8 Data encryption Meno Abels (see projects)
9 Communication encryption Meno Abels
10 Version 1.37 Kern (see below)
11 ========================================================
14 #3 Migration (Move, Copy, Archive Jobs)
15 #4 Embedded Python Scripting
16 (Implemented in Dir/SD)
17 #5 Events that call a Python program
18 (Implemented in Dir/SD)
19 #6 Select one from among Multiple Storage Devices for Job
20 #7 Single Job Writing to Multiple Storage Devices
22 ## Integrate web-bacula into a new Bacula project with
24 ## Consider moving docs to their own project.
26 Suggestions for Preben:
27 - Look at adding Client run command that will use the
28 port opened by the client.
29 - Implement WildFile and WildDir to solve problem of
30 saving only *.doc files.
31 - Bootstrap from JobMedia records.
32 - Optimized bootstrap.
35 - Add a restore directory-x
36 - When labeling tapes, if you enter 000026, Bacula uses
37 the tape index rather than the Volume name 000026.
38 - Max Vols limit in Pool off by one?
39 - Look at Preben's acl.c error handling code.
40 - See multiple-store.txt for Multiple Storage implementation
42 - Create a new GUI chapter explaining all the GUI programs.
43 - Tell the "restore" user when browsing is no longer possible.
44 - Add disk seeking on restore.
45 - Cancel command should include JobId in list of Jobs.
46 - Require restore via the restore command or make a restore Job
47 get the bootstrap file.
48 - Add dump of VolSessionId/Time and FileIndex with bls.
49 - Make bootstrap file handle multiple MediaTypes (SD)
50 - Add offline tape command to Bacula console.
51 - Add performance testing hooks
52 - Add Python writable variable for changing the Priority,
53 Client, Storage, JobStatus (error), ...
54 - Document that Bootstrap files can be written with cataloging
56 - Look at adding full Volume and Pool information to a Volume
57 label so that bscan can get *all* the info.
58 - Scratch Pool where the volumes can be re-assigned to any Pool.
59 - Upgrade to MySQL 4.1.1 See:
60 http://dev.mysql.com/doc/mysql/en/Server_SQL_mode.html
62 Enter MediaId or Volume name: 32
63 Enter new Volume name: DLT-20Dec04
64 Automatically selected Pool: Default
65 Connecting to Storage daemon DLTDrive at 192.168.68.104:9103 ...
66 Sending relabel command from "DLT-28Jun03" to "DLT-20Dec04" ...
67 block.c:552 Write error at 0:0 on device /dev/nst0. ERR=Bad file descriptor.
68 Error writing final EOF to tape. This tape may not be readable.
69 dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
70 askdir.c:219 NULL Volume name. This shouldn't happen!!!
71 3912 Failed to label Volume: ERR=dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
73 Label command failed for Volume DLT-20Dec04.
74 Do not forget to mount the drive!!!
75 - Bug: if a job is manually scheduled to run later, it does not appear
76 in any status report and cannot be cancelled.
78 Regression tests (Jo):
79 - Add Pool/Storage override regression test.
80 - Add delete JobId to regression.
81 - Add a regression test for dbcheck.
82 - New test to add bscan to four-concurrent-jobs regression,
83 i.e. after the four-concurrent jobs zap the
84 database as is done in the bscan-test, then use bscan to
85 restore the database, do a restore and compare with the
87 - Add restore of specific JobId to regression (item 3
88 on the restore prompt)
89 - Add IPv6 to regression
90 - Add database test to regression. Test each function like delete,
93 - AntiVir can slow down backups on Win32 systems.
94 - Win32 systems with FAT32 can be much slower than NTFS for
95 more than 1000 files per directory.
99 === From Carsten Menke <bootsy52@gmx.net>
100 The bacula-devel list is CC'ed (I'm also subscribed to the list right now,
101 so I will also get the responses).
103 Following is a list of what I think in the situations where I'm faced with,
104 could be a usefull enhancement to bacula, which I'm certain other users will
105 benefit from as well.
107 1. NextJob/NextJobs Directive within a Job Resource in the form of
108 NextJobs = job1,job2.
111 I currently solved the problem with running multiple jobs each after each
112 by setting the Max Wait Time for a job to 8 hours, and give
113 the jobs different Priorities. However, there scenarios where
114 1 Job is directly depending on another job, so if the former job fails,
115 the job after it needn't to be run
116 while maybe other jobs should run despite of that
119 A Backup Job and a Verify job, if the backup job fails there is no need to run
120 the verify job, as the backup job already failed. However, one may like
121 to backup the Catalog to disk despite of that the main backup job failed.
124 I see that this is related to the Event Handlers which are on the ToDo
125 list, also it is maybe a good idea to check for the return value and
126 execute different actions based on the return value
128 2. Priorities also for the Schedule Resource
132 With this it is possible to run monthly jobs at the end of the month
133 instead of the beginning of the next month, so one could define
134 weekly and monthly pools to be run on Fridays and if there are
135 2 Jobs resulting on the same time, the one with the higher priority
136 takes precedence. This is also helpfull to implement a daily
137 rotation scheme with an additional 2 weeks changing weekly pool
141 Name = "NightlyBackup"
142 Run = Priority = 1,Pool=Weekly friday at 21:00
143 Run = Priority = 2,Pool=Weekly monthly on friday at 21:00
144 Run = Priority = 3,Pool=Weekly yearly on friday at 21:00
147 3. offline capability to bconsole
150 Currently I use a script which I execute within the last Job via the
151 RunAfterJob Directive, to release and eject the tape.
152 So I have to call bconsole "release=Storage-Name" and afterwards
153 mt -f /dev/nst0 eject to get the tape out.
