2 - SNAPSHOT FORMAT CHANGES:
3 - Timestamps encoded into the top-level snapshot file name are
4 now UTC instead of the local time zone. This should avoid
5 ambiguity and ensure sorted order is chronological order in
6 the face of time zone changes.
7 - The format identifier is now "Format: Cumulus Snapshot v0.nn"
8 instead of "Format: LBS Snapshot v0.nn" (but the updated tools
9 will accept either variation).
10 - Additional checksum algorithms: sha224, sha256 (sha1 is still
12 - LOCAL DATABASE CHANGES:
13 - Segment utilization data (per snapshot) is tracked
14 differently, to allow better segment cleaning decisions.
15 - New, greatly-enhanced file include/exclude filtering language.
16 This is based on the filter language is rsync (though simplified)
17 and allows glob-like patterns. It also supports filter rules
18 merged at backup-time from files in the file system; this allows,
19 for example, users to control which files in their home
20 directories are included in backups. See doc/exclude.rst for
22 - A new Python-based command-line interface to Cumulus. Eventually
23 this should be the only command directly invoked by the user; it
24 will automate backup expiration, segment cleaning, backup
25 mirroring, etc. Configuration options are read from a config file
26 so they can be applied consistently.
29 - Make a release that packages up various long-existing patches.
30 - Add FTP and SFTP storage backends (from Ralf Schlatterbeck and
35 - Rework storage layer, to make it easier to support local files,
36 S3, and in the future more storage backends. cumulus-util now
37 permits URLS for specifying storage locations, including s3:// for
39 - The script-level interface between the main cumulus binary and
40 remote storage has changed. It should not yet be considered
42 - Add a tool (cumulus-sync) for copying snapshots between storage
44 - Add a (mostly proof-of-concept) FUSE interface for accessing
46 - Implement a basic garbage-collection command for deleting unused
47 segments. There still needs to be a way to select snapshots to
48 delete (perhaps based on some type of schedule).
49 - Assorted minor bugfixes.
52 - SNAPSHOT FORMAT CHANGE: The snapshot format has been extended to
53 allow the local database to be partially rebuilt from the contents
54 of the metadata log. This may be useful for disaster recovery and
55 other scenarios. The changes are small, but old tools will not
57 - Add the --rebuild-statcache option which will re-read all files
58 instead of depending on the statcache file to be correct. This
59 will additionally print warnings if a file has changed but would
60 not have been detected as changed based on the statcache, and will
61 add subfile incremental signatures for existing files that do not
65 - LOCAL DATABASE CHANGE: The addition of subfile incrementals has
66 requires an extension to the local database. The upgrade script
67 contrib/upgrade0.7-localdb.sql should be run prior to running
68 backups with this version.
69 - Name change: the system is now known as Cumulus (replacing the old
70 name of "LBS"). Some traces of the old name still remain.
71 - Initial support for direct backups to remote storage. A sample
72 script is provided for backing up to Amazon S3. Other scripts
73 should be simple to write. The interface should not yet be
74 considered stable, and may change in a future release.
75 - Support for efficient subfile incremental backups. Changes to a
76 small portion of a file do not require re-uploading substantially
77 all of the file, even in cases where there is an insertion or
78 deletion that shifts most of the file contents.
79 - Allow selected files/directories to be restored, not simply an
80 entire snapshot. Additionally, restore files in an order that
81 should optimize performance (restore files based on how they are
82 grouped into segments, instead of lexicographic order).
83 Currently, the implementation of these changes requires that all
84 metadata be loaded into memory when the restore tool runs, so it
85 is more memory-intensive than the old version. This may be fixed
86 in a future version; in the meantime, if the current restore tool
87 requires too much memory, try the old restore tool or the
89 - Add a verbose ("-v") flag. By default, files will not be listed
90 as they are backed up. The old behavior can be turned back on
94 - SNAPSHOT FORMAT CHANGE: A few minor tweaks have been made to the
95 snapshot format. There is nothing substantial--this is primarily
96 a cleaning up of the format before use becomes more widespread.
97 An efficient and explicit means for representing sparse files has
98 been added to the format. Updated tools will be needed to
99 properly understand the new snapshots. Tools will continue to be
100 able to read the old snapshot format.
