Changeset 672 in MondoRescue for trunk/mondo/ChangeLog
- Timestamp:
- Jun 12, 2006, 9:55:14 AM (18 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
trunk/mondo/ChangeLog
r649 r672 3 3 MONDO CHANGES 4 4 5 v2.2.0(2006-06-12)5 2.2.0-1 (2006-06-12) 6 6 - indent on source files 7 - asprintf used instead of malloc_string + sprintf/strcpy/strcat without 8 verifications. Optimisations on memory consumption (allocate only size 9 needed + remove memory leaks) 10 - Internationalization (A huge thank to rene-marc dolhen 11 <rmd_at_mecreant.org> who achieved that big task)7 - asprintf used instead of malloc_string + sprintf/strcpy/strcat without verifications. Optimisations on memory consumption (allocate only size needed + remove memory leaks) (Bruno Cornec) 8 - Internationalization (A huge thank to rene-marc dolhen <rmd_at_mecreant.org> who achieved that big task) 9 10 2.0.8-4 (2006-06-10) 11 - nfsmount option added to allow redeployment from another NFS server (Bruno Cornec) 12 12 13 13 2.0.8-3 (2006-06-07) 14 - Fix a bug in -I and -E handling !! 15 (Paolo Bernardoni <bernardoni_at_sysnet.it>/Bruno Cornec) 16 - Fix permissions for autorun 17 (Bruno Cornec) 18 - Fox delivery problems for tar files with too restrictive umask 19 (Bruno Cornec) 20 - Fix parsing of DHCP information in start-nfs script 21 (Andree Leidenfrost) 14 - Fix a bug in -I and -E handling !! (Paolo Bernardoni <bernardoni_at_sysnet.it>/Bruno Cornec) 15 - Fix permissions for autorun (Bruno Cornec) 16 - Fox delivery problems for tar files with too restrictive umask (Bruno Cornec) 17 - Fix parsing of DHCP information in start-nfs script (Andree Leidenfrost) 22 18 23 19 2.0.8-2 (2006-06-02) 24 - PXE mode now supports change of NIC for redeployment 25 (Bruno Cornec) 20 - PXE mode now supports change of NIC for redeployment (Bruno Cornec) 26 21 27 22 2.0.8-1 (2006-05-25) 28 23 - new build process (Bruno Cornec) 29 - Fix a bug in .spec for RPM build (%attr now unused) 30 (Bruno Cornec) 24 - Fix a bug in .spec for RPM build (%attr now unused) (Bruno Cornec) 31 25 - Support of dm and LVM v2 (Andree Leidenfrost) 32 26 - New mr_strtok functionn added and used for dm support (Andree Leidenfrost) 33 - Complete doc is now a separate package. mondo still contains the 34 man pages and howto in minimal useful formats 35 (Bruno Cornec) 27 - Complete doc is now a separate package. mondo still contains the man pages and howto in minimal useful formats (Bruno Cornec) 36 28 - HOWTO now contains a new chapter on unattended support for mondo 37 29 - Increase size (4 times) of include|exclude variables 38 30 - Fix a bug on -I and -E not working with multiple parameters 39 - Fix a bug in verify for NFS by swapping nfs_remote_dir and isodir when 40 assembling name for image file to verify 41 (Andree Leidenfrost) 42 - Fix mondo when restoring filenames containing blanks 43 (still a problem for filenames with ') 44 (Bruno Cornec) 45 - Fix a RPM generation bug for rh7.3 (i386-redhat-linux prefix for binaries) 46 (Bruno Cornec) 31 - Fix a bug in verify for NFS by swapping nfs_remote_dir and isodir when assembling name for image file to verify (Andree Leidenfrost) 32 - Fix mondo when restoring filenames containing blanks (still a problem for filenames with ') (Bruno Cornec) 33 - Fix a RPM generation bug for rh7.3 (i386-redhat-linux prefix for binaries) (Bruno Cornec) 47 34 48 35 2.0.7 (2006-03-10) 49 - useless cat, sort|uniq commands removed 50 (Bruno Cornec/Sébastien Aperghis-Tramoni) 36 - useless cat, sort|uniq commands removed (Bruno Cornec/Sébastien Aperghis-Tramoni) 51 37 - Doc cleanup (Andree Leidenfrost) 52 - Add the actual format to messages after calls to function 53 is_this_a_valid_disk_format() about unsupported formats. 54 (Andree Leidenfrost) 55 - Abort|Warn when one of the include|exclude dirs (-I|-E) does not exist 56 (Bruno Cornec/Jeffs) 57 - Replaced partimagehack with ntfsclone from ntfsprogs package. 58 (Andree Leidenfrost) 38 - Add the actual format to messages after calls to function is_this_a_valid_disk_format() about unsupported formats. (Andree Leidenfrost) 39 - Abort|Warn when one of the include|exclude dirs (-I|-E) does not exist (Bruno Cornec/Jeffs) 40 - Replaced partimagehack with ntfsclone from ntfsprogs package. (Andree Leidenfrost) 59 41 - use df -P everywhere (Bruno Cornec) 60 42 - Paypal incitations removed (Andree Leidenfrost) … … 63 45 - no shared librairies and no X11 anymore (were useless) (Bruno Cornec) 64 46 - files > 2GB are now really supported (Andree Leidenfrost) 65 - new SGML based Mondo Rescue documentation + new Web site 66 (Bruno Cornec/Andree Leidenfrost) 47 - new SGML based Mondo Rescue documentation + new Web site (Bruno Cornec/Andree Leidenfrost) 67 48 - mondoarchive aborts when 'mindi --findkernel' gives a fatal error (See also Debian bug #352323.) (Andree Leidenfrost) 68 49 - /tmp not excluded anymore from backup (Bruno Cornec) … … 78 59 - fix to force growisofs to use speed=1 for DVD burning (Philippe De Muyter) 79 60 - now handles cifs correctly (Bruno Cornec) 80 - fix issue where mondoarchive ejects CD/DVD despite writing iso images 81 (Andree Leidenfrost) 61 - fix issue where mondoarchive ejects CD/DVD despite writing iso images (Andree Leidenfrost) 82 62 - Add -P option to df calls (Andree Leidenfrost/Chuan-kai Lin) 83 63 - fix usage of joint -B and -m options (Andree Leidenfrost/Efraim Feinstein) 84 - Quadrupled ARBITRARY_MAXIMUM from 500 to 2000 for mondorestore's filebrowser 85 (Andree Leidenfrost) 64 - Quadrupled ARBITRARY_MAXIMUM from 500 to 2000 for mondorestore's filebrowser (Andree Leidenfrost) 86 65 - remove the renice of mondoarchive (Hugo Rabson) 87 - relocate what was under /usr/share to /usr/lib (FHS compliance) 88 (Bruno Cornec/Andree Leidenfrost) 89 - manage non ambiguous delivery under /usr (packages) or /usr/local (tar ball) 90 (Bruno Cornec) 66 - relocate what was under /usr/share to /usr/lib (FHS compliance) (Bruno Cornec/Andree Leidenfrost) 67 - manage non ambiguous delivery under /usr (packages) or /usr/local (tar ball) (Bruno Cornec) 91 68 - disable x11 build by default (Bruno Cornec) 92 69 - remove sbminst (Bruno Cornec/Andree Leidenfrost) 93 70 - use parted2fdisk everywhere (Bruno Cornec) 94 71 - exports MONDO_LIB (Bruno Cornec) 95 - RPM build for fedora core 4, sles9, redhat 7.3, rhel 3/4, mandriva 2006.0, 96 mandrake 10.2/10.1 (Bruno Cornec/Gary Granger) 97 - interactive mode now asks for image size and prefix in NFS mode 98 (Gallig Renaud/Bruno Cornec) 72 - RPM build for fedora core 4, sles9, redhat 7.3, rhel 3/4, mandriva 2006.0, mandrake 10.2/10.1 (Bruno Cornec/Gary Granger) 73 - interactive mode now asks for image size and prefix in NFS mode (Gallig Renaud/Bruno Cornec) 99 74 - iso-prefix should be read in iso mode even when -H not given (Stan Benoit) 100 75 - VERSION/RELEASE Tag added (Bruno Cornec) 101 - many code cleanup, small fixes, PXE/NFS code improvements 102 (Sébastien Aperghis-Tramoni/Bruno Cornec) 76 - many code cleanup, small fixes, PXE/NFS code improvements (Sébastien Aperghis-Tramoni/Bruno Cornec) 103 77 104 78 v2.05 (2005-11-19) … … 109 83 110 84 v2.04_berlios (2005-10-30) 111 - Add -p option to generate ISO images file names with prefix. The new default 112 name for ISO images is mondorescue-1.iso, ... For PXE environment, you have 113 to use the prefix option on the command line (read README.pxe) 114 (Bruno Cornec) 85 - Add -p option to generate ISO images file names with prefix. The new default name for ISO images is mondorescue-1.iso, ... For PXE environment, you have to use the prefix option on the command line (read README.pxe) (Bruno Cornec) 115 86 - Mandrake 2005 support (Bruno Cornec) 116 87 - NFS patches (Yann Aubert <technique@alixen.fr>) 117 - mondorestore shouldn't now ask final questions with -H 118 (this is an unattended mode) (Bruno Cornec) 88 - mondorestore shouldn't now ask final questions with -H (this is an unattended mode) (Bruno Cornec) 119 89 120 90 v2.04 (2005-05-03) 121 - made mondo more clever about finding its home. Avoids mondo considering 122 directories like '/usr/share/doc/momdo' as its home. 91 - made mondo more clever about finding its home. Avoids mondo considering directories like '/usr/share/doc/momdo' as its home. 123 92 124 93 v2.03 (2004-09-29) … … 127 96 128 97 v2.02 (2004-07-21) 129 - instead of using 'dd' to erase partition table, delete existing 130 partitions w/ the same call to fdisk that is used to create the 131 new partitions; this should avoids locking up the partition table 132 - set bootable partition in the above same call to fdisk, for 133 the same