1 <? require_once("inc/header.php"); ?>
12 Bacula Projects Roadmap
13 Status updated 7 July 2007
14 After re-ordering in vote priority
17 Item: 2 Implement a Bacula GUI/management tool.
18 Item: 18 Quick release of FD-SD connection after backup.
19 Item: 23 Implement from-client and to-client on restore command line.
20 Item: 25 Implement huge exclude list support using dlist
21 Item: 41 Enable to relocate files and directories when restoring
22 Item: 42 Batch attribute inserts (ten times faster)
23 Item: 43 More concurrency in SD using micro-locking
24 Item: 44 Performance enhancements (POSIX/Win32 OS file access hints).
25 Item: 40 Include JobID in spool file name
28 Item: 1 Accurate restoration of renamed/deleted files
29 Item: 2* Implement a Bacula GUI/management tool.
30 Item: 3 Allow FD to initiate a backup
31 Item: 4 Merge multiple backups (Synthetic Backup or Consolidation).
32 Item: 5 Deletion of Disk-Based Bacula Volumes
33 Item: 6 Implement Base jobs.
34 Item: 7 Implement creation and maintenance of copy pools
35 Item: 8 Directive/mode to backup only file changes, not entire file
36 Item: 9 Implement a server-side compression feature
37 Item: 10 Improve Bacula's tape and drive usage and cleaning management.
38 Item: 11 Allow skipping execution of Jobs
39 Item: 12 Add a scheduling syntax that permits weekly rotations
40 Item: 13 Archival (removal) of User Files to Tape
41 Item: 14 Cause daemons to use a specific IP address to source communications
42 Item: 15 Multiple threads in file daemon for the same job
43 Item: 16 Add Plug-ins to the FileSet Include statements.
44 Item: 17 Restore only file attributes (permissions, ACL, owner, group...)
45 Item: 18* Quick release of FD-SD connection after backup.
46 Item: 19 Implement a Python interface to the Bacula catalog.
48 Item: 21 Split documentation
49 Item: 22 Implement support for stacking arbitrary stream filters, sinks.
50 Item: 23* Implement from-client and to-client on restore command line.
51 Item: 24 Add an override in Schedule for Pools based on backup types.
52 Item: 25* Implement huge exclude list support using hashing.
53 Item: 26 Implement more Python events in Bacula.
54 Item: 27 Incorporation of XACML2/SAML2 parsing
55 Item: 28 Filesystem watch triggered backup.
56 Item: 29 Allow inclusion/exclusion of files in a fileset by creation/mod times
57 Item: 30 Tray monitor window cleanups
58 Item: 31 Implement multiple numeric backup levels as supported by dump
59 Item: 32 Automatic promotion of backup levels
60 Item: 33 Clustered file-daemons
61 Item: 34 Commercial database support
62 Item: 35 Automatic disabling of devices
63 Item: 36 An option to operate on all pools with update vol parameters
64 Item: 37 Add an item to the restore option where you can select a pool
65 Item: 38 Include timestamp of job launch in "stat clients" output
66 Item: 39 Message mailing based on backup types
67 Item: 40* Include JobID in spool file name
68 Item: 41* Enable to relocate files and directories when restoring
69 Item: 42* Batch attribute inserts (ten times faster)
70 Item: 43* More concurrency in SD using micro-locking
71 Item: 44* Performance enhancements (POSIX/Win32 OS file access hints).
73 Item 1: Accurate restoration of renamed/deleted files
74 Date: 28 November 2005
75 Origin: Martin Simmons (martin at lispworks dot com)
76 Status: Robert Nelson will implement this
78 What: When restoring a fileset for a specified date (including "most
79 recent"), Bacula should give you exactly the files and directories
80 that existed at the time of the last backup prior to that date.
82 Currently this only works if the last backup was a Full backup.
83 When the last backup was Incremental/Differential, files and
84 directories that have been renamed or deleted since the last Full
85 backup are not currently restored correctly. Ditto for files with
86 extra/fewer hard links than at the time of the last Full backup.
