3 Bacula Projects Roadmap
4 Status updated 14 April 2007
5 After re-ordering in vote priority
8 Item: 18 Quick release of FD-SD connection after backup.
9 Item: 40 Include JobID in spool file name
10 Item: 25 Implement huge exclude list support using dlist
11 Item: 41 Enable to relocate files and directories when restoring
14 Item: 1 Accurate restoration of renamed/deleted files
15 Item: 2 Implement a Bacula GUI/management tool.
16 Item: 3 Allow FD to initiate a backup
17 Item: 4 Merge multiple backups (Synthetic Backup or Consolidation).
18 Item: 5 Deletion of Disk-Based Bacula Volumes
19 Item: 6 Implement Base jobs.
20 Item: 7 Implement creation and maintenance of copy pools
21 Item: 8 Directive/mode to backup only file changes, not entire file
22 Item: 9 Implement a server-side compression feature
23 Item: 10 Improve Bacula's tape and drive usage and cleaning management.
24 Item: 11 Allow skipping execution of Jobs
25 Item: 12 Add a scheduling syntax that permits weekly rotations
26 Item: 13 Archival (removal) of User Files to Tape
27 Item: 14 Cause daemons to use a specific IP address to source communications
28 Item: 15 Multiple threads in file daemon for the same job
29 Item: 16 Add Plug-ins to the FileSet Include statements.
30 Item: 17 Restore only file attributes (permissions, ACL, owner, group...)
31 Item: 18* Quick release of FD-SD connection after backup.
32 Item: 19 Implement a Python interface to the Bacula catalog.
34 Item: 21 Split documentation
35 Item: 22 Implement support for stacking arbitrary stream filters, sinks.
36 Item: 23 Implement from-client and to-client on restore command line.
37 Item: 24 Add an override in Schedule for Pools based on backup types.
38 Item: 25* Implement huge exclude list support using hashing.
39 Item: 26 Implement more Python events in Bacula.
40 Item: 27 Incorporation of XACML2/SAML2 parsing
41 Item: 28 Filesystem watch triggered backup.
42 Item: 29 Allow inclusion/exclusion of files in a fileset by creation/mod times
43 Item: 30 Tray monitor window cleanups
44 Item: 31 Implement multiple numeric backup levels as supported by dump
45 Item: 32 Automatic promotion of backup levels
46 Item: 33 Clustered file-daemons
47 Item: 34 Commercial database support
48 Item: 35 Automatic disabling of devices
49 Item: 36 An option to operate on all pools with update vol parameters
50 Item: 37 Add an item to the restore option where you can select a pool
51 Item: 38 Include timestamp of job launch in "stat clients" output
52 Item: 39 Message mailing based on backup types
53 Item: 40* Include JobID in spool file name
54 Item: 41* Enable to relocate files and directories when restoring
56 Item 1: Accurate restoration of renamed/deleted files
57 Date: 28 November 2005
58 Origin: Martin Simmons (martin at lispworks dot com)
59 Status: Robert Nelson will implement this
61 What: When restoring a fileset for a specified date (including "most
62 recent"), Bacula should give you exactly the files and directories
63 that existed at the time of the last backup prior to that date.
65 Currently this only works if the last backup was a Full backup.
66 When the last backup was Incremental/Differential, files and
67 directories that have been renamed or deleted since the last Full
68 backup are not currently restored correctly. Ditto for files with
69 extra/fewer hard links than at the time of the last Full backup.
71 Why: Incremental/Differential would be much more useful if this worked.
73 Notes: Merging of multiple backups into a single one seems to
74 rely on this working, otherwise the merged backups will not be
75 truly equivalent to a Full backup.
77 Kern: notes shortened. This can be done without the need for
78 inodes. It is essentially the same as the current Verify job,
79 but one additional database record must be written, which does
80 not need any database change.
82 Kern: see if we can correct restoration of directories if
83 replace=ifnewer is set. Currently, if the directory does not
84 exist, a "dummy" directory is created, then when all the files
85 are updated, the dummy directory is newer so the real values
88 Item 2: Implement a Bacula GUI/management tool.
93 What: Implement a Bacula console, and management tools
94 probably using Qt3 and C++.
96 Why: Don't we already have a wxWidgets GUI? Yes, but
97 it is written in C++ and changes to the user interface
98 must be hand tailored using C++ code. By developing
99 the user interface using Qt designer, the interface
100 can be very easily updated and most of the new Python
101 code will be automatically created. The user interface
102 changes become very simple, and only the new features
103 must be implement. In addition, the code will be in
104 Python, which will give many more users easy (or easier)
105 access to making additions or modifications.
107 Notes: There is a partial Python-GTK implementation
108 Lucas Di Pentima <lucas at lunix dot com dot ar> but
109 it is no longer being developed.
111 Item 3: Allow FD to initiate a backup
112 Origin: Frank Volf (frank at deze dot org)
113 Date: 17 November 2005
116 What: Provide some means, possibly by a restricted console that
117 allows a FD to initiate a backup, and that uses the connection
118 established by the FD to the Director for the backup so that
119 a Director that is firewalled can do the backup.
121 Why: Makes backup of laptops much easier.
124 Item 4: Merge multiple backups (Synthetic Backup or Consolidation).
125 Origin: Marc Cousin and Eric Bollengier
126 Date: 15 November 2005
127 Status: Waiting implementation. Depends on first implementing
128 project Item 2 (Migration) which is now done.
130 What: A merged backup is a backup made without connecting to the Client.
131 It would be a Merge of existing backups into a single backup.
132 In effect, it is like a restore but to the backup medium.
134 For instance, say that last Sunday we made a full backup. Then
135 all week long, we created incremental backups, in order to do
136 them fast. Now comes Sunday again, and we need another full.
