     <b>vfat</b> is the MS-DOS filesystem, updated to allow long filenames. It is
77     also the only filesystem type that the EFI firmware on ia64 systems
78     understands. The boot partition on ia64 systems should always be vfat, but for
79     your data partitions you should use one of the other filesystems listed below.
80     </p>
82     <p>
83     <b>ext2</b> is the tried and true Linux filesystem but doesn't have metadata
84     journaling, which means that routine ext2 filesystem checks at startup time can
85     be quite time-consuming. There is now quite a selection of newer-generation
86     journaled filesystems that can be checked for consistency very quickly and are
87     thus generally preferred over their non-journaled counterparts. Journaled
88     filesystems prevent long delays when you boot your system and your filesystem
89 swift 1.14 happens to be in an inconsistent state.
90 neysx 1.1 </p>
92     <p>
93     <b>ext3</b> is the journaled version of the ext2 filesystem, providing metadata
94     journaling for fast recovery in addition to other enhanced journaling modes like
95     full data and ordered data journaling. It uses an HTree index that enables high
96 nightmorph 1.6 performance in almost all situations. In short, ext3 is a very good and
97 swift 1.14 reliable filesystem.
98 neysx 1.1 </p>
100 swift 1.12 <p test="contains('x86 Alpha MIPS AMD64 arm IA64 SPARC HPPA PPC PPC64',func:keyval('arch'))">
101 swift 1.7 <b>ext4</b> is a filesystem created as a fork of ext3 bringing new features,
102     performance improvements and removal of size limits with moderate changes
103     to the on-disk format. It can span volumes up to 1 EB and with maximum file
104     size of 16 TB. Instead of the classic ext2/3 bitmap block allocation ext4 uses
105     <uri link="http://en.wikipedia.org/wiki/Extent_%28file_systems%29">extents</uri>,
106     which improve large file performance and reduce fragmentation. Ext4 also provides
107     more sophisticated block allocation algorithms (delayed allocation and multiblock
108     allocation) giving the filesystem driver more ways to optimise the layout of data
109     on the disk. The ext4 filesystem is a compromise between production-grade code
110     stability and the desire to introduce extensions to an almost decade old
111 swift 1.13 filesystem. Ext4 is the recommended all-purpose all-platform filesystem.
112 swift 1.7 </p>
114 swift 1.14 <p>
115 swift 1.15 If you intend to install Gentoo on a small partition (less than 8GB), then you'll
116 swift 1.14 need to tell ext2, ext3 or ext4 (if available) to reserve enough inodes when you
117     create the filesystem. The <c>mke2fs</c> application uses the "bytes-per-inode"
118     setting to calculate how many inodes a file system should have. By running
119     <c>mke2fs -T small /dev/&lt;device&gt;</c> (ext2) or <c>mke2fs -j -T small
120     /dev/&lt;device&gt;</c> (ext3/ext4) the number of inodes will generally
121     quadruple for a given file system as its "bytes-per-inode" reduces from
122     one every 16kB to one every 4kB. You can tune this even further by using
123     <c>mke2fs -i &lt;ratio&gt; /dev/&lt;device&gt;</c> (ext2) or <c>mke2fs -j
124     -i &lt;ratio&gt; /dev/&lt;device&gt;</c> (ext3/ext4).
125     </p>
127 neysx 1.1 </body>
128     <body test="not(func:keyval('arch')='SPARC')">
130     <p test="not(func:keyval('arch')='PPC')">
131     <b>JFS</b> is IBM's high-performance journaling filesystem. JFS is a light,
132     fast and reliable B+tree-based filesystem with good performance in various
133     conditions.
134     </p>
136     <p>
137     <b>ReiserFS</b> is a B+tree-based journaled filesystem that has good overall
138     performance, especially when dealing with many tiny files at the cost of more
139     CPU cycles. ReiserFS appears to be less maintained than other filesystems.
140     </p>
142     <p>
143     <b>XFS</b> is a filesystem with metadata journaling which comes with a robust
144     feature-set and is optimized for scalability. XFS seems to be less forgiving to
145     various hardware problems.
146     </p>
148     </body>
149     </section>
150     </included>

  ViewVC Help
Powered by ViewVC 1.1.20