88 Why: Incremental/Differential would be much more useful if this worked.
90 Notes: Merging of multiple backups into a single one seems to
91 rely on this working, otherwise the merged backups will not be
92 truly equivalent to a Full backup.
94 Kern: notes shortened. This can be done without the need for
95 inodes. It is essentially the same as the current Verify job,
96 but one additional database record must be written, which does
97 not need any database change.
99 Kern: see if we can correct restoration of directories if
100 replace=ifnewer is set. Currently, if the directory does not
101 exist, a "dummy" directory is created, then when all the files
102 are updated, the dummy directory is newer so the real values
105 Item 2: Implement a Bacula GUI/management tool.
107 Date: 28 October 2005
110 What: Implement a Bacula console, and management tools
111 probably using Qt3 and C++.
113 Why: Don't we already have a wxWidgets GUI? Yes, but
114 it is written in C++ and changes to the user interface
115 must be hand tailored using C++ code. By developing
116 the user interface using Qt designer, the interface
117 can be very easily updated and most of the new Python
118 code will be automatically created. The user interface
119 changes become very simple, and only the new features
120 must be implement. In addition, the code will be in
121 Python, which will give many more users easy (or easier)
122 access to making additions or modifications.
124 Notes: There is a partial Python-GTK implementation
125 Lucas Di Pentima <lucas at lunix dot com dot ar> but
126 it is no longer being developed.
128 Item 3: Allow FD to initiate a backup
129 Origin: Frank Volf (frank at deze dot org)
130 Date: 17 November 2005
133 What: Provide some means, possibly by a restricted console that
134 allows a FD to initiate a backup, and that uses the connection
135 established by the FD to the Director for the backup so that
136 a Director that is firewalled can do the backup.
138 Why: Makes backup of laptops much easier.
141 Item 4: Merge multiple backups (Synthetic Backup or Consolidation).
142 Origin: Marc Cousin and Eric Bollengier
143 Date: 15 November 2005
144 Status: Waiting implementation. Depends on first implementing
145 project Item 2 (Migration) which is now done.
147 What: A merged backup is a backup made without connecting to the Client.
148 It would be a Merge of existing backups into a single backup.
149 In effect, it is like a restore but to the backup medium.
151 For instance, say that last Sunday we made a full backup. Then
152 all week long, we created incremental backups, in order to do
153 them fast. Now comes Sunday again, and we need another full.
154 The merged backup makes it possible to do instead an incremental
155 backup (during the night for instance), and then create a merged
156 backup during the day, by using the full and incrementals from
157 the week. The merged backup will be exactly like a full made
158 Sunday night on the tape, but the production interruption on the
159 Client will be minimal, as the Client will only have to send
162 In fact, if it's done correctly, you could merge all the
163 Incrementals into single Incremental, or all the Incrementals
164 and the last Differential into a new Differential, or the Full,
165 last differential and all the Incrementals into a new Full
166 backup. And there is no need to involve the Client.
168 Why: The benefit is that :
169 - the Client just does an incremental ;
170 - the merged backup on tape is just as a single full backup,
171 and can be restored very fast.
173 This is also a way of reducing the backup data since the old
174 data can then be pruned (or not) from the catalog, possibly
175 allowing older volumes to be recycled
177 Item 5: Deletion of Disk-Based Bacula Volumes
179 Origin: Ross Boylan <RossBoylan at stanfordalumni dot org> (edited
183 What: Provide a way for Bacula to automatically remove Volumes
184 from the filesystem, or optionally to truncate them.
185 Obviously, the Volume must be pruned prior removal.
187 Why: This would allow users more control over their Volumes and
188 prevent disk based volumes from consuming too much space.
190 Notes: The following two directives might do the trick:
192 Volume Data Retention = <time period>
193 Remove Volume After = <time period>
195 The migration project should also remove a Volume that is
196 migrated. This might also work for tape Volumes.