137 The merged backup makes it possible to do instead an incremental
138 backup (during the night for instance), and then create a merged
139 backup during the day, by using the full and incrementals from
140 the week. The merged backup will be exactly like a full made
141 Sunday night on the tape, but the production interruption on the
142 Client will be minimal, as the Client will only have to send
145 In fact, if it's done correctly, you could merge all the
146 Incrementals into single Incremental, or all the Incrementals
147 and the last Differential into a new Differential, or the Full,
148 last differential and all the Incrementals into a new Full
149 backup. And there is no need to involve the Client.
151 Why: The benefit is that :
152 - the Client just does an incremental ;
153 - the merged backup on tape is just as a single full backup,
154 and can be restored very fast.
156 This is also a way of reducing the backup data since the old
157 data can then be pruned (or not) from the catalog, possibly
158 allowing older volumes to be recycled
160 Item 5: Deletion of Disk-Based Bacula Volumes
162 Origin: Ross Boylan <RossBoylan at stanfordalumni dot org> (edited
166 What: Provide a way for Bacula to automatically remove Volumes
167 from the filesystem, or optionally to truncate them.
168 Obviously, the Volume must be pruned prior removal.
170 Why: This would allow users more control over their Volumes and
171 prevent disk based volumes from consuming too much space.
173 Notes: The following two directives might do the trick:
175 Volume Data Retention = <time period>
176 Remove Volume After = <time period>
178 The migration project should also remove a Volume that is
179 migrated. This might also work for tape Volumes.
181 Item 6: Implement Base jobs.
182 Date: 28 October 2005
186 What: A base job is sort of like a Full save except that you
187 will want the FileSet to contain only files that are
188 unlikely to change in the future (i.e. a snapshot of
189 most of your system after installing it). After the
190 base job has been run, when you are doing a Full save,
191 you specify one or more Base jobs to be used. All
192 files that have been backed up in the Base job/jobs but
193 not modified will then be excluded from the backup.
194 During a restore, the Base jobs will be automatically
195 pulled in where necessary.
197 Why: This is something none of the competition does, as far as
198 we know (except perhaps BackupPC, which is a Perl program that
199 saves to disk only). It is big win for the user, it
200 makes Bacula stand out as offering a unique
201 optimization that immediately saves time and money.
202 Basically, imagine that you have 100 nearly identical
203 Windows or Linux machine containing the OS and user
204 files. Now for the OS part, a Base job will be backed
205 up once, and rather than making 100 copies of the OS,
206 there will be only one. If one or more of the systems
207 have some files updated, no problem, they will be
208 automatically restored.
210 Notes: Huge savings in tape usage even for a single machine.
211 Will require more resources because the DIR must send
212 FD a list of files/attribs, and the FD must search the
213 list and compare it for each file to be saved.
215 Item 7: Implement creation and maintenance of copy pools
216 Date: 27 November 2005
217 Origin: David Boyes (dboyes at sinenomine dot net)
220 What: I would like Bacula to have the capability to write copies
221 of backed-up data on multiple physical volumes selected
222 from different pools without transferring the data
223 multiple times, and to accept any of the copy volumes
224 as valid for restore.
226 Why: In many cases, businesses are required to keep offsite
227 copies of backup volumes, or just wish for simple
228 protection against a human operator dropping a storage
229 volume and damaging it. The ability to generate multiple
230 volumes in the course of a single backup job allows
231 customers to simple check out one copy and send it
232 offsite, marking it as out of changer or otherwise
233 unavailable. Currently, the library and magazine
234 management capability in Bacula does not make this process
237 Restores would use the copy of the data on the first
238 available volume, in order of copy pool chain definition.
240 This is also a major scalability issue -- as the number of
241 clients increases beyond several thousand, and the volume
242 of data increases, transferring the data multiple times to
243 produce additional copies of the backups will become
244 physically impossible due to transfer speed
245 issues. Generating multiple copies at server side will
246 become the only practical option.
248 How: I suspect that this will require adding a multiplexing
249 SD that appears to be a SD to a specific FD, but 1-n FDs
250 to the specific back end SDs managing the primary and copy
251 pools. Storage pools will also need to acquire parameters
252 to define the pools to be used for copies.
254 Notes: I would commit some of my developers' time if we can agree
255 on the design and behavior.
257 Item 8: Directive/mode to backup only file changes, not entire file
258 Date: 11 November 2005
259 Origin: Joshua Kugler <joshua dot kugler at uaf dot edu>
260 Marek Bajon <mbajon at bimsplus dot com dot pl>
263 What: Currently when a file changes, the entire file will be backed up in
264 the next incremental or full backup. To save space on the tapes
265 it would be nice to have a mode whereby only the changes to the
266 file would be backed up when it is changed.
268 Why: This would save lots of space when backing up large files such as
269 logs, mbox files, Outlook PST files and the like.
271 Notes: This would require the usage of disk-based volumes as comparing
272 files would not be feasible using a tape drive.
274 Item 9: Implement a server-side compression feature
275 Date: 18 December 2006
276 Origin: Vadim A. Umanski , e-mail umanski@ext.ru
278 What: The ability to compress backup data on server receiving data
279 instead of doing that on client sending data.
280 Why: The need is practical. I've got some machines that can send
281 data to the network 4 or 5 times faster than compressing
282 them (I've measured that). They're using fast enough SCSI/FC
283 disk subsystems but rather slow CPUs (ex. UltraSPARC II).
284 And the backup server has got a quite fast CPUs (ex. Dual P4
285 Xeons) and quite a low load. When you have 20, 50 or 100 GB
286 of raw data - running a job 4 to 5 times faster - that
287 really matters. On the other hand, the data can be
288 compressed 50% or better - so losing twice more space for
289 disk backup is not good at all. And the network is all mine
290 (I have a dedicated management/provisioning network) and I
291 can get as high bandwidth as I need - 100Mbps, 1000Mbps...
