1 Backup Format Description
2 for an LFS-Inspired Backup Solution
4 NOTE: This format specification is not yet complete. Right now the code
5 provides the best documentation of the format.
7 This document simply describes the snapshot format. It is described
8 from the point of view of a decompressor which wishes to restore the
9 files from a snapshot. It does not specify the exact behavior required
10 of the backup program writing the snapshot.
12 This document does not explain the rationale behind the format; for
19 In several places in the LBS format, a cryptographic checksum may be
20 used to allow data integrity to be verified. At the moment, only the
21 SHA-1 checksum is supported, but it is expected that other algorithms
22 will be supported in the future.
24 When a checksum is called for, the checksum is always stored in a text
25 format. The general format used is
26 <algorithm>=<hexdigits>
28 <algorithm> identifies the checksum algorithm used, and allows new
29 algorithms to be added later. At the moment, the only permissible value
30 is "sha1", indicating a SHA-1 checksum.
32 <hexdigits> is a sequence of hexadecimal digits which encode the
33 checksum value. For sha1, <hexdigits> should be precisely 40 digits
36 A sample checksum string is
37 sha1=67049e7931ad7db37b5c794d6ad146c82e5f3187
40 SEGMENTS & OBJECTS: STORAGE AND NAMING
41 ======================================
43 An LBS snapshot consists, at its base, of a collection of /objects/:
44 binary blobs of data, much like a file. Higher layers interpret the
45 contents of objects in various ways, but the lowest layer is simply
46 concerned with storing and naming these objects.
48 An object is a sequence of bytes (octets) of arbitrary length. An
49 object may contain as few as zero bytes (though such objects are not
50 very useful). Object sizes are potentially unbounded, but it is
51 recommended that the maximum size of objects produced be on the order of
52 megabytes. Files of essentially unlimited size can be stored in an LBS
53 snapshot using objects of modest size, so this should not cause any real
56 For storage purposes, objects are grouped together into /segments/.
57 Segments use the TAR format; each object within a segment is stored as a
58 separate file. Segments are named using UUIDs (Universally Unique
59 Identifiers), which are 128-bit numbers. The textual form of a UUID is
60 a sequence of lowercase hexadecimal digits with hyphens inserted at
61 fixed points; an example UUID is
62 a704eeae-97f2-4f30-91a4-d4473956366b
63 This segment could be stored in the filesystem as a file
64 a704eeae-97f2-4f30-91a4-d4473956366b.tar
65 The UUID used to name a segment is assigned when the segment is created.
67 Filters can be layered on top of the segment storage to provide
68 compression, encryption, or other features. For example, the example
69 segment above might be stored as
70 a704eeae-97f2-4f30-91a4-d4473956366b.tar.bz2
72 a704eeae-97f2-4f30-91a4-d4473956366b.tar.gpg
73 if the file data had been filtered through bzip2 or gpg, respectively,
74 before storage. Filtering of segment data is outside the scope of this
75 format specification, however; it is assumed that if filtering is used,
76 when decompressing the unfiltered data can be recovered (yielding data
79 Objects within a segment are numbered sequentially. This sequence
80 number is then formatted as an 8-digit (zero-padded) hexadecimal
81 (lowercase) value. The fully qualified name of an object consists of
82 the segment name, followed by a slash ("/"), followed by the object
83 sequence number. So, for example
84 a704eeae-97f2-4f30-91a4-d4473956366b/000001ad
87 Within the segment TAR file, the filename used for each object is its
88 fully-qualified name. Thus, when extracted using the standard tar
89 utility, a segment will produce a directory with the same name as the
90 segment itself, and that directory will contain a set of
91 sequentially-numbered files each storing the contents of a single
94 NOTE: When naming an object, the segment portion consists of the UUID
95 only. Any extensions appended to the segment when storing it as a file
96 in the filesystem (for example, .tar.bz2) are _not_ part of the name of
99 There are two additional components which may appear in an object name;
102 First, a checksum may be added to the object name to express an
103 integrity constraint: the referred-to data must match the checksum
104 given. A checksum is enclosed in parentheses and appended to the object
106 a704eeae-97f2-4f30-91a4-d4473956366b/000001ad(sha1=67049e7931ad7db37b5c794d6ad146c82e5f3187)
108 Secondly, an object may be /sliced/: a subset of the bytes actually
109 stored in the object may be selected to be returned. The slice syntax
112 where <start> is the first byte to return (as a decimal offset) and
113 <length> specifies the number of bytes to return (again in decimal). It
114 is invalid to select using the slice syntax a range of bytes that does
115 not fall within the original object. The slice specification should be
116 appended to an object name, for example:
117 a704eeae-97f2-4f30-91a4-d4473956366b/000001ad[264+1000]
118 selects only bytes 264..1263 from the original object.
120 Both a checksum and a slice can be used. In this case, the checksum is
121 given first, followed by the slice. The checksum is computed over the
122 original object contents, before slicing.