198 Item 6: Implement Base jobs.
199 Date: 28 October 2005
203 What: A base job is sort of like a Full save except that you
204 will want the FileSet to contain only files that are
205 unlikely to change in the future (i.e. a snapshot of
206 most of your system after installing it). After the
207 base job has been run, when you are doing a Full save,
208 you specify one or more Base jobs to be used. All
209 files that have been backed up in the Base job/jobs but
210 not modified will then be excluded from the backup.
211 During a restore, the Base jobs will be automatically
212 pulled in where necessary.
214 Why: This is something none of the competition does, as far as
215 we know (except perhaps BackupPC, which is a Perl program that
216 saves to disk only). It is big win for the user, it
217 makes Bacula stand out as offering a unique
218 optimization that immediately saves time and money.
219 Basically, imagine that you have 100 nearly identical
220 Windows or Linux machine containing the OS and user
221 files. Now for the OS part, a Base job will be backed
222 up once, and rather than making 100 copies of the OS,
223 there will be only one. If one or more of the systems
224 have some files updated, no problem, they will be
225 automatically restored.
227 Notes: Huge savings in tape usage even for a single machine.
228 Will require more resources because the DIR must send
229 FD a list of files/attribs, and the FD must search the
230 list and compare it for each file to be saved.
232 Item 7: Implement creation and maintenance of copy pools
233 Date: 27 November 2005
234 Origin: David Boyes (dboyes at sinenomine dot net)
237 What: I would like Bacula to have the capability to write copies
238 of backed-up data on multiple physical volumes selected
239 from different pools without transferring the data
240 multiple times, and to accept any of the copy volumes
241 as valid for restore.
243 Why: In many cases, businesses are required to keep offsite
244 copies of backup volumes, or just wish for simple
245 protection against a human operator dropping a storage
246 volume and damaging it. The ability to generate multiple
247 volumes in the course of a single backup job allows
248 customers to simple check out one copy and send it
249 offsite, marking it as out of changer or otherwise
250 unavailable. Currently, the library and magazine
251 management capability in Bacula does not make this process
254 Restores would use the copy of the data on the first
255 available volume, in order of copy pool chain definition.
257 This is also a major scalability issue -- as the number of
258 clients increases beyond several thousand, and the volume
259 of data increases, transferring the data multiple times to
260 produce additional copies of the backups will become
261 physically impossible due to transfer speed
262 issues. Generating multiple copies at server side will
263 become the only practical option.
265 How: I suspect that this will require adding a multiplexing
266 SD that appears to be a SD to a specific FD, but 1-n FDs
267 to the specific back end SDs managing the primary and copy
268 pools. Storage pools will also need to acquire parameters
269 to define the pools to be used for copies.
271 Notes: I would commit some of my developers' time if we can agree
272 on the design and behavior.
274 Item 8: Directive/mode to backup only file changes, not entire file
275 Date: 11 November 2005
276 Origin: Joshua Kugler <joshua dot kugler at uaf dot edu>
277 Marek Bajon <mbajon at bimsplus dot com dot pl>
280 What: Currently when a file changes, the entire file will be backed up in
281 the next incremental or full backup. To save space on the tapes
282 it would be nice to have a mode whereby only the changes to the
283 file would be backed up when it is changed.
285 Why: This would save lots of space when backing up large files such as
286 logs, mbox files, Outlook PST files and the like.
288 Notes: This would require the usage of disk-based volumes as comparing
289 files would not be feasible using a tape drive.
291 Item 9: Implement a server-side compression feature
292 Date: 18 December 2006
293 Origin: Vadim A. Umanski , e-mail umanski@ext.ru
295 What: The ability to compress backup data on server receiving data
296 instead of doing that on client sending data.
297 Why: The need is practical. I've got some machines that can send
298 data to the network 4 or 5 times faster than compressing
299 them (I've measured that). They're using fast enough SCSI/FC
300 disk subsystems but rather slow CPUs (ex. UltraSPARC II).