101 - LOCAL DATABASE CHANGE: The local database has been changed so that
102 segment utilization data for past snapshots is stored more
103 efficiently. The script in contrib/upgrade0.6-localdb.sql must be
104 run on a database to upgrade it prior to running backups with the
106 - The statcache implementation has been reworked, and the format
107 changed. The first backup will run more slowly since the old
108 statcache data will be ignored. New statcache data will be
109 written to statcache2. The old statcache file can be deleted.
110 - Metadata is shared between snapshots where possible. The
111 --full-metadata option can be used to disable this (completely
112 writing out all metadata with the new snapshot).
113 - Add an "intent" field to snapshots, which specifies informally how
114 long a snapshot is intended to be kept. For example, 1 can be
115 used for daily snapshots, and 7 for weekly snapshots. Segment
116 cleaning is now partly guided by intent values, but tuning of the
117 cleaning algorithms is not yet finished.
120 - Have the lbs-util tool check the version number when reading a
121 snapshot, and signal an error if the format is not recognized.
122 - Include a sample script, contrib/parity-gen, for creating
123 RAID-like parity sets to recover from some segment corruption.
124 The par2 command (http://parchive.sourceforge.net/) is used to
125 actually generate the parity sets; the parity-gen script simply
126 automates maintaining the parity sets.
129 - Much improved Python interface for accessing and manipulating LBS
130 archives and local database information. The interface should not
131 yet be considered completely stable.
132 - Python implementation of lbs-util now includes most of the
133 features of the Perl implementation, plus some other new features.
134 The Perl library and utility are deprecated and have been removed
136 - Preliminary snapshot restore support in lbs-util. This is still
137 not yet extensively tested.
140 - Documentation improvements: a getting started README, and a
141 description of some of the implementation details.
142 - Include an example filter script, lbs-filter-gpg, for encrypting
143 segment data as it is written.
144 - Add support for signed snapshot descriptor files (via the
145 --signature-filter option). Due to the checksums contained in the
146 snapshot descriptor, the signature covers the entire contents of
148 - Preview release of lbs-util.py, a Python-based program for
149 managing LBS snapshots. It currently implements a very simple
150 form of automatic segment cleaning, though this isn't much tested.
151 - Improved segment repacking support; after cleaning segments,
152 objects can be written out in different groups based on age. This
153 may help with long-term segment cleaning efficiency, but probably
157 - LOCAL DATABASE CHANGE: A checksums file is now written out along
158 with the snapshot descriptor that contains checksums of segments
159 used in a snapshot. This will allow quick integrity checks of a
160 snapshot, without needing to decompress or decrypt the segment
161 files. The local database schema was modified to store the
162 checksums for segments.
163 - Build dependence on libtar has been dropped. All necessary
164 support for writing TAR files is now directly included with the
166 - The snapshot format is now documented, at least in part. See
168 - Snapshots now include link count and inode number for files with
169 multiple hard links, so that a restore program could determine
170 which files should be hard linked on restore. The reference
171 restore.pl script does not use this.
172 - Bugfix: Print a help message instead of crashing if no files are
174 - Bugfix: File descriptors of files being backed up were closed
175 twice. In some cases this might have led to an unrelated file
176 being closed (if the file descriptor was quickly re-used).
177 - Preview of a new lbs-util command for maintaining snapshots.
178 Functionality is currently limited.
181 - Update reference restore.pl script to handle octal/hexadecimal
182 (format change introduced in 0.2).
183 - Better support for mixing multiple backup schemes: the statcache
184 file is now per-scheme, so performance should be better in cases
185 where multiple different backups (of different source directories)
186 are made, all sharing the same local database directory.
189 - Bugfix: Do not print an error when only one directory is specified
193 - SNAPSHOT FORMAT CHANGE: Metadata logs in backups can now contain
194 octal and hexadecimal values. Old restore programs may not be
195 able to read these new backups. The reference restore.pl script
196 has not yet been updated.
197 - LOCAL DATABASE CHANGE: Backups may be assigned a name with the
198 --scheme= command-line option, to allow different backup sets to
199 be mixed in the same directory. The database schema for the local
201 - Improved selection (inclusion/exclusion) of files to be included
203 - Device major/minor numbers are dumped for block and character