292 That's why the server-side compression feature is needed!
295 Item 10: Improve Bacula's tape and drive usage and cleaning management.
296 Date: 8 November 2005, November 11, 2005
297 Origin: Adam Thornton <athornton at sinenomine dot net>,
298 Arno Lehmann <al at its-lehmann dot de>
301 What: Make Bacula manage tape life cycle information, tape reuse
302 times and drive cleaning cycles.
304 Why: All three parts of this project are important when operating
306 We need to know which tapes need replacement, and we need to
307 make sure the drives are cleaned when necessary. While many
308 tape libraries and even autoloaders can handle all this
309 automatically, support by Bacula can be helpful for smaller
310 (older) libraries and single drives. Limiting the number of
311 times a tape is used might prevent tape errors when using
312 tapes until the drives can't read it any more. Also, checking
313 drive status during operation can prevent some failures (as I
314 [Arno] had to learn the hard way...)
316 Notes: First, Bacula could (and even does, to some limited extent)
317 record tape and drive usage. For tapes, the number of mounts,
318 the amount of data, and the time the tape has actually been
319 running could be recorded. Data fields for Read and Write
320 time and Number of mounts already exist in the catalog (I'm
321 not sure if VolBytes is the sum of all bytes ever written to
322 that volume by Bacula). This information can be important
323 when determining which media to replace. The ability to mark
324 Volumes as "used up" after a given number of write cycles
325 should also be implemented so that a tape is never actually
326 worn out. For the tape drives known to Bacula, similar
327 information is interesting to determine the device status and
328 expected life time: Time it's been Reading and Writing, number
329 of tape Loads / Unloads / Errors. This information is not yet
330 recorded as far as I [Arno] know. A new volume status would
331 be necessary for the new state, like "Used up" or "Worn out".
332 Volumes with this state could be used for restores, but not
333 for writing. These volumes should be migrated first (assuming
334 migration is implemented) and, once they are no longer needed,
335 could be moved to a Trash pool.
337 The next step would be to implement a drive cleaning setup.
338 Bacula already has knowledge about cleaning tapes. Once it
339 has some information about cleaning cycles (measured in drive
340 run time, number of tapes used, or calender days, for example)
341 it can automatically execute tape cleaning (with an
342 autochanger, obviously) or ask for operator assistance loading
345 The final step would be to implement TAPEALERT checks not only
346 when changing tapes and only sending the information to the
347 administrator, but rather checking after each tape error,
348 checking on a regular basis (for example after each tape
349 file), and also before unloading and after loading a new tape.
350 Then, depending on the drives TAPEALERT state and the known
351 drive cleaning state Bacula could automatically schedule later
352 cleaning, clean immediately, or inform the operator.
354 Implementing this would perhaps require another catalog change
355 and perhaps major changes in SD code and the DIR-SD protocol,
356 so I'd only consider this worth implementing if it would
357 actually be used or even needed by many people.
359 Implementation of these projects could happen in three distinct
360 sub-projects: Measuring Tape and Drive usage, retiring
361 volumes, and handling drive cleaning and TAPEALERTs.
363 Item 11: Allow skipping execution of Jobs
364 Date: 29 November 2005
365 Origin: Florian Schnabel <florian.schnabel at docufy dot de>
368 What: An easy option to skip a certain job on a certain date.
369 Why: You could then easily skip tape backups on holidays. Especially
370 if you got no autochanger and can only fit one backup on a tape
371 that would be really handy, other jobs could proceed normally
372 and you won't get errors that way.
374 Item 12: Add a scheduling syntax that permits weekly rotations
375 Date: 15 December 2006
376 Origin: Gregory Brauer (greg at wildbrain dot com)
379 What: Currently, Bacula only understands how to deal with weeks of the
380 month or weeks of the year in schedules. This makes it impossible
381 to do a true weekly rotation of tapes. There will always be a
382 discontinuity that will require disruptive manual intervention at
383 least monthly or yearly because week boundaries never align with
384 month or year boundaries.
386 A solution would be to add a new syntax that defines (at least)
387 a start timestamp, and repetition period.
389 Why: Rotated backups done at weekly intervals are useful, and Bacula
390 cannot currently do them without extensive hacking.
392 Notes: Here is an example syntax showing a 3-week rotation where full
393 Backups would be performed every week on Saturday, and an
394 incremental would be performed every week on Tuesday. Each
395 set of tapes could be removed from the loader for the following
396 two cycles before coming back and being reused on the third
397 week. Since the execution times are determined by intervals
398 from a given point in time, there will never be any issues with
399 having to adjust to any sort of arbitrary time boundary. In
400 the example provided, I even define the starting schedule
401 as crossing both a year and a month boundary, but the run times
402 would be based on the "Repeat" value and would therefore happen
407 Name = "Week 1 Rotation"
408 #Saturday. Would run Dec 30, Jan 20, Feb 10, etc.
412 Start = 2006-12-30 01:00
416 #Tuesday. Would run Jan 2, Jan 23, Feb 13, etc.
420 Start = 2007-01-02 01:00
427 Name = "Week 2 Rotation"
428 #Saturday. Would run Jan 6, Jan 27, Feb 17, etc.
432 Start = 2007-01-06 01:00
436 #Tuesday. Would run Jan 9, Jan 30, Feb 20, etc.
440 Start = 2007-01-09 01:00
447 Name = "Week 3 Rotation"
448 #Saturday. Would run Jan 13, Feb 3, Feb 24, etc.
452 Start = 2007-01-13 01:00
456 #Tuesday. Would run Jan 16, Feb 6, Feb 27, etc.
460 Start = 2007-01-16 01:00
466 Item 13: Archival (removal) of User Files to Tape
468 Origin: Ray Pengelly [ray at biomed dot queensu dot ca
471 What: The ability to archive data to storage based on certain parameters
472 such as age, size, or location. Once the data has been written to
473 storage and logged it is then pruned from the originating
474 filesystem. Note! We are talking about user's files and not
477 Why: This would allow fully automatic storage management which becomes
478 useful for large datastores. It would also allow for auto-staging
479 from one media type to another.