301 And the backup server has got a quite fast CPUs (ex. Dual P4
302 Xeons) and quite a low load. When you have 20, 50 or 100 GB
303 of raw data - running a job 4 to 5 times faster - that
304 really matters. On the other hand, the data can be
305 compressed 50% or better - so losing twice more space for
306 disk backup is not good at all. And the network is all mine
307 (I have a dedicated management/provisioning network) and I
308 can get as high bandwidth as I need - 100Mbps, 1000Mbps...
309 That's why the server-side compression feature is needed!
312 Item 10: Improve Bacula's tape and drive usage and cleaning management.
313 Date: 8 November 2005, November 11, 2005
314 Origin: Adam Thornton <athornton at sinenomine dot net>,
315 Arno Lehmann <al at its-lehmann dot de>
318 What: Make Bacula manage tape life cycle information, tape reuse
319 times and drive cleaning cycles.
321 Why: All three parts of this project are important when operating
323 We need to know which tapes need replacement, and we need to
324 make sure the drives are cleaned when necessary. While many
325 tape libraries and even autoloaders can handle all this
326 automatically, support by Bacula can be helpful for smaller
327 (older) libraries and single drives. Limiting the number of
328 times a tape is used might prevent tape errors when using
329 tapes until the drives can't read it any more. Also, checking
330 drive status during operation can prevent some failures (as I
331 [Arno] had to learn the hard way...)
333 Notes: First, Bacula could (and even does, to some limited extent)
334 record tape and drive usage. For tapes, the number of mounts,
335 the amount of data, and the time the tape has actually been
336 running could be recorded. Data fields for Read and Write
337 time and Number of mounts already exist in the catalog (I'm
338 not sure if VolBytes is the sum of all bytes ever written to
339 that volume by Bacula). This information can be important
340 when determining which media to replace. The ability to mark
341 Volumes as "used up" after a given number of write cycles
342 should also be implemented so that a tape is never actually
343 worn out. For the tape drives known to Bacula, similar
344 information is interesting to determine the device status and
345 expected life time: Time it's been Reading and Writing, number
346 of tape Loads / Unloads / Errors. This information is not yet
347 recorded as far as I [Arno] know. A new volume status would
348 be necessary for the new state, like "Used up" or "Worn out".
349 Volumes with this state could be used for restores, but not
350 for writing. These volumes should be migrated first (assuming
351 migration is implemented) and, once they are no longer needed,
352 could be moved to a Trash pool.
354 The next step would be to implement a drive cleaning setup.
355 Bacula already has knowledge about cleaning tapes. Once it
356 has some information about cleaning cycles (measured in drive
357 run time, number of tapes used, or calender days, for example)
358 it can automatically execute tape cleaning (with an
359 autochanger, obviously) or ask for operator assistance loading
362 The final step would be to implement TAPEALERT checks not only
363 when changing tapes and only sending the information to the
364 administrator, but rather checking after each tape error,
365 checking on a regular basis (for example after each tape
366 file), and also before unloading and after loading a new tape.
367 Then, depending on the drives TAPEALERT state and the known
368 drive cleaning state Bacula could automatically schedule later
369 cleaning, clean immediately, or inform the operator.
371 Implementing this would perhaps require another catalog change
372 and perhaps major changes in SD code and the DIR-SD protocol,
373 so I'd only consider this worth implementing if it would
374 actually be used or even needed by many people.
376 Implementation of these projects could happen in three distinct
377 sub-projects: Measuring Tape and Drive usage, retiring
378 volumes, and handling drive cleaning and TAPEALERTs.
380 Item 11: Allow skipping execution of Jobs
381 Date: 29 November 2005
382 Origin: Florian Schnabel <florian.schnabel at docufy dot de>
385 What: An easy option to skip a certain job on a certain date.
386 Why: You could then easily skip tape backups on holidays. Especially
387 if you got no autochanger and can only fit one backup on a tape
388 that would be really handy, other jobs could proceed normally
389 and you won't get errors that way.
391 Item 12: Add a scheduling syntax that permits weekly rotations
392 Date: 15 December 2006
393 Origin: Gregory Brauer (greg at wildbrain dot com)
396 What: Currently, Bacula only understands how to deal with weeks of the
397 month or weeks of the year in schedules. This makes it impossible
398 to do a true weekly rotation of tapes. There will always be a
399 discontinuity that will require disruptive manual intervention at
400 least monthly or yearly because week boundaries never align with
401 month or year boundaries.