reason (avoids locking up the partition table) 98 - instead of using 'dd' to erase partition table, delete existing partitions w/ the same call to fdisk that is used to create the new partitions; this should avoids locking up the partition table 99 - set bootable partition in the above same call to fdisk, for the same reason (avoids locking up the partition table) 134 100 - better software RAID support 135 101 - mount ext3 partitions as ext2 when restoring - better for Debian … … 139 105 v2.01 (2004-06-21) 140 106 - fixed cvs for SuSE systems 141 - fixed NTFS backup/restore bug relating to partimagehack 142 log file overflow and NTFS v non-NTFS differentiation 107 - fixed NTFS backup/restore bug relating to partimagehack log file overflow and NTFS v non-NTFS differentiation 143 108 - more reliable extraction of config info from CDs, floppies 144 109 - better support of ISO dirs at restore-time (Conor Daly) … … 165 130 - added support for boot/root multi floppies 166 131 - call 'mt' to set block size to 32K before opening in/out tape 167 - updated mondo-prep.c to create each disk's partitions all at once 168 (one call per drive) instead of one call to fdisk per partition 169 - when extracting cfg file and mountlist from all.tar.gz (tape copy), 170 use block size of INTERNAL_TAPE_BLK_SIZE, not TAPE_BLOCK_SIZE 132 - updated mondo-prep.c to create each disk's partitions all at once (one call per drive) instead of one call to fdisk per partition 133 - when extracting cfg file and mountlist from all.tar.gz (tape copy), use block size of INTERNAL_TAPE_BLK_SIZE, not TAPE_BLOCK_SIZE 171 134 - added star and rudimentary SELinux support 172 135 - fixed lots of bugs 173 - all logging now goes to /var/log/mondo-archive.log, with symlink to 174 /tmp/mondo-restore.log for restore-time log-tracking 136 - all logging now goes to /var/log/mondo-archive.log, with symlink to /tmp/mondo-restore.log for restore-time log-tracking 175 137 - added grub-install.patched 176 138 - removed embleer & other binaries … … 181 143 - fixed chmod/chown bug (Jens Richter) 182 144 - ask user to confirm NFS mountpoint in Interactive Mode 183 - rewritten format_everything() to make sure LVMs, RAIDs and 184 regular partitions are prepped in the correct order 145 - rewritten format_everything() to make sure LVMs, RAIDs and regular partitions are prepped in the correct order 185 146 - better magicdev support 186 147 - rewritten external binary caller subroutine … … 189 150 - fixed multi-tape support 190 151 - re-implemented -D and -J 191 - fixed bug in extract_config_file_from_ramdisk() which 192 affected tape+floppy users 193 - updated is_incoming_block_valid() to make it 194 return end-of-tape if >300 flotsam blocks 152 - fixed bug in extract_config_file_from_ramdisk() which affected tape+floppy users 153 - updated is_incoming_block_valid() to make it return end-of-tape if >300 flotsam blocks 195 154 - unmount CD-ROM before burning (necessary for RH8/9) 196 155 - fixed some stray assert()'s … … 198 157 - make user remove floppy/CD before restoring interactively from tape 199 158 - fixed bug in am_I_in_disaster_recovery_mode() 200 - added code to nuke_mode() to make sure NFS 201 (backup) share is mounted in Nuke Mode 159 - added code to nuke_mode() to make sure NFS (backup) share is mounted in Nuke Mode 202 160 - improved tape device detection code 203 161 - better GRUB support … … 219 177 - fixed RAID-related bug in where_is_root_mounted() 220 178 - ISO tweaks 221 - fixed silly bug in load_filelist() which stopped 222 funny German filenames from being handled properly 179 - fixed silly bug in load_filelist() which stopped funny German filenames from being handled properly 223 180 - misc fixes (Michael Hanscho's friend) 224 181 - added rudimentary support for SME … … 234 191 - post_param_configuration() --- store iso-dev and isodir 235 192 - added post-nuke-sample.tgz to package 236 - Nuke Mode now checks mountlist against hardware; offer user 237 opportunity to edit mountlist if insane; if user declines, abort 193 - Nuke Mode now checks mountlist against hardware; offer user opportunity to edit mountlist if insane; if user declines, abort 238 194 - added lots of assert()'s and other checks 239 195 - ran code thru Valgrind to catch & fix some memory leaks 240 - made mondo-restore.c smaller by moving some subroutines to 241 common/libmondo-raid.c and mondorestore/mondo-rstr-compare.c 242 - added '-Q' flag, to let user test mondoarchive's ability 243 to find their boot loader and type 196 - made mondo-restore.c smaller by moving some subroutines to common/libmondo-raid.c and mondorestore/mondo-rstr-compare.c 197 - added '-Q' flag, to let user test mondoarchive's ability to find their boot loader and type 244 198 - improved which_boot_loader() 245 - when burning or comparing to a CD, defeat autorun if it is 246 running, to avoid confusing mondoarchive and the user 247 - if original backup media no longer available at boot-time 248 then offer user chance to choose another media source 249 - when booting, type 'nuke noresize' to nuke w/o resizing 250 mountlist to fill your drives 199 - when burning or comparing to a CD, defeat autorun if it is running, to avoid confusing mondoarchive and the user 200 - if original backup media no longer available at boot-time then offer user chance to choose another media source 201 - when booting, type 'nuke noresize' to nuke w/o resizing mountlist to fill your drives 251 202 - add 'textonly' when booting, to avoid using Newt gui 252 203 - run nice(20) to prioritize mondoarchive at start 253 - don't pause and wait for next blank CD at backup-time 254 unless necessary (e.g. previous CD has not been removed) 204 - don't pause and wait for next blank CD at backup-time unless necessary (e.g. previous CD has not been removed) 255 205 - get_phys_size_of_drive() --- better support of older drives 256 206 - don't eject if "donteject" is in kernel's command line … … 276 226 - add '/' to custom filelist as workaround for obscure bug 277 227 - ask user speed of CDRW if writing to CD 278 - find_cdrom_device() --- if nonexistent/not found then 279 make sure to return '' as dev str and 1 as res 228 - find_cdrom_device() --- if nonexistent/not found then make sure to return '' as dev str and 1 as res 280 229 - tweaked restore scripts tgz 281 230 - cleaned up find_cdrom_device() 282 - if user creates /usr/share/mondo/payload.tgz then untar 283 payload to CD at backup-time 231 - if user creates /usr/share/mondo/payload.tgz then untar payload to CD at backup-time 284 232 - fixed insist_on_this_cd_number() 285 233 - fixed am_i_in_disaster_recovery_mode() … … 289 237 - fixed nfs support 290 238 - fixed iso support 291 - at restore-time, only sort mountlist internally, 292 in mount_all_devices() and unmount_all_devices() 239 - at restore-time, only sort mountlist internally, in mount_all_devices() and unmount_all_devices() 293 240 - fixed cosmetic bug in label-partitions-as-necessary 294 241 - updated documentation … … 306 253 - sensibly_set_tmpdir_and_scratchdir() --- exclude smb and smbfs 307 254 - better logging by eval_call_to_make_ISO() 308 - accept -J <fname> to let user provide their own fs catalog 309 instead of -I <paths> to backup 310 - if dir excluded with -E or included with -I and dir is actually 311 a softlink then exclude/include the dir pointed to, as well 255 - accept -J <fname> to let user provide their own fs catalog instead of -I <paths> to backup 256 - if dir excluded with -E or included with -I and dir is actually a softlink then exclude/include the dir pointed to, as well 312 257 - better location for manpage 313 - adjusted block size of tarball at start of tape, to help 314 users w/ broken tape driver firmware 258 - adjusted block size of tarball at start of tape, to help users w/ broken tape driver firmware 315 259 - sort -u fstab after modifying it 316 260 - if backup type is nfs then don't estimate noof media … … 327 271 - wipe only the partition table (not MBR) when repartitioning drives 328 272 - ignore lilo.conf.anaconda when looking for lilo.conf file 329 - accepts '-l RAW' to backup/restore original boot sector instead 330 of running grub or lilo to init it after restoring 273 - accepts '-l RAW' to backup/restore original boot sector instead of running grub or lilo to init it after restoring 331 274 - fixed&updated stabgrub-me script; software RAID + GRUB work now 332 275 - mount/unmount /boot partition for Gentoo 1.2 users … … 334 277 - disabled spurious warnings 335 278 - unmount/remount supermounts at start/end of live restore, if nec. 