155 If I have multiple Storage Devices, than these may not be /dev/nst0 and
156 I have to modify the script or call it with parameters etc.
157 This would actually not be needed, as everything is already defined
158 in bacula-sd.conf and if I can invoke bconsole with the
159 storage name via $1 in the script than I'm done and information is
162 4. %s for Storage Name added to the chars being substituted in "RunAfterJob"
166 For the reason mentioned in 3. to have the ability to call a
167 script with /scripts/foobar %s and in the script use $1
168 to pass the Storage Name to bconsole
170 5. Setting Volume State within a Job Resource
173 Instead of using "Maximum Volume Jobs" in the Pool Resource,
174 I would have the possibilty to define
175 in a Job Resource that after this certain job is run, the Volume State
176 should be set to "Volume State = Used", this give more flexibility (IMHO).
178 6. Localization of Bacula Messages
181 Unfortunatley many,many people I work with don't speak english very well.
182 So if at least the Reporting messages would be localized then they
183 would understand that they have to change the tape,etc. etc.
185 I volunteer to do the german translations, and if I can convince my wife also
186 french and Morre (western african language).
188 7. OK, this is evil, probably bound to security risks and maybe not possible
189 due to the design of bacula.
191 Implementation of Backtics ( `command` ) for shell comand execution to
192 the "Label Format" Directive.
196 Currently I have defined BACULA_DAY_OF_WEEK="day1|day2..." resulting in
197 Label Format = "HolyBackup-${BACULA_DAY_OF_WEEK[${WeekDay}]}". If I could
198 use backticks than I could use "Label Format = HolyBackup-`date +%A` to have
199 the localized name for the day of the week appended to the
200 format string. Then I have the tape labeled automatically with weekday
201 name in the correct language.
203 - Can one write tapes faster with 8192 byte block sizes?
204 - Specify a single directory to restore.
205 - Implement Maximum Job Spool Size
206 - Document security problems with the same password for everyone in
207 rpm and Win32 releases.
208 - Browse generations of files.
209 - I've seen an error when my catalog's File table fills up. I
210 then have to recreate the File table with a larger maximum row
211 size. Relevant information is at
212 http://dev.mysql.com/doc/mysql/en/Full_table.html ; I think the
213 "Installing and Configuring MySQL" chapter should talk a bit
214 about this potential problem, and recommend a solution.
215 - For Solaris must use POSIX awk.
216 - Want speed of writing to tape while despooling.
217 - Supported autochanger:
225 Wangtek 6525ES (SCSI-1 QIC drive, 525MB), under Linux 2.4.something,
226 bacula 1.36.0/1 works with blocksize 16k INSIDE bacula-sd.conf.
227 - Add regex from http://www.pcre.org to Bacula for Win32.
228 - Use only shell tools no make in CDROM package.
229 - Document PostgreSQL performance problems bug 131.
231 - Document that ChangerDevice is used for Alert command.
232 - Include within include does it work?
233 - Implement a Pool of type Cleaning?
234 - Implement VolReadTime and VolWriteTime in SD
235 - Modify Backing up Your Database to include a bootstrap file.
236 - Think about making certain database errors fatal.
237 - Look at correcting the time jump in the scheduler for daylight
238 savings time changes.
239 - Add a "real" timer to network connections.
240 - Write non-optimized bsrs from the JobMedia and Media records,
241 even after Files are pruned.
242 - Promote to Full = Time period
243 - Despool attributes simultaneously with data in a separate
244 thread, rejoined at end of data spooling.
245 - Implement Files/Bytes,... stats for restore job.
246 - Implement Total Bytes Written, ... for restore job.
247 - Check dates entered by user for correctness (month/day/... ranges)
248 - Compress restore Volume listing by date and first file.
249 - Look at patches/bacula_db.b2z postgresql that loops during restore.
251 - Perhaps add read/write programs and/or plugins to FileSets.
252 - How to handle backing up portables ...
253 - Add some sort of guaranteed Interval for upgrading jobs.
254 - Can we write the state file after every job terminates? On Win32
255 the system crashes and the state file is not updated.
258 Documentation to do: (any release a little bit at a time)
259 - Doc to do unmount before removing magazine.
260 - Alternative to static linking "ldd prog" save all binaries listed,
261 restore them and point LD_LIBRARY_PATH to them.
262 - Document add "</dev/null >/dev/null 2>&1" to the bacula-fd command line
263 - Document query file format.
264 - Add more documentation for bsr files.
265 - Document problems with Verify and pruning.
266 - Document how to use multiple databases.
267 - VXA drives have a "cleaning required"
268 indicator, but Exabyte recommends preventive cleaning after every 75
271 In this context, it should be noted that Exabyte has a command-line
272 vxatool utility available for free download. (The current version is
273 vxatool-3.72.) It can get diagnostic info, read, write and erase tapes,
274 test the drive, unload tapes, change drive settings, flash new firmware,
276 Of particular interest in this context is that vxatool <device> -i will
277 report, among other details, the time since last cleaning in tape motion
278 minutes. This information can be retrieved (and settings changed, for
279 that matter) through the generic-SCSI device even when Bacula has the
280 regular tape device locked. (Needless to say, I don't recommend
281 changing tape settings while a job is running.)
282 - Lookup HP cleaning recommendations.
283 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
284 - Document doing table repair
287 Testing to do: (painful)
290 For 1.37 Testing/Documentation:
291 - Add bandwidth limiting.
292 - Add acks every once and a while from the SD to keep
293 the line from timing out.
294 - When an error in input occurs and conio beeps, you can back
295 up through the prompt.
296 - Detect fixed tape block mode during positioning by looking at
297 block numbers in btape "test". Possibly adjust in Bacula.
298 - Fix list volumes to output volume retention in some other
299 units, perhaps via a directive.
300 - If opening a tape in read/write mode fails attempt to open
301 it in read-only mode, and mark the tape for read only.
302 - Add a read-only mode to the mount option.
303 - Allow Simultaneous Priorities = yes => run up to Max concurrent jobs even
304 with multiple priorities.
305 - If you use restore replace=never, the directory attributes for
306 non-existent directories will not be restored properly.
308 - see lzma401.zip in others directory for new compression
310 - Minimal autochanger handling in Bacula and in btape.
311 - Look into how tar does not save sockets and the possiblity of
312 not saving them in Bacula (Martin Simmons reported this).
