Fix restoring of files splitted on several volumes and starting on 2nd or later tapes.
[dump.git] / CHANGES
1 $Id: CHANGES,v 1.319 2010/12/06 14:26:50 stelian Exp $
2
3
4 Changes between versions 0.4b43 and 0.4b44 (released ?????????????)
5 ===================================================================
6
7 1. Recognize ext4 partitions in dump -w commands (Sourceforge
8 bug #3125587, RedHat bug #658890). Thanks to Jan Görig
9 <jgorig@users.sourceforge.net> for reporting the bug and
10 submitting the patch.
11
12 2. Fix an issue with multi volume backups, which breaks
13 restoring of files which are splitted on two or more
14 volumes and are starting on 2nd tape or later (RedHat
15 bug #507948).
16
17 Changes between versions 0.4b42 and 0.4b43 (released June 11, 2010)
18 ===================================================================
19
20 1. Fix a bug in dump making impossible to handle large toc files
21 (> 2 GB). Thanks to X DUGi <xdugi@users.sourceforge.net> for
22 reporting the bug (Sourceforge bug #2820629)
23
24 2. Fix 'restore -x' in multi-volume mode, which caused files being
25 spanned on the first and second volumes to be incorrectly extracted.
26 Thanks to Adam Tkac <atkac@redhat.com> for reporting the bug.
27 (there may be several bug reports about this issue, including
28 Sourceforge bug #2117008, RedHat bugs #232415 and #444958).
29
30 3. Remove -lselinux -lsepol from the standard libraries list,
31 they shouldn't be needed if selinux is not enabled in dump
32 (this also fixes a build failure when selinux libraries are
33 not installed).
34
35 4. Fix restoration of extended attributes of fifos and device nodes.
36 Thanks to Igor Zhbanov <izh1979@gmail.com> for the bug report.
37
38 5. Don't attempt to set extended attributes on files that were not
39 extracted in 'restore -N' mode. Thanks to Jan Görig
40 <jgorig@users.sourceforge.net> for the patch.
41
42 6. Make the dump man page explicit about ext4 support.
43
44 7. Fix a bug in dump which caused EA entries to be unrecognizable by
45 restore in some corner cases. Add a workaround to restore making
46 it able to properly deal with those corrupted EA entries. Thanks
47 to John Austin <jaustin1@users.sourceforge.net> for reporting
48 the bug.
49
50 8. Fix some issues when restoring a dump which was generated using
51 exclusion patterns (either via -e or via the nodump attribute)
52 (see the Debian bug #574667 for details. Thanks to Frédéric
53 Brière <fbriere@fbriere.net> for the bug report and the associated
54 patch.
55
56 9. Fix progress/estimated blocks display in dump when doing really
57 huge backups. Thanks to Steve Bonds <sbonds@users.sourceforge.net>
58 for the bug report and tests. (Sourceforge bug #2987758)
59
60 10. Improve level 1 dump speed by rearanging the "mapdirs" code.
61 Thanks to Andreas Kies <andikies@t-online.de> for finding the
62 bottleneck and sending a patch (Sourceforge bug #2998119).
63
64 11. Use only ctime in order to compare newness instead of both
65 ctime and mtime. This should both speed up dumping time and
66 correct some side-effect bugs (like dumping files with an mtime
67 in the future in every incremental backup). Thanks to
68 Kieran Clancy <codebeard@users.sourceforge.net> for reporting the
69 bug (Sourceforge bug #2999207).
70
71 12. Extract dumped UNIX sockets instead of ignoring them.
72 (Sourceforge bug #3007216).
73
74 13. Compiler warning fixes (mainly warn_unused_result ones).
75
76 Changes between versions 0.4b41 and 0.4b42 (released June 18, 2009)
77 ===================================================================
78
79 1. Fix printout of the 'Connection to' message in dump/restore.
80 Thanks to Dale Wiles <dwiles@users.sourceforge.net> for the
81 bug report and its fix.
82
83 2. Fix dump -w/-W output for filesystems having the last backup
84 level equal to 0. Thanks to Pascal Bouchareine
85 <kalou@users.sourceforge.net> for reporting the bug.
86
87 3. Enable restore to handle restoring onto a different SELinux
88 policy from the dump. Thanks to Tony Nelson
89 <tony_nelson@users.sourceforge.net> for the patch.
90 (see http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=189845
91 for details)
92
93 4. Fix EA set failures when restoring immutable files. Thanks
94 to Andrew Kroeger for the patch.
95
96 5. Fix "mode file too large" errors when restoring huge backups.
97 Thanks to Kenneth Porter <shiva@sewingwitch.com> for the bug
98 report.
99
100 6. Add the ability to link against libtinfo library if present,
101 instead of libtermcap or libncurses libraries. Thanks to
102 Adam Tkac <atkac@redhat.com> for the patch.
103
104 7. When comparing a backup, do not consider to be an error the fact
105 that we encounter files created while dump was in progress.
106 Thanks to Kenneth Porter <shiva@sewingwitch.com> for the bug
107 report.
108
109 8. Force '-a' to be the default value when running 'restore -C'.
110 Thanks to Kenneth Porter <shiva@sewingwitch.com> for the bug
111 report.
112
113 9. Many improvements to the 'cron_dump_to_disk' example.
114 Thanks to Aaron S. Hawley <ashawley@users.sourceforge.net> for
115 the patch.
116
117 10. Fix the QFA generation when extended attributes are backuped
118 along with the inodes. Prior to this fix, the entries for
119 some inodes may get corrupted (duplicated entries or incorrect
120 ones), making the QFA file unusable for repositionning in
121 restore. Many thanks to Kenneth Porter <shiva@sewingwitch.com>
122 for the bug report and the testing of the fix.
123
124 11. Fix EA restoration in debug mode (Sourceforge bug #1986216).
125
126 12. Stefan Auracher <star7@onlinehome.de> noticed that the Tower
127 of Hanoi backup strategy documented in the man page wasn't
128 actually based on the Tower of Hanoi algorithm. The man page
129 was updated accordingly.
130
131 13. Fix missing level information from dump output when using
132 the default level (RedHat bug #493635). Thanks to Adam Tkac
133 <atkac@redhat.com> for the patch.
134
135 14. Fix the dump man page to reflect the fact that the default
136 dump level is 0 not 9 (RedHat bug #356121). Thanks to Adam Tkac
137 <atkac@redhat.com> for the patch.
138
139 15. Fixed a few spelling errors in the man pages (RedHat bug #489853).
140 Thanks to Adam Tkac <atkac@redhat.com> for the patch.
141
142 16. Fixed restore man page and program usage which incorrectly stated
143 that -P and -A were both allowed in the same invocation (RedHat
144 bug #490627). Thanks to Adam Tkac <atkac@redhat.com> for the patch.
145
146 17. Use sys_clone under Linux to share I/O contexts between dump
147 processes, thus drastically increasing the performance of dump
148 under CFQ (which is the default I/O scheduler used in a number of
149 distributions). Thanks to Jeff Moyer <jmoyer@redhat.com> for the
150 patch.
151
152 18. Add (preliminary) ext4 support - thanks to libext2fs which does
153 all the job for us. Thanks to Gertjan van Wingerde
154 <gwingerde@gmail.com> for the patch.
155
156 Changes between versions 0.4b40 and 0.4b41 (released January 2, 2006)
157 =====================================================================
158
159 1. Fix restore of dumped Access Control Lists. The previous
160 code for EA works fine, but ACLs needed conversion from the
161 ext2/3 disk format to posix_acl format before restoring.
162
163 2. Fix some issues with restoration of EA on big endian
164 platforms.
165
166 3. Fix restore when the symtab is over 2GB in size.
167
168 4. Made the directory hash indexing an optional feature,
169 accessible by the '-H' option of restore, and disable it
170 by default.
171
172 5. Fixed dump to not include extended attributes information
173 in the toc (archive) file which confused restore -t.
174
175 Changes between versions 0.4b39 and 0.4b40 (released May 2, 2005)
176 =================================================================
177
178 1. Changed restore to emit warnings (instead of emitting a fatal
179 error) if a file (or a directory) is unavailable for a
180 comparision (if the user doesn't have the necessary permissions
181 to access it for example). Thanks to Kenneth Porter
182 <shiva@sewingwitch.com> for the bug report.
183
184 2. Re-done the 'do not save directory entries to non-dumped inodes
185 (excluded from dump)' feature. The previous implementation
186 worked well for excluded directories but not for regular files.
187 Thanks to Kenneth Porter <shiva@sewingwitch.com> for the bug
188 report.
189
190 3. Fixed a bug in dump where the tape size was miscalculated when
191 the user used -d/-s to specify the tape characteristics. Thanks
192 to Philip Goisman <goisman@physics.arizona.edu> for reporting
193 the bug.
194
195 4. Fixed another bug introduced in restore with the hashtree
196 implementation. This one caused restore to stop saying
197 "removenode: non-empty directory" in some cases.
198
199 5. Added support for dumping and restoring ext2/3 extended
200 attributes (EA), like the access control lists (ACL) or
201 the security labels used by SELinux.