336 - cleaned up mondo's tape block handling (now, TAPE_BLOCK_SIZE=128K 337 and I've added INTERNAL_TAPE_BLK_SIZE=32K variable for buffering) 279 - cleaned up mondo's tape block handling (now, TAPE_BLOCK_SIZE=128K and I've added INTERNAL_TAPE_BLK_SIZE=32K variable for buffering) 338 280 - added Makefile 339 281 - added -l RAW option, to backup and restore original MBR … … 372 314 - fixed "Can't find first ISO when verifying nonbootable ISO" bug 373 315 - multithreaded make_afioballs_and_images() 374 - tmpdir and scratchdir set sensibly whether mondoarchive is called with 375 command-line parameters or not 376 - fixed bug in strip_spaces() which stopped it from handling 377 small strings correctly - affected mountlist editor 378 - create a repaired copy of grub-install which is RAID-friendly; 379 use it when initializing boot sector with run_grub() 316 - tmpdir and scratchdir set sensibly whether mondoarchive is called with command-line parameters or not 317 - fixed bug in strip_spaces() which stopped it from handling small strings correctly - affected mountlist editor 318 - create a repaired copy of grub-install which is RAID-friendly; use it when initializing boot sector with run_grub() 380 319 - fixed bug in mondo-makefilelist 381 320 … … 395 334 - unmount/remount /mnt/floppy before/after backing up, if Mandrake 396 335 - restructured the source files 397 - fixed serious bug in line 1546 - should have been !=, not ==; stopped 398 mondorestore from correctly restoring big files 336 - fixed serious bug in line 1546 - should have been !=, not ==; stopped mondorestore from correctly restoring big files 399 337 - added '#include <signal.h>' to my-stuff.h 400 338 - exclude "incheckentry xwait()" from changed.files … … 403 341 - insist on tape #1 when start verifying 404 342 - added internal buffering, replacing the external 'buffer' exe 405 - if differential backup then don't permit formatting or fdisking, 406 whether Interactive or Nuke mode 407 - if mondorestore is run on live filesystem (or from ramdisk) without 408 parameters then mondorestore will ask which backup media (tape, CD, etc.) 409 was used; it will read the config file from the media and proceed from there 343 - if differential backup then don't permit formatting or fdisking, whether Interactive or Nuke mode 344 - if mondorestore is run on live filesystem (or from ramdisk) without parameters then mondorestore will ask which backup media (tape, CD, etc.) was used; it will read the config file from the media and proceed from there 410 345 - if tape streamer is softlink then resolve it first 411 346 - incorporate post-nuke tarball 412 - if user doesn't specify tape size, proceed anyway; behave intelligently 413 in the event of end-of-tape 414 - prefix bkpinfo->restore_path to biggiefile fname before generating 415 checksum & comparing to archived biggiefile 416 - if /etc/lilo.conf not found not /etc/lilo.conf.anaconda found 417 then create a softlink from the former to the latter, to work 418 around RH7.3's b0rken LILO support 347 - if user doesn't specify tape size, proceed anyway; behave intelligently in the event of end-of-tape 348 - prefix bkpinfo->restore_path to biggiefile fname before generating checksum & comparing to archived biggiefile 349 - if /etc/lilo.conf not found not /etc/lilo.conf.anaconda found then create a softlink from the former to the latter, to work around RH7.3's b0rken LILO support 419 350 - LFS support (mharris, michele, hugo) 420 351 - fixed verify bug --- CD#1 was being verified again & again & ... … … 424 355 - better handling of bigfiles' checksums, perms and owns 425 356 - delete final filelist if <=2 bytes long 426 - if kernel not found and mondo in graphics mode then popup and ask 427 for kernel path+filename 357 - if kernel not found and mondo in graphics mode then popup and ask for kernel path+filename 428 358 - misc clean-ups (Troff) 429 359 - noninteractivity-related & misc clean-ups (Carl Wilhelm Soderstrom) … … 439 369 - cleaned up live filesystem restoration 440 370 - permit '-0' (no compression) flag 441 - if user restores selectively (picks some files to be restored) then 442 large files archived during '-L' (LZO) backup will now be restored properly 371 - if user restores selectively (picks some files to be restored) then large files archived during '-L' (LZO) backup will now be restored properly 443 372 - fixed bug in make-me-bootable 444 373 - don't change /tmp's permissions unless it doesn't exist & has to be created 445 374 - changed $i to $j in mondo-makefilelist (Karsten) 446 - fixed 'I refuse to write floppy images to /dev/fd0 because its size is 447 not apparent from its filename' bug 375 - fixed 'I refuse to write floppy images to /dev/fd0 because its size is not apparent from its filename' bug 448 376 - disabled 'nr-failed-disks' entry in raidtab generator 449 377 - added '-o' flag to let user choose LILO as CD/floppy's boot loader … … 454 382 - mondorestore makes /var/run/console in case user excluded it 455 383 - the -s switch now accepts multiple sizes, one per tape/CD 456 - added -u switch for streaming to device; like -t (tape) but without 457 various error checks; use with caution 384 - added -u switch for streaming to device; like -t (tape) but without various error checks; use with caution 458 385 - bugfixes; cleaned up code (Troff) 459 386 - changed lockfiles from static to random filenames … … 462 389 - added -follow to search for home (Bruno Vidal) 463 390 - unmount/remount CD-ROM at start/end if supermounted 464 - delete old /tmp/filelist.full,biggielist.txt if restoring to 465 live filesystem 391 - delete old /tmp/filelist.full,biggielist.txt if restoring to live filesystem 466 392 - if Mindi aborted then show last 6 lines of its log before mondoarchive aborts 467 393 - reformatted, redistributed source code (Stan Benoit) … … 475 401 476 402 v1.42-1 (2002-04-07) 477 - mondorestore may now be run on live filesystem, i.e. user does not 478 have to reboot in order to restore data; works for CD's and tapes; 479 NFS and ISO restores will be supported at a later date 403 - mondorestore may now be run on live filesystem, i.e. user does not have to reboot in order to restore data; works for CD's and tapes; NFS and ISO restores will be supported at a later date 480 404 - fixed restore-time RAID-handling bugs 481 405 - added -E patch to exclude files and/or dirs properly (Andy Glass) 482 406 - cat /proc/cpuinfo and run uname -r; save output to logfile at backup-time 483 407 - fixed minor bug in stablilo-me 484 - if /tmp/fstab.new exists then use _it_ instead of /tmp/fstab to label 485 ext2 or ext3 partitions 408 - if /tmp/fstab.new exists then use _it_ instead of /tmp/fstab to label ext2 or ext3 partitions 486 409 - bootable CD's now use SYSLINUX instead of LILO (Andy Glass) 487 410 - don't pclose() tape after calling closein_tape(): the latter does it already … … 495 418 v1.40-7 (2002-03-10) 496 419 - if '-L' then test for lzop; abort if not present 497 - if Mondo cannot figure out which boot loader the user is using, 498 tell the user to read the manual and mention -l, -f switches 420 - if Mondo cannot figure out which boot loader the user is using, tell the user to read the manual and mention -l, -f switches 499 421 - abort if ramdisk not available 500 422 - updated man page (Mikael Hultgren) … … 517 439 - added updated docs (Cafeole, Randy Delfs, Stan Benoit et al) 518 440 - added pico to restore-scripts.tgz; deprecated some scripts 519 - if restoring to /dev/ida/* or /dev/cciss/* then drop the 'p' 520 before trying to discover the size of the partition itself 441 - if restoring to /dev/ida/* or /dev/cciss/* then drop the 'p' before trying to discover the size of the partition itself 521 442 - support pathnames containing spaces (KirkJT) 522 443 - added -C switch, to treat CD burner almost as a tape streamer … … 548 469 - changed get_phys_size_of_drive() to read value from fdisk, not dmesg 549 470 - if hard disk does not exist then warn user when editing mountlist 550 - if boot+floppy disk images were not created then don't offer to write them 551 to disks; instead, mention the mindi.iso bootable CD images 471 - if boot+floppy disk images were not created then don't offer to write them to disks; instead, mention the mindi.iso bootable CD images 552 472 - exclude each partition's .autofsck from backup 553 473 - fixed bug in CD restoration progress form … … 564 484 - source code has been beautified (Stan Benoit) 565 485 - makefile, Config implemented (Stan Benoit) 566 - use /tmp/mondo-restore.cfg instead of all kinds of