Debian Codegraph @

upstream/debian-unstable/man8/xfs_fsr.8

raw
1 .TH xfs_fsr 8
2 .SH NAME
3 xfs_fsr \- filesystem reorganizer for XFS
4 .SH SYNOPSIS
5 .nf
6 \f3xfs_fsr\f1 [\f3\-vdg\f1] \c
7 [\f3\-t\f1 seconds] [\f3\-p\f1 passes] [\f3\-f\f1 leftoff] [\f3\-m\f1 mtab]
8 \f3xfs_fsr\f1 [\f3\-vdg\f1] \c
9 [xfsdev | file] ...
10 .br
11 .B xfs_fsr \-V
12 .fi
13 .SH DESCRIPTION
14 .I xfs_fsr
15 is applicable only to XFS filesystems.
16 .PP
17 .I xfs_fsr
18 improves the organization of mounted filesystems.
19 The reorganization algorithm operates on one file at a time,
20 compacting or otherwise improving the layout of
21 the file extents (contiguous blocks of file data).
22 .PP
23 The following options are accepted by
24 .IR xfs_fsr .
25 The
26 .BR \-m ,
27 .BR \-t ,
28 and
29 .B \-f
30 options have no meaning if any filesystems
31 or files are specified on the command line.
32 .TP 13
33 .BI \-m " mtab"
34 Use this file for the list of filesystems to reorganize.
35 The default is to use
36 .IR /etc/mtab .
37 .TP
38 .BI \-t " seconds"
39 How long to reorganize.
40 The default is 7200 seconds (2 hours).
41 .TP
42 .BI \-p " passes"
43 Number of passes before terminating global re-org.
44 The default is 10 passes.
45 .TP
46 .BI \-f " leftoff"
47 Use this file instead of
48 .I /var/tmp/.fsrlast
49 to read the state of where to start and as the file
50 to store the state of where reorganization left off.
51 .TP
52 .B \-v
53 Verbose.
54 Print cryptic information about
55 each file being reorganized.
56 .TP
57 .B \-d
58 Debug. Print even more cryptic information.
59 .TP
60 .B \-g
61 Print to syslog (default if stdout not a tty).
62 .TP
63 .B \-V
64 Prints the version number and exits.
65
66 .PP
67 When invoked with no arguments
68 .I xfs_fsr
69 reorganizes all regular files in all mounted filesystems.
70 .I xfs_fsr
71 makes many cycles over
72 .I /etc/mtab
73 each time making a single pass over each XFS filesystem.
74 Each pass goes through and selects files
75 that have the largest number of extents. It attempts
76 to defragment the top 10% of these files on each pass.
77 .PP
78 It runs for up to two hours after which it records the filesystem
79 where it left off, so it can start there the next time.
80 This information is stored in the file
81 .I /var/tmp/.fsrlast_xfs.
82 If the information found here
83 is somehow inconsistent or out of date
84 it is ignored
85 and reorganization starts at the beginning of the first
86 filesystem found in
87 .IR /etc/mtab .
88 .PP
89 .I xfs_fsr
90 can be called with one or more arguments
91 naming filesystems (block device name),
92 and files to reorganize.
93 In this mode
94 .I xfs_fsr
95 does not read or write
96 .I /var/tmp/.fsrlast_xfs
97 nor does it run for a fixed time interval.
98 It makes one pass through each specified regular file and
99 all regular files in each specified filesystem.
100 A command line name referring to a symbolic link
101 (except to a file system device),
102 FIFO, or UNIX domain socket
103 generates a warning message, but is otherwise ignored.
104 While traversing the filesystem these types
105 of files are silently skipped.
106 .SH FILES
107 .PD 0
108 .TP 21
109 /etc/mtab
110 contains default list of filesystems to reorganize.
111 .TP 21
112 /var/tmp/.fsrlast_xfs
113 records the state where reorganization left off.
114 .PD
115 .SH "SEE ALSO"
116 xfs_fsr(8),
117 mkfs.xfs(8),
118 xfs_ncheck(8),
119 xfs(5).
120 .SH "NOTES"
121 .I xfs_fsr
122 improves the layout of extents for each file by copying the entire
123 file to a temporary location and then interchanging the data extents
124 of the target and temporary files in an atomic manner.
125 This method requires that enough free disk space be available to copy
126 any given file and that the space be less fragmented than the original
127 file.
128 It also requires the owner of the file to have enough remaining
129 filespace quota to do the copy on systems running quotas.
130 .I xfs_fsr
131 generates a warning message if space is not sufficient to improve
132 the target file.
133 .PP
134 A temporary file used in improving a file given on the command line
135 is created in the same parent directory of the target file and
136 is prefixed by the string '\f3.fsr\f1'.
137 The temporary files used in improving an entire XFS device are stored
138 in a directory at the root of the target device and use the same
139 naming scheme.
140 The temporary files are unlinked upon creation so data will not be
141 readable by any other process.
142 .PP
143 .I xfs_fsr
144 does not operate on files that are currently mapped in memory.
145 A 'file busy' error can be seen for these files if the verbose
146 flag (\f3-v\f1) is set.
147 .PP
148 Files marked as no\-defrag will be skipped. The
149 .IR xfs_io (8)
150 chattr command with the f attribute can be used to set or clear
151 this flag. Files and directories created in a directory with the
152 no\-defrag flag will inherit the attribute.
153 .PP
154 An entry in
155 .I /etc/mtab
156 or the file specified using the
157 .B \-m
158 option must have the
159 .B rw
160 option specified for read and write access.
161 If this option is not present, then
162 .I xfs_fsr
163 skips the
164 filesystem described by that line.
165 See the
166 .IR fstab (5)
167 reference page for
168 more details.
169 .PP
170 In general we do not foresee the need to run
171 .I xfs_fsr
172 on system partitions such as
173 .IR / ,
174 .I /boot
175 and
176 .I /usr
177 as in general these will not suffer from fragmentation.
178 There are also issues with defragmenting files
179 .IR lilo (8)
180 uses to boot your system. It is recommended that these files
181 should be flagged as no\-defrag with the
182 .IR xfs_io (8)
183 chattr command. Should these files be moved by
184 .I xfs_fsr
185 then you must rerun
186 .I lilo
187 before you reboot or you may have an unbootable system.