481 Example 1) Medical imaging needs to store large amounts of data.
482 They decide to keep data on their servers for 6 months and then put
483 it away for long term storage. The server then finds all files
484 older than 6 months writes them to tape. The files are then removed
487 Example 2) All data that hasn't been accessed in 2 months could be
488 moved from high-cost, fibre-channel disk storage to a low-cost
489 large-capacity SATA disk storage pool which doesn't have as quick of
490 access time. Then after another 6 months (or possibly as one
491 storage pool gets full) data is migrated to Tape.
493 Item 14: Cause daemons to use a specific IP address to source communications
494 Origin: Bill Moran <wmoran@collaborativefusion.com>
497 What: Cause Bacula daemons (dir, fd, sd) to always use the ip address
498 specified in the [DIR|DF|SD]Addr directive as the source IP
499 for initiating communication.
500 Why: On complex networks, as well as extremely secure networks, it's
501 not unusual to have multiple possible routes through the network.
502 Often, each of these routes is secured by different policies
503 (effectively, firewalls allow or deny different traffic depending
504 on the source address)
505 Unfortunately, it can sometimes be difficult or impossible to
506 represent this in a system routing table, as the result is
507 excessive subnetting that quickly exhausts available IP space.
508 The best available workaround is to provide multiple IPs to
509 a single machine that are all on the same subnet. In order
510 for this to work properly, applications must support the ability
511 to bind outgoing connections to a specified address, otherwise
512 the operating system will always choose the first IP that
513 matches the required route.
514 Notes: Many other programs support this. For example, the following
515 can be configured in BIND:
516 query-source address 10.0.0.1;
517 transfer-source 10.0.0.2;
518 Which means queries from this server will always come from
519 10.0.0.1 and zone transfers will always originate from
522 Item 15: Multiple threads in file daemon for the same job
523 Date: 27 November 2005
524 Origin: Ove Risberg (Ove.Risberg at octocode dot com)
527 What: I want the file daemon to start multiple threads for a backup
528 job so the fastest possible backup can be made.
530 The file daemon could parse the FileSet information and start
531 one thread for each File entry located on a separate
534 A confiuration option in the job section should be used to
535 enable or disable this feature. The confgutration option could
536 specify the maximum number of threads in the file daemon.
538 If the theads could spool the data to separate spool files
539 the restore process will not be much slower.
541 Why: Multiple concurrent backups of a large fileserver with many
542 disks and controllers will be much faster.
544 Item 16: Add Plug-ins to the FileSet Include statements.
545 Date: 28 October 2005
547 Status: Partially coded in 1.37 -- much more to do.
549 What: Allow users to specify wild-card and/or regular
550 expressions to be matched in both the Include and
551 Exclude directives in a FileSet. At the same time,
552 allow users to define plug-ins to be called (based on
553 regular expression/wild-card matching).
555 Why: This would give the users the ultimate ability to control
556 how files are backed up/restored. A user could write a
557 plug-in knows how to backup his Oracle database without
558 stopping/starting it, for example.
560 Item 17: Restore only file attributes (permissions, ACL, owner, group...)
561 Origin: Eric Bollengier
565 What: The goal of this project is to be able to restore only rights
566 and attributes of files without crushing them.
568 Why: Who have never had to repair a chmod -R 777, or a wild update
569 of recursive right under Windows? At this time, you must have
570 enough space to restore data, dump attributes (easy with acl,
571 more complex with unix/windows rights) and apply them to your
572 broken tree. With this options, it will be very easy to compare
573 right or ACL over the time.
575 Notes: If the file is here, we skip restore and we change rights.
576 If the file isn't here, we can create an empty one and apply
577 rights or do nothing.
579 Item 18: Quick release of FD-SD connection after backup.
580 Origin: Frank Volf (frank at deze dot org)
581 Date: 17 November 2005
582 Status: Done -- implemented by Kern -- in CVS 26Jan07
584 What: In the Bacula implementation a backup is finished after all data
585 and attributes are successfully written to storage. When using a
586 tape backup it is very annoying that a backup can take a day,
587 simply because the current tape (or whatever) is full and the
588 administrator has not put a new one in. During that time the
589 system cannot be taken off-line, because there is still an open
590 session between the storage daemon and the file daemon on the
593 Although this is a very good strategy for making "safe backups"
594 This can be annoying for e.g. laptops, that must remain
595 connected until the backup is completed.
597 Using a new feature called "migration" it will be possible to
598 spool first to harddisk (using a special 'spool' migration
599 scheme) and then migrate the backup to tape.
601 There is still the problem of getting the attributes committed.
602 If it takes a very long time to do, with the current code, the
603 job has not terminated, and the File daemon is not freed up. The
604 Storage daemon should release the File daemon as soon as all the
605 file data and all the attributes have been sent to it (the SD).
606 Currently the SD waits until everything is on tape and all the
607 attributes are transmitted to the Director before signaling
608 completion to the FD. I don't think I would have any problem
609 changing this. The reason is that even if the FD reports back to
610 the Dir that all is OK, the job will not terminate until the SD
611 has done the same thing -- so in a way keeping the SD-FD link
612 open to the very end is not really very productive ...
614 Why: Makes backup of laptops much faster.
616 Item 19: Implement a Python interface to the Bacula catalog.
617 Date: 28 October 2005
621 What: Implement an interface for Python scripts to access
622 the catalog through Bacula.