202
203 Changes between versions 0.4b38 and 0.4b39 (released January 21, 2005)
204 ======================================================================
205
206 1. The newly added dump_on_cd_3 example was buggy, replace it
207 with an updated version from Andrew Basterfield
208 <bob@cemetery.homeunix.org>.
209
210 2. Made restore to chdir() back into the initial directory when
211 dumping core while aborting a comparision operation. The
212 previous behaviour was to write the corefile at the root of
213 the directory being compared, which could very well be
214 read only and preventing the corefile generation. Thanks
215 to Kenneth Porter <shiva@sewingwitch.com> for the bug report.
216
217 3. Silenced the failure to call fgetflags() when comparing an
218 entry which has no ext2 attributes (as in lsattr()).
219
220 4. Fix a brown paper bug in restore -C which broke restore and
221 caused modifications on the filesystem being compared
222 (directories containing a file with the same name as the
223 directory get renamed to RSTTMP...). Thanks to Kenneth Porter
224 <shiva@sewingwitch.com> for finding the bug and helping me
225 reproduce it.
226
227 5. Made restore -C force the -N flag (no writing allowed on
228 the disk) in order to prevent more bugs like the above one.
229
230 Changes between versions 0.4b37 and 0.4b38 (released January 7, 2005)
231 =====================================================================
232
233 1. Fix a couple of troff syntax bugs in the man pages.
234 Thanks to Eric Raymond <esr@thyrsus.com> for the patch.
235
236 2. Made restore use either libncurses or libtermcap, depending
237 on which one is available at configure time.
238
239 3. Fixed restore negative size display bug when comparing a
240 dump containing files over 2GB. Thanks to Steve Bonds
241 <sbonds@users.sourceforge.net> for the bug report.
242
243 4. Do not save directory entries to non-dumped inodes
244 (excluded from dump). This will eliminate the 'missing
245 file' warnings when doing 'restore -C'.
246
247 5. Fix dump crash when backuping a huge (2TB) filesystem,
248 due to a bogus calculation on the inode map size.
249 Thanks to Kevin B. Haines <K.B.Haines@rl.ac.uk> for
250 submitting the bug and testing the fix.
251
252 6. Fix a problem in restore where the final \0 in the symbolic
253 link names could have been lost, generating corrupt filenames.
254 Thanks to Kyle Wilson <kyle.wilson@amd.com> for reporting the
255 bug.
256
257 7. Implemented a hash list for the directory names in restore.
258 The linear list used before caused problems in interactive
259 restores when dealing with directories having thousands of
260 entries. Thanks to Brian Ristuccia <bristuccia@starentnetworks.com>
261 for reporting the bug.
262
263 8. Improved restore -C, this time including the directory
264 attributes into the comparision.
265
266 9. Made restore understand tapes containing EA/ACLs (which will
267 be dumped by the next version of dump). In this version
268 extended attributes on the tape are ignored, for full EA/ACL
269 support wait for the next version or try the experimental EA
270 patch.
271
272 Changes between versions 0.4b36 and 0.4b37 (released July 7, 2004)
273 ==================================================================
274
275 1. Added the --enable-staticz configure option which enables
276 dump and restore to be linked with static versions of
277 libz and libbz2 (and dynamic versions of all other libraries).
278 This will make Debian users happy, because libz and libbz2
279 were the only needed libraries living in /usr, all the
280 others live in /lib. In case of system emergency, it is
281 better not to have to rely on an extra filesystem.
282 Thanks to Bdale Garbee <bdale@gag.com> for the suggestion.
283
284 2. Fix compilation on (at least the Linux Debian port to) AMD64.
285 (<ext2fs/ext2_types.h> defines some types (__s64 and __u64)
286 that are also defined by <linux/types.h> (<asm/types.h>) and
287 they conflict).
288
289 3. Make dump's reading of the dumpdates file a bit more robust,
290 preventing dump from crashing when the dumpdates file has
291 been modified by hand.
292
293 4. Fixed some offset calculations in dump code which could
294 lead to "bread lseek errors" on large filesystems. Thanks
295 to Bruce Lowekamp <lowekamp@users.sourceforge.net> for
296 reporting this bug and debugging the issue.
297
298 5. Made dump use the blkid library when searching for devices
299 by label or uuid instead of dump's own routines.
300
301 6. Corrected a bug in dump where a wrong LABEL=... line in
302 /etc/fstab could prevent dump from dumping unrelated
303 filesystems. Thanks to Bruce Lowekamp
304 <lowekamp@users.sourceforge.net> for reporting the bug.
305
306 Changes between versions 0.4b35 and 0.4b36 (released April 21, 2004)
307 ====================================================================
308
309 1. Fixed dump compilation with old gcc versions. Thanks to
310 Mike Castle <dalgoda@users.sourceforge.net> for the patch.
311
312 2. Fixed some warnings (howmany, roundup, powerof2 redefined)
313 when compiling against a recent glibc version.
314
315 3. Fixed a bug in restore preventing the read of a dump tape
316 written with Solaris 7 ufsdump. Thanks to Patrick Higgins
317 <phiggins@transzap.com> for reporting the bug and providing
318 the test case.
319
320 4. Changed dump to enable the creation of volumes of different
321 sizes in a single run (make -B accept a list of values).
322 Patch contributed by Florian Zumbiehl <florz@gmx.de>.
323
324 5. Use the glibc provided minor() and major() macros instead
325 of our own bitmask implementation. This should be safe for
326 when the major/minor namespace will migrate to 32 bits.
327 Thanks to Zhang Jun <zhangjun@nanjing-fnst.com> for reporting
328 the bug.
329
330 6. Made explicit in the dump man page that dump will not create
331 a remote file, it will only write to an already existing one.
332
333 7. Another try at making size estimates better again.
334
335 8. Put back the inconditional running of the end-of-tape
336 script which was changed in 0.434 to be run only when -M
337 or multiple -f were NOT specified. Some users rely on this
338 feature even when it is combined with -M/-f.
339
340 9. Fixed restore when restoring huge backups (where rstdir...
341 temporary files are over 2GB). Thanks to
342 Raphael Jaffey <rjaffey@artic.edu> for reporting this,
343 debugging the issue and making the patch.
344
345 10. Made restore understand FreeBSD UFS2 tapes. Thanks to
346 David <vrtin@users.sourceforge.net> for submitting the bug
347 and providing a test case.
348
349 11. Made dump work with an arbitrary integer level (no more 10
350 levels only limitation). Thanks to Matthew
351 <msvincen@midway.uchicago.edu> for the patch.
352
353 Changes between versions 0.4b34 and 0.4b35 (released December 21, 2003)
354 =======================================================================
355
356 1. Added a note in the dump man page saying that the default
357 blocksize can be 32 if -d is used with a high density argument.
358 Thanks to Antonios Christofides <A.Christofides@itia.ntua.gr>
359 for the patch.
360
361 2. Fixed configure to correctly understand CPPFLAGS, CFLAGS,
362 LDFLAGS environment variables. Thanks to Arcady Genkin
363 <antipode@users.sourceforge.net> for reporting the bug.
364
365 3. Made -e/-E options of dump accept an unlimited number of inodes
366 to be excluded, rather than a hardcoded maximum. Thanks to
367 Dietrich Rothe <d-rothe@users.sourceforge.net> for the patch.
368
369 4. Updated the autoconf system to 2.50. Forced the -D_BSD_SOURCE
370 and -D_USE_BSD_SIGNAL defines in configure in order to solve
371 64bit build problems because quad_t is redifined with a
372 different signature. Thanks to Mike Harris <mharris@redhat.com>
373 for reporting this bug.
374
375 5. Made restore build on Solaris, making possible to
376 restore Linux's "enhanced" tapes. Thanks to Uwe Gohlke
377 <uwe@ugsoft.de> for the patch.
378
379 6. Made an extension in the dump tape format capable of saving
380 MacOSX specific inode extensions. Uwe Gohlke <uwe@ugsoft.de>
381 wrote the extension and contributed the restore code back
382 into this codebase. The same extension mechanism will be
383 used in the future to save ACLs...
384
385 7. Made rmt work correctly with regard to QFA and local/remote
386 files and tapes. The remote access will however work only
387 when the dump provided rmt version is used. If you want to
388 use another rmt server, please do not use the QFA feature.
389 Thanks to Uwe Gohlke <uwe@ugsoft.de> for the patch.
390
391 Changes between versions 0.4b33 and 0.4b34 (released April 18, 2003)
392 ====================================================================
393
394 1. Fixed the INSTALL file to reflect the actual install paths.
395 Thanks to David Ronis <ronis@ronispc.chem.mcgill.ca> for
396 reporting the bug.
397
398 2. Fixed the configure script to only check for headers presence
399 instead of trying to compile them. This should fix issues
400 with old build environments. Thanks to Kari Lempiainen
401 <kari@funky.pp.fi> for reporting the bug.
402
403 3. Fixed restore to correctly ignore sockets when comparing
404 a dump (as socket cannot be properly restored anyway). Thanks
405 to Gunther Reiszig <gunther@mit.edu> for reporting the bug.
406
407 4. Fixed restore to correctly access the archive file (-A argument)
408 even when using a remote tape. Thanks to Fabrice Bellet
409 <fabrice@bellet.info> for reporting the bug.
410
411 5. Fixed (again) handling of long (largefile) seeks in rmt.