313 - Fix restore jobs so that multiple jobs can run if they
314 are not using the same tape(s).
315 - Allow the user to select JobType for manual pruning/purging.
316 - bscan does not put first of two volumes back with all info in
318 - Implement the FreeBSD nodump flag in chflags.
319 - Figure out how to make named console messages go only to that
320 console and to the non-restricted console (new console class?).
321 - Make restricted console prompt for password if *ask* is set or
322 perhaps if password is undefined.
323 - Implement "from ISO-date/time every x hours/days/weeks/months" in
328 MA = (last_MA * 3 + rate) / 4
329 rate = (bytes - last_bytes) / (runtime - last_runtime)
331 ==== from Marc Schoechlin
332 - the help-command should be more verbose
333 (it should explain the paramters of the different
335 -> it´s time-comsuming to consult the manual anytime
336 you need a special parameter
337 -> maybe it´s more easy to maintain this, if the
338 descriptions of that commands are outsourced to
340 - the cd-command should allow complete paths
341 i.e. cd /foo/bar/foo/bar
342 -> if a customer mails me the path to a certain file,
343 it´s faster to enter the specified directory
344 - if the password is not configured in bconsole.conf
345 you should be asked for it.
346 -> sometimes you like to do restore on a customer-machine
347 which shouldn´t know the password for bacula.
348 -> adding the password to the file favours admins
349 to forget to remove the password after usage
351 the protection of that file is less important
352 - long-listed-output of commands should be scrollable
353 like the unix more/less-command does
354 -> if someone runs 200 and more machines, the lists could
355 be a little long and complex
356 - command-output should be shown column by column
357 to reduce scrolling and to increase clarity
359 - lsmark should list the selected files with full
361 - wildcards for selecting and file and directories would be nice
362 - any actions should be interuptable with STRG+C
363 - command-expansion would be pretty cool
365 - Compaction of Disk space by "migrating" Volumes that have pruned
366 Jobs (what criteria? size, #jobs, time).
367 - Add prune all command
368 - Document fact that purge can destroy a part of a restore by purging
369 one volume while others remain valid -- perhaps mark Jobs.
370 - Add multiple-media-types.txt
371 - look at mxt-changer.html
372 - Make ? do a help command (no return needed).
373 - Implement restore directory.
374 - Document streams and how to implement them.
375 - Try not to re-backup a file if a new hard link is added.
376 - Add feature to backup hard links only, but not the data.
377 - Fix stream handling to be simpler.
378 - Add Priority and Bootstrap to Run a Job.
379 - Eliminate Restore "Run Restore Job" prompt by allowing new "run command
381 - Remove View FileSet button from Run a Job dialog.
382 - Handle prompt for restore job at end of Restore command.
383 - Add display of total selected files to Restore window.
384 - Add tree pane to left of window.
385 - Add progress meter.
386 - Max wait time or max run time causes seg fault -- see runtime-bug.txt
387 - Document writing to a CD/DVD with Bacula.
388 - Add a "base" package to the window installer for pthreadsVCE.dll
389 which is needed by all packages.
390 - Add message to user to check for fixed block size when the forward
391 space test fails in btape.
392 - When unmarking a directory check if all files below are unmarked and
393 then remove the + flag -- in the restore tree.
394 - Possibly implement: Action = Unmount Device="TapeDrive1" in Admin jobs.
395 - Setup lrrd graphs: (http://www.linpro.no/projects/lrrd/) Mike Acar.
396 - Revisit the question of multiple Volumes (disk) on a single device.
397 - Add a block copy option to bcopy.
398 - Investigate adding Mac Resource Forks.
399 - Finish work on Gnome restore GUI.
400 - Fix "llist jobid=xx" where no fileset or client exists.
402 it seems to be complaining about 12:00pm which should be a valid 12
403 hour time. I changed the time to 11:59am and everything works fine.
404 Also 12:00am works fine. 0:00pm also works (which I don't think
405 should). None of the values 12:00pm - 12:59pm work for that matter.
406 - Add level to estimate command.
407 - For each job type (Admin, Restore, ...) require only the really necessary
409 - Fix option 2 of restore -- list where file is backed up -- require Client,
410 then list last 20 backups.
411 - Pass Director resource name as an option to the Console.
412 - Add a "batch" mode to the Console (no unsolicited queries, ...).
413 - Add a .list all files in the restore tree (probably also a list all files)
414 Do both a long and short form.
415 - Allow browsing the catalog to see all versions of a file (with
416 stat data on each file).
417 - Restore attributes of directory if replace=never set but directory
419 - Use SHA1 on authentication if possible.
420 - See comtest-xxx.zip for Windows code to talk to USB.
421 - Add Events and Perl scripting.
422 - Add John's appended files:
423 Appended = { /files/server/logs/http/*log }
424 and such files would be treated as follows.On a FULL backup, they would
425 be backed up like any other file.On an INCREMENTAL backup, where a
426 previous INCREMENTAL or FULL was already in thecatalogue and the length
427 of the file wasgreater than the length of the last backup, only thedata
428 added since the last backup will be dumped.On an INCREMENTAL backup, if
429 the length of the file is less than thelength of the file with the same
430 name last backed up, the completefile is dumped.On Windows systems, with
431 creation date of files, we can be evensmarter about this and not count
432 entirely upon the length.On a restore, the full and all incrementals
433 since it will beapplied in sequence to restore the file.
434 - Allow for optional cancelling of SD and FD in case DIR
435 gets a fatal error. Requested by Jesse Guardiani <jesse@wingnet.net>
436 - Add "limit=n" for "list jobs"
437 - Check new HAVE_WIN32 open bits.
438 - Check if the tape has moved before writing.
439 - Handling removable disks -- see below:
440 - Keep track of tape use time, and report when cleaning is necessary.
441 - Add FromClient and ToClient keywords on restore command (or
442 BackupClient RestoreClient).
443 - Implement a JobSet, which groups any number of jobs. If the
444 JobSet is started, all the jobs are started together.
445 Allow Pool, Level, and Schedule overrides.
446 - Enhance cancel to timeout BSOCK packets after a specific delay.