624 Why: This will permit users to customize Bacula through
627 Item 20: Archive data
629 Origin: calvin streeting calvin at absentdream dot com
632 What: The abilty to archive to media (dvd/cd) in a uncompressed format
633 for dead filing (archiving not backing up)
635 Why: At my works when jobs are finished and moved off of the main file
636 servers (raid based systems) onto a simple linux file server (ide based
637 system) so users can find old information without contacting the IT
640 So this data dosn't realy change it only gets added to,
641 But it also needs backing up. At the moment it takes
642 about 8 hours to back up our servers (working data) so
643 rather than add more time to existing backups i am trying
644 to implement a system where we backup the acrhive data to
645 cd/dvd these disks would only need to be appended to
646 (burn only new/changed files to new disks for off site
647 storage). basialy understand the differnce between
648 achive data and live data.
650 Notes: Scan the data and email me when it needs burning divide
651 into predifind chunks keep a recored of what is on what
652 disk make me a label (simple php->mysql=>pdf stuff) i
653 could do this bit ability to save data uncompresed so
654 it can be read in any other system (future proof data)
655 save the catalog with the disk as some kind of menu
658 Item 21: Split documentation
659 Origin: Maxx <maxxatworkat gmail dot com>
663 What: Split documentation in several books
665 Why: Bacula manual has now more than 600 pages, and looking for
666 implementation details is getting complicated. I think
667 it would be good to split the single volume in two or
670 1) Introduction, requirements and tutorial, typically
671 are useful only until first installation time
673 2) Basic installation and configuration, with all the
674 gory details about the directives supported 3)
675 Advanced Bacula: testing, troubleshooting, GUI and
676 ancillary programs, security managements, scripting,
680 Item 22: Implement support for stacking arbitrary stream filters, sinks.
681 Date: 23 November 2006
682 Origin: Landon Fuller <landonf@threerings.net>
683 Status: Planning. Assigned to landonf.
685 What: Implement support for the following:
686 - Stacking arbitrary stream filters (eg, encryption, compression,
687 sparse data handling))
688 - Attaching file sinks to terminate stream filters (ie, write out
689 the resultant data to a file)
690 - Refactor the restoration state machine accordingly
692 Why: The existing stream implementation suffers from the following:
693 - All state (compression, encryption, stream restoration), is
694 global across the entire restore process, for all streams. There are
695 multiple entry and exit points in the restoration state machine, and
696 thus multiple places where state must be allocated, deallocated,
697 initialized, or reinitialized. This results in exceptional complexity
698 for the author of a stream filter.
699 - The developer must enumerate all possible combinations of filters
700 and stream types (ie, win32 data with encryption, without encryption,
701 with encryption AND compression, etc).
703 Notes: This feature request only covers implementing the stream filters/
704 sinks, and refactoring the file daemon's restoration implementation
705 accordingly. If I have extra time, I will also rewrite the backup
706 implementation. My intent in implementing the restoration first is to
707 solve pressing bugs in the restoration handling, and to ensure that
708 the new restore implementation handles existing backups correctly.
710 I do not plan on changing the network or tape data structures to
711 support defining arbitrary stream filters, but supporting that
712 functionality is the ultimate goal.
714 Assistance with either code or testing would be fantastic.
716 Item 23: Implement from-client and to-client on restore command line.
717 Date: 11 December 2006
718 Origin: Discussion on Bacula-users entitled 'Scripted restores to
719 different clients', December 2006
720 Status: New feature request
722 What: While using bconsole interactively, you can specify the client
723 that a backup job is to be restored for, and then you can
724 specify later a different client to send the restored files
725 back to. However, using the 'restore' command with all options
726 on the command line, this cannot be done, due to the ambiguous
727 'client' parameter. Additionally, this parameter means different
728 things depending on if it's specified on the command line or
729 afterwards, in the Modify Job screens.
731 Why: This feature would enable restore jobs to be more completely
732 automated, for example by a web or GUI front-end.
734 Notes: client can also be implied by specifying the jobid on the command
737 Item 24: Add an override in Schedule for Pools based on backup types.
739 Origin: Chad Slater <chad.slater@clickfox.com>
742 What: Adding a FullStorage=BigTapeLibrary in the Schedule resource
743 would help those of us who use different storage devices for different
744 backup levels cope with the "auto-upgrade" of a backup.
746 Why: Assume I add several new device to be backed up, i.e. several
747 hosts with 1TB RAID. To avoid tape switching hassles, incrementals are
748 stored in a disk set on a 2TB RAID. If you add these devices in the
749 middle of the month, the incrementals are upgraded to "full" backups,
750 but they try to use the same storage device as requested in the
751 incremental job, filling up the RAID holding the differentials. If we
752 could override the Storage parameter for full and/or differential
753 backups, then the Full job would use the proper Storage device, which
754 has more capacity (i.e. a 8TB tape library.
756 Item 25: Implement huge exclude list support using hashing (dlists).
757 Date: 28 October 2005
759 Status: Done in 2.1.2 but was done with dlists (doubly linked lists
760 since hashing will not help. The huge list also supports
761 large include lists).
763 What: Allow users to specify very large exclude list (currently
764 more than about 1000 files is too many).
766 Why: This would give the users the ability to exclude all
767 files that are loaded with the OS (e.g. using rpms
768 or debs). If the user can restore the base OS from
769 CDs, there is no need to backup all those files. A
770 complete restore would be to restore the base OS, then
771 do a Bacula restore. By excluding the base OS files, the
772 backup set will be *much* smaller.
774 Item 26: Implement more Python events in Bacula.
775 Date: 28 October 2005
779 What: Allow Python scripts to be called at more places
780 within Bacula and provide additional access to Bacula
783 Why: This will permit users to customize Bacula through
791 Also add a way to get a listing of currently running
792 jobs (possibly also scheduled jobs).
795 Item 27: Incorporation of XACML2/SAML2 parsing
796 Date: 19 January 2006
797 Origin: Adam Thornton <athornton@sinenomine.net>
800 What: XACML is "eXtensible Access Control Markup Language" and
801 "SAML is the "Security Assertion Markup Language"--an XML standard
802 for making statements about identity and authorization. Having these
803 would give us a framework to approach ACLs in a generic manner, and
804 in a way flexible enough to support the four major sorts of ACLs I
805 see as a concern to Bacula at this point, as well as (probably) to
806 deal with new sorts of ACLs that may appear in the future.