412 Thanks to Fabrice Bellet <fabrice@bellet.info> for reporting
413 the bug.
414
415 6. Fixed restore corner case when dealing with large block sizes
416 dump is able to write now (-b 1024). Thanks to Fabrice Bellet
417 <fabrice@bellet.info> for reporting the bug.
418
419 7. Fixed a bug preventing dump to access a filesystem having
420 a label exactly 16 bytes in length. Thanks to <minduni@ti-edu.ch>
421 for reporting the bug.
422
423 8. Made dump store dump dates using explicit timezones, fixing a
424 problem with incremental dumps if the timezone is modified
425 between the dumps. Thanks to Philippe Troin <phil@fifi.org> for
426 the bug report and the patch.
427
428 9. Fixed a bug encountered when dumping individual files (not full
429 filesystems or directories) and dangling symbolic links happen
430 to be in the list of files. For as far as dump is concerned,
431 dangling symbolic links are allowed, and are dumped as is.
432 Thanks to Jin-su Ahn <jsahn@ee.snu.ac.kr> for reporting the
433 bug and providing the fix.
434
435 10. Fixed open and creation modes and permissions for QFA and
436 table-of-contents files in dump and restore. Thanks to
437 Philippe Troin <phil@fifi.org> for the patch.
438
439 11. Fixed the archive file descriptor handling enabling it to be 0.
440 This can happen in some cases when shell redirections are used.
441 Thanks to Philippe Troin <phil@fifi.org> for the patch.
442
443 12. Delayed the opening of archive file until after suid had been
444 dropped (fixing a possible security issue if dump is suid).
445 Thanks to Philippe Troin <phil@fifi.org> for the patch.
446
447 13. Fixed the 'S' command handling in the rmt client part.
448 Thanks to Philippe Troin <phil@fifi.org> for the patch.
449
450 14. Modified the end-of-tape script handling to print out statistics
451 (and stop the timer) before launching the eot script. Also, the eot
452 script does not get run anymore when using -M (which makes sense) or
453 when multiple tapes are listed on the command line
454 (-f tape0,tape1,tapen) (which also makes sense).
455 Thanks to Philippe Troin <phil@fifi.org> for the patch.
456
457 15. Relicensed dump/restore under the 'revised' BSD license, as per
458 ftp://ftp.cs.berkeley.edu/ucb/4bsd/README.Impt.License.Change.
459
460 16. Added LZO compression to dump. This new compression method has
461 the advantage of being super fast, thus not killing tape streaming
462 on slow machines. Thanks to Helmut Jarausch
463 <jarausch@igpm.rwth-aachen.de> for the patch and to
464 Markus Oberhumer <markus@oberhumer.com> for giving special permission
465 to include his miniLZO project (GPL licensed) in dump/restore.
466
467 17. Some small buffer overruns fixes in rmt. Thanks to Antonomasia
468 <ant@notatla.demon.co.uk> for reporting the bugs.
469
470 18. Added a special rmt version which can do encryption when writing
471 to tape. Read examples/encrypted_rmt/README for details on
472 how to enable and configure it. Thanks to Ken Lalonde
473 <ken@globalremit.com> for the patch.
474
475 19. Made dump work with 2.5 kernel end of tape early warning semantics.
476 Thanks to Kai Makisara <Kai.Makisara@kolumbus.fi> for the patch.
477
478 20. Fixed a bug which caused dump -w|-W not to work anymore, because
479 the fs_freq and fs_passno fields in /etc/mtab are always set
480 to 0 0. Thanks to Trent Piepho <xyzzy@speakeasy.org> for
481 reporting the bug.
482
483 Changes between versions 0.4b32 and 0.4b33 (released February 10, 2003)
484 =======================================================================
485
486 1. Added a note in the restore man page clarifying the question
487 'set the permissions on the current directory ?' asked by
488 restore at the end of treatment in -i and -x modes.
489
490 2. Fixed the endianess issues when reading compressed tapes.
491 Thanks to Dark Force <daq4th@users.sourceforge.net> for
492 reporting this bug and providing test cases.
493
494 3. Fixed the "ACL won't be dumped" warning message (which showed
495 an extra, unrelated error message). Thanks to Dragan Krnic
496 <dkrnic@lycos.com> for reporting this bug.
497
498 4. Made dump look first into /etc/mtab, then into /etc/fstab
499 when searching for filesystem paths. Also fixed some problems
500 caused by binding mounts or multiple block device mounts.
501 Thanks to Matus Uhlar <uhlar@fantomas.sk>, Elliott Mitchell
502 <ehem@m5p.com>, Greg Edwards <gedwards@users.sourceforge.net>,
503 Brian Hoy <brian.hoy@opus.co.nz>. (fixes Debian bugs #147086
504 and #160305, Sourceforge bugs #618699 and #679832).
505
506 5. Made dump's -I option accept the value '0' meaning all the
507 read errors will be ignored. This can be useful when running
508 dump from unattended sessions (like cron jobs). Thanks to
509 John I Wang <jiwang@users.sourceforge.net> for the suggestion.
510
511 6. Fixed the output of dump to indicate 'blocks' instead of
512 'tape blocks' in the various messages (blocks are always
513 1 Kilobyte, tape blocks are 1 BK * '-b' argument), and
514 made it clearly print the current blocksize at the start of
515 a dump. Thanks to Michal Szymanski <msz@astrouw.edu.pl> for
516 the suggestions.
517
518 7. Made rmt understand long (largefiles) seeks.
519
520 8. Fixed build with very old versions of libext2fs, where
521 EXT2_FT_* constants were undefined.
522
523 9. Made dump accept the dumpdates path on the command line
524 (-D file option) instead of using only the hardcoded one.
525 Thanks to Piete Brooks <pb22@users.sourceforge.net> for the
526 suggestion.
527
528 10. Enabled rmt, LFS, readline, QFA options by default in
529 ./configure. Updated the configure process (new versions
530 of config.guess, config.sub etc).
531
532 Changes between versions 0.4b31 and 0.4b32 (released November 15, 2002)
533 =======================================================================
534
535 1. Changed dump to use fcntl(F_SETLK) style locking instead
536 of flock() when locking the dumpdates file. With the old
537 locking scheme, a local user having read rights on the
538 dumpdates file could be able to do a Denial of Service attack
539 on dump. In order to lock the dumpdates file with the new
540 scheme, the user would need to have write access on the file.
541 Thanks to Richard Johnson <Richard.Johnson3@ey.com> for
542 reporting the bug (originally a bugtraq post).
543
544 2. Fixed interactive 'ls' which caused spurious errors warnings
545 about 'undefined filetypes' detected. Thanks to Jorgen Ostling
546 <jorgen_ostling@users.sourceforge.net> for reporting this
547 bug.
548
549 3. Fixed dump's estimate when dealing with sparse inodes.
550
551 4. Modified dump to allow setting a blocksize bigger than 32kB
552 (raised the limit to 1024kB), because newer hardware needs
553 this for proper operation. Thanks to Dirk Traenapp
554 <dtraenapp@users.sourceforge.net> for reporting this.
555
556 5. Fixed a bug causing Dump to stop and report an error if an
557 inode number in the exclude file was followed by some amount
558 of whitespace. Thanks to Jeffrey Sofferin
559 <sofferin@users.sourceforge.net> for reporting this bug.
560
561 6. Fixed a bug which caused restore, in some particular cases,
562 to ask some 'scary' questions and leave a bunch of RSTTMP
563 directories behind when restoring incremental tapes. Thanks
564 to Philippe Troin <phil@fifi.org> for reporting this bug and
565 providing the test cases.
566
567 7. Changed the wording when inodes are excluded from dump:
568 replaced 'Added inode 7 to exclude list' with
569 'Excluding inode 7 (resize inode) from dump', as suggested
570 by Elliott Mitchell <ehem@m5p.com> in a Debian bug report.
571
572 Changes between versions 0.4b30 and 0.4b31 (released July 30, 2002)
573 ===================================================================
574
575 1. Fixed rmt open flags transmission (GNU's symbolic syntax over
576 rmt) which I broke in 0.4b29. Thanks to Eros Albertazzi
577 <eros@lamel.bo.cnr.it> for reporting the bug.
578
579 Changes between versions 0.4b29 and 0.4b30 (released July 25, 2002)
580 ===================================================================
581
582 1. Made dump print out the ext2 logical block number in case of a read
583 error, which can be given as an argument to debugfs' ncheck command,
584 in order to find out the affected inode. Added note about this
585 usage in dump's man page.
586
587 2. Fixed a problem in restore when reading tapes written on big
588 endian machines with very old versions of dump. The patch was
589 contributed by George Helffrich <george@geology.bristol.ac.uk>.
590
591 3. Fixed the tape length calculation when using large tapes
592 and compression. Thanks to Georg Lippold
593 <g_lippold@sourceforge.net> for reporting the bug.
594
595 4. Added a new set of examples in dump_on_cd_2 directory, based
596 on dump_on_cd examples but somewhat enhanced, supporting DVD
597 media, and localized in english and german. Thanks to
598 Georg Lippold <g_lippold@sourceforge.net> for the new scripts.
599
600 5. Made dump save 32 bit UID/GID. Internally, this was achieved
601 by switching from the old BSD inode format to the new BSD
602 inode format, which means that the tape format was changed.