447 - Do scheduling by UTC using gmtime_r() in run_conf, scheduler, and
448 ua_status.!!! Thanks to Alan Brown for this tip.
449 - Look at updating Volume Jobs so that Max Volume Jobs = 1 will work
450 correctly for multiple simultaneous jobs.
451 - Correct code so that FileSet MD5 is calculated for < and | filename
453 - Implement the Media record flag that indicates that the Volume does disk
455 - Implement VolAddr, which is used when Volume is addressed like a disk,
456 and form it from VolFile and VolBlock.
457 - Make multiple restore jobs for multiple media types specifying
458 the proper storage type.
459 - Implement MediaType keyword in bsr?
460 - Fix fast block rejection (stored/read_record.c:118). It passes a null
461 pointer (rec) to try_repositioning().
462 - Look at extracting Win data from BackupRead.
463 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
464 which would take precidence over the Catalog "JobRetention".
465 - Implement Label Format in Add and Label console commands.
466 - Possibly up network buffers to 65K. Put on variable.
467 - Put email tape request delays on one or more variables. User wants
468 to cancel the job after a certain time interval. Maximum Mount Wait?
469 - Job, Client, Device, Pool, or Volume?
470 Is it possible to make this a directive which is *optional* in multiple
471 resources, like Level? If so, I think I'd make it an optional directive
472 in Job, Client, and Pool, with precedence such that Job overrides Client
473 which in turn overrides Pool.
475 - New Storage specifications:
476 - Want to write to multiple storage devices simultaneously
477 - Want to write to multiple storage devices sequentially (in one job)
478 - Want to read/write simultaneously
479 - Key is MediaType -- it must match
481 Passed to SD as a sort of BSR record called Storage Specification
485 MediaType -> Next MediaType
487 Device -> Next Device
489 Allow multiple Storage specifications
497 Allow Multiple Pool specifications (note, Pool currently
499 Allow Multiple MediaType specifications in Dir conf
500 Allow Multiple Device specifications in Dir conf
501 Perhaps keep this in a single SSR
502 Tie a Volume to a specific device by using a MediaType that
503 is contained in only one device.
504 In SD allow Device to have Multiple MediaTypes
506 - Ideas from Jerry Scharf:
507 First let's point out some big pluses that bacula has for this
509 more importantly it's active. Thank you so much for that
510 even more important, it's not flaky
511 it has an open access catalog, opening many possibilities
512 it's pushing toward heterogeneous systems capability
514 Macintosh file client
515 macs are an interesting niche, but I fear a server is a rathole
516 working bare iron recovery for windows
517 the option for inc/diff backups not reset on fileset revision
518 a) use both change and inode update time against base time
519 b) do the full catalog check (expensive but accurate)
520 sizing guide (how much system is needed to back up N systems/files)
521 consultants on using bacula in building a disaster recovery system
523 or how to get at fancy things that one could do with bacula
524 logwatch code for bacula logs (or similar)
525 linux distro inclusion of bacula (brings good and bad, but necessary)
526 win2k/XP server capability (icky but you asked)
527 support for Oracle database ??
529 - Look at adding SQL server and Exchange support for Windows.
530 - Each DVD-RAM disk would be a volume, just like each tape is
531 a volume. It's a 4.7GB media with random access, but there's nothing about
532 it that I can see that makes it so different than a tape from bacula's
533 perspective. Why couldn't I back up to a bare floppy as a volume (ignoring
535 - Make dev->file and dev->block_num signed integers so that -1 can
536 be an invalid value which happens with BSR.
537 - Create VolAddr for disk files in place of VolFile and VolBlock. This
538 is needed to properly specify ranges.
539 - Add progress of files/bytes to SD and FD.
540 - Print warning message if FileId > 4 billion
541 - do a "messages" before the first prompt in Console
542 - Add a date and time stamp at the beginning of every line in the
543 Job report (Volker Sauer).
544 - Client does not show busy during Estimate command.
545 - Implement Console mtx commands.
546 - Implement a Mount Command and an Unmount Command where
547 the users could specify a system command to be performed
548 to do the mount, after which Bacula could attempt to
549 read the device. This is for Removeable media such as a CDROM.
550 - Most likely, this mount command would be invoked explicitly
551 by the user using the current Console "mount" and "unmount"
552 commands -- the Storage Daemon would do the right thing
553 depending on the exact nature of the device.
554 - As with tape drives, when Bacula wanted a new removable
555 disk mounted, it would unmount the old one, and send a message
556 to the user, who would then use "mount" as described above
557 once he had actually inserted the disk.
558 - Implement dump/print label to UA
559 - Spool to disk only when the tape is full, then when a tape is hung move
561 - bextract is sending everything to the log file ****FIXME****
562 - Allow multiple Storage specifications (or multiple names on
563 a single Storage specification) in the Job record. Thus a job
564 can be backed up to a number of storage devices.
565 - Implement some way for the File daemon to contact the Director
566 to start a job or pass its DHCP obtained IP number.
567 - Implement a query tape prompt/replace feature for a console
568 - Copy console @ code to gnome2-console
569 - Make AES the only encryption algorithm see
570 http://csrc.nist.gov/CryptoToolkit/aes/). It's
571 an officially adopted standard, has survived peer
572 review, and provides keys up to 256 bits.
573 - Take a careful look at SetACL http://setacl.sourceforge.net
574 - Take a careful look at Level for the estimate command, maybe make
575 it a command line option.
576 - Add Volume name to "I cannot write on this volume because"
577 - Make tree walk routines like cd, ls, ... more user friendly
578 by handling spaces better.
579 - Make sure that Bacula rechecks the tape after the 20 min wait.
580 - Set IO_NOWAIT on Bacula TCP/IP packets.
581 - Try doing a raw partition backup and restore by mounting a
584 Yes, it would allow to highly automatic the request for new tapes. If a
585 tape is empty, bacula reads the barcodes (native or simulated), and if
586 an unused tape is found, it runs the label command with all the
587 necessary parameters.