silly /tmp files 567 at boot-time; slowly migrating to a proper /tmp/mondo-restore.cfg file 486 - use /tmp/mondo-restore.cfg instead of all kinds of silly /tmp files at boot-time; slowly migrating to a proper /tmp/mondo-restore.cfg file 568 487 - changed 'retval+=load_raidtab()' to 'load_raidtab()' 569 488 - patch to make it easier to install Mondo almost anywhere (Ian Mortimer) 570 - proper, one-char switches replace the long-winded multi-char switches 571 of 1.36 and earlier 572 - migrating from lots of /tmp files in ramdisk to a single configuration 573 file on ramdisk, at /tmp/mondo-restore.cfg 489 - proper, one-char switches replace the long-winded multi-char switches of 1.36 and earlier 490 - migrating from lots of /tmp files in ramdisk to a single configuration file on ramdisk, at /tmp/mondo-restore.cfg 574 491 - estimates the number of media required by this backup session 575 492 - changed 'retval+=load_raidtab()' to 'load_raidtab()' … … 580 497 - if the user specifies boot loader or device, record those settings 581 498 - fixed cdrecord-related bug in .spec file 582 - if backing up specific paths - e.g. /usr/local/bin - then include 583 /usr and /usr/local/ in filelist; don't backup /usr/* but do recreate 584 /usr and /usr/local/; this is to help the selective restore subroutine 499 - if backing up specific paths - e.g. /usr/local/bin - then include /usr and /usr/local/ in filelist; don't backup /usr/* but do recreate /usr and /usr/local/; this is to help the selective restore subroutine 585 500 - user's BIOS is assumed not to be able to handle LILO's sector-seek merging 586 501 - switches --broken-bios, --debug and --chunk-size have been removed … … 618 533 - if bzip is <v0.99 then abort 619 534 - if hard disk fills to capacity then abort 620 - --post-nuke <tarball> will incorporate a tarball of tools and data files 621 of the user's choosing in the boot CD/floppies; the tarball must contain 622 an executable, 'post-nuke', which will be called at the end of Nuke Mode 535 - --post-nuke <tarball> will incorporate a tarball of tools and data files of the user's choosing in the boot CD/floppies; the tarball must contain an executable, 'post-nuke', which will be called at the end of Nuke Mode 623 536 - --text-mode forces Mondo Archive to use text only (no newt or ncurses) 624 537 - the deprecated scripts have now been disabled 625 - boot floppy will use 'safe' LILO calls; boot CD will use 'fast' LILO calls 626 unless --broken-bios is supplied in call to Mondo 538 - boot floppy will use 'safe' LILO calls; boot CD will use 'fast' LILO calls unless --broken-bios is supplied in call to Mondo 627 539 - cdrecord's buffer is now 4 MB (was 16 MB) 628 540 - tape support has been rewritten from ground up … … 642 554 - fixed bugs in actually_verify_the_tapes() 643 555 - fixed some popen/pclose mismatches 644 - no longer run compare_a_biggiefile() right after restore_a_biggiefile(), 645 as it can upset the restoration of biggiefiles which are spread across CD's 556 - no longer run compare_a_biggiefile() right after restore_a_biggiefile(), as it can upset the restoration of biggiefiles which are spread across CD's 646 557 647 558 v1.33-8 (2001-12-19) … … 687 598 - cosmetic fixes to mondo-prep.c and mondo-newt.c 688 599 - put quotes around chmod's parameter 689 - subtract 5% from reported maximum storage capacity of drives, when checking 690 that the parttions will fit; some drives lie about their max cap 600 - subtract 5% from reported maximum storage capacity of drives, when checking that the parttions will fit; some drives lie about their max cap 691 601 - stablilo-me no longer tries to mark partitions as bootable 692 602 - better exclusion of /var/log/pacct 693 603 - tape users can now use --verify to verify archives w/o rebooting 694 604 - checksum calculated for every file backed up - for future expansion 695 - implemented a _real_ --differential switch, to backup all files changed 696 or added since the last complete backup 605 - implemented a _real_ --differential switch, to backup all files changed or added since the last complete backup 697 606 698 607 v1.32-2 (2001-12-09) 699 608 - if no RAID devices after editing mountlist then delete raidtab 700 609 - don't try to mount RAID partitions (components of RAID dev) 701 - don't add partitions to mountlist when extrapolating from mountlist 702 unless the partitions are not already present in mountlist 610 - don't add partitions to mountlist when extrapolating from mountlist unless the partitions are not already present in mountlist 703 611 - don't format partitions if partitioning failed 704 612 - when user selects RAID level, establish whether said level is available … … 708 616 - log_it() makes sure it only writes one \n on the end of each logged line 709 617 - detects if user is using GRUB/LILO/other; abort if neither LILO nor GRUB 710 - passes the boot loader's name and device to Mindi, to be available at 711 boot-time 712 - if user successfully restores using Nuke or Interactive w/o error then Mondo 713 generates a random message, basically saying, "Cool." 618 - passes the boot loader's name and device to Mindi, to be available at boot-time 619 - if user successfully restores using Nuke or Interactive w/o error then Mondo generates a random message, basically saying, "Cool." 714 620 - enlarged afio's "-M" parameter from 12M to 16M 715 - if stablilo-me fails then restore the original lilo.conf and fstab before 716 offering user the opportunity to edit them (Dean Carpenter) 621 - if stablilo-me fails then restore the original lilo.conf and fstab before offering user the opportunity to edit them (Dean Carpenter) 717 622 - added ISO Mode, to let user restore from ISO images in directory 718 623 - cosmetic fix to compare/restore phase windows … … 726 631 - exclude each device's lost+found directory 727 632 - mountlist editor now permits advanced editing of RAID devices 728 - fixed a silly bug that was causing freeze-up on <1% of systems; if biggielist 729 was empty then sometimes mondo-archive would freeze when backing up big files 730 (cos there weren't any *g*) 633 - fixed a silly bug that was causing freeze-up on <1% of systems; if biggielist was empty then sometimes mondo-archive would freeze when backing up big files (cos there weren't any *g*) 731 634 - partition and set type in 2-stage process (was 1-stage) 732 635 - cleaned up Compare Mode's logging a bit, to make it more legible … … 734 637 - exclude ext3's .journal files from backup (yes, this time I really do) 735 638 - '--my-kernel SUCKS' has been replaced with '--my-kernel FAILSAFE' 736 - softlink mondo-restore to /usr/bin/ before calling Mindi, so that Mindi can 737 pick up mondo-restore and its depdendencies easily 639 - softlink mondo-restore to /usr/bin/ before calling Mindi, so that Mindi can pick up mondo-restore and its depdendencies easily 738 640 - copy mondo-restore to Mondo's temp dir before calling Mindi 739 641 - don't try to compress MP3's, MPG's, WMV's or RA's 740 - cleaned up mondo-newt.c; alphabetized the subroutines; moved some tools from 741 there to mondo-tools.c 742 - mondo-tarme and mondo-restore are now compiled as dynamic executables; I 743 have amended install.sh and mondo-1.31.spec accordingly 642 - cleaned up mondo-newt.c; alphabetized the subroutines; moved some tools from there to mondo-tools.c 643 - mondo-tarme and mondo-restore are now compiled as dynamic executables; I have amended install.sh and mondo-1.31.spec accordingly 744 644 - now make boot/data floppies at end of backup, not beginning 745 645 - backup in alphabetical order now, not reverse-alpha 746 - if Compare Mode reports changed files /var/* and /etc/mtab only then tell 747 the user their archives are good despite the differneces 646 - if Compare Mode reports changed files /var/* and /etc/mtab only then tell the user their archives are good despite the differneces 748 647 - cut back on excessively verbose logging at backup-time 749 648 - when vacuum-packing a partition, describe it as "(max size)", not "(0 MB)" 750 649 - if swap is already mounted, don't abort: just say, 'swap already mounted' 751 - when unmounting all devices, don't try to unmount something if it's not even 752 mounted; if it's not mounted, say so but don't try to unmount it: you'll 753 just end up looking very silly :) 650 - when unmounting all devices, don't try to unmount something if it's not even mounted; if it's not mounted, say so but don't try to unmount it: you'll just end up looking very silly :) 754 651 755 652 v1.30 (2001-11-28) 756 - I have reorganized