8 Data encryption Meno Abels (see projects)
9 Communication encryption Meno Abels
10 Version 1.35 Kern (see below)
11 ========================================================
13 1.35 Items to do for release:
15 - Add new DCR calling sequences everywhere in SD. This will permit
16 simultaneous use of multiple devices by a single job.
17 - Look at patches/bacula_db.b2z postgresql that loops during restore.
19 - Perhaps add read/write programs and/or plugins to FileSets.
20 - Implement Ignore FileSet Change.
22 - Add bscan to four-concurrent-jobs regression.
24 - Add IPv6 to regression
25 - Alternative to static linking "ldd prog" save all binaries listed,
26 restore them and point LD_LIBRARY_PATH to them.
27 - Document a get out of jail procedure if everything breaks if
28 you lost/broke the Catalog -- do the same for "I know my
29 file is there how do I get it back?".
30 - Test/doc Tape Alerts
31 - Doc update AllFromVol
32 - Doc dbcheck eliminate orphaned clients.
33 - Doc new duration time input editing.
34 - Doc -p option in stored
35 - Document that console commands can be abbreviated.
38 Documentation to do: (any release a little bit at a time)
39 - Document query file format.
40 - Add more documentation for bsr files.
41 - Document problems with Verify and pruning.
42 - Document how to use multiple databases.
43 - VXA drives have a "cleaning required"
44 indicator, but Exabyte recommends preventive cleaning after every 75
47 In this context, it should be noted that Exabyte has a command-line
48 vxatool utility available for free download. (The current version is
49 vxatool-3.72.) It can get diagnostic info, read, write and erase tapes,
50 test the drive, unload tapes, change drive settings, flash new firmware,
52 Of particular interest in this context is that vxatool <device> -i will
53 report, among other details, the time since last cleaning in tape motion
54 minutes. This information can be retrieved (and settings changed, for
55 that matter) through the generic-SCSI device even when Bacula has the
56 regular tape device locked. (Needless to say, I don't recommend
57 changing tape settings while a job is running.)
58 - Lookup HP cleaning recommendations.
59 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
60 - Add more obvious documentation of @ for conf files.
61 - Document doing table repair
63 Testing to do: (painful)
66 For 1.37 Testing/Documentation:
67 - Fix find_device in stored/dircmd.c:462 (see code)
68 - Add db check test to regression. Test each function like delete,
70 - If you use restore replace=never, the directory attributes for
71 non-existent directories will not be restored properly.
74 - see lzma401.zip in others directory for new compression
76 - Minimal autochanger handling in Bacula and in btape.
77 - Look into how tar does not save sockets and the possiblity of
78 not saving them in Bacula (Martin Simmons reported this).
79 The next two lines will show them.
80 localmounts=`awk '/ext/ { print $2 }' /proc/mounts` # or whatever
81 find $localmounts -xdev -type s -ls
82 - Fix restore jobs so that multiple jobs can run if they
83 are not using the same tape(s).
84 - Allow the user to select JobType for manual pruning/purging.
85 - Look at adding Client run command that will use the
86 port opened by the client.
87 - bscan does not put first of two volumes back with all info in
89 - Implement the FreeBSD nodump flag in chflags.
90 - Figure out how to make named console messages go only to that
91 console and to the non-restricted console (new console class?).
92 - Make restricted console prompt for password if *ask* is set or
93 perhaps if password is undefined.
94 - Implement "from ISO-date/time every x hours/days/weeks/months" in
96 ==== from Marc Schoechlin
97 - the help-command should be more verbose
98 (it should explain the paramters of the different
100 -> it´s time-comsuming to consult the manual anytime
101 you need a special parameter
102 -> maybe it´s more easy to maintain this, if the
103 descriptions of that commands are outsourced to
105 - the cd-command should allow complete paths
106 i.e. cd /foo/bar/foo/bar
107 -> if a customer mails me the path to a certain file,
108 it´s faster to enter the specified directory
109 - if the password is not configured in bconsole.conf
110 you should be asked for it.
111 -> sometimes you like to do restore on a customer-machine
112 which shouldn´t know the password for bacula.
113 -> adding the password to the file favours admins
114 to forget to remove the password after usage
116 the protection of that file is less important
117 - long-listed-output of commands should be scrollable
118 like the unix more/less-command does
119 -> if someone runs 200 and more machines, the lists could
120 be a little long and complex
121 - command-output should be shown column by column
122 to reduce scrolling and to increase clarity
124 - lsmark should list the selected files with full
126 - wildcards for selecting and file and directories would be nice
127 - any actions should be interuptable with STRG+C
128 - command-expansion would be pretty cool
130 - Compaction of Disk space by "migrating" Volumes that have pruned
131 Jobs (what criteria? size, #jobs, time).
132 - Add prune all command
133 - Document fact that purge can destroy a part of a restore by purging
134 one volume while others remain valid -- perhaps mark Jobs.
135 - Add multiple-media-types.txt
136 - look at mxt-changer.html
137 - Make ? do a help command (no return needed).
138 - Implement restore directory.
139 - Add All Local Partitions = yes to new style saves.
140 - Document streams and how to implement them.
141 - Possibly implement "Ensure Full Backup = yes" looks for a failed full backup
142 and upgrades the current backup if one exists.
143 - Check that barcode reading and update slots scan works.
144 - Try not to re-backup a file if a new hard link is added.
145 - Add feature to backup hard links only, but not the data.
146 - Add "All Local = yes" option to save to include all local partitions.
147 - Fix stream handling to be simpler.
148 - Add Priority and Bootstrap to Run a Job.
149 - Eliminate Restore "Run Restore Job" prompt by allowing new "run command
151 - Remove View FileSet button from Run a Job dialog.
152 - Don't count higher level non-saved directories in Files expected
154 - Handle prompt for restore job at end of Restore command.
155 - Add display of total selected files to Restore window.
156 - Add tree pane to left of window.
157 - Add progress meter.
158 - Polling does not work for restore. It tries a number of times,
159 gives up, and crashes the SD.