589 By the way can bacula automatically "move" an empty/purged volume say
590 in the "short" pool to the "long" pool if this pool runs out of volume
592 - What to do about "list files job=xxx".
593 - Make Dmsg look at global before calling subroutine.
594 - Get and test MySQL 4.0
595 - Look at how fuser works and /proc/PID/fd that is how Nic found the
596 file descriptor leak in Bacula.
597 - Implement WrapCounters in Counters.
598 - Add heartbeat from FD to SD if hb interval expires.
599 - Can we dynamically change FileSets?
600 - If pool specified to label command and Label Format is specified,
601 automatically generate the Volume name.
602 - Why can't SQL do the filename sort for restore?
603 - Look into Microsoft Volume Shadowcopy Service VSS for backing
604 up system state components (Active Directory, System Volume, ...)
605 - Add ExhautiveRestoreSearch
606 - Look at the possibility of loading only the necessary
607 data into the restore tree (i.e. do it one directory at a
608 time as the user walks through the tree).
609 - Possibly use the hash code if the user selects all for a restore command.
610 - Fix "restore all" to bypass building the tree.
611 - Prohibit backing up archive device (findlib/find_one.c:128)
612 - Implement Release Device in the Job resource to unmount a drive.
613 - Implement Acquire Device in the Job resource to mount a drive,
614 be sure this works with admin jobs so that the user can get
615 prompted to insert the correct tape. Possibly some way to say to
616 run the job but don't save the files.
617 - Make things like list where a file is saved case independent for
620 - Use autochanger to handle multiple devices.
621 - Add SuSE install doc to list.
622 - Check and rechedk "Invalid block number"
623 - User wants to NOT backup up certain big files (email files).
624 - Maybe remove multiple simultaneous devices code in SD.
625 - On Windows with very long path names, it may be impossible to create
626 a file (and thus restore it) because the total length is too long.
627 We must cd into the directory then create the file without the
629 - Implement a Recycle command
630 - Add client name to cram-md5 challenge so Director can immediately
631 verify if it is the correct client.
632 - Audit all UA commands to ensure that we always prompt where possible.
633 - Make bootstrap filename unique.
634 - Test a second language e.g. french.
635 - Start working on Base jobs.
636 - Implement UnsavedFiles DB record.
637 - Implement argc/argv for daemon command line scanning using table driven
639 - From Phil Stracchino:
640 It would probably be a per-client option, and would be called
641 something like, say, "Automatically purge obsoleted jobs". What it
642 would do is, when you successfully complete a Differential backup of a
643 client, it would automatically purge all Incremental backups for that
644 client that are rendered redundant by that Differential. Likewise,
645 when a Full backup on a client completed, it would automatically purge
646 all Differential and Incremental jobs obsoleted by that Full backup.
647 This would let people minimize the number of tapes they're keeping on
648 hand without having to master the art of retention times.
649 - Implement a M_SECURITY message class.
650 - When doing a Backup send all attributes back to the Director, who
651 would then figure out what files have been deleted.
652 - Currently in mount.c:236 the SD simply creates a Volume. It should have
653 explicit permission to do so. It should also mark the tape in error
654 if there is an error.
655 - SD Bytes Read is wrong.
656 - Investigate doing RAW backup of Win32 partition.
657 - Cancel waiting for Client connect in SD if FD goes away.
659 - Implement timeout in response() when it should come quickly.
660 - Implement console @echo command.
661 - Implement a Slot priority (loaded/not loaded).
662 - Implement "vacation" Incremental only saves.
663 - Implement single pane restore (much like the Gftp panes).
664 - Implement Automatic Mount even in operator wait.
665 - Implement create "FileSet"?
666 - Implement restore "current system", but take all files without
667 doing selection tree -- so that jobs without File records can
669 - Add prefixlinks to where or not where absolute links to FD.
670 - Issue message to mount a new tape before the rewind.
671 - Simplified client job initiation for portables.
672 - If SD cannot open a drive, make it periodically retry.
673 - Add more of the config info to the tape label.
675 - If tape is marked read-only, then try opening it read-only rather than
676 failing, and remember that it cannot be written.
677 - Refine SD waiting output:
678 Device is being positioned
679 > Device is being positioned for append
680 > Device is being positioned to file x
682 - Figure out some way to estimate output size and to avoid splitting
683 a backup across two Volumes -- this could be useful for writing CDROMs
684 where you really prefer not to have it split -- not serious.
685 - Have SD compute MD5 or SHA1 and compare to what FD computes.
686 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
687 actual data on the Volume and compare it.
688 - Implement Bacula plugins -- design API
689 - Make bcopy read through bad tape records.
690 - Program files (i.e. execute a program to read/write files).
691 Pass read date of last backup, size of file last time.
692 - Add Signature type to File DB record.
693 - CD into subdirectory when open()ing files for backup to
694 speed up things. Test with testfind().
695 - Priority job to go to top of list.
696 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
697 - Implement some way for the Console to dynamically create a job.
698 - Restore to a particular time -- e.g. before date, after date.
699 - Solaris -I on tar for include list
700 - Need a verbose mode in restore, perhaps to bsr.
701 - bscan without -v is too quiet -- perhaps show jobs.
702 - Add code to reject whole blocks if not wanted on restore.
703 - Check if we can increase Bacula FD priorty in Win2000
704 - Make sure the MaxVolFiles is fully implemented in SD
705 - Check if both CatalogFiles and UseCatalog are set to SD.
706 - Figure out how to do a bare metal Windows restore
707 - Possibly add email to Watchdog if drive is unmounted too
708 long and a job is waiting on the drive.
709 - After unmount, if restore job started, ask to mount.
710 - Convert all %x substitution variables, which are hard to remember
711 and read to %(variable-name). Idea from TMDA.
712 - Remove NextId for SQLite. Optimize.
713 - Move all SQL statements into a single location.
714 - Add UA rc and history files.
715 - put termcap (used by console) in ./configure and
716 allow -with-termcap-dir.