403 A solution would be to add a new syntax that defines (at least)
404 a start timestamp, and repetition period.
406 Why: Rotated backups done at weekly intervals are useful, and Bacula
407 cannot currently do them without extensive hacking.
409 Notes: Here is an example syntax showing a 3-week rotation where full
410 Backups would be performed every week on Saturday, and an
411 incremental would be performed every week on Tuesday. Each
412 set of tapes could be removed from the loader for the following
413 two cycles before coming back and being reused on the third
414 week. Since the execution times are determined by intervals
415 from a given point in time, there will never be any issues with
416 having to adjust to any sort of arbitrary time boundary. In
417 the example provided, I even define the starting schedule
418 as crossing both a year and a month boundary, but the run times
419 would be based on the "Repeat" value and would therefore happen
424 Name = "Week 1 Rotation"
425 #Saturday. Would run Dec 30, Jan 20, Feb 10, etc.
429 Start = 2006-12-30 01:00
433 #Tuesday. Would run Jan 2, Jan 23, Feb 13, etc.
437 Start = 2007-01-02 01:00
444 Name = "Week 2 Rotation"
445 #Saturday. Would run Jan 6, Jan 27, Feb 17, etc.
449 Start = 2007-01-06 01:00
453 #Tuesday. Would run Jan 9, Jan 30, Feb 20, etc.
457 Start = 2007-01-09 01:00
464 Name = "Week 3 Rotation"
465 #Saturday. Would run Jan 13, Feb 3, Feb 24, etc.
469 Start = 2007-01-13 01:00
473 #Tuesday. Would run Jan 16, Feb 6, Feb 27, etc.
477 Start = 2007-01-16 01:00
483 Item 13: Archival (removal) of User Files to Tape
485 Origin: Ray Pengelly [ray at biomed dot queensu dot ca
488 What: The ability to archive data to storage based on certain parameters
489 such as age, size, or location. Once the data has been written to
490 storage and logged it is then pruned from the originating
491 filesystem. Note! We are talking about user's files and not
494 Why: This would allow fully automatic storage management which becomes
495 useful for large datastores. It would also allow for auto-staging
496 from one media type to another.
498 Example 1) Medical imaging needs to store large amounts of data.
499 They decide to keep data on their servers for 6 months and then put
500 it away for long term storage. The server then finds all files
501 older than 6 months writes them to tape. The files are then removed
504 Example 2) All data that hasn't been accessed in 2 months could be
505 moved from high-cost, fibre-channel disk storage to a low-cost
506 large-capacity SATA disk storage pool which doesn't have as quick of
507 access time. Then after another 6 months (or possibly as one
508 storage pool gets full) data is migrated to Tape.
510 Item 14: Cause daemons to use a specific IP address to source communications
511 Origin: Bill Moran <wmoran@collaborativefusion.com>
514 What: Cause Bacula daemons (dir, fd, sd) to always use the ip address
515 specified in the [DIR|DF|SD]Addr directive as the source IP
516 for initiating communication.
517 Why: On complex networks, as well as extremely secure networks, it's
518 not unusual to have multiple possible routes through the network.
519 Often, each of these routes is secured by different policies
520 (effectively, firewalls allow or deny different traffic depending
521 on the source address)
522 Unfortunately, it can sometimes be difficult or impossible to
523 represent this in a system routing table, as the result is
524 excessive subnetting that quickly exhausts available IP space.
525 The best available workaround is to provide multiple IPs to
526 a single machine that are all on the same subnet. In order
527 for this to work properly, applications must support the ability
528 to bind outgoing connections to a specified address, otherwise
529 the operating system will always choose the first IP that
530 matches the required route.