the C code, sorted the subroutines alphabetically, and 757 re-jigged Mondo's installation routines (inc. the RPM spec file) to reflect 758 the changes 759 - check for mondo-restore's presence; if absent then abort & ask user to check 760 to see if Mondo was installed properly 653 - I have reorganized the C code, sorted the subroutines alphabetically, and re-jigged Mondo's installation routines (inc. the RPM spec file) to reflect the changes 654 - check for mondo-restore's presence; if absent then abort & ask user to check to see if Mondo was installed properly 761 655 - always exclude /var/log/pacct from backup 762 656 - optimal set size is now 24MB 763 657 - install.sh will abort w/error if compilation of mondo-* fails 764 658 - after creating the mountlist, put a copy in Mondo's tmp directory 765 - each ext3 filesystem has a .journal file in its base directory; Mondo will 766 now exclude each of these journal files automatically 767 - use the archived copy of LILO, not the copy from the restore-time ramdisk, 768 to configure the MBR 659 - each ext3 filesystem has a .journal file in its base directory; Mondo will now exclude each of these journal files automatically 660 - use the archived copy of LILO, not the copy from the restore-time ramdisk, to configure the MBR 769 661 - size_of_file() now returns long long, not long 770 662 - find mondo-archive and set MONDO_PATH accordingly 771 663 - scratchdir is /home/mondo.scratch by default 772 - at backup-time, run mondo-restore in test mode to make sure that the 773 executable is sane & will run OK at restore-time 664 - at backup-time, run mondo-restore in test mode to make sure that the executable is sane & will run OK at restore-time 774 665 775 666 v1.19 (2001-11-14) … … 778 669 - at restore-time, always check for presence of decompressor; abort if notfound 779 670 - run 'chmod -r 1777 /mnt/RESTORING/tmp' after restoring 780 - fixed bugs in Compare Mode:- 781 - missing else{} 782 - erroneous 'cannot add archive' error; res+=retval should have read 783 retval+=res (Tony Kim) 784 - use 'umount /mnt/cdrom' instead of 'umount /dev/cdrom', to make it compatible 785 with busybox's umount command 786 - fixed bug in mondo-archive which stopped Mondo from pausing to retry if the 787 CD/ISO fails the verification phase 671 - fixed bugs in Compare Mode:- - missing else{} - erroneous 'cannot add archive' error; res+=retval should have read retval+=res (Tony Kim) 672 - use 'umount /mnt/cdrom' instead of 'umount /dev/cdrom', to make it compatible with busybox's umount command 673 - fixed bug in mondo-archive which stopped Mondo from pausing to retry if the CD/ISO fails the verification phase 788 674 - modified compare_all_tarballs() in mondo-restore.c 789 - if running interactively then offer to modify fstab and lilo.conf if user 790 says s/he has modified the mountlist 791 - if mondo-tarme segfaults (as it does, on a _very_ few systems) then Mondo 792 picks up on the fact & reports it 793 - when restoring, create /mnt/RESTORING/tmp before _and_ after restoring data, 794 just in case it is necessary; also set permissions okay 675 - if running interactively then offer to modify fstab and lilo.conf if user says s/he has modified the mountlist 676 - if mondo-tarme segfaults (as it does, on a _very_ few systems) then Mondo picks up on the fact & reports it 677 - when restoring, create /mnt/RESTORING/tmp before _and_ after restoring data, just in case it is necessary; also set permissions okay 795 678 - C code is now compiled 'O0' (without optimizations) 796 679 - each CD's slices were being listed to screen; now, piped to /dev/null … … 802 685 v1.17 (2001-10-20) 803 686 - the deprecated scripts now work with LZO 804 - if mondo-restore cannot mount all devices, it will _unmount_ whatever it did 805 mount, before it returns an error 687 - if mondo-restore cannot mount all devices, it will _unmount_ whatever it did mount, before it returns an error 806 688 - works with Red Hat 6.2 and 2.2.19 kernel 807 689 - cosmetic change to output (no more 'burning first/last CD' msgs) 808 690 - abort if user uses '--use-lzo' but doesn't install LZO(P) first 809 691 - changed --differential's \( and \) to ( and ) <-- thanks, Michael Moellney) 810 - added Marcus Oberhumer's LZO to Mindi and Mondo; use '--use-lzo' switch 811 to activate this groovy new compression algorithm; it is not as efficient 812 as bzip2 but it is can take up to 50% less time than bzip2 to compress data 692 - added Marcus Oberhumer's LZO to Mindi and Mondo; use '--use-lzo' switch to activate this groovy new compression algorithm; it is not as efficient as bzip2 but it is can take up to 50% less time than bzip2 to compress data 813 693 - incorporated Bruno Cornec's "--cd-recovery" patch 814 694 - at restore-time, set biggiefiles to +x … … 816 696 - don't report non-existent differences/errors between archives and filesystem 817 697 - fixed bug in --differential switch (Michael Moellney) 818 - scripts will no longer function unless gawk and printf are included in 819 Mindi's deplist.txt; add them if you want to use the scripts 698 - scripts will no longer function unless gawk and printf are included in Mindi's deplist.txt; add them if you want to use the scripts 820 699 821 700 v1.16 (2001-10-10) 822 701 - default scratchdir is `pwd`/mondo.scratch.$$ (was mondo.scratch) 823 - when asking user to send me a log file, warn them that the log file may 824 contain information which they do not want me to see but point out that I 825 cannot help much unless I have a copy of their log file 826 - when comparing to or restoring from tape, only update the progree form 827 if/when I've just restored files; otherwise, I would screw up the time est.s 702 - when asking user to send me a log file, warn them that the log file may contain information which they do not want me to see but point out that I cannot help much unless I have a copy of their log file 703 - when comparing to or restoring from tape, only update the progree form if/when I've just restored files; otherwise, I would screw up the time est.s 828 704 - increase afio's block count from 256 to 1024 when restoring from tape 829 705 - plays nicely with devfs-enabled kernels and distros 830 - when restoring or comparing, track progress by the archive# (N out of M), 831 not by the file#; the former is more reliable and useful than the latter 832 - has been tested successfully with the following distros 833 - Linux-Mandrake 8.1 and its 2.4.8, devfs-enabled stock kernel --- tape / CDR 834 - Red Hat 7.1 and its 2.4.2 stock kernel --- tape / CDR 835 - tape streamers are now supported; use the '--write-tapes <dev> <sizeMB>' 836 switch, e.g. --write-tapes /dev/st0 4096 706 - when restoring or comparing, track progress by the archive# (N out of M), not by the file#; the former is more reliable and useful than the latter 707 - has been tested successfully with the following distros - Linux-Mandrake 8.1 and its 2.4.8, devfs-enabled stock kernel --- tape / CDR - Red Hat 7.1 and its 2.4.2 stock kernel --- tape / CDR 708 - tape streamers are now supported; use the '--write-tapes <dev> <sizeMB>' switch, e.g. --write-tapes /dev/st0 4096 837 709 - LVM support is stabilizing 838 - when mounting/unmounting partitions to be restored, don't forget to 839 mount/unmount the swap partitions 710 - when mounting/unmounting partitions to be restored, don't forget to mount/unmount the swap partitions 840 711 - when zeroing a drive, use plain old C fopen/fputc/fclose, not 'dd' 841 - after restoring data and running LILO, unmount (as usual) and then run 842 label-partitions-as-necessary, to set the ext2fs LABELs if it is appropriate 712 - after restoring data and running LILO, unmount (as usual) and then run label-partitions-as-necessary, to set the ext2fs LABELs if it is appropriate 843 713 - fixed bugs in selective restoration scripts and executable 844 - fixed silly bug in hack-lilo; bug was triggered when I moved to Mandrake 8.1; 845 bug caused '/dev/fd0' entry in lilo.conf to become corrupted 846 - if user tries to make a bootable rescue CD with the '--no-bkpath' switch then 847 politely direct them to Mindi :-) 714 - fixed silly bug in hack-lilo; bug was triggered when I moved to Mandrake 8.1; bug caused '/dev/fd0' entry in lilo.conf to become corrupted 715 - if user tries to make a bootable rescue CD with the '--no-bkpath' switch then politely direct them to Mindi :-) 848 716 849 717 v1.15 (2001-09-28) 850 - serialized the call to mondo-filelistchop.c; helps to improve visual feedback 851 (yeah, sounds weird but it's true) 