603 However, since all restore versions out there should
604 transparently support both inode formats, the change should
605 have no side effects. Thanks to John Yu <jky@cs.bu.edu> for
606 reporting the bug.
607
608 6. Fixed a lot of warnings in the code shown when compiling
609 with 'gcc -W'. Thanks to Matthias Andree
610 <matthias.andree@stud.uni-dortmund.de> for reporting this.
611
612 7. Fixed a small markup bug in the dump man page. Thanks to
613 Eric S. Raymond <esr@minx.thyrsus.com> for submitting the
614 patch.
615
616 8. Rewrote entirely the man pages using the tmac.an macro
617 package (Linux man page format) instead of the original BSD
618 format. They should be now cleaner and easier to modify.
619
620 Changes between versions 0.4b28 and 0.4b29 (released June 8, 2002)
621 ==================================================================
622
623 1. Fixed a problem in the rmt ioctl command, where ioctl's issued from
624 non Linux clients were misinterpreted. The description of the problem
625 (incompatible numbering in Linux mtio opcodes) is documented at
626 ftp://ftp.fokus.gmd.de/pub/unix/star/README.mtio . Thanks to
627 Jörg Schilling <schilling@fokus.gmd.de> for reporting this bug and
628 providing an excellent, cross-platform replacement for rmt in his
629 star package.
630
631 2. Fixed a bug reported by Andy Mentges <amentges@jumpline.com>
632 which caused restore to fail when the inode maps were not entirely
633 contained on the first volume (when using really small volumes or
634 when dumping a huge number of inodes).
635
636 3. Fixed a problem in dump, where files in subdirectories of directories
637 that have the nodump flag set (or being excluded with -e/-E) were
638 sometimes incorrectly dumped (instead of being excluded). The original
639 fix is from David C Lawrence <tale@dd.org> in the FreeBSD version
640 of dump: http://www.FreeBSD.org/cgi/query-pr.cgi?pr=32414
641 Thanks to Ted Grzesik <tedgyz@roostme.com> for reporting the bug and
642 help testing the patch.
643
644 4. Added some example scripts from Gerd Bavendiek <bav@epost.de>
645 which makes one able to pipe the output of dump, by the net, to
646 a remote CD-burner server.
647
648 5. Made dump use O_CREAT|O_TRUNC both locally and remotely (over rmt),
649 and use GNU's symbolic syntax over rmt instead of numerical values
650 to assure multiple platform compatibility.
651
652 6. Documented the -d option in restore.
653
654 7. Added a -v (verbose) mode to dump. For now it just prints the number
655 of the inode being dumped, but this could evolve in future versions
656 to include interesting debugging output.
657
658 8. Added a -o flag to restore, which automatically restores the current
659 directory permissions in -i or -x mode, without asking the operator
660 whether to do so. Patch submitted by Tony Clayton <tonyc@e-smith.com>
661 and Peter Samuel <peters@e-smith.com>.
662
663 Changes between versions 0.4b27 and 0.4b28 (released April 12, 2002)
664 ====================================================================
665
666 1. Fixed a bug in the QFA feature which could have caused
667 a bad tape positionning on restore (causing extra delays in
668 restore). Credits go to Uwe Gohlke <uwe@ugsoft.de>.
669
670 2. Added a small note in the dump man page specifying that
671 there should be no white space between the option letter and
672 the -j or -z parameter, thanks to Kenneth Porter <shiva@well.com>
673
674 3. Made restore work with older versions of the readline library,
675 by checking in configure for several extended symbols. Restore
676 can now be compiled with a readline as old as the 2.0 release
677 (though it may be a good idea to upgrade it to a more recent
678 version...). Thanks to Andrew Donkin <ard@waikato.ac.nz> for
679 reporting the build failures.
680
681 4. Fixed a performance problem with the QFA file creation in
682 dump, which made unnecessary seeks on the tape slowing down
683 the dump. Thanks to Andrew Donkin <ard@waikato.ac.nz> for
684 reporting this issue.
685
686 5. Removed the inclusion of some kernel headers in the dump
687 source, which prevented the compile in some kernel/glibc
688 headers/architecture combination. Thanks to Bdale Garbee
689 <bdale@gag.com> for reporting the bug.
690
691 6. Added the appropriate error message when dump fails to
692 open the output file for writing. Thanks to Amith Varghese
693 <amithv@yahoo.com> for reporting this bug.
694
695 7. Made restore able to understand large Solaris ufsdump tapes
696 (containing inodes bigger than 4194304). Sun have introduced
697 an "extension" to the dump tape format when dealing with
698 those inodes, which was uncorrectly handled by Linux restore.
699 Thanks to Uwe Gohlke <uwe@ugsoft.de> for reporting the bug and
700 providing a test case.
701
702 8. Added the -m parameter to dump which optimises the output for
703 inodes having been changed but not modified since the last dump
704 ('changed' and 'modified' have the meaning defined in stat(2)).
705 For those inodes, dump will save only the metadata, instead of
706 saving the entire inode contents. Inodes which are either
707 directories or have been modified since the last dump are saved
708 in a regular way. Uses of this flag must be consistent, meaning
709 that either every dump in an incremental dump set have the flag,
710 or no one has it.
711
712 Changes between versions 0.4b26 and 0.4b27 (released February 15, 2002)
713 =======================================================================
714
715 1. Fixed behaviour of dump when exceeding resource limits
716 (SIGXFSZ treatment).
717
718 2. Added the -L flag to restore to allow the user to specify a
719 maximal allowed number of miscompares when using restore
720 with the -C option to check the backup.
721
722 3. Detailed the manual entry for the -N option of restore.
723
724 4. Added the -a flag to restore to make able doing unattended
725 restores in -i or -x mode (automatically walks through the
726 multiple dump volumes).
727
728 5. Extended the QFA mode to work with local files and/or
729 remote tapes and files. This way, restore can know in advance
730 the tape number and the offset for the inodes to extract and
731 can minimize the extraction time by seeking directly to the
732 good tape/offset.
733
734 6. Added the -A <archive> option to both dump and restore,
735 which makes dump to archive a dump table-of-contents in
736 the specified file to be used by restore to determine
737 whether a file is in the dump file that is being restored.
738 (the archive file syntax is also compatible with the
739 Solaris ufsdump generated one).
740
741 7. Small fix in restore making it able to read some (broken ?)
742 Solaris ufsdump tapes.
743
744 8. Fixed dump to correctly recognise the root filesystem when using
745 ext2 disk labels (LABEL=/). Thanks to John Yu <jky@it.bu.edu>
746 for reporting this bug.
747
748 9. Added the -P <file> option to restore to create a
749 Quick File Access file from an already made dump. Patch
750 contributed by Uwe Gohlke <uwe@ugsoft.de>.
751
752 10. Made restore compile and run on Solaris, making it a
753 possible replacement for the standard ufsrestore. Port was
754 contributed by Uwe Gohlke <uwe@ugsoft.de>.
755
756 Changes between versions 0.4b25 and 0.4b26 (released January 7, 2002)
757 =====================================================================
758
759 1. Added a set of backup scripts from Eugenio Diaz
760 <getnito@yahoo.com> in the examples section. It features
761 automatic (cron based) full and incremental dumping of
762 several filesystems on a separate filesystem.
763
764 2. Fixed a off-by-one miscalculation which disabled dumping
765 a one letter subdirectory. Thanks to Chris Siebenmann
766 <cks@utcc.utoronto.ca> for reporting the bug.
767
768 3. Fixed several restore bugs occuring when trying to
769 restore some missing files on the tape. Thanks to Chris
770 Siebenmann <cks@utcc.utoronto.ca> for reporting the bug.
771
772 4. Fixed --with-ldopts configure argument passing, installing from
773 a separate object directory, makefile cleanups contributed
774 by <splite@purdue.edu>.
775
776 5. Fix a bug which could caused, in some conditions, the highest
777 number inode of a filesystem, to not be dumped. Many thanks
778 to Chris Siebenmann <cks@utcc.utoronto.ca> for helping me
779 chase this bug.
780
781 Changes between versions 0.4b24 and 0.4b25 (released November 17, 2001)
782 =======================================================================
783
784 1. Added a mini howto from Patrick Walsh in the examples
785 directory.
786
787 2. Minor man pages syntax corrections. Thanks to
788 Chris Danis <screechco@home.com> for reporting the bugs.
789
790 3. Added a script from David B. Peterson <dave@toppledwagon.com>
791 to the examples section. It features dumping several
792 filesystems to a remote tape drive upon ssh.
793
794 4. Added a patch provided by Richard Jones <rich@annexia.org>
795 which allows BRADEMAX (number of read errors tolerated by
796 dump) to be adjusted using the -I option.
797
798 5. Fixed a bug which disabled doing "restore -C -f -". Thanks
799 to Clemens Stiglechner <a9401816@unet.univie.ac.at> for the
800 patch.