531 Notes: Many other programs support this. For example, the following
532 can be configured in BIND:
533 query-source address 10.0.0.1;
534 transfer-source 10.0.0.2;
535 Which means queries from this server will always come from
536 10.0.0.1 and zone transfers will always originate from
539 Item 15: Multiple threads in file daemon for the same job
540 Date: 27 November 2005
541 Origin: Ove Risberg (Ove.Risberg at octocode dot com)
544 What: I want the file daemon to start multiple threads for a backup
545 job so the fastest possible backup can be made.
547 The file daemon could parse the FileSet information and start
548 one thread for each File entry located on a separate
551 A confiuration option in the job section should be used to
552 enable or disable this feature. The confgutration option could
553 specify the maximum number of threads in the file daemon.
555 If the theads could spool the data to separate spool files
556 the restore process will not be much slower.
558 Why: Multiple concurrent backups of a large fileserver with many
559 disks and controllers will be much faster.
561 Item 16: Add Plug-ins to the FileSet Include statements.
562 Date: 28 October 2005
564 Status: Partially coded in 1.37 -- much more to do.
566 What: Allow users to specify wild-card and/or regular
567 expressions to be matched in both the Include and
568 Exclude directives in a FileSet. At the same time,
569 allow users to define plug-ins to be called (based on
570 regular expression/wild-card matching).
572 Why: This would give the users the ultimate ability to control
573 how files are backed up/restored. A user could write a
574 plug-in knows how to backup his Oracle database without
575 stopping/starting it, for example.
577 Item 17: Restore only file attributes (permissions, ACL, owner, group...)
578 Origin: Eric Bollengier
582 What: The goal of this project is to be able to restore only rights
583 and attributes of files without crushing them.
585 Why: Who have never had to repair a chmod -R 777, or a wild update
586 of recursive right under Windows? At this time, you must have
587 enough space to restore data, dump attributes (easy with acl,
588 more complex with unix/windows rights) and apply them to your
589 broken tree. With this options, it will be very easy to compare
590 right or ACL over the time.
592 Notes: If the file is here, we skip restore and we change rights.
593 If the file isn't here, we can create an empty one and apply
594 rights or do nothing.
596 Item 18: Quick release of FD-SD connection after backup.
597 Origin: Frank Volf (frank at deze dot org)
598 Date: 17 November 2005
599 Status: Done -- implemented by Kern -- in CVS 26Jan07
601 What: In the Bacula implementation a backup is finished after all data
602 and attributes are successfully written to storage. When using a
603 tape backup it is very annoying that a backup can take a day,
604 simply because the current tape (or whatever) is full and the
605 administrator has not put a new one in. During that time the
606 system cannot be taken off-line, because there is still an open
607 session between the storage daemon and the file daemon on the
610 Although this is a very good strategy for making "safe backups"
611 This can be annoying for e.g. laptops, that must remain
612 connected until the backup is completed.
614 Using a new feature called "migration" it will be possible to
615 spool first to harddisk (using a special 'spool' migration
616 scheme) and then migrate the backup to tape.
618 There is still the problem of getting the attributes committed.
619 If it takes a very long time to do, with the current code, the
620 job has not terminated, and the File daemon is not freed up. The
621 Storage daemon should release the File daemon as soon as all the
622 file data and all the attributes have been sent to it (the SD).
623 Currently the SD waits until everything is on tape and all the
624 attributes are transmitted to the Director before signaling
625 completion to the FD. I don't think I would have any problem
626 changing this. The reason is that even if the FD reports back to
627 the Dir that all is OK, the job will not terminate until the SD
628 has done the same thing -- so in a way keeping the SD-FD link
629 open to the very end is not really very productive ...
631 Why: Makes backup of laptops much faster.
633 Item 19: Implement a Python interface to the Bacula catalog.
634 Date: 28 October 2005
638 What: Implement an interface for Python scripts to access
639 the catalog through Bacula.