718 - serialized the call to mondo-filelistchop.c; helps to improve visual feedback (yeah, sounds weird but it's true) 852 719 - the switches --differential and --exclude-paths should play nicely now 853 720 - improved the logging a bit, esp. when ejecting or unmounting CD's 854 - improved the visual feedback (specifically, the bar and percentage figures 855 indicating time spent & remaining) 721 - improved the visual feedback (specifically, the bar and percentage figures indicating time spent & remaining) 856 722 - experimental LVM support 857 723 - fixed 'cannot eject CD' bug, which was caused by busybox's sync & umount … … 860 726 - fail if user runs mondo-archive _from CD_; this is for the user's protection 861 727 - fixed silly bug in mount-me (a crucial line was disabled) 862 - working on stablilo-me, hack-lilo and hack-fstab to make sure they work 863 with the new, busybox-based Mindi v0.41 864 - Mindi v0.41 is very different from v0.40, internally anyway; so, I had to do 865 a lot of testing to make sure Mondo would play nicely with the new Mindi 728 - working on stablilo-me, hack-lilo and hack-fstab to make sure they work with the new, busybox-based Mindi v0.41 729 - Mindi v0.41 is very different from v0.40, internally anyway; so, I had to do a lot of testing to make sure Mondo would play nicely with the new Mindi 866 730 867 731 v1.14 (2001-09-18) 868 732 - if verify is enabled then verify each CD/ISO after writing it 869 - fixed a quasi-bug in mondo/mount-me; this script used 'sort +1' which is 870 not supported by busybox; I've changed it to just plain old 'sort' 733 - fixed a quasi-bug in mondo/mount-me; this script used 'sort +1' which is not supported by busybox; I've changed it to just plain old 'sort' 871 734 - Mondo now tells Mindi whether or not to use its stock kernel 872 735 - fixed bug in 'biggify-me' script; now, checksums will be verified 873 736 - fixed bug in checksum comparison code (mondo-restore.c) 874 - reimplemented '--verify' switch; now, user can verify ISO's or CD's against 875 the live filesystem, just for 'belt and braces' emotional security 737 - reimplemented '--verify' switch; now, user can verify ISO's or CD's against the live filesystem, just for 'belt and braces' emotional security 876 738 - improved the error-reporting at restore-time (if tarball cannot be accessed) 877 - if user does not specify CD-R(W) drive's speed or device then Mondo will 878 try to locate the drive and will assume 4x speed 879 - user may now type 'edit-mountlist' after booting, if they want to edit the 880 mountlist without risking accidentally wiping their hard disks 739 - if user does not specify CD-R(W) drive's speed or device then Mondo will try to locate the drive and will assume 4x speed 740 - user may now type 'edit-mountlist' after booting, if they want to edit the mountlist without risking accidentally wiping their hard disks 881 741 - fixed a few silly typos in the error messages in mondo-tarme 882 742 - updated the FAQ 883 743 - added code to verify ISO images 884 - used 'blank=fast' instead of '-blank fast' for backward-compatibility with 885 older versions of cdrecord 886 - if nonzero value returned by Mondo, then Mondo creates a mondo.err.$$.tgz 887 tarball containing log files & stuff, which it tells the user to e-mail me 744 - used 'blank=fast' instead of '-blank fast' for backward-compatibility with older versions of cdrecord 745 - if nonzero value returned by Mondo, then Mondo creates a mondo.err.$$.tgz tarball containing log files & stuff, which it tells the user to e-mail me 888 746 - when debugging, don't pause before running mondo-tarme; just run it 889 747 - automatically exclude any file including '/vmmhiber.w9x' in its pathname 890 748 - added Minix support (although I doubt anyone needs it) 891 - ported mondo-tarme to C 892 - prettier interface at run-time (backup), giving better feedback 893 - obviates a mysterious, intermittent bug in bash itself 894 - shows '% done' when burning CD 749 - ported mondo-tarme to C - prettier interface at run-time (backup), giving better feedback - obviates a mysterious, intermittent bug in bash itself - shows '% done' when burning CD 895 750 - afio now has a 12MB FIFO buffer (was 16MB) 896 751 - cdrecord now has a 20MB FIFO buffer (was 24MB) … … 905 760 - mondo-tarme now uses -M 16m -t 4k in its call to afio, to improve performance 906 761 - fixed mondo-restore.c's mkreiserfs call; ditto, prep-me's mkreiserfs call 907 - fixed mondo-restore.c's compare_a_biggiefile() subroutine, which would 908 generate erroneous error messages 909 - fixed mondo-archive.c's restore_a_biggiefile() and restore_all_biggiefiles() 910 subroutines; I can't get them to break but others can, so I've streamlined 911 the code to make them easier to debug and in theory more stable too 912 - fixed 'hack-fstab' to make sure that non-LABELed partitions are correctly 913 moved/changed, etc. 762 - fixed mondo-restore.c's compare_a_biggiefile() subroutine, which would generate erroneous error messages 763 - fixed mondo-archive.c's restore_a_biggiefile() and restore_all_biggiefiles() subroutines; I can't get them to break but others can, so I've streamlined the code to make them easier to debug and in theory more stable too 764 - fixed 'hack-fstab' to make sure that non-LABELed partitions are correctly moved/changed, etc. 914 765 - fixed the 'space in big filename' bug (I think) 915 - make Interactive Mode sort the mountlist (by partition size) before creating 916 the dirs and mounting the partitions to them 766 - make Interactive Mode sort the mountlist (by partition size) before creating the dirs and mounting the partitions to them 917 767 - fixed a slew of bugs related to Red Hat's LABEL= extensions 918 - cleaned up mondo-restore.c and mondo-newt.c; replaced lots of arbitrary 919 limits and lengths with a small number of #define's 920 - added some sample RAID0/1/5 raidtab (config) files to restore-scripts.tgz, 921 which will be restored to the ramdisk's /etc dir when the user boots from 922 the CD; this will only take up an extra 50K or so & will help any user 923 who is moving from non-RAID to RAID 768 - cleaned up mondo-restore.c and mondo-newt.c; replaced lots of arbitrary limits and lengths with a small number of #define's 769 - added some sample RAID0/1/5 raidtab (config) files to restore-scripts.tgz, which will be restored to the ramdisk's /etc dir when the user boots from the CD; this will only take up an extra 50K or so & will help any user who is moving from non-RAID to RAID 924 770 - changed FIFO size to 20MB 925 - only use the VFAT kludge on /dev/hda1 or /dev/sda1; otherwise, even if 926 the partition type is VFAT, do a regular mkfs.vfat 771 - only use the VFAT kludge on /dev/hda1 or /dev/sda1; otherwise, even if the partition type is VFAT, do a regular mkfs.vfat 927 772 - if mkfs.vfat doesn't exist, use mkfs.dos 928 773 - only try to add 'LABEL=' to ext2 and ext3 partitions 929 774 - moved a bunch of scripts from Mindi's rootfs.tgz to our restore-scripts.tgz 930 - fixed bugs in mondo-restore.c, which would sometimes just sit there when 931 trying to compare or restore big files that were spread across CD's 932 - Interactive Mode goes straight to 'mondo-restore --interactive' now, i.e. 933 the user doesn't have to choose Interactive once at boot and once within 934 mondo-restore 775 - fixed bugs in mondo-restore.c, which would sometimes just sit there when trying to compare or restore big files that were spread across CD's 776 - Interactive Mode goes straight to 'mondo-restore --interactive' now, i.e. the user doesn't have to choose Interactive once at boot and once within mondo-restore 935 777 - better post-run clean-up 936 778 - improved mondo-restore.c's error-reporting a little bit 937 - stablilo-me (the script) now supports Red Hat's "LABEL=" extensions to 938 /etc/fstab, via the script, 'label-partitions-as-necessary' 779 - stablilo-me (the script) now supports Red Hat's "LABEL=" extensions to /etc/fstab, via the script, 'label-partitions-as-necessary' 939 780 - added a user-friendly mountlist editor to mondo-restore.c 940 781 - added lots of sanity-checking to mountlist editor … … 945 786 - amended the .spec file to be saner when building RPM's 946 787 - archives are named '.afio.bz2' instead of '.tar.bz2' 947 - mondo-archive, mondo-tarme, mondo-restore.c and the DEPRECATED scripts have 948 been amended accordingly 788 - mondo-archive, mondo-tarme, mondo-restore.c and the DEPRECATED scripts have been amended accordingly 949 789 - should be able to create up to 20 CD-R's per