160 - Lock jcr_chain when doing attach/detach in acquire.c
161 - Add assert in free_jcr if attach/detach chain active.
162 - Max wait time or max run time causes seg fault -- see runtime-bug.txt
163 - Document writing to a CD/DVD with Bacula.
164 - Add check for tape alerts.
165 - Add a "base" package to the window installer for pthreadsVCE.dll
166 which is needed by all packages.
167 - Add message to user to check for fixed block size when the forward
168 space test fails in btape.
169 - When unmarking a directory check if all files below are unmarked and
170 then remove the + flag -- in the restore tree.
171 - Possibly implement: Action = Unmount Device="TapeDrive1" in Admin jobs.
172 - Setup lrrd graphs: (http://www.linpro.no/projects/lrrd/) Mike Acar.
173 - Revisit the question of multiple Volumes (disk) on a single device.
174 - Finish SIGHUP work.
175 - Check that all change in wait status in the SD are
176 signaled to the Director.
177 - Add a block copy option to bcopy.
178 - Investigate adding Mac Resource Forks.
179 - Finish work on Gnome restore GUI.
180 - Fix "llist jobid=xx" where no fileset or client exists.
181 - Check pruning of restore jobs.
183 it seems to be complaining about 12:00pm which should be a valid 12
184 hour time. I changed the time to 11:59am and everything works fine.
185 Also 12:00am works fine. 0:00pm also works (which I don't think
186 should). None of the values 12:00pm - 12:59pm work for that matter.
187 - Add level to estimate command.
188 - For each job type (Admin, Restore, ...) require only the really necessary
190 - Fix option 2 of restore -- list where file is backed up -- require Client,
191 then list last 20 backups.
192 - Add all pools in Dir conf to DB also update them to catch changed
193 LabelFormats and such.
194 - Update volumes FromPool=xxx does all volumes.
195 - Pass Director resource name as an option to the Console.
196 - Add a "batch" mode to the Console (no unsolicited queries, ...).
197 - Add code to check for tape alerts -- tapeinfo.
198 - Make sure list of Volumes needed is in correct order for restore.
200 - Remove paths (and files that reference them) that have no trailing slash
201 in dbcheck -- or add a trailing slash.
202 - Remove Filenames (and files that reference them) that have a trailing
203 slash in dbcheck -- or remove the trailing slash.
204 - Remove orphaned paths/filenames by copying them to a new table with a
205 reference count, then mark all referenced files/paths and remove unreferenced
207 - Add a .list all files in the restore tree (probably also a list all files)
208 Do both a long and short form.
209 - Allow browsing the catalog to see all versions of a file (with
210 stat data on each file).
211 - Restore attributes of directory if replace=never set but directory
213 - Allow "delete job jobid=xxx,yyy,aaa-bbb" i.e. list + ranges.
214 - Use SHA1 on authentication if possible.
215 - See comtest-xxx.zip for Windows code to talk to USB.
216 - Make btape accept Device Names in addition to Archive names.
217 - Add Events and Perl scripting.
218 - Add John's appended files:
219 Appended = { /files/server/logs/http/*log }
220 and such files would be treated as follows.On a FULL backup, they would
221 be backed up like any other file.On an INCREMENTAL backup, where a
222 previous INCREMENTAL or FULL was already in thecatalogue and the length
223 of the file wasgreater than the length of the last backup, only thedata
224 added since the last backup will be dumped.On an INCREMENTAL backup, if
225 the length of the file is less than thelength of the file with the same
226 name last backed up, the completefile is dumped.On Windows systems, with
227 creation date of files, we can be evensmarter about this and not count
228 entirely upon the length.On a restore, the full and all incrementals
229 since it will beapplied in sequence to restore the file.
230 - Add a regression test for dbcheck.
231 - Add disk seeking on restore. - Allow
232 for optional cancelling of SD and FD in case DIR
233 gets a fatal error. Requested by Jesse Guardiani <jesse@wingnet.net>
234 - Bizarre message: Error: Could not open WriteBootstrap file:
235 - Build console in client only build.
236 - Add "limit=n" for "list jobs"
237 - Check new HAVE_WIN32 open bits.
238 - Check if the tape has moved before writing.
239 - Handling removable disks -- see below:
240 - Multiple drive autochanger support -- see below.
241 - Keep track of tape use time, and report when cleaning is necessary.
242 - Fix FreeBSD mt_count problem.
243 - Add FromClient and ToClient keywords on restore command (or
244 BackupClient RestoreClient).
245 - Automatic "update slots" on user configuration directive when a
247 - Implement a JobSet, which groups any number of jobs. If the
248 JobSet is started, all the jobs are started together.
249 Allow Pool, Level, and Schedule overrides.
250 - Enhance cancel to timeout BSOCK packets after a specific delay.
251 - When I restore to Windows the Created, Accessed and Modifiedtimes are
252 those of the time of the restore, not those of the originalfile.
253 The dates you will find in your restore log seem to be the original
255 - Volume "add"ed to Pool gets recycled in first use. VolBytes=0
256 - If a tape is recycled while it is mounted, Stanislav Tvrudy must do an
257 additional mount to deblock the job.
259 bscan: sql_update.c:65 UPDATE File SET MD5='Ij+5kwN6TFIxK+8l8+/I+A' WHERE FileId=0
260 bscan: bscan.c:1074 Could not add MD5/SHA1 to File record. ERR=sql_update.c:65 Update problem: affected_rows=0
261 - Do scheduling by UTC using gmtime_r() in run_conf, scheduler, and
262 ua_status.!!! Thanks to Alan Brown for this tip.
263 - Look at updating Volume Jobs so that Max Volume Jobs = 1 will work
264 correctly for multiple simultaneous jobs.
265 - Correct code so that FileSet MD5 is calculated for < and | filename
267 - Mark Volume in error on error from WEOF.
268 - Implement the Media record flag that indicates that the Volume does disk
270 - Implement VolAddr, which is used when Volume is addressed like a disk,
271 and form it from VolFile and VolBlock.
272 - Make multiple restore jobs for multiple media types specifying
273 the proper storage type.