641 Why: This will permit users to customize Bacula through
644 Item 20: Archive data
646 Origin: calvin streeting calvin at absentdream dot com
649 What: The abilty to archive to media (dvd/cd) in a uncompressed format
650 for dead filing (archiving not backing up)
652 Why: At my works when jobs are finished and moved off of the main file
653 servers (raid based systems) onto a simple linux file server (ide based
654 system) so users can find old information without contacting the IT
657 So this data dosn't realy change it only gets added to,
658 But it also needs backing up. At the moment it takes
659 about 8 hours to back up our servers (working data) so
660 rather than add more time to existing backups i am trying
661 to implement a system where we backup the acrhive data to
662 cd/dvd these disks would only need to be appended to
663 (burn only new/changed files to new disks for off site
664 storage). basialy understand the differnce between
665 achive data and live data.
667 Notes: Scan the data and email me when it needs burning divide
668 into predifind chunks keep a recored of what is on what
669 disk make me a label (simple php->mysql=>pdf stuff) i
670 could do this bit ability to save data uncompresed so
671 it can be read in any other system (future proof data)
672 save the catalog with the disk as some kind of menu
675 Item 21: Split documentation
676 Origin: Maxx <maxxatworkat gmail dot com>
680 What: Split documentation in several books
682 Why: Bacula manual has now more than 600 pages, and looking for
683 implementation details is getting complicated. I think
684 it would be good to split the single volume in two or
687 1) Introduction, requirements and tutorial, typically
688 are useful only until first installation time
690 2) Basic installation and configuration, with all the
691 gory details about the directives supported 3)
692 Advanced Bacula: testing, troubleshooting, GUI and
693 ancillary programs, security managements, scripting,
697 Item 22: Implement support for stacking arbitrary stream filters, sinks.
698 Date: 23 November 2006
699 Origin: Landon Fuller <landonf@threerings.net>
700 Status: Planning. Assigned to landonf.
702 What: Implement support for the following:
703 - Stacking arbitrary stream filters (eg, encryption, compression,
704 sparse data handling))
705 - Attaching file sinks to terminate stream filters (ie, write out
706 the resultant data to a file)
707 - Refactor the restoration state machine accordingly
709 Why: The existing stream implementation suffers from the following:
710 - All state (compression, encryption, stream restoration), is
711 global across the entire restore process, for all streams. There are
712 multiple entry and exit points in the restoration state machine, and
713 thus multiple places where state must be allocated, deallocated,
714 initialized, or reinitialized. This results in exceptional complexity
715 for the author of a stream filter.
716 - The developer must enumerate all possible combinations of filters
717 and stream types (ie, win32 data with encryption, without encryption,
718 with encryption AND compression, etc).
720 Notes: This feature request only covers implementing the stream filters/
721 sinks, and refactoring the file daemon's restoration implementation
722 accordingly. If I have extra time, I will also rewrite the backup
723 implementation. My intent in implementing the restoration first is to
724 solve pressing bugs in the restoration handling, and to ensure that
725 the new restore implementation handles existing backups correctly.
727 I do not plan on changing the network or tape data structures to
728 support defining arbitrary stream filters, but supporting that
729 functionality is the ultimate goal.
731 Assistance with either code or testing would be fantastic.
733 Item 23: Implement from-client and to-client on restore command line.
734 Date: 11 December 2006
735 Origin: Discussion on Bacula-users entitled 'Scripted restores to
736 different clients', December 2006
737 Status: New feature request
739 What: While using bconsole interactively, you can specify the client
740 that a backup job is to be restored for, and then you can
741 specify later a different client to send the restored files
742 back to. However, using the 'restore' command with all options
743 on the command line, this cannot be done, due to the ambiguous
744 'client' parameter. Additionally, this parameter means different
745 things depending on if it's specified on the command line or
746 afterwards, in the Modify Job screens.
748 Why: This feature would enable restore jobs to be more completely
749 automated, for example by a web or GUI front-end.
751 Notes: client can also be implied by specifying the jobid on the command
754 Item 24: Add an override in Schedule for Pools based on backup types.
756 Origin: Chad Slater <chad.slater@clickfox.com>
759 What: Adding a FullStorage=BigTapeLibrary in the Schedule resource
760 would help those of us who use different storage devices for different
761 backup levels cope with the "auto-upgrade" of a backup.
