]> git.wh0rd.org - dump.git/blob - CHANGES
Problem when restoring a hard link to an inode having the immutable
[dump.git] / CHANGES
1 $Id: CHANGES,v 1.41 2000/02/08 12:22:42 stelian Exp $
2
3 Changes between versions 0.4b13 and 0.4b14 (released ????????????????)
4 ======================================================================
5
6 1. Fixed a bug in dump which may have caused invalid deleted
7 directories to be dumped out if they were deleted after the
8 mapping pass. This could occure on active filesystem and lead
9 to heap corruption (causing dump malfunction in many possible ways).
10 Thanks to Jan Sanislo <oystr@cs.washington.edu> for finding this
11 bug and submitting the patch.
12
13 2. Fixed the handling of the filesystem parameter in dump. A
14 '/mnt/disk' parameter caused the disk contents to be dumped,
15 but a '/mnt/disk/' parameter caused the mountpoint directory
16 to be dumped (generally an empty directory).
17
18 3. Improved the output of dump in order to tell which directory
19 it is currently dumping (when dumping a subtree).
20
21 4. Added the '-e' exclude inode option to dump. Thanks to
22 Isaac Chuang <ike@isl.stanford.edu> for contributing with the patch.
23
24 5. Added a REPORTING-BUGS file in order to provide a guide
25 on how to correctly report a bug in dump/restore.
26
27 6. Fixed a restore problem when restoring a hard link to an inode
28 having the immutable or append-only attribute set. Thanks to
29 Ambrose Li <acli@mingpaoxpress.com> for submitting the patch.
30
31 Changes between versions 0.4b12 and 0.4b13 (released January 21, 2000)
32 ======================================================================
33
34 1. Small Y2K fix in the man pages :). Thanks to Bernhard Sadlowski
35 <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug.
36
37 2. Removed the requirement to build the RPM as root from the
38 spec file. Thanks to Christian Weisgerber
39 <naddy@mips.rhein-neckar.de> for submitting this.
40
41 3. Fixed a bug in dump related to the 'filetype' feature of ext2fs,
42 causing dump to block when dumping really huge filesystems.
43 Many thanks to Patrik Schindler <poc@pocnet.net> for
44 helping me find this bug.
45
46 4. Fixed the treatment for an interrupt signal when dump access
47 the remote tape through RSH. Thanks to Christian Weisgerber
48 <naddy@mips.rhein-neckar.de> for providing the patch.
49
50 5. Fixed a bug which was causing dump/restore to display
51 garbage characters instead of the remote host name.
52
53 Changes between versions 0.4b11 and 0.4b12 (released January 8, 2000)
54 =====================================================================
55
56 1. Small fix in the dump man page. Thanks to Thorsten Kukuk
57 <kukuk@suse.de> for submitting the patch.
58
59 2. Fix for the exit code when using the size estimate option of
60 dump. Thanks to Matti Taskinen <mkt@rni.helsinki.fi> for
61 submitting the patch.
62
63 3. Handle EINTR in atomical reads/writes in dump, which was causing
64 dump to fail on some systems. Thanks to Eric Jergensen
65 <eric@dvns.com> for reporting the bug and submitting the patch.
66
67 4. Handle more than 16 characters for the device names in dumpdates.
68 (up to 255 now). Thanks to Rainer Clasen <bj@ncc.cicely.de> for
69 tracking down the problem and proposing the solution.
70
71 5. Fixed a bug in dump which prevented the creation of the
72 dumpdates file when doing a 0-level dump without already
73 having a dumpdates file. Thanks to Patrik Schindler
74 <poc@pocnet.net> for reporting the bug.
75
76 6. Changed the way dump 'S' flag reports the size estimate
77 from number of blocks into bytes (making it compatible
78 with the Solaris version, and simplifying things for
79 amanda users). Thanks to Jason L Tibbitts III
80 <tibbs@math.uh.edu> for reporting the bug.
81
82 7. Fixed a compatibility problem in linux/alpha dump tape format.