274 - Implement MediaType keyword in bsr?
275 - Fix fast block rejection (stored/read_record.c:118). It passes a null
276 pointer (rec) to try_repositioning().
277 - Look at extracting Win data from BackupRead.
278 - Having dashes in filenames apparently creates problems for restore
279 by filename??? hard to believe.
280 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
281 which would take precidence over the Catalog "JobRetention".
282 - Implement Label Format in Add and Label console commands.
283 - Possibly up network buffers to 65K. Put on variable.
284 - Put email tape request delays on one or more variables. User wants
285 to cancel the job after a certain time interval. Maximum Mount Wait?
286 - Job, Client, Device, Pool, or Volume?
287 Is it possible to make this a directive which is *optional* in multiple
288 resources, like Level? If so, I think I'd make it an optional directive
289 in Job, Client, and Pool, with precedence such that Job overrides Client
290 which in turn overrides Pool.
291 - Print a message when a job starts if the conf file is not current.
293 - Spooling ideas taken from Volker Sauer's and other's emails:
294 > IMHO job spooling should be turned on
300 > where and 2) overrides 1) and 3) is independent.
302 Yes, this is the minimum that I think is necessary.
306 > It should be switched by job, because the job that backs up the machine
307 > with the bacula-sd on doesn't need spooling.
308 > It should be switched by schedule, because for full-backups I don't need
309 > spooling, so I can switch it off (because the network faster then the
312 True, with the exception that if you have enough disk spool space,
313 and you want to run concurrent jobs, spooling can eliminate the block
314 interleaving restore inefficiencies.
316 > And you should be able to turn it of by sd for sd-machines with low disk
317 > capacity or if you just don't need or want this feature.
320 > - definitly the possibility for multipe spool direcories
322 Having multiple directories is no problem -- having different maximum
323 sizes creates specification problems. At some point, I will probably
324 have a common SD pool of spool directories as well as a set of
325 private spool directories for each device. The first implementation
326 will be a set of private spool directories for each device since
327 managing a global pool with a bunch of threads writing into the same
328 directory is *much* more complicated and prone to error.
330 > - the ability to spool parts of a backup (not the whole client)
332 This may change in the future, but for the moment, it will spool
333 either to a job high water mark, or until the directory is full
334 (reaches max spool size or I/O error). It will then write to tape,
335 truncate the spool file, and begin spooling again.
337 > - spooling while writing to tape
339 Not within a job, but yes, if you run concurrent jobs -- each is a
340 different thread. Within a job could be a feature, but *much* later.
342 > - parallel spooling (like parallel jobs/ concurrent jobs) of clients
344 Yes, this is one of my main motivations for doing it (aside from
345 eliminating tape "shoe shine" during incremental backups.
347 > - flushing a backup that only went to disk (like amflush in amanda)
349 This will be a future feature, since spooling is different from backing
350 up to disk. The future feature will be "migration" which will move a job
351 from one backup Volume to another.
353 - New Storage specifications:
354 Passed to SD as a sort of BSR record called Storage Specification
358 MediaType -> Next MediaType
360 Device -> Next Device
361 Write Copy Resource that makes a copy of a resource.
363 Allow multiple Storage specifications
371 Allow Multiple Pool specifications (note, Pool currently
373 Allow Multiple MediaType specifications
374 Allow Multiple Device specifications
375 Perhaps keep this in a single SSR
376 Tie a Volume to a specific device by using a MediaType that
377 is contained in only one device.
378 In SD allow Device to have Multiple MediaTypes
381 - Look at www.nu2.nu/pebuilder as a helper for full windows
383 Ideas from Jerry Scharf:
384 First let's point out some big pluses that bacula has for this
386 more importantly it's active. Thank you so much for that
387 even more important, it's not flaky
388 it has an open access catalog, opening many possibilities
389 it's pushing toward heterogeneous systems capability
391 I don't remember an include file directive for config files
392 (not filesets, actual config directives)
393 can you check the configs without starting the daemon?
394 some warnings about possible common mistakes
396 doing the testing and blessing of concurrent backup writes
397 this is absolutely necessary in the enterprise
398 easy user recovery GUI with full access checking
399 Macintosh file client
400 macs are an interesting niche, but I fear a server is a rathole
401 working bare iron recovery for windows
402 much better handling on running config changes
403 thinking through the logic of what happens to jobs in progress
404 the option for inc/diff backups not reset on fileset revision
405 a) use both change and inode update time against base time
406 b) do the full catalog check (expensive but accurate)
407 sizing guide (how much system is needed to back up N systems/files)
408 consultants on using bacula in building a disaster recovery system
410 or how to get at fancy things that one could do with bacula
411 logwatch code for bacula logs (or similar)
412 linux distro inclusion of bacula (brings good and bad, but necessary)
413 win2k/XP server capability (icky but you asked)
414 support for Oracle database ??
416 - Look at adding SQL server and Exchange support for Windows.
417 - Restore: Enter Filename: 'C:/Documents and Settings/Comercial/My
418 Documents/MOP/formulário de registro BELAS ARTES.doc' causes Bacula to
420 - Each DVD-RAM disk would be a volume, just like each tape is
421 a volume. It's a 4.7GB media with random access, but there's nothing about
422 it that I can see that makes it so different than a tape from bacula's
423 perspective. Why couldn't I back up to a bare floppy as a volume (ignoring
425 - Make dev->file and dev->block_num signed integers so that -1 can
426 be an invalid value which happens with BSR.
427 - Create VolAddr for disk files in place of VolFile and VolBlock. This
428 is needed to properly specify ranges.
429 - Print bsmtp output to job report so that problems will be seen.
430 - Pass the number of files to be restored to the FD for reporting
431 - Add progress of files/bytes to SD and FD.
432 - Don't continue Restore if no files selected.
433 - Print warning message if FileId > 4 billion
434 - do a "messages" before the first prompt in Console
435 - Add a date and time stamp at the beginning of every line in the
436 Job report (Volker Sauer).
437 - Client does not show busy during Estimate command.
438 - Implement Console mtx commands.