808 Why: Bacula is beginning to need to back up systems with ACLs
809 that do not map cleanly onto traditional Unix permissions. I see
810 four sets of ACLs--in general, mutually incompatible with one
811 another--that we're going to need to deal with. These are: NTFS
812 ACLs, POSIX ACLs, NFSv4 ACLS, and AFS ACLS. (Some may question the
813 relevance of AFS; AFS is one of Sine Nomine's core consulting
814 businesses, and having a reputable file-level backup and restore
815 technology for it (as Tivoli is probably going to drop AFS support
816 soon since IBM no longer supports AFS) would be of huge benefit to
817 our customers; we'd most likely create the AFS support at Sine Nomine
818 for inclusion into the Bacula (and perhaps some changes to the
819 OpenAFS volserver) core code.)
821 Now, obviously, Bacula already handles NTFS just fine. However, I
822 think there's a lot of value in implementing a generic ACL model, so
823 that it's easy to support whatever particular instances of ACLs come
824 down the pike: POSIX ACLS (think SELinux) and NFSv4 are the obvious
825 things arriving in the Linux world in a big way in the near future.
826 XACML, although overcomplicated for our needs, provides this
827 framework, and we should be able to leverage other people's
828 implementations to minimize the amount of work *we* have to do to get
829 a generic ACL framework. Basically, the costs of implementation are
830 high, but they're largely both external to Bacula and already sunk.
832 Item 28: Filesystem watch triggered backup.
834 Origin: Jesper Krogh <jesper@krogh.cc>
835 Status: Unimplemented, depends probably on "client initiated backups"
837 What: With inotify and similar filesystem triggeret notification
838 systems is it possible to have the file-daemon to monitor
839 filesystem changes and initiate backup.
841 Why: There are 2 situations where this is nice to have.
842 1) It is possible to get a much finer-grained backup than
843 the fixed schedules used now.. A file created and deleted
844 a few hours later, can automatically be caught.
846 2) The introduced load on the system will probably be
847 distributed more even on the system.
849 Notes: This can be combined with configration that specifies
850 something like: "at most every 15 minutes or when changes
853 Kern Notes: I would rather see this implemented by an external program
854 that monitors the Filesystem changes, then uses the console
855 to start the appropriate job.
857 Item 29: Allow inclusion/exclusion of files in a fileset by creation/mod times
858 Origin: Evan Kaufman <evan.kaufman@gmail.com>
859 Date: January 11, 2006
862 What: In the vein of the Wild and Regex directives in a Fileset's
863 Options, it would be helpful to allow a user to include or exclude
864 files and directories by creation or modification times.
866 You could factor the Exclude=yes|no option in much the same way it
867 affects the Wild and Regex directives. For example, you could exclude
868 all files modified before a certain date:
872 Modified Before = ####
875 Or you could exclude all files created/modified since a certain date:
879 Created Modified Since = ####
882 The format of the time/date could be done several ways, say the number
883 of seconds since the epoch:
884 1137008553 = Jan 11 2006, 1:42:33PM # result of `date +%s`
886 Or a human readable date in a cryptic form:
887 20060111134233 = Jan 11 2006, 1:42:33PM # YYYYMMDDhhmmss
889 Why: I imagine a feature like this could have many uses. It would
890 allow a user to do a full backup while excluding the base operating
891 system files, so if I installed a Linux snapshot from a CD yesterday,
892 I'll *exclude* all files modified *before* today. If I need to
893 recover the system, I use the CD I already have, plus the tape backup.
894 Or if, say, a Windows client is hit by a particularly corrosive
895 virus, and I need to *exclude* any files created/modified *since* the
898 Notes: Of course, this feature would work in concert with other
899 in/exclude rules, and wouldnt override them (or each other).
901 Notes: The directives I'd imagine would be along the lines of
902 "[Created] [Modified] [Before|Since] = <date>".
903 So one could compare against 'ctime' and/or 'mtime', but ONLY 'before'
907 Item 30: Tray monitor window cleanups
908 Origin: Alan Brown ajb2 at mssl dot ucl dot ac dot uk
911 What: Resizeable and scrollable windows in the tray monitor.
913 Why: With multiple clients, or with many jobs running, the displayed
914 window often ends up larger than the available screen, making
915 the trailing items difficult to read.
918 Item 31: Implement multiple numeric backup levels as supported by dump
920 Origin: Daniel Rich <drich@employees.org>
922 What: Dump allows specification of backup levels numerically instead of just
923 "full", "incr", and "diff". In this system, at any given level, all
924 files are backed up that were were modified since the last backup of a
925 higher level (with 0 being the highest and 9 being the lowest). A
926 level 0 is therefore equivalent to a full, level 9 an incremental, and
927 the levels 1 through 8 are varying levels of differentials. For
928 bacula's sake, these could be represented as "full", "incr", and
929 "diff1", "diff2", etc.