83 Now the linux/alpha dump are (again) compatible with the
84 other dump formats. But this breaks compatibility with
85 older dumps made on alpha. Thanks to Mike Tibor
86 <tibor@lib.uaa.alaska.edu> for helping me in finding this bug.
87
88 Changes between versions 0.4b10 and 0.4b11 (released December 5, 1999)
89 ======================================================================
90
91 1. Added a '--enable-kerberos' to configure.
92
93 2. Added a 'S' option to dump which determines the amount of space
94 that is needed to perform the dump without actually doing it, similar
95 to the Sun's ufsdump 'S' option. Patch contributed by Rob Cermak
96 <cermak@ahab.rutgers.edu>.
97
98 3. Added a 'M' multi-volume option to dump and restore which enables
99 dumping to multiple files (useful when dumping to an ext2
100 partition to make several dump files in order to bypass the 2GB
101 file size limitation). The 'f' argument is treated as a prefix and
102 the output files will be named <prefix>001, <prefix>002 etc. With
103 the 'M' flag, restore automatically selects the right file without
104 asking to enter a new tape each time.
105
106 4. Fixed a memory leak which was causing dump to grow very big
107 (270MB when dumping a 10GB filesystem...). Thanks to Jason
108 Fearon <jasonf@netrider.org.au> for reporting the bug.
109
110 Changes between versions 0.4b9 and 0.4b10 (released November 21, 1999)
111 ======================================================================
112
113 1. Make configure test if the system glob routines support
114 extended syntax (ALTDIRFUNC). If not, use the internal glob
115 routines instead of system ones. Thanks to Bernhard Sadlowski
116 <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug
117 and helping me resolve this and other minor libc5 compiling
118 glitches.
119
120 2. Fix a problem when dumping a ext2fs with the 'filetype'
121 feature enabled. Thanks to Patrick J. LoPresti
122 <patl@cag.lcs.mit.edu> for reporting the bug and to
123 Theodore Y. Ts'o <tytso@mit.edu> for providing the patch.
124
125 3. Made the nodump flag work on directories. A directory which
126 has the nodump flag gets never dumped, regardless of its
127 contents.
128
129 4. Integrate a patch from Jeremy Fitzhardinge <jeremy@goop.org>
130 which allows dump on an active ext3 filesystem. However, this
131 is a "quick and dirty" patch which enables backup of an ext3
132 filesystem through the ext2 compatibility (by ignoring the
133 NEEDS_RECOVERY bit). The journal file is not recognized and
134 it is dumped (it should not).
135
136 5. Test the superblock compatibility flags when dumping, in order
137 to be sure that we know how to deal with specific features.
138
139 Changes between versions 0.4b8 and 0.4b9 (released November 5, 1999)
140 ====================================================================
141
142 1. Use lchown instead of chown, fixing a possible security problem
143 when restoring symlinks (a malicious user could use this
144 to deliberately corrupt the ownership of important system files).
145 Thanks to Chris Siebenmann <cks@utcc.utoronto.ca> for detecting
146 this and providing the patch.
147
148 Changes between versions 0.4b7 and 0.4b8 (released November 3, 1999)
149 ====================================================================
150
151 1. Put dump sources under CVS, added Id tags in all files so
152 one can use 'ident' on binary files.
153
154 2. Added the dump/restore version in the usage text so one can
155 easily verify the version he is using.
156
157 3. Small patch from Nuno Oliveira <nuno@eq.uc.pt> which fixes
158 a va_start/va_end problem on linux-ppc (always call va_start
159 va_end in pairs each time we use a vararg function).
160
161 4. Added again the DT_* constants because old libc does not
162 contain them :(. Thanks to Eric Maisonobe <virnet@nat.fr>
163 for submitting the bug report.
164
165 5. Use ext2fs_llseek instead of llseek. With recent e2fsprogs
166 this should enable dumping big (huge) filesystems.
167
168 6. Added the RSH environment variable in order to be able to
169 use a rsh replacement like ssh when doing remote backups (and
170 bypass the security limitations of rcmd). Now you can do remote
171 backups without being root (or making dump setuid root).
172
173 7. Modified again the way dumpdates works. For incremental dumps,
174 we need to read dumpdates even if we are not using 'u' option.