439 - Add a default DB password to MySQL.
440 GRANT all privileges ON bacula.* TO bacula@localhost IDENTIFIED BY
443 - Implement a Mount Command and an Unmount Command where
444 the users could specify a system command to be performed
445 to do the mount, after which Bacula could attempt to
446 read the device. This is for Removeable media such as a CDROM.
447 - Most likely, this mount command would be invoked explicitly
448 by the user using the current Console "mount" and "unmount"
449 commands -- the Storage Daemon would do the right thing
450 depending on the exact nature of the device.
451 - As with tape drives, when Bacula wanted a new removable
452 disk mounted, it would unmount the old one, and send a message
453 to the user, who would then use "mount" as described above
454 once he had actually inserted the disk.
455 - Implement dump/print label to UA
456 - Implement disk spooling. Two parts: 1. Spool to disk then
457 immediately to tape to speed up tape operations. 2. Spool to
458 disk only when the tape is full, then when a tape is hung move
460 - Scratch Pool where the volumes can be re-assigned to any Pool.
461 - bextract is sending everything to the log file ****FIXME****
462 - Add Progress command that periodically reports the progress of
464 - Restrict characters permitted in a Resource name, and don't permit
466 - Allow multiple Storage specifications (or multiple names on
467 a single Storage specification) in the Job record. Thus a job
468 can be backed up to a number of storage devices.
469 - Implement some way for the File daemon to contact the Director
470 to start a job or pass its DHCP obtained IP number.
471 - Implement multiple Consoles.
472 - Implement a query tape prompt/replace feature for a console
474 Two jobs ready to go, first one blocked waiting for media
475 Cancel 2nd job ("waiting execution" one)
478 - Copy console @ code to gnome2-console
479 - Make AES the only encryption algorithm see
480 http://csrc.nist.gov/CryptoToolkit/aes/). It's
481 an officially adopted standard, has survived peer
482 review, and provides keys up to 256 bits.
483 - Think about how space could be freed up on a tape -- perhaps this
484 is a Merge or Compact feature that is needed.
485 - Modify FileSet, did not upgrade the current Increment job, but
486 waited for the next job to be upgraded.
487 - Take a careful look at SetACL http://setacl.sourceforge.net
488 - Implement a where command for the tree telling where a file
490 - Take a careful look at Level for the estimate command, maybe make
491 it a command line option.
492 - Add Volume name to "I cannot write on this volume because"
493 - Make tree walk routines like cd, ls, ... more user friendly
494 by handling spaces better.
495 - Write your PID file and chown root:wheel before drop.
496 - Make sure there is no symlink in a file before creating a
498 - Look at mktemp or mkstemp(3).
499 mktemp and mkstemp create files with predictable names too. That's
500 not the vulnerability. The vulnerability is in creating files without
501 using the O_EXCL flag, which means "only create this file if it doesn't
502 exist, including if the file is a dangling symlink."
504 It is *NOT* enough to do the equivalent of
506 if doesn't exist $filename
507 then create $filename
509 because between the test and the create another process could have
510 gotten the CPU and created the file. You must use atomic functions
511 (those that don't get interrupted by other processes) and O_EXCL is
512 the only way for this particular example.
513 - Automatically create pools, but instead of looking for what
514 in in Job records, walk through the pool resources.
515 - Check and double check tree code, why does it take so long?
516 - Add device name to "Current Volume not acceptable because ..."
517 - Make sure that Bacula rechecks the tape after the 20 min wait.
518 - Set IO_NOWAIT on Bacula TCP/IP packets.
519 - Try doing a raw partition backup and restore by mounting a
522 Yes, it would allow to highly automatic the request for new tapes. If a
523 tape is empty, bacula reads the barcodes (native or simulated), and if
524 an unused tape is found, it runs the label command with all the
525 necessary parameters.
527 By the way can bacula automatically "move" an empty/purged volume say
528 in the "short" pool to the "long" pool if this pool runs out of volume
530 - Eliminate orphaned jobs: dbcheck, normal pruning, delete job command.
531 Hm. Well, there are the remaining orphaned job records:
533 | 105 | Llioness Save | 0000-00-00 00:00:00 | B | D | 0 | 0 | f |
534 | 110 | Llioness Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
535 | 115 | Llioness Save | 2003-09-10 02:22:03 | B | I | 0 | 0 | A |
536 | 128 | Catalog Save | 2003-09-11 03:53:32 | B | I | 0 | 0 | C |
537 | 131 | Catalog Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
539 As you can see, three of the five are failures. I already deleted the
540 one restore and one other failure using the by-client option. Deciding
541 what is an orphaned job is a tricky problem though, I agree. All these
542 records have or had 0 files/ 0 bytes, except for the restore. With no
543 files, of course, I don't know of the job ever actually becomes
544 associated with a Volume.
546 (I'm not sure if this is documented anywhere -- what are the meanings of
547 all the possible JobStatus codes?)
549 Looking at my database, it appears to me as though all the "orphaned"
550 jobs fit into one of two categories:
552 1) The Job record has a StartTime but no EndTime, and the job is not
555 2) The Job record has an EndTime, indicating that it completed, but
556 it has no associated JobMedia record.
559 This does suggest an approach. If failed jobs (or jobs that, for some
560 other reason, write no files) are associated with a volume via a
561 JobMedia record, then they should be purged when the associated volume
562 is purged. I see two ways to handle jobs that are NOT associated with a
565 1) purge them automatically whenever any volume is manually purged;
567 2) add an option to the purge command to manually purge all jobs with
568 no associated volume.
570 I think Restore jobs also fall into category 2 above .... so one might
571 want to make that "The Job record has an EndTime,, but no associated
572 JobMedia record, and is not a Restore job."
573 - make "btape /tmp" work.
574 - Make sure a rescheduled job is properly reported by status.
575 - Walk through the Pool records rather than the Job records
576 in dird.c to create/update pools.
577 - What to do about "list files job=xxx".
578 - When job rescheduled, status gives is waiting for Client Rufus
579 to connect to Storage File. Dir needs to inform SD that job
581 - Make Dmsg look at global before calling subroutine.