931 Why: Support of multiple backup levels would provide for more advanced backup
932 rotation schemes such as "Towers of Hanoi". This would allow better
933 flexibility in performing backups, and can lead to shorter recover
936 Notes: Legato Networker supports a similar system with full, incr, and 1-9 as
939 Item 32: Automatic promotion of backup levels
940 Date: 19 January 2006
941 Origin: Adam Thornton <athornton@sinenomine.net>
944 What: Amanda has a feature whereby it estimates the space that a
945 differential, incremental, and full backup would take. If the
946 difference in space required between the scheduled level and the next
947 level up is beneath some user-defined critical threshold, the backup
948 level is bumped to the next type. Doing this minimizes the number of
949 volumes necessary during a restore, with a fairly minimal cost in
952 Why: I know at least one (quite sophisticated and smart) user
953 for whom the absence of this feature is a deal-breaker in terms of
954 using Bacula; if we had it it would eliminate the one cool thing
955 Amanda can do and we can't (at least, the one cool thing I know of).
957 Item 33: Clustered file-daemons
958 Origin: Alan Brown ajb2 at mssl dot ucl dot ac dot uk
961 What: A "virtual" filedaemon, which is actually a cluster of real ones.
963 Why: In the case of clustered filesystems (SAN setups, GFS, or OCFS2, etc)
964 multiple machines may have access to the same set of filesystems
966 For performance reasons, one may wish to initate backups from
967 several of these machines simultaneously, instead of just using
968 one backup source for the common clustered filesystem.
970 For obvious reasons, normally backups of $A-FD/$PATH and
971 B-FD/$PATH are treated as different backup sets. In this case
972 they are the same communal set.
974 Likewise when restoring, it would be easier to just specify
975 one of the cluster machines and let bacula decide which to use.
977 This can be faked to some extent using DNS round robin entries
978 and a virtual IP address, however it means "status client" will
979 always give bogus answers. Additionally there is no way of
980 spreading the load evenly among the servers.
982 What is required is something similar to the storage daemon
983 autochanger directives, so that Bacula can keep track of
984 operating backups/restores and direct new jobs to a "free"
989 Item 34: Commercial database support
990 Origin: Russell Howe <russell_howe dot wreckage dot org>
994 What: It would be nice for the database backend to support more
995 databases. I'm thinking of SQL Server at the moment, but I guess Oracle,
996 DB2, MaxDB, etc are all candidates. SQL Server would presumably be
997 implemented using FreeTDS or maybe an ODBC library?
999 Why: We only really have one database server, which is MS SQL Server
1000 2000. Maintaining a second one for the backup software (we grew out of
1001 SQLite, which I liked, but which didn't work so well with our database
1002 size). We don't really have a machine with the resources to run
1003 postgres, and would rather only maintain a single DBMS. We're stuck with
1004 SQL Server because pretty much all the company's custom applications
1005 (written by consultants) are locked into SQL Server 2000. I can imagine
1006 this scenario is fairly common, and it would be nice to use the existing
1007 properly specced database server for storing Bacula's catalog, rather
1008 than having to run a second DBMS.
1010 Item 35: Automatic disabling of devices
1012 Origin: Peter Eriksson <peter at ifm.liu dot se>
1015 What: After a configurable amount of fatal errors with a tape drive
1016 Bacula should automatically disable further use of a certain
1017 tape drive. There should also be "disable"/"enable" commands in
1018 the "bconsole" tool.
1020 Why: On a multi-drive jukebox there is a possibility of tape drives
1021 going bad during large backups (needing a cleaning tape run,
1022 tapes getting stuck). It would be advantageous if Bacula would
1023 automatically disable further use of a problematic tape drive
1024 after a configurable amount of errors has occurred.
1026 An example: I have a multi-drive jukebox (6 drives, 380+ slots)
1027 where tapes occasionally get stuck inside the drive. Bacula will
1028 notice that the "mtx-changer" command will fail and then fail
1029 any backup jobs trying to use that drive. However, it will still
1030 keep on trying to run new jobs using that drive and fail -
1031 forever, and thus failing lots and lots of jobs... Since we have
1032 many drives Bacula could have just automatically disabled
1033 further use of that drive and used one of the other ones
1036 Item 36: An option to operate on all pools with update vol parameters
1037 Origin: Dmitriy Pinchukov <absh@bossdev.kiev.ua>
1038 Date: 16 August 2006
1041 What: When I do update -> Volume parameters -> All Volumes
1042 from Pool, then I have to select pools one by one. I'd like
1043 console to have an option like "0: All Pools" in the list of
1046 Why: I have many pools and therefore unhappy with manually
1047 updating each of them using update -> Volume parameters -> All
1048 Volumes from Pool -> pool #.
1050 Item 37: Add an item to the restore option where you can select a pool
1051 Origin: kshatriyak at gmail dot com
1055 What: In the restore option (Select the most recent backup for a
1056 client) it would be useful to add an option where you can limit
1057 the selection to a certain pool.
1059 Why: When using cloned jobs, most of the time you have 2 pools - a
1060 disk pool and a tape pool. People who have 2 pools would like to
1061 select the most recent backup from disk, not from tape (tape
1062 would be only needed in emergency). However, the most recent
1063 backup (which may just differ a second from the disk backup) may
1064 be on tape and would be selected. The problem becomes bigger if
1065 you have a full and differential - the most "recent" full backup
1066 may be on disk, while the most recent differential may be on tape
1067 (though the differential on disk may differ even only a second or
1068 so). Bacula will complain that the backups reside on different
1069 media then. For now the only solution now when restoring things
1070 when you have 2 pools is to manually search for the right
1071 job-id's and enter them by hand, which is a bit fault tolerant.
1073 Item 38: Include timestamp of job launch in "stat clients" output
1074 Origin: Mark Bergman <mark.bergman@uphs.upenn.edu>
1075 Date: Tue Aug 22 17:13:39 EDT 2006
1078 What: The "stat clients" command doesn't include any detail on when
1079 the active backup jobs were launched.
1081 Why: Including the timestamp would make it much easier to decide whether
1082 a job is running properly.
1084 Notes: It may be helpful to have the output from "stat clients" formatted
1085 more like that from "stat dir" (and other commands), in a column
1086 format. The per-client information that's currently shown (level,
1087 client name, JobId, Volume, pool, device, Files, etc.) is good, but
1088 somewhat hard to parse (both programmatically and visually),
1089 particularly when there are many active clients.
