1BTRFS-CONVERT(8) BTRFS BTRFS-CONVERT(8)
2
3
4
6 btrfs-convert - convert from ext2/3/4 or reiserfs filesystem to btrfs
7 in-place
8
10 btrfs-convert [options] <device>
11
13 The btrfs-convert tool can be used to convert existing source filesys‐
14 tem image to a btrfs filesystem in-place. The original filesystem im‐
15 age is accessible in subvolume named like ext2_saved as file image.
16
17 Supported filesystems:
18
19 • ext2, ext3, ext4 -- original feature, always built in
20
21 • reiserfs -- since version 4.13, optionally built, requires libreis‐
22 erfscore 3.6.27
23
24 • ntfs -- external tool https://github.com/maharmstone/ntfs2btrfs
25
26 The list of supported source filesystem by a given binary is listed at
27 the end of help (option --help).
28
29 WARNING:
30 If you are going to perform rollback to the original filesystem, you
31 should not execute btrfs balance command on the converted filesys‐
32 tem. This will change the extent layout and make btrfs-convert un‐
33 able to rollback.
34
35 The conversion utilizes free space of the original filesystem. The ex‐
36 act estimate of the required space cannot be foretold. The final btrfs
37 metadata might occupy several gigabytes on a hundreds-gigabyte filesys‐
38 tem.
39
40 If the ability to rollback is no longer important, the it is recom‐
41 mended to perform a few more steps to transition the btrfs filesystem
42 to a more compact layout. This is because the conversion inherits the
43 original data blocks' fragmentation, and also because the metadata
44 blocks are bound to the original free space layout.
45
46 Due to different constraints, it is only possible to convert filesys‐
47 tems that have a supported data block size (ie. the same that would be
48 valid for mkfs.btrfs). This is typically the system page size (4KiB on
49 x86_64 machines).
50
51 BEFORE YOU START
52
53 The source filesystem must be clean, eg. no journal to replay or no re‐
54 pairs needed. The respective fsck utility must be run on the source fi‐
55 lessytem prior to conversion. Please refer to the manual pages in case
56 you encounter problems.
57
58 For ext2/3/4:
59
60 # e2fsck -fvy /dev/sdx
61
62 For reiserfs:
63
64 # reiserfsck -fy /dev/sdx
65
66 Skipping that step could lead to incorrect results on the target
67 filesystem, but it may work.
68
69 REMOVE THE ORIGINAL FILESYSTEM METADATA
70
71 By removing the subvolume named like ext2_saved or reiserfs_saved, all
72 metadata of the original filesystem will be removed:
73
74 # btrfs subvolume delete /mnt/ext2_saved
75
76 At this point it is not possible to do a rollback. The filesystem is
77 usable but may be impacted by the fragmentation inherited from the
78 original filesystem.
79
80 MAKE FILE DATA MORE CONTIGUOUS
81
82 An optional but recommended step is to run defragmentation on the en‐
83 tire filesystem. This will attempt to make file extents more contigu‐
84 ous.
85
86 # btrfs filesystem defrag -v -r -f -t 32M /mnt/btrfs
87
88 Verbose recursive defragmentation (-v, -r), flush data per-file (-f)
89 with target extent size 32MiB (-t).
90
91 ATTEMPT TO MAKE BTRFS METADATA MORE COMPACT
92
93 Optional but recommended step.
94
95 The metadata block groups after conversion may be smaller than the de‐
96 fault size (256MiB or 1GiB). Running a balance will attempt to merge
97 the block groups. This depends on the free space layout (and fragmen‐
98 tation) and may fail due to lack of enough work space. This is a soft
99 error leaving the filesystem usable but the block group layout may re‐
100 main unchanged.
101
102 Note that balance operation takes a lot of time, please see also
103 btrfs-balance(8).
104
105 # btrfs balance start -m /mnt/btrfs
106
108 --csum <type>, --checksum <type>
109 Specify the checksum algorithm. Default is crc32c. Valid values
110 are crc32c, xxhash, sha256 or blake2. To mount such filesystem
111 kernel must support the checksums as well.
112
113 -d|--no-datasum
114 disable data checksum calculations and set the NODATASUM file
115 flag, this can speed up the conversion
116
117 -i|--no-xattr
118 ignore xattrs and ACLs of files
119
120 -n|--no-inline
121 disable inlining of small files to metadata blocks, this will
122 decrease the metadata consumption and may help to convert a
123 filesystem with low free space
124
125 -N|--nodesize <SIZE>
126 set filesystem nodesize, the tree block size in which btrfs
127 stores its metadata. The default value is 16KiB (16384) or the
128 page size, whichever is bigger. Must be a multiple of the sec‐
129 torsize, but not larger than 65536. See mkfs.btrfs(8) for more
130 details.
131
132 -r|--rollback
133 rollback to the original ext2/3/4 filesystem if possible
134
135 -l|--label <LABEL>
136 set filesystem label during conversion
137
138 -L|--copy-label
139 use label from the converted filesystem
140
141 -O|--features <feature1>[,<feature2>...]
142 A list of filesystem features enabled the at time of conversion.
143 Not all features are supported by old kernels. To disable a fea‐
144 ture, prefix it with ^. Description of the features is in sec‐
145 tion FILESYSTEM FEATURES of mkfs.btrfs(8).
146
147 To see all available features that btrfs-convert supports run:
148
149 btrfs-convert -O list-all+
150
151 -p|--progress
152 show progress of conversion (a heartbeat indicator and number of
153 inodes processed), on by default
154
155 --no-progress
156 disable progress and show only the main phases of conversion
157
158 --uuid <SPEC>
159 set the FSID of the new filesystem based on 'SPEC':
160
161 • new - (default) generate UUID for the FSID of btrfs
162
163 • copy - copy UUID from the source filesystem
164
165 • UUID - a conforming UUID value, the 36 byte string representa‐
166 tion
167
169 btrfs-convert will return 0 if no error happened. If any problems hap‐
170 pened, 1 will be returned.
171
173 mkfs.btrfs(8)
174
176 2022
177
178
179
180
1815.18 May 25, 2022 BTRFS-CONVERT(8)