582 - Enable trace output at runtime for Win32
583 - Available volumes for autochangers (see patrick@baanboard.com 3 Sep 03
584 and 4 Sep) scan slots.
585 - Get and test MySQL 4.0
586 - Do a complete audit of all pthreads_mutex, cond, ... to ensure that
587 any that are dynamically initialized are destroyed when no longer used.
588 - Look at how fuser works and /proc/PID/fd that is how Nic found the
589 file descriptor leak in Bacula.
590 - Implement WrapCounters in Counters.
591 - Turn on SIGHUP in dird.c and test.
592 - Use system dependent calls to get more precise info on tape errors.
593 - Add heartbeat from FD to SD if hb interval expires.
594 - Suppress read error on blank tape when doing a label.
595 - Can we dynamically change FileSets?
596 - If pool specified to label command and Label Format is specified,
597 automatically generate the Volume name.
598 - Why can't SQL do the filename sort for restore?
599 - Look at libkse (man kse) for FreeBSD threading.
600 - Look into Microsoft Volume Shadowcopy Service VSS for backing
601 up system state components (Active Directory, System Volume, ...)
602 - Add ExhautiveRestoreSearch
603 - Look at the possibility of loading only the necessary
604 data into the restore tree (i.e. do it one directory at a
605 time as the user walks through the tree).
606 - Possibly use the hash code if the user selects all for a restore command.
607 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
608 - Fix "restore all" to bypass building the tree.
609 - Prohibit backing up archive device (findlib/find_one.c:128)
610 - Implement Release Device in the Job resource to unmount a drive.
611 - Implement Acquire Device in the Job resource to mount a drive,
612 be sure this works with admin jobs so that the user can get
613 prompted to insert the correct tape. Possibly some way to say to
614 run the job but don't save the files.
615 - Implement FileOptions (see end of this document)
616 - Make things like list where a file is saved case independent for
619 - Bacula needs to propagate SD errors.
620 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
621 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
622 /dev/nst0. ERR=Input/output error.
623 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
624 Last block=5162 Current block=5164.
625 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
627 - Use autochanger to handle multiple devices.
628 - Add SuSE install doc to list.
629 - Check and rechedk "Invalid block number"
630 - Make bextract release the drive properly between tapes
631 so that an autochanger can be made to work.
632 - User wants to NOT backup up certain big files (email files).
633 - Maybe remove multiple simultaneous devices code in SD.
634 - On Windows with very long path names, it may be impossible to create
635 a file (and thus restore it) because the total length is too long.
636 We must cd into the directory then create the file without the
638 - lstat() is not going to work on Win32 for testing date.
639 - Implement a Recycle command
640 - Add client name to cram-md5 challenge so Director can immediately
641 verify if it is the correct client.
642 - Add JobLevel in FD status (but make sure it is defined).
643 - Audit all UA commands to ensure that we always prompt where possible.
644 - Check Jmsg in bnet, may not work, must dup bsock.
645 - Suppress Job Name in Jmsg for console
646 - Create Pools that are referenced in a Run statement at startup if possible.
647 - Use runbeforejob to unload, then reload a volume previously used,
648 then the next job run gets an error reading the drive.
649 - Make bootstrap filename unique.
650 - Test a second language e.g. french.
651 - Start working on Base jobs.
652 - Make "make binary-release" work from any directory.
653 - Implement UnsavedFiles DB record.
654 - Implement argc/argv for daemon command line scanning using table driven
656 - Implement table driven single argc/argv scanner to pickup all arguments.
657 Much like xxx_conf.c scan table.
658 keyword, handler(store_routine), store_address, code, flags, default.
659 - From Phil Stracchino:
660 It would probably be a per-client option, and would be called
661 something like, say, "Automatically purge obsoleted jobs". What it
662 would do is, when you successfully complete a Differential backup of a
663 client, it would automatically purge all Incremental backups for that
664 client that are rendered redundant by that Differential. Likewise,
665 when a Full backup on a client completed, it would automatically purge
666 all Differential and Incremental jobs obsoleted by that Full backup.
667 This would let people minimize the number of tapes they're keeping on
668 hand without having to master the art of retention times.
669 - Implement a M_SECURITY message class.
670 - When doing a Backup send all attributes back to the Director, who
671 would then figure out what files have been deleted.
672 - Currently in mount.c:236 the SD simply creates a Volume. It should have
673 explicit permission to do so. It should also mark the tape in error
674 if there is an error.
675 - Make sure all restore counters are working correctly in the FD.
676 - SD Bytes Read is wrong.
677 - Look at ALL higher level routines that call block.c to be sure
678 they don't expect something in errmsg.
679 - Investigate doing RAW backup of Win32 partition.
680 - Add thread specific data to hold the jcr -- send error messages from
681 low level routines by accessing it and using Jmsg().
682 - Cancel waiting for Client connect in SD if FD goes away.
684 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
685 - Implement timeout in response() when it should come quickly.
686 - Implement console @echo command.
687 - Implement a Slot priority (loaded/not loaded).
688 - Implement "vacation" Incremental only saves.
689 - Implement single pane restore (much like the Gftp panes).
690 - Implement Automatic Mount even in operator wait.
691 - Implement create "FileSet"?
692 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
693 - Implement "scratch pool" where tapes are defined and can be
694 taken by any pool that needs them.
695 - Implement restore "current system", but take all files without
696 doing selection tree -- so that jobs without File records can
698 - Add prefixlinks to where or not where absolute links to FD.
699 - Issue message to mount a new tape before the rewind.
700 - Simplified client job initiation for portables.
701 - If SD cannot open a drive, make it periodically retry.
702 - Add more of the config info to the tape label.
704 - If tape is marked read-only, then try opening it read-only rather than
705 failing, and remember that it cannot be written.
706 - Refine SD waiting output:
707 Device is being positioned
708 > Device is being positioned for append
709 > Device is being positioned to file x
711 - Figure out some way to estimate output size and to avoid splitting
712 a backup across two Volumes -- this could be useful for writing CDROMs
713 where you really prefer not to have it split -- not serious.