175 Thanks to Bdale Garbee <bdale@gag.com> for his ideas on how
176 this should work.
177
178 Changes between versions 0.4b6 and 0.4b7 (released October 8, 1999)
179 ===================================================================
180
181 1. Removed the 'k' flag from the restore 'about' text if kerberos
182 was not compiled in.
183
184 2. Prototyped (f)setflags from e2fsprogs and corrected the calls
185 to them (fsetflags takes a char*, setflags an open fd!).
186
187 3. (f)setflags is called only if the flags aren't empty. If the
188 file is a special file, a warning is printed, because changing
189 flags implies opening the device. Normally, a special file
190 should not have any flag... (Debian bug #29775, patch provided
191 by Abhijit Dasgupta <abhijit@ans.net>).
192
193 4. Made possible to dump a file system not mentioned in /etc/fstab.
194 (Debian bug #11904, patch provided by Eirik Fuller <eirik@netcom.com>).
195
196 5. Changed the default behaviour to not create dumpdates
197 unless 'u' option is specified. Removed the old "debian-patch"
198 which provided the same thing. (Debian bug #38136, #33818).
199
200 6. Removed all those dump*announce, since they were getting old...
201
202 7. Added warning messages if dumpdates does not exist and
203 when an inferior level dump does not exist (except for a level 0
204 dump).
205
206 8. Debugged the glob calls in interactive mode: restore used a
207 dirent struct which was different from the /usr/include/dirent.h
208 one (this used to work, is it a glibc2 change?), so none of the
209 compat glob (which used /usr/include/dirent.h) or the system glob
210 worked. Restore use now the system dirent (and the system
211 DT_* constants), which are compatible with BSD ones.
212
213 9. Added a configure flag (--with-dumpdatespath) to specify
214 the location of dumpdates file. By default, it is
215 /etc/dumpdates.
216
217 10. Added the "AUTHOR" and "AVAILABILITY" sections and
218 included the current date/version in man pages.
219
220 11. Corrected the estimation of remaining time when
221 the operator doesn't change the tapes quickly enough. This
222 was an old bug, I thought I corrected it, and discovered
223 that in fact it was corrected in two different places, so
224 the results canceled each other...
225
226 Changes between versions 0.4b5 and 0.4b6 (released October 1, 1999)
227 ===================================================================
228
229 1. Integrated multiple patches from RedHat, Debian and SuSE:
230
231 - tweak dump/itime.c to not try to read dumpdates if the 'u' option
232 isn't specified.
233 - several fixes in the man pages.
234 - update the default tape device to /dev/st0.
235 - many updates for Linux Alpha (byte ordering, size_t etc).
236 - buffer overruns.
237 - use environment variable for TMPDIR (instead of /tmp).
238 - use sigjmp_buf instead of jmp_buf (RedHat bug #3260).
239 - workaround egcs bug (RedHat bugs #4281 and #2989).
240 - wire $(OPT) throughout Makefile's.
241
242 2. Upgrade the dump revision to 1, making possible to dump filesystems
243 made with e2fsprogs-1.15 or newer. Nothing seems to break...
244
245 3. Fix some compile warnings, prototype all functions.
246
247 4. Use glibc err/glob instead of internal compatibility
248 routines (only if available).
249
250 5. Fix a compile error on Linux 2.2.7 / libc5 (5.4.44) (patch provided
251 by Bernhard Sadlowski <sadlowsk@mathematik.uni-bielefeld.de>).
252
253 Changes between versions 0.4b4 and 0.4b5 (released September 22, 1999)
254 ======================================================================
255
256 1. Integrated the changes from FreeBSD-3.1-RELEASE
257 (mostly bug fixes, buffer overruns, dump has now an "automatic
258 tape length calculation" flag, dump/restore can use kerberos now
259 (this is NOT tested), use environment variables for TAPE and
260 RMT etc.).
261
262 2. Integrated three RedHat patches ("glibc", "kernel" and "bread" patches)
263
264 3. Corrected a bug in restore when using 'C' option with multi-volumes
265 tapes (files splited accros two tapes give "size changed" errors
266 when comparing).