backup set now 950 790 - moved a line within prep-me that was in the wrong place (thanks, Fran Boon!) … … 952 792 - added 10-second delay before mondo-restore's Nuke Mode, to let the user abort 953 793 - really fixed the '/mnt/RESTORING/tmp' permissions bug (thanks, Ang Tan!) 954 - sometimes, bash gets all screwed up & won't let the user type 'continue' when 955 Mondo asks them to; if that happens, Mondo will detect it & use a different 956 kind of mondo-askme; the new mondo-askme will check every 10 minutes for a 957 blank CD; if it's there, then the burn continues 794 - sometimes, bash gets all screwed up & won't let the user type 'continue' when Mondo asks them to; if that happens, Mondo will detect it & use a different kind of mondo-askme; the new mondo-askme will check every 10 minutes for a blank CD; if it's there, then the burn continues 958 795 - abort if isodir or scratchdir is a softlink 959 - replaced mondo-askme with a good shell script, to work around flaws in 960 bash and/or kernel (don't know which) 961 - added a failsafe kernel to Mindi, which means you don't have to make sure 962 your kernel is suitable for a boot disk; you can just use Mindi's instead 796 - replaced mondo-askme with a good shell script, to work around flaws in bash and/or kernel (don't know which) 797 - added a failsafe kernel to Mindi, which means you don't have to make sure your kernel is suitable for a boot disk; you can just use Mindi's instead 963 798 - make+chop filelist _after_ calling Mindi 964 799 - updated the FAQ … … 968 803 - added -pad -sort to the call to mkisofs 969 804 - stop saying, "N files differ so far" after comparing each tarball 970 - updated the FAQ, to remind users that they can boot from floppy images if 971 their kernel does not have the right support of CD-ROM's, filesystems, etc. 805 - updated the FAQ, to remind users that they can boot from floppy images if their kernel does not have the right support of CD-ROM's, filesystems, etc. 972 806 - erroneous error-reporting by /mondo/biggify-me; fixed (David Granz) 973 807 - don't estimate the time remaining unless >20 sets processed already 974 - default set size (uncompressed) is now 8MB (was 12MB); when 8MB, seems 975 to run more smoothly & a little quicker, too, at restore-time 808 - default set size (uncompressed) is now 8MB (was 12MB); when 8MB, seems to run more smoothly & a little quicker, too, at restore-time 976 809 - chmod 1777 /mnt/RESTORING/tmp after restoring/creating /tmp 977 810 - dropped mondo-askme.c; using regular 'read' instead … … 979 812 - cdrecord now uses a 16MB FIFO buffer (was 8MB) 980 813 - fixed floating point error in mondo-restore (when showing progress) 981 - when mondo-restore is called without params (e.g. by Interactive Mode), it 982 lets the user choose between nuke, interactive or compare (or abort) 814 - when mondo-restore is called without params (e.g. by Interactive Mode), it lets the user choose between nuke, interactive or compare (or abort) 983 815 - updated the FAQ 984 816 - working on Cuckoo Mode... … … 987 819 - default compression level is now 3 (was 2) 988 820 - retry/fail/abort loop bug fixed 989 - calls mindi with "--custom $TMP $scratchdir/images" to force mindi to use 990 my temp dir and to put the images in my 'images' folder 821 - calls mindi with "--custom $TMP $scratchdir/images" to force mindi to use my temp dir and to put the images in my 'images' folder 991 822 992 823 v1.08 (2001-07-28) … … 997 828 v1.07 (2001-07-22) 998 829 - fixed cosmetic mistake in an error msg in mondo-tarme ('continue'-related) 999 - hack-fstab now uses "defaults 0 0" instead of "default 1 3" when making 1000 a new entry in /etc/fstab 830 - hack-fstab now uses "defaults 0 0" instead of "default 1 3" when making a new entry in /etc/fstab 1001 831 - /mondo/restore-me always makes /mnt/RESTORING/tmp; so does mondo-restore.c 1002 832 - cosmetic changes to mondo-restore.c's screen output 1003 - RAID-related enhancements to mondo-restore.c 1004 - RAID bkp/restore appears to be stable 1005 - it is possible to move from non-RAID to RAID by backing up, wiping, 1006 creating /etc/raidtab, modifying /tmp/mountlist.txt and then restoring 1007 - mondo-restore.c works well 1008 - the only weakness is that the kernel's messages are written to the 1009 current console, not to a separate console... HELP? :-) 833 - RAID-related enhancements to mondo-restore.c - RAID bkp/restore appears to be stable - it is possible to move from non-RAID to RAID by backing up, wiping, creating /etc/raidtab, modifying /tmp/mountlist.txt and then restoring - mondo-restore.c works well - the only weakness is that the kernel's messages are written to the current console, not to a separate console... HELP? :-) 1010 834 - updated README (esp. FAQ) and TODO 1011 835 - tested with Linux 2.4.7; works fine 1012 - mondo-1.07.spec and install.sh compile every C program with -lncurses, just 1013 in case some programs (besides mondo-restore.c) need the NCURSES library 836 - mondo-1.07.spec and install.sh compile every C program with -lncurses, just in case some programs (besides mondo-restore.c) need the NCURSES library 1014 837 - mondo-restore.c is now compiled statically 1015 838 … … 1018 841 - install.sh only compiles mondo-restore.c once now 1019 842 - if scratchdir or isodir contain more than one path (each), abort 1020 - mondo-restore.c creates dummy partitions if the user starts the official 1021 etc/fstab table at /dev/hda7 or something; so, partitioning should work, 1022 even if the mountlist has some gaps 843 - mondo-restore.c creates dummy partitions if the user starts the official etc/fstab table at /dev/hda7 or something; so, partitioning should work, even if the mountlist has some gaps 1023 844 - for each /dev/md* device in the mountlist, run raidstop before mkraid 1024 845 … … 1040 861 - added 'mondo-restore.c' 1041 862 - C program, to replace the old restore scripts 1042 - Mindi will incorporate it in the data disks, if it can find Mondo 1043 - Mindi will also incorporate the old restore scripts (from Mondo) 1044 just in case the new C program doesn't work 863 - Mindi will incorporate it in the data disks, if it can find Mondo - Mindi will also incorporate the old restore scripts (from Mondo) just in case the new C program doesn't work 1045 864 - calls the script, 'stablilo-me' to run LILO if desired 1046 - added 'restore-scripts.tgz', which contains the /mondo scripts that used 1047 to live in Mindi; now, Mindi grabs them from here if necessary 865 - added 'restore-scripts.tgz', which contains the /mondo scripts that used to live in Mindi; now, Mindi grabs them from here if necessary 1048 866 - cleared up Mode-related scripts & the executable to replace them 1049 867 - if mkisofs fails, ABORT, don't just warn 1050 - added '--differential' option, to backup the files changed during the 1051 last N days (N specified by user) 868 - added '--differential' option, to backup the files changed during the last N days (N specified by user) 1052 869 1053 870 v1.02 (2001-07-07) 1054 871 - retract the CD tray after asking user to enter 'continue' (PauseAndAskForCDR) 1055 - uses C program instead of bash loop for PauseAndAskForCDR; some bash versions 1056 go into an infinite loop for no reason; hopefully, outsourcing to C will help 872 - uses C program instead of bash loop for PauseAndAskForCDR; some bash versions go into an infinite loop for no reason; hopefully, outsourcing to C will help 1057 873 - when slicing, indicate progress ("N slices...") 