714 - Have SD compute MD5 or SHA1 and compare to what FD computes.
715 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
716 actual data on the Volume and compare it.
717 - Implement Bacula plugins -- design API
718 - Make bcopy read through bad tape records.
719 - Program files (i.e. execute a program to read/write files).
720 Pass read date of last backup, size of file last time.
721 - Add Signature type to File DB record.
722 - CD into subdirectory when open()ing files for backup to
723 speed up things. Test with testfind().
724 - Priority job to go to top of list.
725 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
726 - Implement some way for the Console to dynamically create a job.
727 - Restore to a particular time -- e.g. before date, after date.
728 - Solaris -I on tar for include list
729 - Need a verbose mode in restore, perhaps to bsr.
730 - bscan without -v is too quiet -- perhaps show jobs.
731 - Add code to reject whole blocks if not wanted on restore.
732 - Check if we can increase Bacula FD priorty in Win2000
733 - Make sure the MaxVolFiles is fully implemented in SD
734 - Check if both CatalogFiles and UseCatalog are set to SD.
735 - Figure out how to do a bare metal Windows restore
736 - Possibly add email to Watchdog if drive is unmounted too
737 long and a job is waiting on the drive.
738 - Restore program that errs in SD due to no tape, reports
739 OK incorrectly in output.
740 - After unmount, if restore job started, ask to mount.
741 - Convert all %x substitution variables, which are hard to remember
742 and read to %(variable-name). Idea from TMDA.
743 - Remove NextId for SQLite. Optimize.
744 - Move all SQL statements into a single location.
745 - Add UA rc and history files.
746 - put termcap (used by console) in ./configure and
747 allow -with-termcap-dir.
748 - Fix Autoprune for Volumes to respect need for full save.
749 - Fix Win32 config file definition name on /install
750 - Compare tape to Client files (attributes, or attributes and data)
751 - Make all database Ids 64 bit.
752 - Write an applet for Linux.
753 - Allow console commands to detach or run in background.
754 - Fix status delay on storage daemon during rewind.
755 - Add SD message variables to control operator wait time
756 - Maximum Operator Wait
757 - Minimum Message Interval
758 - Maximum Message Interval
759 - Send Operator message when cannot read tape label.
760 - Verify level=Volume (scan only), level=Data (compare of data to file).
761 Verify level=Catalog, level=InitCatalog
763 - Add keyword search to show command in Console.
764 - Events : tape has more than xxx bytes.
765 - Complete code in Bacula Resources -- this will permit
766 reading a new config file at any time.
767 - Handle ctl-c in Console
768 - Implement script driven addition of File daemon to config files.
769 - Think about how to make Bacula work better with File (non-tape) archives.
770 - Write Unix emulator for Windows.
771 - Put memory utilization in Status output of each daemon
772 if full status requested or if some level of debug on.
773 - Make database type selectable by .conf files i.e. at runtime
774 - Set flag for uname -a. Add to Volume label.
775 - Implement throttled work queue.
776 - Restore files modified after date
777 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
778 - Implement Restore FileSet=
779 - Create a protocol.h and protocol.c where all protocol messages
781 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
782 - Timout a job or terminate if link goes down, or reopen link and query.
783 - Concept of precious tapes (cannot be reused).
784 - Make bcopy copy with a single tape drive.
785 - Permit changing ownership during restore.
788 > My suggestion: Add a feature on the systray menu-icon menu to request
789 > an immediate backup now. This would be useful for laptop users who may
790 > not be on the network when the regular scheduled backup is run.
792 > My wife's suggestion: Add a setting to the win32 client to allow it to
793 > shut down the machine after backup is complete (after, of course,
794 > displaying a "System will shut down in one minute, click here to cancel"
795 > warning dialog). This would be useful for sites that want user
796 > woorkstations to be shut down overnight to save power.
799 - Autolabel should be specified by DIR instead of SD.
802 - AutoScan (check checksum of tape)
803 - Format command = "format /dev/nst0"
807 - Seek resolution (usually corresponds to buffer size)
808 - EODErrorCode=ENOSPC or code
809 - Partial Read error code
810 - Partial write error code
811 - Nonformatted read error
812 - Nonformatted write error
813 - WriteProtected error
817 - IgnoreCloseErrors=yes
827 - FD sends unsaved file list to Director at end of job (see
829 - File daemon should build list of files skipped, and then
830 at end of save retry and report any errors.
831 - Write a Storage daemon that uses pipes and
832 standard Unix programs to write to the tape.
834 - Need something that monitors the JCR queue and
835 times out jobs by asking the deamons where they are.
836 - Enhance Jmsg code to permit buffering and saving to disk.
837 - device driver = "xxxx" for drives.
839 - Ensure that /dev/null works
840 - Need report class for messages. Perhaps
841 report resource where report=group of messages
842 - enhance scan_attrib and rename scan_jobtype, and
843 fill in code for "since" option
844 - Director needs a time after which the report status is sent
845 anyway -- or better yet, a retry time for the job.
846 - Don't reschedule a job if previous incarnation is still running.
847 - Some way to automatically backup everything is needed????
848 - Need a structure for pending actions:
850 - termination status (part of buffered msgs?)
852 Read, Write, Clean, Delete
853 - Login to Bacula; Bacula users with different permissions:
854 owner, group, user, quotas
855 - Store info on each file system type (probably in the job header on tape.
856 This could be the output of df; or perhaps some sort of /etc/mtab record.
859 - Design at hierarchial storage for Bacula. Migration and Clone.
860 - Implement FSM (File System Modules).
861 - Audit M_ error codes to ensure they are correct and consistent.
862 - Add variable break characters to lex analyzer.
863 Either a bit mask or a string of chars so that
864 the caller can change the break characters.
865 - Make a single T_BREAK to replace T_COMMA, etc.
866 - Ensure that File daemon and Storage daemon can
867 continue a save if the Director goes down (this
868 is NOT currently the case). Must detect socket error,
869 buffer messages for later.
870 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
871 - Add ability to backup to two Storage devices (two SD sessions) at
872 the same time -- e.g. onsite, offsite.