717 - Fix Autoprune for Volumes to respect need for full save.
718 - Fix Win32 config file definition name on /install
719 - Compare tape to Client files (attributes, or attributes and data)
720 - Make all database Ids 64 bit.
721 - Write an applet for Linux.
722 - Allow console commands to detach or run in background.
723 - Fix status delay on storage daemon during rewind.
724 - Add SD message variables to control operator wait time
725 - Maximum Operator Wait
726 - Minimum Message Interval
727 - Maximum Message Interval
728 - Send Operator message when cannot read tape label.
729 - Verify level=Volume (scan only), level=Data (compare of data to file).
730 Verify level=Catalog, level=InitCatalog
732 - Add keyword search to show command in Console.
733 - Events : tape has more than xxx bytes.
734 - Complete code in Bacula Resources -- this will permit
735 reading a new config file at any time.
736 - Handle ctl-c in Console
737 - Implement script driven addition of File daemon to config files.
738 - Think about how to make Bacula work better with File (non-tape) archives.
739 - Write Unix emulator for Windows.
740 - Put memory utilization in Status output of each daemon
741 if full status requested or if some level of debug on.
742 - Make database type selectable by .conf files i.e. at runtime
743 - Set flag for uname -a. Add to Volume label.
744 - Restore files modified after date
745 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
746 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
747 - Timout a job or terminate if link goes down, or reopen link and query.
748 - Concept of precious tapes (cannot be reused).
749 - Make bcopy copy with a single tape drive.
750 - Permit changing ownership during restore.
753 > My suggestion: Add a feature on the systray menu-icon menu to request
754 > an immediate backup now. This would be useful for laptop users who may
755 > not be on the network when the regular scheduled backup is run.
757 > My wife's suggestion: Add a setting to the win32 client to allow it to
758 > shut down the machine after backup is complete (after, of course,
759 > displaying a "System will shut down in one minute, click here to cancel"
760 > warning dialog). This would be useful for sites that want user
761 > woorkstations to be shut down overnight to save power.
764 - Autolabel should be specified by DIR instead of SD.
767 - AutoScan (check checksum of tape)
768 - Format command = "format /dev/nst0"
772 - Seek resolution (usually corresponds to buffer size)
773 - EODErrorCode=ENOSPC or code
774 - Partial Read error code
775 - Partial write error code
776 - Nonformatted read error
777 - Nonformatted write error
778 - WriteProtected error
782 - IgnoreCloseErrors=yes
792 - FD sends unsaved file list to Director at end of job (see
794 - File daemon should build list of files skipped, and then
795 at end of save retry and report any errors.
796 - Write a Storage daemon that uses pipes and
797 standard Unix programs to write to the tape.
799 - Need something that monitors the JCR queue and
800 times out jobs by asking the deamons where they are.
801 - Enhance Jmsg code to permit buffering and saving to disk.
802 - device driver = "xxxx" for drives.
804 - Ensure that /dev/null works
805 - Need report class for messages. Perhaps
806 report resource where report=group of messages
807 - enhance scan_attrib and rename scan_jobtype, and
808 fill in code for "since" option
809 - Director needs a time after which the report status is sent
810 anyway -- or better yet, a retry time for the job.
811 - Don't reschedule a job if previous incarnation is still running.
812 - Some way to automatically backup everything is needed????
813 - Need a structure for pending actions:
815 - termination status (part of buffered msgs?)
817 Read, Write, Clean, Delete
818 - Login to Bacula; Bacula users with different permissions:
819 owner, group, user, quotas
820 - Store info on each file system type (probably in the job header on tape.
821 This could be the output of df; or perhaps some sort of /etc/mtab record.
823 ========= ideas ===============
824 From: "Jerry K. Schieffer" <jerry@skylinetechnology.com>
825 To: <kern@sibbald.com>
826 Subject: RE: [Bacula-users] future large programming jobs
827 Date: Thu, 26 Feb 2004 11:34:54 -0600
829 I noticed the subject thread and thought I would offer the following
830 merely as sources of ideas, i.e. something to think about, not even as
831 strong as a request. In my former life (before retiring) I often
832 dealt with backups and storage management issues/products as a
833 developer and as a consultant. I am currently migrating my personal
834 network from amanda to bacula specifically because of the ability to
835 cross media boundaries during storing backups.
836 Are you familiar with the commercial product called ADSM (I think IBM
837 now sells it under the Tivoli label)? It has a couple of interesting
838 ideas that may apply to the following topics.
840 1. Migration: Consider that when you need to restore a system, there
841 may be pressure to hurry. If all the information for a single client
842 can eventually end up on the same media (and in chronological order),
843 the restore is facillitated by not having to search past information
844 from other clients. ADSM has the concept of "client affinity" that
845 may be associated with it's storage pools. It seems to me that this
846 concept (as an optional feature) might fit in your architecture for
849 ADSM also has the concept of defining one or more storage pools as
850 "copy pools" (almost mirrors, but only in the sense of contents).
851 These pools provide the ability to have duplicte data stored both
852 onsite and offsite. The copy process can be scheduled to be handled
853 by their storage manager during periods when there is no backup
854 activity. Again, the migration process might be a place to consider
855 implementing something like this.
858 > It strikes me that it would be very nice to be able to do things
860 > have the Job(s) backing up the machines run, and once they have all
861 > completed, start a migration job to copy the data from disks Volumes
863 > a tape library and then to offsite storage. Maybe this can already
865 > done with some careful scheduling and Job prioritzation; the events
866 > mechanism described below would probably make it very easy.
868 This is the goal. In the first step (before events), you simply
870 the Migration to tape later.
872 2. Base jobs: In ADSM, each copy of each stored file is tracked in
873 the database. Once a file (unique by path and metadata such as dates,
874 size, ownership, etc.) is in a copy pool, no more copies are made. In
875 other words, when you start ADSM, it begins like your concept of a
876 base job. After that it is in the "incremental" mode. You can
877 configure the number of "generations" of files to be retained, plus a
878 retention date after which even old generations are purged. The
879 database tracks the contents of media and projects the percentage of
880 each volume that is valid. When the valid content of a volume drops
881 below a configured percentage, the valid data are migrated to another
882 volume and the old volume is marked as empty. Note, this requires
883 ADSM to have an idea of the contents of a client, i.e. marking the
884 database when an existing file was deleted, but this would solve your
885 issue of restoring a client without restoring deleted files.