267
268 4. Corrected the long standing bug when dumping multiple tapes.
269 This works for me, needs further testing.
270
271 Changes between versions 0.4b3 and 0.4b4 (released January 17, 1997)
272 ====================================================================
273
274 1. Dump now runs correctly on kernels 2.1.x
275 Fix made by Gerald Peters <gapeters@worldnet.att.net>
276
277 Changes between versions 0.4b2 and 0.4b3
278 ========================================
279
280 1. Use realpath() if available
281
282 2. Report statistics
283
284 Changes between versions 0.4b1 and 0.4b2
285 ========================================
286
287 1. Fixed the bug fix from Greg Lutz (I had made a mistake when integrating
288 the patch)
289
290 2. Fixed restore to make it able to read FreeBSD 2.x dumps again
291
292 3. Fixed configure.in to correctly handle --enable-rmt
293
294 Changes between versions 0.3 and 0.4b1
295 ======================================
296
297 1. Integrated the changes from 4.4BSD-Lite2
298
299 2. Integrated the patches from Debian and Red Hat
300
301 3. Portability changes: use the __u32, __u16, __s32, and __s16 types
302
303 4. Changed dump to use the Ext2fs library to get block addresses. This
304 should solve the endianness problem on SparcLinux.
305
306 5. Created a configure.in file (shamelessly stolen from the e2fsprogs
307 distribution's one) to use autoconf
308
309 6. Fixed a few minor bugs
310
311 Changes between versions 0.2e and 0.2f
312 ======================================
313
314 1. Added the creation of named pipes (fifos) in restore.
315
316 2. Added the -N flag in the restore manual page.
317
318 3. Added the file kernel.patch which contains the llseek() optimization
319 patch for 1.2.x kernels.
320
321 4. Fixed a bug in the restoration of symbolic links: owner and group were
322 not restored.
323
324 5. Integrated some changes from FreeBSD 2.2.
325
326 6. Added a call to ftruncate() after restoring each file to restore
327 correctly files ending by a hole.
328
329 Changes between versions 0.2d and 0.2e
330 ======================================
331
332 1. Fixed a bug in the "set owner/mode" process. Every file was restored
333 with owner = root (0) and group = root/wheel/whatever (0).
334
335 Changes between versions 0.2c and 0.2d
336 ======================================
337
338 1. Dump is now able to backup 2GB+ filesystems.
339
340 2. Dump and restore can now be linked as static binaries.
341
342 Changes between versions 0.2b and 0.2c
343 ======================================
344
345 1. Fixed a bug when dumping ``slow'' (i.e. normal) symbolic links.
346
347 Changes between versions 0.2a and 0.2b
348 ======================================
349
350 1. Really fixed the bug that I should have corrected in 0.2a.
351
352 2. Enabled optimization again.
353
354 Changes between versions 0.2 and 0.2a
355 =====================================
356
357 1. Disabled the optimization during compilation.
358
359 Changes between versions 0.1 and 0.2
360 ====================================
361
362 1. Fixed a bug in fstab.c which caused a null pointer to be stored in
363 the fs_type field (actually, I modified the file fstab.c to make it
364 use the mntent functions).
365
366 2. Dump and restore now use a 4.3 BSD compatible dump format. Backups
367 made by dump should be readable by the BSD restore and backups made
368 by the BSD dump should be readable by restore. Unfortunately, this
369 means that the dump format has changed between version 0.1 and version
370 0.2 :-(
371
372 3. Dump is now able to backup a subtree, it is no longer limited to whole
373 filesystems like the BSD version.
374
375 4. Dump now uses ext2_llseek() so it is able to backup filesystems bigger
376 than 2 GB.
377
378 Changes between versions 0.0 and 0.1
379 ====================================
380
381 1. Now create links rdump and rrestore during the `make install' step.
382
383 2. Linux port specific bugs added to the manual pages
384
385 3. Incorrect estimation of the number of tapes blocks fixed when doing
386 incremental backups.
387
388 4. Better ls-like format in restore in interactive mode.