873 - Add the ability to consolidate old backup sets (basically do a restore
874 to tape and appropriately update the catalog). Compress Volume sets.
875 Might need to spool via file is only one drive is available.
876 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
877 someway to do so with every volume that has less than x% valid
881 Migration: Move a backup from one Volume to another
882 Clone: Copy a backup -- two Volumes
884 Bacula Migration is based on Jobs (apparently Networker is file by file).
886 Migration triggered by:
890 Highwater mark (keep total size)
895 ======================================================
897 It is somewhat like a Full save becomes an incremental since
898 the Base job (or jobs) plus other non-base files.
900 - A Base backup is same as Full backup, just different type.
901 - New BaseFiles table that contains:
903 BaseJobId - Base JobId referenced for this FileId (needed ???)
904 JobId - JobId currently running
905 FileId - File not backed up, exists in Base Job
906 FileIndex - FileIndex from Base Job.
907 i.e. for each base file that exists but is not saved because
908 it has not changed, the File daemon sends the JobId, BaseId,
909 FileId, FileIndex back to the Director who creates the DB entry.
910 - To initiate a Base save, the Director sends the FD
911 the FileId, and full filename for each file in the Base.
912 - When the FD finds a Base file, he requests the Director to
913 send him the full File entry (stat packet plus MD5), or
914 conversely, the FD sends it to the Director and the Director
915 says yes or no. This can be quite rapid if the FileId is kept
916 by the FD for each Base Filename.
917 - It is probably better to have the comparison done by the FD
918 despite the fact that the File entry must be sent across the
920 - An alternative would be to send the FD the whole File entry
921 from the start. The disadvantage is that it requires a lot of
922 space. The advantage is that it requires less communications
924 - The Job record must be updated to indicate that one or more
926 - At end of Job, FD returns:
927 1. Count of base files/bytes not written to tape (i.e. matches)
928 2. Count of base file that were saved i.e. had changed.
929 - No tape record would be written for a Base file that matches, in the
930 same way that no tape record is written for Incremental jobs where
931 the file is not saved because it is unchanged.
932 - On a restore, all the Base file records must explicitly be
933 found from the BaseFile tape. I.e. for each Full save that is marked
934 to have one or more Base Jobs, search the BaseFile for all occurrences
936 - An optimization might be to make the BaseFile have:
942 This would avoid the need to explicitly fetch each File record for
943 the Base job. The Base Job record will be fetched to get the
944 VolSessionId and VolSessionTime.
945 =========================================================
948 ==========================================================
950 For each Incremental job that is run, there may be files that
951 were found but not saved because they were locked (this applies
952 only to Windows). Such a system could send back to the Director
953 a list of Unsaved files.
955 - New UnSavedFiles table that contains:
959 - Then in the next Incremental job, the list of Unsaved Files will be
960 feed to the FD, who will ensure that they are explicitly chosen even
961 if standard date/time check would not have selected them.
962 =============================================================
965 Done: (see kernsdone for more)
968 - John's Full save failed with 1.32c FD and 1.31 Dir no FD status,
969 and no error message.
970 - Add fd and st as Console keywords.
971 - Recycling volume with a Slot requires an operator intervention:
972 rufus-dir: Start Backup JobId 18, Job=kernsave.2003-11-01_21.23.52
973 rufus-dir: Pruned 1 Job on Volume Vol01 from catalog.
974 rufus-dir: There are no Jobs associated with Volume Vol01. Marking it purged.
975 rufus-dir: Recycled volume "Vol01"
976 rufus-sd: Please mount Volume "Vol01" on Storage Device "DDS-4" for Job kernsave.2003-11-01_21.23.52
977 Use "mount" command to release Job.
978 - Implement Dan's bacula script (email of 26 Oct).
979 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
980 - Edit the Client/Storage name into authentication failure messages.
981 - Fix packet too big problem. This is most likely a Windows TCP stack
983 - Implement ClientRunBeforeJob and ClientRunAfterJob.
984 - Implement forward spacing block/file: position_device(bsr) --
985 just before read_block_from_device();
988 Multiple drive autochanger data: see Alan Brown
989 mtx -f xxx unloadStorage Element 1 is Already Full(drive 0 was empty)
990 Unloading Data Transfer Element into Storage Element 1...source Element
993 (drive 0 was empty and so was slot 1)
994 > mtx -f xxx load 15 0
995 no response, just returns to the command prompt when complete.
996 > mtx -f xxx status Storage Changer /dev/changer:2 Drives, 60 Slots ( 2 Import/Export )
997 Data Transfer Element 0:Full (Storage Element 15 Loaded):VolumeTag = HX001
998 Data Transfer Element 1:Empty
999 Storage Element 1:Empty
1000 Storage Element 2:Full :VolumeTag=HX002
1001 Storage Element 3:Full :VolumeTag=HX003
1002 Storage Element 4:Full :VolumeTag=HX004
1003 Storage Element 5:Full :VolumeTag=HX005
1004 Storage Element 6:Full :VolumeTag=HX006
1005 Storage Element 7:Full :VolumeTag=HX007
1006 Storage Element 8:Full :VolumeTag=HX008
1007 Storage Element 9:Full :VolumeTag=HX009
1008 Storage Element 10:Full :VolumeTag=HX010
1009 Storage Element 11:Empty
1010 Storage Element 12:Empty
1011 Storage Element 13:Empty
1012 Storage Element 14:Empty
1013 Storage Element 15:Empty
1014 Storage Element 16:Empty....
1015 Storage Element 28:Empty
1016 Storage Element 29:Full :VolumeTag=CLNU01L1
1017 Storage Element 30:Empty....
1018 Storage Element 57:Empty
1019 Storage Element 58:Full :VolumeTag=NEX261L2
1020 Storage Element 59 IMPORT/EXPORT:Empty
1021 Storage Element 60 IMPORT/EXPORT:Empty
1023 Unloading Data Transfer Element into Storage Element 15...done
1025 (just to verify it remembers where it came from, however it can be
1026 overrriden with mtx unload {slotnumber} to go to any storage slot.)