887 This is pretty far from what bacula now does, but if you are going to
888 rip things up for Base jobs,.....
889 Also, the benefits of this are huge for very large shops, especially
890 with media robots, but are a pain for shops with manual media
894 > Base jobs sound pretty useful, but I'm not dying for them.
896 Nobody is dying for them, but when you see what it does, you will die
899 3. Restoring deleted files: Since I think my comments in (2) above
900 have low probability of implementation, I'll also suggest that you
901 could approach the issue of deleted files by a mechanism of having the
902 fd report to the dir, a list of all files on the client for every
903 backup job. The dir could note in the database entry for each file
904 the date that the file was seen. Then if a restore as of date X takes
905 place, only files that exist from before X until after X would be
906 restored. Probably the major cost here is the extra date container in
907 each row of the files table.
909 Thanks for "listening". I hope some of this helps. If you want to
910 contact me, please send me an email - I read some but not all of the
911 mailing list traffic and might miss a reply there.
913 Please accept my compliments for bacula. It is doing a great job for
914 me!! I sympathize with you in the need to wrestle with excelence in
915 execution vs. excelence in feature inclusion.
920 ==============================
923 - Design at hierarchial storage for Bacula. Migration and Clone.
924 - Implement FSM (File System Modules).
925 - Audit M_ error codes to ensure they are correct and consistent.
926 - Add variable break characters to lex analyzer.
927 Either a bit mask or a string of chars so that
928 the caller can change the break characters.
929 - Make a single T_BREAK to replace T_COMMA, etc.
930 - Ensure that File daemon and Storage daemon can
931 continue a save if the Director goes down (this
932 is NOT currently the case). Must detect socket error,
933 buffer messages for later.
934 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
935 - Add ability to backup to two Storage devices (two SD sessions) at
936 the same time -- e.g. onsite, offsite.
937 - Add the ability to consolidate old backup sets (basically do a restore
938 to tape and appropriately update the catalog). Compress Volume sets.
939 Might need to spool via file is only one drive is available.
940 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
941 someway to do so with every volume that has less than x% valid
945 Migration: Move a backup from one Volume to another
946 Clone: Copy a backup -- two Volumes
948 Bacula Migration is based on Jobs (apparently Networker is file by file).
950 Migration triggered by:
954 Highwater mark (keep total size)
959 ======================================================
961 It is somewhat like a Full save becomes an incremental since
962 the Base job (or jobs) plus other non-base files.
964 - A Base backup is same as Full backup, just different type.
965 - New BaseFiles table that contains:
967 BaseJobId - Base JobId referenced for this FileId (needed ???)
968 JobId - JobId currently running
969 FileId - File not backed up, exists in Base Job
970 FileIndex - FileIndex from Base Job.
971 i.e. for each base file that exists but is not saved because
972 it has not changed, the File daemon sends the JobId, BaseId,
973 FileId, FileIndex back to the Director who creates the DB entry.
974 - To initiate a Base save, the Director sends the FD
975 the FileId, and full filename for each file in the Base.
976 - When the FD finds a Base file, he requests the Director to
977 send him the full File entry (stat packet plus MD5), or
978 conversely, the FD sends it to the Director and the Director
979 says yes or no. This can be quite rapid if the FileId is kept
980 by the FD for each Base Filename.
981 - It is probably better to have the comparison done by the FD
982 despite the fact that the File entry must be sent across the
984 - An alternative would be to send the FD the whole File entry
985 from the start. The disadvantage is that it requires a lot of
986 space. The advantage is that it requires less communications
988 - The Job record must be updated to indicate that one or more
990 - At end of Job, FD returns:
991 1. Count of base files/bytes not written to tape (i.e. matches)
992 2. Count of base file that were saved i.e. had changed.
993 - No tape record would be written for a Base file that matches, in the
994 same way that no tape record is written for Incremental jobs where
995 the file is not saved because it is unchanged.
996 - On a restore, all the Base file records must explicitly be
997 found from the BaseFile tape. I.e. for each Full save that is marked
998 to have one or more Base Jobs, search the BaseFile for all occurrences
1000 - An optimization might be to make the BaseFile have:
1006 This would avoid the need to explicitly fetch each File record for
1007 the Base job. The Base Job record will be fetched to get the
1008 VolSessionId and VolSessionTime.
1009 =========================================================
1012 ==========================================================
1014 For each Incremental job that is run, there may be files that
1015 were found but not saved because they were locked (this applies
1016 only to Windows). Such a system could send back to the Director
1017 a list of Unsaved files.
1019 - New UnSavedFiles table that contains:
1023 - Then in the next Incremental job, the list of Unsaved Files will be
1024 feed to the FD, who will ensure that they are explicitly chosen even
1025 if standard date/time check would not have selected them.
1026 =============================================================
1029 Done: (see kernsdone for more)
1032 - John's Full save failed with 1.32c FD and 1.31 Dir no FD status,
1033 and no error message.
1034 - Add fd and st as Console keywords.
1035 - Recycling volume with a Slot requires an operator intervention:
1036 rufus-dir: Start Backup JobId 18, Job=kernsave.2003-11-01_21.23.52
1037 rufus-dir: Pruned 1 Job on Volume Vol01 from catalog.
1038 rufus-dir: There are no Jobs associated with Volume Vol01. Marking it purged.
1039 rufus-dir: Recycled volume "Vol01"
1040 rufus-sd: Please mount Volume "Vol01" on Storage Device "DDS-4" for Job kernsave.2003-11-01_21.23.52
1041 Use "mount" command to release Job.
1042 - Implement Dan's bacula script (email of 26 Oct).
1043 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
1044 - Edit the Client/Storage name into authentication failure messages.