801
802 6. Add the -l option to restore to specify if, when doing a
803 remote restore, the file used to access the backup is a
804 regular file (the defaults being a tape drive). Restore needs
805 to know this information when reading a remote compressed
806 dump. Previously, this information was autodetected, but
807 the autodetection code fails (with ioctl: Inappropriate ioctl
808 for device) when using a non Linux remote box. Thanks to
809 many users and especially to Eros Albertazzi
810 <eros@lamel.bo.cnr.it> for reporting this.
811
812 7. Found a workaroung for the dump deadlock problem (3 childs
813 stuck in pause(), father in read()). The workaround seems
814 to work for me and several beta-testers. If it doesn't work
815 for you, please report back.
816
817 8. Updated the RPM spec file (BuildPrereq, URL etc).
818
819 Changes between versions 0.4b23 and 0.4b24 (released September 12, 2001)
820 ========================================================================
821
822 1. Fixed the permissions of a newly created QFA file by dump.
823
824 2. Cleaned up the source of dump (the external variables
825 definition was a complete mess, making possible to have
826 objects overlap).
827
828 3. Fixed restore to use the full tape volume path when doing
829 a compare (since it changes the working directory to the
830 filesystem being compared in the process).
831
832 4. Added the -q option to dump which makes dump abort
833 whenever it needs operator attention. It should be
834 easier to use dump in scripts with this option.
835
836 5. Detect the use of incompatible options to dump and
837 refuse them (like -a and -B options together).
838
839 6. Added bzip2 compression to dump/restore (use option -j level
840 to select it). Note that you will need a restore version
841 >= 0.4b24 in order to restore a bzip2 compressed dump.
842 The same warning as for the zlib compression applies:
843 the tape format of a bzip2 dump is not compatible with the
844 original BSD tape format.
845
846 7. Fixed a overflow problem in dump corrupting the dump when
847 very large files were encountered. Thanks to Vladimir Nadvornik
848 for the bug report.
849
850 8. Added a ioctl(BLKFLSBUF) in dump which should flush the
851 kernel buffer/page cache before starting the dump, helping
852 a bit those who use dump on mounted filesystems. Thanks to
853 John Yu <jky@it.bu.edu> and to Theodore T'so <tytso@mit.edu>
854 for this suggestion.
855
856 9. Updated the RPM spec file following the RedHat changes
857 (dynamically linked binaries now in /usr/sbin etc).
858
859 10. Added a patch from Helmut Jarausch <jarausch@igpm.rwth-aachen.de>
860 which enables restore to recognise multi volume compressed dumps
861 done on CD. Included his scripts for dump (which pipe the dump
862 output directly into cdrecord) and restore. There is now
863 possible to to backups to CD on the fly!
864
865 Changes between versions 0.4b22 and 0.4b23 (released July 20, 2001)
866 ===================================================================
867
868 1. Fixed a buffer overflow in restore/tape.c. Patch provided by
869 Marcus Meissner (Caldera International Security Dept.).
870
871 2. Implement the Sun rmt extended protocol. Patch submitted
872 by Ian Gordon <iangordon@users.sourceforge.net>.
873
874 3. Check for the e2fsprogs header <ext2fs/ext2_fs.h> instead of
875 the linux kernel header. This ensures that dump always has the
876 latest version of this file. Patch submitted by
877 Andreas Dilger <adilger@turbolinux.com>.
878
879 4. Report any filesystem present in either /etc/fstab with a
880 positive passno or /etc/dumpdates in dump -w output.
881 Patch submitted by Andreas Dilger <adilger@turbolinux.com>.
882
883 5. Fixed the looping problem in dump introduced in the
884 previous version.
885
886 6. Changed the -B option of dump to limit the size of
887 _compressed_ output per volume if compression is on.
888 Patch contributed by Helmut Jarausch
889 <jarausch@igpm.rwth-aachen.de>. Note however that, since
890 it is impossible to predict the size of the compressed
891 data before doing the compression, the -B limit is a bit
892 conservative.
893
894 7. Fixed a bug in reading the operator typed file/tape path for
895 the next volume in restore.
896
897 8. Implemented a "-F script" option to restore which permits the
898 user to specify a script which will be launched at the
899 beginning of each tape, useful for automatic programming of
900 tape changers for example. See the restore man page for the
901 script parameters and return codes.
902
903 9. Small fix for the QFA routines provided by Uwe Gohlke
904 <uwe@ugsoft.de>, and some recommendations for QFA uses in
905 the man pages.
906
907 10. Fixed the multivolume restoring where making a mistake
908 to the 'Mount next tape' prompt caused several blocks to
909 be lost.
910
911 11. Enhanced the -e option of dump to take as a parameter a
912 comma separated list of inode numbers.
913
914 12. Added the -E option to dump which specify a file containing
915 inode numbers to exclude from the dump.
916
917 13. Fixed the compressed multi-volume dump + restore.
918
919 Changes between versions 0.4b21 and 0.4b22 (released May 12, 2001)
920 ==================================================================
921
922 1. Made dump report the number of blocks dumped per volume.
923 Thanks to Kenneth Porter <shiva@well.com> for the suggestion.
924
925 2. Fix a bug in dump where waiting too much at the 'change volumes'
926 question caused the volume to be rewritten. Thanks to
927 Graham Reed <greed@users.sourceforge.net> for reporting the
928 bug and providing a patch.
929
930 3. Added a compression option to dump, contributed by Jerry
931 Peters <gapeters@worldnet.att.net>.
932
933 WARNING: the compression works only on files/pipes and
934 drives supporting variable block size.
935
936 WARNING2: tapes created using the compression feature are
937 incompatible with the standard dump tape format, and a
938 version of restore >= 0.4b22 will be needed for extraction.
939
940 4. Fixed some compilation problems for glibc 2.2.2 and 64 bit
941 architectures. Thanks to Paul Slootman <paul@debian.org> for
942 the patch and to Bdale Garbee <bdale@gag.com> for forwarding
943 it upstream.
944
945 5. Many cleanups (CPP defines, const char warnings, check of
946 ext2fs COMPAT flags, time_t cleanups, added libext2 version
947 in dump usage text) by Andreas Dilger <adilger@turbolinux.com>.
948
949 6. Made --prefix option work in configure. All the install path
950 are now based on the configure parameters.
951
952 7. Added the Quick File Access mode in dump/restore, contributed
953 by Uwe Gohlke <uwe@ugsoft.de>. In this mode, dump stores in
954 a file tape position for each inode, and this file is used by
955 restore (if called with parameter Q and the filename)
956 to directly position the tape at the file restore is currently
957 working on. This saves hours when restoring single files from
958 large backups, saves the tapes and the drive's head. Use
959 --enable-qfa option of configure to compile in the QFA support.
960
961 8. Added the possibility to dump several files and directories
962 in a single invocation of dump. Thanks to Uwe Gohlke
963 <uwe@ugsoft.de> for implementing this option.
964
965 9. Fixed the dumping and restoring of files with holes
966 (files ending with holes didn't get dumped or restored
967 correctly).
968
969 10. Fixed a socket descriptor leak in dump, which leaved opened
970 3 file descriptors per dump process (and there is one dump
971 process per tape).
972
973 11. Fixed dump large file system support, by explicit use of
974 open64/lseek64/etc functions (explicit use needed because
975 e2fsprogs libraries don't behave well when compiled with
976 FILE_OFFSET_BITS=64).
977
978 Changes between versions 0.4b20 and 0.4b21 (released January 13, 2001)
979 ======================================================================
980
981 1. Fixed some bugs in the dump -w|-W logic introduced by
982 the previous version. Thanks to Andreas Dilger
983 <adilger@turbolinux.com> for his help on this one.
984
985 2. Fixed again a compilation problem when using old e2fs
986 headers (filesystem label related). Thanks to many users
987 who reported this stupid error.
988
989 3. Fixed a build problem on old lib5 based systems dealing with
990 _PATH_MNTTAB being undefined. Thanks to John Adams
991 <johna@onevista.com> for reporting the bug.
992
993 4. Improved the error detection in restore when creating
994 the temporary files in TMPDIR. Now restore will corectly
995 report a 'no space left on device' error instead of
996 strange errors which could imply an archive corruption.
997 Thanks to Gabriel Somlo <somlo@cs.colostate.edu> and
998 bgjenero <bgjenero@sympatico.ca> for reporting the bug.
999
1000 5. Added the throughput information in dump's progression
1001 messages. Thanks to Andreas Dilger <adilger@turbolinux.com>
1002 for the patch.
1003
1004 6. Use libext2fs's inode scanning routines, which are particularly
1005 robust when dealing with errors (especially when having some
1006 bad blocks on the filesystem). Many thanks to Colin
1007 <colin@horizon.com> for reporting the bug, investigating
1008 the issues, finding a workaround, writing the patch and
1009 fully testing it... (of course, if this patch will break
1010 anything, he is to blame for :-)).
1011
1012 7. Made dump and restore LFS aware. Dump can dump a filesystem
1013 containing large files, generate a large file on output and
1014 restore can restore them. This feature is not enabled by
1015 default, use --enable-largefile option of configure to enable
1016 it (you need to have a LFS aware glibc though). Thanks to
1017 Andrea Arcangeli <andrea@suse.de> for submitting the patch,
1018 and to Theodore T'so <tytso@valinux.com> for his always
1019 useful thoughts.