1028 There needs to be a table of drive # to devices somewhere - If there are
1029 multiple changers or drives there may not be a 1:1 correspondance between
1030 changer drive number and system device name - and depending on the way the
1031 drives are hooked up to scsi busses, they may not be linearly numbered
1032 from an offset point either.something like
1034 Autochanger drives = 2
1035 Autochanger drive 0 = /dev/nst1
1036 Autochanger drive 1 = /dev/nst2
1037 IMHO, it would be _safest_ to use explicit mtx unload commands at all
1038 times, not just for multidrive changers. For a 1 drive changer, that's
1044 MTX's manpage (1.2.15):
1045 unload [<slotnum>] [ <drivenum> ]
1046 Unloads media from drive <drivenum> into slot
1047 <slotnum>. If <drivenum> is omitted, defaults to
1048 drive 0 (as do all commands). If <slotnum> is
1049 omitted, defaults to the slot that the drive was
1050 loaded from. Note that there's currently no way
1051 to say 'unload drive 1's media to the slot it
1052 came from', other than to explicitly use that
1053 slot number as the destination.AB
1059 undef# camcontrol devlist
1060 <WANGTEK 51000 SCSI M74H 12B3> at scbus0 target 2 lun 0 (pass0,sa0)
1061 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 0 (pass1,sa1)
1062 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 1 (pass2)
1064 tapeinfo -f /dev/sg0 with a bad tape in drive 1:
1065 [kern@rufus mtx-1.2.17kes]$ ./tapeinfo -f /dev/sg0
1066 Product Type: Tape Drive
1068 Product ID: 'C5713A '
1070 Attached Changer: No
1071 TapeAlert[3]: Hard Error: Uncorrectable read/write error.
1072 TapeAlert[20]: Clean Now: The tape drive neads cleaning NOW.
1079 Medium Type: Not Loaded
1082 DataCompEnabled: yes
1083 DataCompCapable: yes
1084 DataDeCompEnabled: yes
1091 Handling removable disks
1093 From: Karl Cunningham <karlc@keckec.com>
1095 My backups are only to hard disk these days, in removable bays. This is my
1096 idea of how a backup to hard disk would work more smoothly. Some of these
1097 things Bacula does already, but I mention them for completeness. If others
1098 have better ways to do this, I'd like to hear about it.
1100 1. Accommodate several disks, rotated similar to how tapes are. Identified
1101 by partition volume ID or perhaps by the name of a subdirectory.
1102 2. Abort & notify the admin if the wrong disk is in the bay.
1103 3. Write backups to different subdirectories for each machine to be backed
1105 4. Volumes (files) get created as needed in the proper subdirectory, one
1107 5. When a disk is recycled, remove or zero all old backup files. This is
1108 important as the disk being recycled may be close to full. This may be
1109 better done manually since the backup files for many machines may be
1110 scattered in many subdirectories.
1115 - To pass Include 1 or two letter commands
1116 I Name Include name - first record
1117 B Name Base name - repeat
1120 E zzz Exclude expression (wild card)
1122 D "reader" Reader program
1123 T "writer" Writer program
1124 O Options In current commpressed format
1125 (compression, signature, onefs, recurse, sparse,
1126 replace, verify options, ...)
1128 B BaseName Start second option set any letter
1131 F Number Number of filenames to follow
1136 F Number Number of filenames to follow
1139 - Add regression of btape "fill"
1141 - Implement "label slots=1,2... pool=yyy barcodes"
1142 - Mark a non-removable disk Volume in error rather than asking sys op to
1144 - Implement Fixed storage LabelFormat test.
1145 - Add reporting in attr despooling.
1146 - Figure out how to use ssh or stunnel to protect Bacula communications.
1147 Add Dan's work to manual See ssl.txt
1150 - Implement fast tree insert (doubly linked list)
1151 - Disallow using Internal database
1152 - Feedback while the restore tree is being built.
1153 - Add detection of Whitebox release in ./configure
1154 - Add InChanger to "list media"
1155 - Implement second cut of SIGHUP in Dir
1156 (Implement resources on a single pointer)
1157 - Add regular expressions to FileSets (needs testing).
1158 - Test new despooling error recovery code when partition fills.
1159 - Sort Scheduled jobs status listing by start time.
1160 - Add priority to Scheduled jobs status listing.
1161 - Look at Win32 inc problem when new directory added files not saved???
1162 - Require resource names to be unique.
1163 - Make sure that all errors from libacl are printed.
1164 - Make FD run on Win95 if possible (not really possible).
1165 - Fix GROUP BYs in all SQL (done by Dan -- thanks)
1166 - Make entering multiple Storage names in Dir illegal.
1167 - Make btape release the drive during the "test" append.
1169 You have selected the following JobIds: 739,856,797
1170 Building directory tree for JobId 739 ...
1171 Building directory tree for JobId 856 ...
1172 Building directory tree for JobId 797 ...3
1173 Jobs inserted into the tree.
1174 - Fix errno handling in win32 compat routines.
1175 - Add better error codes to run_program (10000+)
1176 - Revisit and revise Disaster Recovery (fix SCSI and RAID
1178 - When passwords do not match, print message that points the
1180 - Do tape alerts -- see tapealert.txt
1181 - Find memory leaks in address_conf.c
1182 - Network order of port and ip addresses.
1183 - SetIP in dird/fd_cmds.c is broken.
1184 - don't use inet_addr() it is deprecated.
1185 - Fix code in filed.c for init_bsock();
1186 - Fix code in console.c for init_bsock();
1187 - Solve the termcap.h problem on Solaris configure.
1188 - Make Verify jobs require exclusive use of Volume as Restore
1190 - Fix restore ++++ that get intermingled with "Building directory tree"
1191 - Fix doc in dirdconf.html for @, where the file must
1192 have not a list of files, but
1195 - Test Win32 errno handling.
1196 - port 1.34.6-block.patch to 1.35
1197 - Look at Chris' patch for bscan bug.
1198 - Fix error handling in spooling both data and attribute.