1092 Item 39: Message mailing based on backup types
1093 Origin: Evan Kaufman <evan.kaufman@gmail.com>
1094 Date: January 6, 2006
1097 What: In the "Messages" resource definitions, allowing messages
1098 to be mailed based on the type (backup, restore, etc.) and level
1099 (full, differential, etc) of job that created the originating
1102 Why: It would, for example, allow someone's boss to be emailed
1103 automatically only when a Full Backup job runs, so he can
1104 retrieve the tapes for offsite storage, even if the IT dept.
1105 doesn't (or can't) explicitly notify him. At the same time, his
1106 mailbox wouldnt be filled by notifications of Verifies, Restores,
1107 or Incremental/Differential Backups (which would likely be kept
1110 Notes: One way this could be done is through additional message types, for example:
1113 # email the boss only on full system backups
1114 Mail = boss@mycompany.com = full, !incremental, !differential, !restore,
1116 # email us only when something breaks
1117 MailOnError = itdept@mycompany.com = all
1121 Item 40: Include JobID in spool file name ****DONE****
1122 Origin: Mark Bergman <mark.bergman@uphs.upenn.edu>
1123 Date: Tue Aug 22 17:13:39 EDT 2006
1124 Status: Done. (patches/testing/project-include-jobid-in-spool-name.patch)
1125 No need to vote for this item.
1127 What: Change the name of the spool file to include the JobID
1129 Why: JobIDs are the common key used to refer to jobs, yet the
1130 spoolfile name doesn't include that information. The date/time
1131 stamp is useful (and should be retained).
1133 ============= New Freature Requests after vote of 26 Jan 2007 ========
1134 Item 41: Enable to relocate files and directories when restoring
1136 Origin: Eric Bollengier <eric@eb.homelinux.org>
1139 What: The where= option is not powerful enough. It will be
1140 a great feature if bacula can restore a file in the
1141 same directory, but with a different name, or in
1142 an other directory without recreating the full path.
1144 Why: When i want to restore a production environment to a
1145 development environment, i just want change the first
1146 directory. ie restore /prod/data/file.dat to /rect/data/file.dat.
1147 At this time, i have to move by hand files. You must have a big
1148 dump space to restore and move data after.
1150 When i use Linux or SAN snapshot, i mount them to /mnt/snap_xxx
1151 so, when a restore a file, i have to move by hand
1152 from /mnt/snap_xxx/file to /xxx/file. I can't replace a file
1155 When a user ask me to restore a file in its personal folder,
1156 (without replace the existing one), i can't restore from
1157 my_file.txt to my_file.txt.old witch is very practical.
1160 Notes: I think we can enhance the where= option very easily by
1161 allowing regexp expression.
1163 Since, many users think that regexp are not user friendly, i think
1164 that bat, bconsole or brestore must provide a simple way to
1165 configure where= option (i think to something like in
1166 openoffice "search and replace").
1168 Ie, if user uses where=/tmp/bacula-restore, we keep the old
1171 If user uses something like where=s!/prod!/test!, files will
1172 be restored from /prod/xxx to /test/xxx.
1174 If user uses something like where=s/$/.old/, files will
1175 be restored from /prod/xxx.txt to /prod/xxx.txt.old.
1177 If user uses something like where=s/txt$/old.txt/, files will
1178 be restored from /prod/xxx.txt to /prod/xxx.old.txt
1180 if user uses something like where=s/([a-z]+)$/old.$1/, files will
1181 be restored from /prod/xxx.ext to /prod/xxx.old.ext
1183 Item n: Implement Catalog directive for Pool resource in Director
1185 Origin: Alan Davis adavis@ruckus.com
1189 What: The current behavior is for the director to create all pools
1190 found in the configuration file in all catalogs. Add a
1191 Catalog directive to the Pool resource to specify which
1192 catalog to use for each pool definition.
1194 Why: This allows different catalogs to have different pool
1195 attributes and eliminates the side-effect of adding
1196 pools to catalogs that don't need/use them.
1201 Item n: Implement NDMP protocol support
1206 What: Network Data Management Protocol is implemented by a number of
1207 NAS filer vendors to enable backups using third-party
1210 Why: This would allow NAS filer backups in Bacula without incurring
1211 the overhead of NFS or SBM/CIFS.
1213 Notes: Further information is available:
1215 http://www.ndmp.org/wp/wp.shtml
1216 http://www.traakan.com/ndmjob/index.html
1218 There are currently no viable open-source NDMP
1219 implementations. There is a reference SDK and example
1220 app available from ndmp.org but it has problems
1221 compiling on recent Linux and Solaris OS'. The ndmjob
1222 reference implementation from Traakan is known to
1223 compile on Solaris 10.
1225 Notes (Kern): I am not at all in favor of this until NDMP becomes
1226 an Open Standard or until there are Open Source libraries
1227 that interface to it.
1229 Item n: make changing "spooldata=yes|no" possible for
1230 manual/interactive jobs
1232 Origin: Marc Schiffbauer <marc@schiffbauer.net>
1234 Date: 12 April 2007)
1238 What: Make it possible to modify the spooldata option
1239 for a job when being run from within the console.
1240 Currently it is possible to modify the backup level
1241 and the spooldata setting in a Schedule resource.
1242 It is also possible to modify the backup level when using
1243 the "run" command in the console.
1244 But it is currently not possible to to the same
1245 with "spooldata=yes|no" like:
1247 run job=MyJob level=incremental spooldata=yes
1249 Why: In some situations it would be handy to be able to switch
1250 spooldata on or off for interactive/manual jobs based on
1251 which data the admin expects or how fast the LAN/WAN
1252 connection currently is.
1256 ============= Empty Feature Request form ===========
1257 Item n: One line summary ...
1258 Date: Date submitted
1259 Origin: Name and email of originator.
1262 What: More detailed explanation ...
1264 Why: Why it is important ...
1266 Notes: Additional notes or features (omit if not used)
1267 ============== End Feature Request form ==============