1020
1021 8. Made dump ask upon a tape write error if it should rewrite
1022 the current volume (assume this is a bad tape) or if it should
1023 assume an end-of-tape condition (useful for tape drives which
1024 behaves badly at the end of the tape). Thanks to Andreas
1025 Dilger <adilger@turbolinux.com> for the suggestion.
1026
1027 Changes between versions 0.4b19 and 0.4b20 (released November 10, 2000)
1028 =======================================================================
1029
1030 1. Fixed a small compilation problem due to a change
1031 in the definintion of the struct sigaction in
1032 glibc 2.0 / libc5. Thanks to Gunther Schlegel
1033 <schlegel@riege.de> for reporting the bug and to
1034 Dave Platt <dplatt@snulbug.mtview.ca.us> for suggesting
1035 a fix.
1036
1037 2. Modified the label and uuid analysis in order to be
1038 self-contained (independant of kernel/libc headers). This
1039 should fix the compile with older kernel/libc headers and
1040 will preserve the functionality. Thanks to Bernhard Erdmann
1041 <bernhard.erdmann@gmx.de> for reporting the bug.
1042
1043 3. The 'exclude inode' option, if applied to a directory
1044 inode, excludes now correctly all the leaves of this
1045 directory. Thanks to John R. Dennison
1046 <gerdesas@users.sourceforge.net> for reporting the bug.
1047
1048 4. Fixed the '-e' option to disable the possibility
1049 to exclude the root inode (which causes the dumps to
1050 be unreadable by restore). Prevented array overflow
1051 when multiple -e options are used.
1052
1053 5. Fixed dump to correctly interpret a filesystem argument
1054 which is a mountpoint and it is not an absolute path
1055 (as specified in the fstab). Thanks to Bernhard R. Erdmann
1056 <be@berdmann.de> for reporting the bug.
1057
1058 6. Made dump able to backup files larger than 2 GB. Note that
1059 dump still doesn't cope with files larger than 4 GB.
1060
1061 7. Restore the real uid and gid before invoking an external
1062 RSH program (big hole when dump or restore is suid root!).
1063
1064 8. Get the values for the default tape device and the location
1065 of fstab file from the system headers. Thanks to
1066 Andreas Dilger <adilger@turbolinux.com> for the patch.
1067
1068 9. Made dump -w|-W report all recognized filesystems
1069 present in either /etc/fstab or /etc/dumpdates, and present
1070 the list in the same order as in fstab file. Thanks
1071 to Andreas Dilger <adilger@turbolinux.com> for the patch.
1072
1073 10. Made dump's -a (automatic end of tape detection) the
1074 default. Specifying one of -B, -c, -d or -s options will
1075 override the EOM detection. Thanks to Andreas Dilger
1076 <adilger@turbolinux.com> for the patch.
1077
1078 11. Save the ext2 filesystem label into the dump volume label.
1079 Specifying a volume label on the command line (-L option)
1080 will override this feature. Thanks to Andreas Dilger
1081 <adilger@turbolinux.com> for the patch.
1082
1083 Changes between versions 0.4b18 and 0.4b19 (released August 20, 2000)
1084 =====================================================================
1085
1086 1. Fixed the signal handling in dump (which I broke in 0.4b17)
1087 which was causing several strange dump failures (dump
1088 hanged or segmentation faults etc).
1089
1090 2. Specified the default blocksize in the dump man page.
1091
1092 3. Changed two info messages of restore to be written on stdout
1093 instead of stderr in order to leave stderr only for errors
1094 or warnings. Thanks to Stephen Atwell
1095 <satwell@urbana.css.mot.com> for the suggestion.
1096
1097 4. Corrected an off by one calculation which prevented
1098 dumping a directory having a 1 character name length.
1099 Thanks to Bernhard Erdmann <bernhard.erdmann@gmx.de>
1100 for reporting the bug.
1101
1102 5. Reinforce dump in order to allow only level 0 and no
1103 -u option when dumping a subdirectory, and document
1104 this in the man page. Thanks to Bernhard Erdmann
1105 <bernhard.erdmann@gmx.de> for reporting the bug.
1106
1107 6. Fixed a small harmless bug in restore which caused
1108 in some conditions hard links to be restored several
1109 times (therefore generation some warning messages).
1110 Thanks to Kenneth Porter <shiva@well.com> for
1111 reporting the bug.
1112
1113 7. Updated the RPM spec file to the latest RedHat version,
1114 providing FHS packaging and other cosmetic changes.
1115 You will need RPM version >= 3.0.5 in order to build the RPM.
1116
1117 8. Updated the configure script to check for libtermcap
1118 before checking for libreadline, because we need this
1119 library in order to compile the readline support.
1120
1121 9. Made dump understand the LABEL= and UUID= notation
1122 both in /etc/fstab and on the command line. Note that
1123 you will need the /proc filesystem in order to use
1124 these notations. Thanks to Erik Troan <ewt@redhat.com>
1125 for providing the patch.
1126
1127 Changes between versions 0.4b17 and 0.4b18 (released June 30, 2000)
1128 ===================================================================
1129
1130 1. Fixed a potential buffer overflow in restore. Thanks
1131 to Stan Bubrouski <satan@fastdial.net> for reporting
1132 the bug.
1133
1134 2. Fixed a readline-related bug which prevented
1135 'cat DUMPFILE | restore -i -f -' from working. Thanks
1136 to Charles Karney <karney@users.sourceforge.net>
1137 for the bug report.
1138
1139 3. Changed a few "panic" into "exit", causing restore to
1140 be more stable against some attacks (like the last one
1141 reported on Bugtraq, although the last version of restore
1142 was not vulnerable - just dumped core). Thanks to
1143 Andreas Hasenack <andreas@conectiva.com.br> for reporting
1144 the bugs to me.
1145
1146 4. Removed the suid-root bit on dump and restore in the
1147 default build (and generated RPMs). It should be safer
1148 now. Users who need the suid-root capabilities in order
1149 to make network backups should read first the man page
1150 of dump and enable the suid bit manually.
1151
1152 5. Added -ltermcap to the compile parameters for restore
1153 when using readline support, in order to make the compile
1154 process work on some systems (SuSE ?). Thanks to
1155 Patrik Schindler <poc@pocnet.net> for reporting the bug.
1156
1157 Changes between versions 0.4b16 and 0.4b17 (released June 1st, 2000)
1158 ====================================================================
1159
1160 1. The -F script is called now *only* at the end of a tape,
1161 not at the end of the dump. Thanks to Christian Haul
1162 <haul@informatik.tu-darmstadt.de> for the bug report.
1163
1164 Normally, the device name and the current volume number
1165 are passed on the command line of the script. However,
1166 if you want the old style script (with no arguments
1167 passed on the command line), you can enable it in
1168 configure with the --enable-oldstylefscript.
1169
1170 2. Use posix signal handling to preserve dump functionality
1171 with libc5. Thanks to Jeff Johnson <jbj@redhat.com> for
1172 the patch.
1173
1174 3. Made the exit code of restore in case of a 'C'ompare
1175 command reflect the result of the compare. An exit status
1176 of 0 means the dump archive is correct, 1 means tape errors,
1177 2 means that some files were modified. Thanks to Kenneth Porter
1178 <shiva@well.com> for the suggestion.
1179
1180 4. Made (finally) quotes work in interactive restore.
1181
1182 5. Small fixes in order to allow dump to compile with a
1183 really old e2fsprogs version. Thanks to Ian Zimmerman
1184 <itz@speakeasy.org> for the bug report.
1185
1186 6. Add GNU readline capabilities to interactive restore.
1187 Use configure's --enable-readline flag to enable this feature.
1188 Thanks to Patrik Schindler <poc@pocnet.net> for the
1189 suggestion.
1190
1191 7. Do the compare on the fly in restore 'C' mode (this will
1192 allow not to exhaust the available /tmp space when
1193 ccmparing large files). Thanks to Kenneth Porter
1194 <shiva@well.com> for the suggestion.
1195
1196 Changes between versions 0.4b15 and 0.4b16 (released March 11, 2000)
1197 ====================================================================
1198
1199 1. Fixed some several duplicate 'DUMP: DUMP:' in the
1200 output of dump.
1201
1202 2. Corrected the estimation of blocks to dump. Note that
1203 this is still wrong for large directory inodes, due
1204 to the size of a BSD directory entry on the tape
1205 (estimation is based on the size of the ext2 directory,
1206 which is smaller because it doesn't contain the null
1207 character at the end).
1208
1209 3. Made dump report the total number of MB written to
1210 tape at the end of run. Thanks to W. Reilly Cooley
1211 <wcooley@nakedape.cc> for the patch.
1212
1213 4. Added the -X option to restore giving the possibility
1214 to read the names of the files to be extracted/listed
1215 from a text file (in addition of the command line).
1216 Thanks to Dejan Muhamedagic <dejan@quant-x.com> for the
1217 patch.
1218
1219 5. Added the device name and the current volume number
1220 as arguments to the end of tape script (dump -F option).
1221
1222 6. Made the multi-volume dump work again (after having
1223 broken it in 0.4b15).
1224
1225 Changes between versions 0.4b14 and 0.4b15 (released March 2, 2000)
1226 ===================================================================
1227
1228 1. Added a prompt command in interactive restore mode. Thanks
1229 to Andreas Dilger <adilger@home.com> for the patch.