1045 - Fix packet too big problem. This is most likely a Windows TCP stack
1047 - Implement ClientRunBeforeJob and ClientRunAfterJob.
1048 - Implement forward spacing block/file: position_device(bsr) --
1049 just before read_block_from_device();
1052 Multiple drive autochanger data: see Alan Brown
1053 mtx -f xxx unloadStorage Element 1 is Already Full(drive 0 was empty)
1054 Unloading Data Transfer Element into Storage Element 1...source Element
1055 Address 480 is Empty
1057 (drive 0 was empty and so was slot 1)
1058 > mtx -f xxx load 15 0
1059 no response, just returns to the command prompt when complete.
1060 > mtx -f xxx status Storage Changer /dev/changer:2 Drives, 60 Slots ( 2 Import/Export )
1061 Data Transfer Element 0:Full (Storage Element 15 Loaded):VolumeTag = HX001
1062 Data Transfer Element 1:Empty
1063 Storage Element 1:Empty
1064 Storage Element 2:Full :VolumeTag=HX002
1065 Storage Element 3:Full :VolumeTag=HX003
1066 Storage Element 4:Full :VolumeTag=HX004
1067 Storage Element 5:Full :VolumeTag=HX005
1068 Storage Element 6:Full :VolumeTag=HX006
1069 Storage Element 7:Full :VolumeTag=HX007
1070 Storage Element 8:Full :VolumeTag=HX008
1071 Storage Element 9:Full :VolumeTag=HX009
1072 Storage Element 10:Full :VolumeTag=HX010
1073 Storage Element 11:Empty
1074 Storage Element 12:Empty
1075 Storage Element 13:Empty
1076 Storage Element 14:Empty
1077 Storage Element 15:Empty
1078 Storage Element 16:Empty....
1079 Storage Element 28:Empty
1080 Storage Element 29:Full :VolumeTag=CLNU01L1
1081 Storage Element 30:Empty....
1082 Storage Element 57:Empty
1083 Storage Element 58:Full :VolumeTag=NEX261L2
1084 Storage Element 59 IMPORT/EXPORT:Empty
1085 Storage Element 60 IMPORT/EXPORT:Empty
1087 Unloading Data Transfer Element into Storage Element 15...done
1089 (just to verify it remembers where it came from, however it can be
1090 overrriden with mtx unload {slotnumber} to go to any storage slot.)
1092 There needs to be a table of drive # to devices somewhere - If there are
1093 multiple changers or drives there may not be a 1:1 correspondance between
1094 changer drive number and system device name - and depending on the way the
1095 drives are hooked up to scsi busses, they may not be linearly numbered
1096 from an offset point either.something like
1098 Autochanger drives = 2
1099 Autochanger drive 0 = /dev/nst1
1100 Autochanger drive 1 = /dev/nst2
1101 IMHO, it would be _safest_ to use explicit mtx unload commands at all
1102 times, not just for multidrive changers. For a 1 drive changer, that's
1108 MTX's manpage (1.2.15):
1109 unload [<slotnum>] [ <drivenum> ]
1110 Unloads media from drive <drivenum> into slot
1111 <slotnum>. If <drivenum> is omitted, defaults to
1112 drive 0 (as do all commands). If <slotnum> is
1113 omitted, defaults to the slot that the drive was
1114 loaded from. Note that there's currently no way
1115 to say 'unload drive 1's media to the slot it
1116 came from', other than to explicitly use that
1117 slot number as the destination.AB
1123 undef# camcontrol devlist
1124 <WANGTEK 51000 SCSI M74H 12B3> at scbus0 target 2 lun 0 (pass0,sa0)
1125 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 0 (pass1,sa1)
1126 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 1 (pass2)
1128 tapeinfo -f /dev/sg0 with a bad tape in drive 1:
1129 [kern@rufus mtx-1.2.17kes]$ ./tapeinfo -f /dev/sg0
1130 Product Type: Tape Drive
1132 Product ID: 'C5713A '
1134 Attached Changer: No
1135 TapeAlert[3]: Hard Error: Uncorrectable read/write error.
1136 TapeAlert[20]: Clean Now: The tape drive neads cleaning NOW.
1143 Medium Type: Not Loaded
1146 DataCompEnabled: yes
1147 DataCompCapable: yes
1148 DataDeCompEnabled: yes
1155 Handling removable disks
1157 From: Karl Cunningham <karlc@keckec.com>
1159 My backups are only to hard disk these days, in removable bays. This is my
1160 idea of how a backup to hard disk would work more smoothly. Some of these
1161 things Bacula does already, but I mention them for completeness. If others
1162 have better ways to do this, I'd like to hear about it.
1164 1. Accommodate several disks, rotated similar to how tapes are. Identified
1165 by partition volume ID or perhaps by the name of a subdirectory.
1166 2. Abort & notify the admin if the wrong disk is in the bay.
1167 3. Write backups to different subdirectories for each machine to be backed
1169 4. Volumes (files) get created as needed in the proper subdirectory, one
1171 5. When a disk is recycled, remove or zero all old backup files. This is
1172 important as the disk being recycled may be close to full. This may be
1173 better done manually since the backup files for many machines may be
1174 scattered in many subdirectories.
1179 - Save mount point for directories not traversed with onefs=yes.
1180 - Add seconds to start and end times in the Job report output.
1181 - if 2 concurrent backups are attempted on the same tape
1182 drive (autoloader) into different tape pools, one of them will exit
1183 fatally instead of halting until the drive is idle
1184 - Update StartTime if job held in Job Queue.
1185 - Look at www.nu2.nu/pebuilder as a helper for full windows
1186 bare metal restore. (done by Scott)
1187 - Fix orphanned buffers:
1188 Orphaned buffer: 24 bytes allocated at line 808 of rufus-dir job.c
1189 Orphaned buffer: 40 bytes allocated at line 45 of rufus-dir alist.c
1190 - Implement Preben's suggestion to add
1191 File System Types = ext2, ext3
1192 to FileSets, thus simplifying backup of *all* local partitions.