125 FILE METADATA LISTING
126 =====================
128 A snapshot stores two distinct types of data into the object store
129 described above: data and metadata. Data for a file may be stored as a
130 single object, or the data may be broken apart into blocks which are
131 stored as separate objects. The file /metadata/ log (which may be
132 spread across multiple objects) specifies the names of the files in a
133 snapshot, metadata about them such as ownership and timestamps, and
134 gives the list of objects that contain the data for the file.
136 The metadata log consists of a set of stanzas, each of which are
137 formatted somewhat like RFC 822 (email) headers. An example is:
140 checksum: sha1=11bd6ec140e4ec3110a91e1dd0f02b63b701421f
141 data: 2f46bce9-4554-4a60-a4a2-543637bd3989/000001f7
149 The meanings of all the fields are described later. A blank line
150 separates stanzas with information about different files. In addition
151 to regular stanzas, the metadata listing may contain a line containing
152 an object reference prefixed with "@". Such a line indicates that the
153 contents of the referenced object should be fetched and parsed as a
154 metadata listing at this point, prior to continuing to parse the current
157 Several common encodings are used for various fields. The encoding used
158 for each field is specified in the field listing that follows.
159 encoded string: An arbitrary string (octet sequence), with bytes
160 optionally escaped by replacing a byte with %xx, where "xx" is a
161 hexadecimal representation of the byte replaced. For example,
162 space can be replaced with "%20". This is the same escaping
163 mechanism as used in URLs.
164 integer: An integer, which may be written in decimal, octal, or
165 hexadecimal. Strings starting with 0 are interpreted as octal,
166 and those starting with 0x are intepreted as hexadecimal.
168 Common fields (required in all stanzas):
169 name [encoded string]: Full path of the file archived.
170 user [special]: The user ID of the file, as an integer, optionally
171 followed by a space and the corresponding username, as an
172 escaped string enclosed in parentheses.
173 group [special]: The group ID which owns the file. Encoding is the
174 same as for the user field: an integer, with an optional name in
175 parentheses following.
176 mode [integer]: Unix mode bits for the file.
177 type [special]: A single character which indicates the type of file.
178 The type indicators are meant to be consistent with the
179 characters used to indicate file type in a directory listing:
187 mtime [integer]: Modification time of the file.
189 Optional common fields:
190 links [integer]: Number of hard links to this file, generally only
191 reported if greater than 1.
192 inode [string]: String specifying the inode number of this file when
193 it was dumped. If "links" is greater than 1, then searching for
194 other files that have an identical "inode" value can be used to
195 determine which files should be hard-linked together when
196 restoring. The inode field should be treated as an opaque
197 string and compared for equality as such; an implementation may
198 choose whatever representation is convenient. The format
199 produced by the standard tool is <major>/<minor>/<inode> (where
200 <major> and <minor> specify the device of the containing
201 filesystem and <inode> is the inode number of the file).
203 Special fields used for regular files:
204 checksum [string]: Checksum of the file contents.
205 size [integer]: Size of the file, in bytes.
206 data [reference list]: Whitespace-separated list of object
207 references. The referenced data, when concatenated in the
208 listed order, will reconstruct the file data. Any reference
209 that begins with a "@" character is an indirect reference--the
210 given object includes a whitespace-separated list of object
211 references which should be parsed in the same manner as the data
218 The snapshot descriptor is a small file which describes a single
219 snapshot. It is one of the few files which is not stored as an object
220 in the segment store. It is stored as a separate file, in plain text,
221 but in the same directory as segments are stored.
223 The name of snapshot descriptor file is
224 snapshot-<scheme>-<timestamp>.lbs
225 <scheme> is a descriptive text which can be used to distinguish several
226 logically distinct sets of snapshots (such as snapshots for two
227 different directory trees) that are being stored in the same location.
228 <timestamp> gives the date and time the snapshot was taken; the format
229 is %Y%m%dT%H%M%S (20070806T092239 means 2007-08-06 09:22:39).
231 The contents of the descriptor are a set of RFC 822-style headers (much
232 like the metadata listing). The fields which are defined are:
233 Format: The string "LBS Snapshot v0.2" which identifies this file as
234 an LBS backup descriptor. The version number (v0.2) might
235 change if there are changes to the format. It is expected that
236 at some point, once the format is stabilized, the version
237 identifier will be changed to v1.0.
238 Producer: A informative string which identifies the program that
240 Date: The date the snapshot was produced. This matches the
241 timestamp encoded in the filename, but is written out in full.
242 A timezone is given. For example: "2007-08-06 09:22:39 -0700".
243 Scheme: The <scheme> field from the descriptor filename.
244 Segments: A whitespace-seprated list of segment names. Any segment
245 which is referenced by this snapshot must be included in the
246 list, since this list can be used in garbage-collecting old
247 segments, determining which segments need to be downloaded to
248 completely reconstruct a snapshot, etc.
249 Root: A single object reference which points to the metadata
250 listing for the snapshot.