1230
1231 2. Fixed a buffer overflow problem in dump (caused by
1232 not checking the size of the filesystem parameter).
1233 Thanks to Kim Yong-jun <loveyou@hackerslab.org> for
1234 reporting this on Bugtraq (and to several dump users
1235 who forwarded me his mail).
1236
1237 3. Added the '-F script' option to dump in order to
1238 launch a script at the end of each tape (to be used
1239 with a tape changer, or to notify the sysadmin by
1240 pager etc.).
1241
1242 4. Fixed a bug in restore compare code caused by the changes
1243 I made in 0.4b14.
1244
1245 5. Fixed the treatment of options using the old BSD syntax
1246 in both dump and restore.
1247
1248 Changes between versions 0.4b13 and 0.4b14 (released February 10, 2000)
1249 =======================================================================
1250
1251 1. Fixed a bug in dump which may have caused invalid deleted
1252 directories to be dumped out if they were deleted after the
1253 mapping pass. This could occure on active filesystem and lead
1254 to heap corruption (causing dump malfunction in many possible ways).
1255 Thanks to Jan Sanislo <oystr@cs.washington.edu> for finding this
1256 bug and submitting the patch.
1257
1258 2. Fixed the handling of the filesystem parameter in dump. A
1259 '/mnt/disk' parameter caused the disk contents to be dumped,
1260 but a '/mnt/disk/' parameter caused the mountpoint directory
1261 to be dumped (generally an empty directory).
1262
1263 3. Improved the output of dump in order to tell which directory
1264 it is currently dumping (when dumping a subtree).
1265
1266 4. Added the '-e' exclude inode option to dump. Thanks to
1267 Isaac Chuang <ike@isl.stanford.edu> for contributing with the patch.
1268
1269 5. Added a REPORTING-BUGS file in order to provide a guide
1270 on how to correctly report a bug in dump/restore.
1271
1272 6. Fixed a restore problem when restoring a hard link to an inode
1273 having the immutable or append-only attribute set. Thanks to
1274 Ambrose Li <acli@mingpaoxpress.com> for submitting the patch.
1275
1276 7. Fixed a compatibility problem between dumps made on little
1277 endian machines (the format was correct) and big endian
1278 machines (the format was incorrect). This fix break the
1279 compatibility with the older dumps made on big endian
1280 machines (sparc, m86k, ppc etc). For the first time in
1281 linux dump's history (I believe), the dumps made by *BSD,
1282 Linux/alpha, Linux/sparc and Linux/x86 are compatible,
1283 so interchangeable. Thanks to Rob Cermak
1284 <cermak@ahab.rutgers.edu> for submitting the bug and
1285 helping me test the fix.
1286
1287 8. Fixed the way dump reports the remaining percent/time, if
1288 the number of blocks actually dumped exceeds the estimated
1289 number of blocks. Thanks to Jean-Paul van der Jagt
1290 <jeanpaul@dutepp0.et.tudelft.nl> for reporting the bug.
1291
1292 Changes between versions 0.4b12 and 0.4b13 (released January 21, 2000)
1293 ======================================================================
1294
1295 1. Small Y2K fix in the man pages :). Thanks to Bernhard Sadlowski
1296 <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug.
1297
1298 2. Removed the requirement to build the RPM as root from the
1299 spec file. Thanks to Christian Weisgerber
1300 <naddy@mips.rhein-neckar.de> for submitting this.
1301
1302 3. Fixed a bug in dump related to the 'filetype' feature of ext2fs,
1303 causing dump to block when dumping really huge filesystems.
1304 Many thanks to Patrik Schindler <poc@pocnet.net> for
1305 helping me find this bug.
1306
1307 4. Fixed the treatment for an interrupt signal when dump access
1308 the remote tape through RSH. Thanks to Christian Weisgerber
1309 <naddy@mips.rhein-neckar.de> for providing the patch.
1310
1311 5. Fixed a bug which was causing dump/restore to display
1312 garbage characters instead of the remote host name.
1313
1314 Changes between versions 0.4b11 and 0.4b12 (released January 8, 2000)
1315 =====================================================================
1316
1317 1. Small fix in the dump man page. Thanks to Thorsten Kukuk
1318 <kukuk@suse.de> for submitting the patch.
1319
1320 2. Fix for the exit code when using the size estimate option of
1321 dump. Thanks to Matti Taskinen <mkt@rni.helsinki.fi> for
1322 submitting the patch.
1323
1324 3. Handle EINTR in atomical reads/writes in dump, which was causing
1325 dump to fail on some systems. Thanks to Eric Jergensen
1326 <eric@dvns.com> for reporting the bug and submitting the patch.
1327
1328 4. Handle more than 16 characters for the device names in dumpdates.
1329 (up to 255 now). Thanks to Rainer Clasen <bj@ncc.cicely.de> for
1330 tracking down the problem and proposing the solution.
1331
1332 5. Fixed a bug in dump which prevented the creation of the
1333 dumpdates file when doing a 0-level dump without already
1334 having a dumpdates file. Thanks to Patrik Schindler
1335 <poc@pocnet.net> for reporting the bug.
1336
1337 6. Changed the way dump 'S' flag reports the size estimate
1338 from number of blocks into bytes (making it compatible
1339 with the Solaris version, and simplifying things for
1340 amanda users). Thanks to Jason L Tibbitts III
1341 <tibbs@math.uh.edu> for reporting the bug.
1342
1343 7. Fixed a compatibility problem in linux/alpha dump tape format.
1344 Now the linux/alpha dump are (again) compatible with the
1345 other dump formats. But this breaks compatibility with
1346 older dumps made on alpha. Thanks to Mike Tibor
1347 <tibor@lib.uaa.alaska.edu> for helping me in finding this bug.
1348
1349 Changes between versions 0.4b10 and 0.4b11 (released December 5, 1999)
1350 ======================================================================
1351
1352 1. Added a '--enable-kerberos' to configure.
1353
1354 2. Added a 'S' option to dump which determines the amount of space
1355 that is needed to perform the dump without actually doing it, similar
1356 to the Sun's ufsdump 'S' option. Patch contributed by Rob Cermak
1357 <cermak@ahab.rutgers.edu>.
1358
1359 3. Added a 'M' multi-volume option to dump and restore which enables
1360 dumping to multiple files (useful when dumping to an ext2
1361 partition to make several dump files in order to bypass the 2GB
1362 file size limitation). The 'f' argument is treated as a prefix and
1363 the output files will be named <prefix>001, <prefix>002 etc. With
1364 the 'M' flag, restore automatically selects the right file without
1365 asking to enter a new tape each time.
1366
1367 4. Fixed a memory leak which was causing dump to grow very big
1368 (270MB when dumping a 10GB filesystem...). Thanks to Jason
1369 Fearon <jasonf@netrider.org.au> for reporting the bug.
1370
1371 Changes between versions 0.4b9 and 0.4b10 (released November 21, 1999)
1372 ======================================================================
1373
1374 1. Make configure test if the system glob routines support
1375 extended syntax (ALTDIRFUNC). If not, use the internal glob
1376 routines instead of system ones. Thanks to Bernhard Sadlowski
1377 <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug
1378 and helping me resolve this and other minor libc5 compiling
1379 glitches.
1380
1381 2. Fix a problem when dumping a ext2fs with the 'filetype'
1382 feature enabled. Thanks to Patrick J. LoPresti
1383 <patl@cag.lcs.mit.edu> for reporting the bug and to
1384 Theodore Y. Ts'o <tytso@mit.edu> for providing the patch.
1385
1386 3. Made the nodump flag work on directories. A directory which
1387 has the nodump flag gets never dumped, regardless of its
1388 contents.
1389
1390 4. Integrate a patch from Jeremy Fitzhardinge <jeremy@goop.org>
1391 which allows dump on an active ext3 filesystem. However, this
1392 is a "quick and dirty" patch which enables backup of an ext3
1393 filesystem through the ext2 compatibility (by ignoring the
1394 NEEDS_RECOVERY bit). The journal file is not recognized and
1395 it is dumped (it should not).
1396
1397 5. Test the superblock compatibility flags when dumping, in order
1398 to be sure that we know how to deal with specific features.
1399
1400 Changes between versions 0.4b8 and 0.4b9 (released November 5, 1999)
1401 ====================================================================
1402
1403 1. Use lchown instead of chown, fixing a possible security problem
1404 when restoring symlinks (a malicious user could use this
1405 to deliberately corrupt the ownership of important system files).
1406 Thanks to Chris Siebenmann <cks@utcc.utoronto.ca> for detecting
1407 this and providing the patch.
1408
1409 Changes between versions 0.4b7 and 0.4b8 (released November 3, 1999)
1410 ====================================================================
1411
1412 1. Put dump sources under CVS, added Id tags in all files so
1413 one can use 'ident' on binary files.
1414
1415 2. Added the dump/restore version in the usage text so one can
1416 easily verify the version he is using.
1417
1418 3. Small patch from Nuno Oliveira <nuno@eq.uc.pt> which fixes
1419 a va_start/va_end problem on linux-ppc (always call va_start
1420 va_end in pairs each time we use a vararg function).