1058 - when told to '--burn-cds', Mondo pipes mkisofs's output to cdrecord, instead 1059 of saving the ISO file to disk first; this change will save time and space 1060 (not in the Star Trek sense, however *smile*) 874 - when told to '--burn-cds', Mondo pipes mkisofs's output to cdrecord, instead of saving the ISO file to disk first; this change will save time and space (not in the Star Trek sense, however *smile*) 1061 875 - default set size (native) is now 12MB 1062 876 - copy /usr/share/mondo to correct directory within CD … … 1067 881 1068 882 v1.00 (2001-07-04) 1069 - thanks to recent changes to Mindi, Mondo is now compatible with the 1070 filesystem formats JFS, ext3, ReiserFS, XFS and ext2 (if your own kernel is) 1071 - CDs [2-N] do not have boot floppy images on them; only CD#1 does (although 1072 CDs [2-N] do have the El Torito boot image, just in case...) 1073 - modified the --burn-cds switch; now it takes 3rd param, 'cdrw' (optional) to 1074 let you wipe and rewrite CDRWs instead of wasting CDRs ;-) 883 - thanks to recent changes to Mindi, Mondo is now compatible with the filesystem formats JFS, ext3, ReiserFS, XFS and ext2 (if your own kernel is) 884 - CDs [2-N] do not have boot floppy images on them; only CD#1 does (although CDs [2-N] do have the El Torito boot image, just in case...) 885 - modified the --burn-cds switch; now it takes 3rd param, 'cdrw' (optional) to let you wipe and rewrite CDRWs instead of wasting CDRs ;-) 1075 886 - removed that PID thing; doesn't work; please don't run Mondo twice at once :) 1076 887 - removed that "filenames contain '|'" error & associated problems … … 1083 894 - don't make MD5 checksums if verify_level==0 1084 895 - when archiving, make N.tar.bz2 and cklist.txt (tarball+checksums) in parallel 1085 - at end of mondo-archive, remove scratchdir and TMP completely IF they were 1086 created by Mondo (e.g. /home/mondo.scratch or /home/tmp.mondo.12345) 1087 - fill CD/image up to 1MB less than its max size; was being filled up to about 1088 8MB less than its max size, which wasted 8-10MB of storage space per disk 896 - at end of mondo-archive, remove scratchdir and TMP completely IF they were created by Mondo (e.g. /home/mondo.scratch or /home/tmp.mondo.12345) 897 - fill CD/image up to 1MB less than its max size; was being filled up to about 8MB less than its max size, which wasted 8-10MB of storage space per disk 1089 898 1090 899 v0.99 (2001-03-15) … … 1112 921 1113 922 v0.976 (2000-09-01) 1114 - technical changes 1115 - 12MB ramdisk 1116 - repackaged: now, Mondo comes in big tarball/RPM with all overhead-files 1117 - each Mondo CD has a 15MB administrative overhead; used to be 50-60MB 1118 - definitely 386-compatible now (RH 6.2 binaries) 1119 - Linux-Mandrake 7.1 (2.2.16) kernel w/ReiserFS & Supermount 1120 - modules.conf / conf.modules support (whichever you've got) 1121 - user may include their own copies of modules, if Mondo doesn't come with 1122 the requisite modules already 1123 - extra facilities 1124 - ReiserFS support 1125 - '--load-modules <foo> <bar> ...' will allow you to force Mondo to load 1126 key modules at boot-time, e.g. aic7xxx for SCSI-only users 1127 - bugfixes 1128 - path-exclusion code (mondo-backup) 1129 - bogus 'flaws=1' message has been eradicated 1130 - CDRW-finding code is now much better at finding the CDRW correctly 1131 - dummy burn mode is working again 1132 - fixed security hole (thanks, Matija); now, user will be warned if Mondo 1133 is configured to execute /tmp/mondo-tarme 1134 - changed lots of 'grep " $string "' to 'grep -w "$string"' (thanks, Matija) 923 - technical changes - 12MB ramdisk - repackaged: now, Mondo comes in big tarball/RPM with all overhead-files - each Mondo CD has a 15MB administrative overhead; used to be 50-60MB - definitely 386-compatible now (RH 6.2 binaries) - Linux-Mandrake 7.1 (2.2.16) kernel w/ReiserFS & Supermount - modules.conf / conf.modules support (whichever you've got) - user may include their own copies of modules, if Mondo doesn't come with the requisite modules already 924 - extra facilities - ReiserFS support - '--load-modules <foo> <bar> ...' will allow you to force Mondo to load key modules at boot-time, e.g. aic7xxx for SCSI-only users 925 - bugfixes - path-exclusion code (mondo-backup) - bogus 'flaws=1' message has been eradicated - CDRW-finding code is now much better at finding the CDRW correctly - dummy burn mode is working again - fixed security hole (thanks, Matija); now, user will be warned if Mondo is configured to execute /tmp/mondo-tarme - changed lots of 'grep " $string "' to 'grep -w "$string"' (thanks, Matija) 1135 926 1136 927 v0.975 (2000-07-03) … … 1148 939 1149 940 v0.972 (2000-06-16) 1150 - auto-archive, mondo-archive and mondo-tarme updated to make them more 1151 friendly towards mult-disk sets 941 - auto-archive, mondo-archive and mondo-tarme updated to make them more friendly towards mult-disk sets 1152 942 - if multiple CDRW drives, user is asked which mondo-archive should use 1153 943 1154 944 v0.971 (2000-06-14) 1155 945 - auto-archive now passes CD size to mondo-archive 1156 - mondo-tarme, in conjunction with auto-archive, now handles CDRW boot/burn 1157 sharing much better than before 946 - mondo-tarme, in conjunction with auto-archive, now handles CDRW boot/burn sharing much better than before 1158 947 1159 948 v0.970 (2000-06-10) … … 1161 950 - lots of silly bugs & overhead problems fixed; 2nd floppy closer to working 1162 951 - restore scripts all log to /var/log/mondo-restore.log 1163 - auto-archive:- 1164 - offers Low, Medium or High compression 1165 - no longer needs a spare hard drive for scratch space; uses loopfs on 1166 existing hard drive (if possible) instead 1167 - handles weird (multi-CDROM) PCs more gracefully now 1168 - handles PCs with CDRW but without CDROM more gracefully now 1169 - deletes each ISO after burning it to a CD-R (if it _is_ burning them) 1170 - asks if PC is old & crappy; if it is, IDE optimizations are disabled 1171 at backup-time and restore-time 1172 - lets user use CDRW for booting _and_ burning, thanks to reroute-softlinks 952 - auto-archive:- - offers Low, Medium or High compression - no longer needs a spare hard drive for scratch space; uses loopfs on existing hard drive (if possible) instead - handles weird (multi-CDROM) PCs more gracefully now - handles PCs with CDRW but without CDROM more gracefully now - deletes each ISO after burning it to a CD-R (if it _is_ burning them) - asks if PC is old & crappy; if it is, IDE optimizations are disabled at backup-time and restore-time - lets user use CDRW for booting _and_ burning, thanks to reroute-softlinks 1173 953 - selective restore is finally fixed 1174 954 - nicer help screens on boot floppy … … 1190 970 v0.963 (2000-05-31) 1191 971 - ISO images are no longer padded 1192 - all currently-loaded modules are backed up to custom hboot.img & installed 1193 at restore-time before CD is accessed 972 - all currently-loaded modules are backed up to custom hboot.img & installed at restore-time before CD is accessed 1194 973 - '--burn-cd <speed>' switch added 1195 - '--no-opt' switch added; use it in the call to mondo-archive if you don't 1196 want IDE optimization when restoring 974 - '--no-opt' switch added; use it in the call to mondo-archive if you don't want IDE optimization when restoring 1197 975 - lots of work done on auto-archive 1198 976 - mondo-extras-1.5 created; excludes mondo-zfile but includes hlib.tar.bz2 1199 977 - boot floppy image now includes kernel w/lots of SCSI stuff 1200 978 - slice* deleted (WipeArchives); was 'slices*' (bug!) 1201 - restore scripts all use 'lastbit-me' to find the 'X' part of '/dev/X' device; 1202 this will make it easier to port Mondo to kernel 2.4.x 1203 - I cannot make SCSI-based CDROMs boot from Mondo CDs properly... but SCSI 1204 hard drives and ZIP drives should be handled ok from now on 1205 - with a little coaxing, Mondo will backup Windows-only PCs (as well as 1206 Lin/Win and Linux-only PCs) 979 - restore scripts all use 'lastbit-me' to find the 'X' part of '/dev/X' device; this will make it easier to port Mondo to kernel 2.4.x 980 - I cannot make SCSI-based CDROMs boot from Mondo CDs properly... but SCSI hard drives and ZIP drives should be handled ok from now on 981 - with a little coaxing, Mondo will backup Windows-only PCs (as well as Lin/Win and Linux-only PCs) 1207 982 1208 983 v0.962 (2000-05-25) … … 1232 1007 - mondo-vfat-1.0 package released, to provide kludging for VFAT bootables 1233 1008 - logfile now saved at /var/log/mondo-archive.log (not `pwd`/mondo-archive.log) 1234 - pid registered/checked _after_ command line processed... so you can do --help 1235 without being shouted at ;) 1009 - pid registered/checked _after_ command line processed... so you can do --help without being shouted at ;) 1236 1010 - restore-me broken into 2 scripts: untar-me and biggify-me 1237 1011 - restore-subroutine-me now renamed untar-subroutine-me 1238 - bzip2 decompression should work again ;) ... a missing '-Z' was added to 1239 the afio call, and the usr/lib/libbz* libraries are now included on CD ISOs 1012 - bzip2 decompression should work again ;) ... a missing '-Z' was added to the afio call, and the usr/lib/libbz* libraries are now included on CD ISOs 1240 1013 - fixed hack-lilo to ignore blank lines 1241 - really ugly boot kludge has been implemented, using embleer files 1242 to make newly-formatted VFAT partitions bootable 1014 - really ugly boot kludge has been implemented, using embleer files to make newly-formatted VFAT partitions bootable
Note:
See TracChangeset
for help on using the changeset viewer.