1421
1422 4. Added again the DT_* constants because old libc does not
1423 contain them :(. Thanks to Eric Maisonobe <virnet@nat.fr>
1424 for submitting the bug report.
1425
1426 5. Use ext2fs_llseek instead of llseek. With recent e2fsprogs
1427 this should enable dumping big (huge) filesystems.
1428
1429 6. Added the RSH environment variable in order to be able to
1430 use a rsh replacement like ssh when doing remote backups (and
1431 bypass the security limitations of rcmd). Now you can do remote
1432 backups without being root (or making dump setuid root).
1433
1434 7. Modified again the way dumpdates works. For incremental dumps,
1435 we need to read dumpdates even if we are not using 'u' option.
1436 Thanks to Bdale Garbee <bdale@gag.com> for his ideas on how
1437 this should work.
1438
1439 Changes between versions 0.4b6 and 0.4b7 (released October 8, 1999)
1440 ===================================================================
1441
1442 1. Removed the 'k' flag from the restore 'about' text if kerberos
1443 was not compiled in.
1444
1445 2. Prototyped (f)setflags from e2fsprogs and corrected the calls
1446 to them (fsetflags takes a char*, setflags an open fd!).
1447
1448 3. (f)setflags is called only if the flags aren't empty. If the
1449 file is a special file, a warning is printed, because changing
1450 flags implies opening the device. Normally, a special file
1451 should not have any flag... (Debian bug #29775, patch provided
1452 by Abhijit Dasgupta <abhijit@ans.net>).
1453
1454 4. Made possible to dump a file system not mentioned in /etc/fstab.
1455 (Debian bug #11904, patch provided by Eirik Fuller <eirik@netcom.com>).
1456
1457 5. Changed the default behaviour to not create dumpdates
1458 unless 'u' option is specified. Removed the old "debian-patch"
1459 which provided the same thing. (Debian bug #38136, #33818).
1460
1461 6. Removed all those dump*announce, since they were getting old...
1462
1463 7. Added warning messages if dumpdates does not exist and
1464 when an inferior level dump does not exist (except for a level 0
1465 dump).
1466
1467 8. Debugged the glob calls in interactive mode: restore used a
1468 dirent struct which was different from the /usr/include/dirent.h
1469 one (this used to work, is it a glibc2 change?), so none of the
1470 compat glob (which used /usr/include/dirent.h) or the system glob
1471 worked. Restore use now the system dirent (and the system
1472 DT_* constants), which are compatible with BSD ones.
1473
1474 9. Added a configure flag (--with-dumpdatespath) to specify
1475 the location of dumpdates file. By default, it is
1476 /etc/dumpdates.
1477
1478 10. Added the "AUTHOR" and "AVAILABILITY" sections and
1479 included the current date/version in man pages.
1480
1481 11. Corrected the estimation of remaining time when
1482 the operator doesn't change the tapes quickly enough. This
1483 was an old bug, I thought I corrected it, and discovered
1484 that in fact it was corrected in two different places, so
1485 the results canceled each other...
1486
1487 Changes between versions 0.4b5 and 0.4b6 (released October 1, 1999)
1488 ===================================================================
1489
1490 1. Integrated multiple patches from RedHat, Debian and SuSE:
1491
1492 - tweak dump/itime.c to not try to read dumpdates if the 'u' option
1493 isn't specified.
1494 - several fixes in the man pages.
1495 - update the default tape device to /dev/st0.
1496 - many updates for Linux Alpha (byte ordering, size_t etc).
1497 - buffer overruns.
1498 - use environment variable for TMPDIR (instead of /tmp).
1499 - use sigjmp_buf instead of jmp_buf (RedHat bug #3260).
1500 - workaround egcs bug (RedHat bugs #4281 and #2989).
1501 - wire $(OPT) throughout Makefile's.
1502
1503 2. Upgrade the dump revision to 1, making possible to dump filesystems
1504 made with e2fsprogs-1.15 or newer. Nothing seems to break...
1505
1506 3. Fix some compile warnings, prototype all functions.
1507
1508 4. Use glibc err/glob instead of internal compatibility
1509 routines (only if available).
1510
1511 5. Fix a compile error on Linux 2.2.7 / libc5 (5.4.44) (patch provided
1512 by Bernhard Sadlowski <sadlowsk@mathematik.uni-bielefeld.de>).
1513
1514 Changes between versions 0.4b4 and 0.4b5 (released September 22, 1999)
1515 ======================================================================
1516
1517 1. Integrated the changes from FreeBSD-3.1-RELEASE
1518 (mostly bug fixes, buffer overruns, dump has now an "automatic
1519 tape length calculation" flag, dump/restore can use kerberos now
1520 (this is NOT tested), use environment variables for TAPE and
1521 RMT etc.).
1522
1523 2. Integrated three RedHat patches ("glibc", "kernel" and "bread" patches)
1524
1525 3. Corrected a bug in restore when using 'C' option with multi-volumes
1526 tapes (files splited accros two tapes give "size changed" errors
1527 when comparing).
1528
1529 4. Corrected the long standing bug when dumping multiple tapes.
1530 This works for me, needs further testing.
1531
1532 Changes between versions 0.4b3 and 0.4b4 (released January 17, 1997)
1533 ====================================================================
1534
1535 1. Dump now runs correctly on kernels 2.1.x
1536 Fix made by Gerald Peters <gapeters@worldnet.att.net>
1537
1538 Changes between versions 0.4b2 and 0.4b3
1539 ========================================
1540
1541 1. Use realpath() if available
1542
1543 2. Report statistics
1544
1545 Changes between versions 0.4b1 and 0.4b2
1546 ========================================
1547
1548 1. Fixed the bug fix from Greg Lutz (I had made a mistake when integrating
1549 the patch)
1550
1551 2. Fixed restore to make it able to read FreeBSD 2.x dumps again
1552
1553 3. Fixed configure.in to correctly handle --enable-rmt
1554
1555 Changes between versions 0.3 and 0.4b1
1556 ======================================
1557
1558 1. Integrated the changes from 4.4BSD-Lite2
1559
1560 2. Integrated the patches from Debian and Red Hat
1561
1562 3. Portability changes: use the __u32, __u16, __s32, and __s16 types
1563
1564 4. Changed dump to use the Ext2fs library to get block addresses. This
1565 should solve the endianness problem on SparcLinux.
1566
1567 5. Created a configure.in file (shamelessly stolen from the e2fsprogs
1568 distribution's one) to use autoconf
1569
1570 6. Fixed a few minor bugs
1571
1572 Changes between versions 0.2e and 0.2f
1573 ======================================
1574
1575 1. Added the creation of named pipes (fifos) in restore.
1576
1577 2. Added the -N flag in the restore manual page.
1578
1579 3. Added the file kernel.patch which contains the llseek() optimization
1580 patch for 1.2.x kernels.
1581
1582 4. Fixed a bug in the restoration of symbolic links: owner and group were
1583 not restored.
1584
1585 5. Integrated some changes from FreeBSD 2.2.
1586
1587 6. Added a call to ftruncate() after restoring each file to restore
1588 correctly files ending by a hole.
1589
1590 Changes between versions 0.2d and 0.2e
1591 ======================================
1592
1593 1. Fixed a bug in the "set owner/mode" process. Every file was restored
1594 with owner = root (0) and group = root/wheel/whatever (0).
1595
1596 Changes between versions 0.2c and 0.2d
1597 ======================================
1598
1599 1. Dump is now able to backup 2GB+ filesystems.
1600
1601 2. Dump and restore can now be linked as static binaries.
1602
1603 Changes between versions 0.2b and 0.2c
1604 ======================================
1605
1606 1. Fixed a bug when dumping ``slow'' (i.e. normal) symbolic links.
1607
1608 Changes between versions 0.2a and 0.2b
1609 ======================================
1610
1611 1. Really fixed the bug that I should have corrected in 0.2a.
1612
1613 2. Enabled optimization again.
1614
1615 Changes between versions 0.2 and 0.2a
1616 =====================================
1617
1618 1. Disabled the optimization during compilation.
1619
1620 Changes between versions 0.1 and 0.2
1621 ====================================
1622
1623 1. Fixed a bug in fstab.c which caused a null pointer to be stored in
1624 the fs_type field (actually, I modified the file fstab.c to make it
1625 use the mntent functions).
1626
1627 2. Dump and restore now use a 4.3 BSD compatible dump format. Backups
1628 made by dump should be readable by the BSD restore and backups made
1629 by the BSD dump should be readable by restore. Unfortunately, this
1630 means that the dump format has changed between version 0.1 and version
1631 0.2 :-(
1632
1633 3. Dump is now able to backup a subtree, it is no longer limited to whole
1634 filesystems like the BSD version.
1635
1636 4. Dump now uses ext2_llseek() so it is able to backup filesystems bigger
1637 than 2 GB.
1638
1639 Changes between versions 0.0 and 0.1
1640 ====================================
1641
1642 1. Now create links rdump and rrestore during the `make install' step.
1643
1644 2. Linux port specific bugs added to the manual pages
1645
1646 3. Incorrect estimation of the number of tapes blocks fixed when doing
1647 incremental backups.
1648
1649 4. Better ls-like format in restore in interactive mode.