763 Why: Assume I add several new device to be backed up, i.e. several
764 hosts with 1TB RAID. To avoid tape switching hassles, incrementals are
765 stored in a disk set on a 2TB RAID. If you add these devices in the
766 middle of the month, the incrementals are upgraded to "full" backups,
767 but they try to use the same storage device as requested in the
768 incremental job, filling up the RAID holding the differentials. If we
769 could override the Storage parameter for full and/or differential
770 backups, then the Full job would use the proper Storage device, which
771 has more capacity (i.e. a 8TB tape library.
773 Item 25: Implement huge exclude list support using hashing (dlists).
774 Date: 28 October 2005
776 Status: Done in 2.1.2 but was done with dlists (doubly linked lists
777 since hashing will not help. The huge list also supports
778 large include lists).
780 What: Allow users to specify very large exclude list (currently
781 more than about 1000 files is too many).
783 Why: This would give the users the ability to exclude all
784 files that are loaded with the OS (e.g. using rpms
785 or debs). If the user can restore the base OS from
786 CDs, there is no need to backup all those files. A
787 complete restore would be to restore the base OS, then
788 do a Bacula restore. By excluding the base OS files, the
789 backup set will be *much* smaller.
791 Item 26: Implement more Python events in Bacula.
792 Date: 28 October 2005
796 What: Allow Python scripts to be called at more places
797 within Bacula and provide additional access to Bacula
800 Why: This will permit users to customize Bacula through
808 Also add a way to get a listing of currently running
809 jobs (possibly also scheduled jobs).
812 Item 27: Incorporation of XACML2/SAML2 parsing
813 Date: 19 January 2006
814 Origin: Adam Thornton <athornton@sinenomine.net>
817 What: XACML is "eXtensible Access Control Markup Language" and
818 "SAML is the "Security Assertion Markup Language"--an XML standard
819 for making statements about identity and authorization. Having these
820 would give us a framework to approach ACLs in a generic manner, and
821 in a way flexible enough to support the four major sorts of ACLs I
822 see as a concern to Bacula at this point, as well as (probably) to
823 deal with new sorts of ACLs that may appear in the future.
825 Why: Bacula is beginning to need to back up systems with ACLs
826 that do not map cleanly onto traditional Unix permissions. I see
827 four sets of ACLs--in general, mutually incompatible with one
828 another--that we're going to need to deal with. These are: NTFS
829 ACLs, POSIX ACLs, NFSv4 ACLS, and AFS ACLS. (Some may question the
830 relevance of AFS; AFS is one of Sine Nomine's core consulting
831 businesses, and having a reputable file-level backup and restore
832 technology for it (as Tivoli is probably going to drop AFS support
833 soon since IBM no longer supports AFS) would be of huge benefit to
834 our customers; we'd most likely create the AFS support at Sine Nomine
835 for inclusion into the Bacula (and perhaps some changes to the
836 OpenAFS volserver) core code.)
838 Now, obviously, Bacula already handles NTFS just fine. However, I
839 think there's a lot of value in implementing a generic ACL model, so
840 that it's easy to support whatever particular instances of ACLs come
841 down the pike: POSIX ACLS (think SELinux) and NFSv4 are the obvious
842 things arriving in the Linux world in a big way in the near future.
843 XACML, although overcomplicated for our needs, provides this
844 framework, and we should be able to leverage other people's
845 implementations to minimize the amount of work *we* have to do to get
846 a generic ACL framework. Basically, the costs of implementation are
847 high, but they're largely both external to Bacula and already sunk.
849 Item 28: Filesystem watch triggered backup.
851 Origin: Jesper Krogh <jesper@krogh.cc>
852 Status: Unimplemented, depends probably on "client initiated backups"
854 What: With inotify and similar filesystem triggeret notification
855 systems is it possible to have the file-daemon to monitor
856 filesystem changes and initiate backup.
858 Why: There are 2 situations where this is nice to have.
859 1) It is possible to get a much finer-grained backup than
860 the fixed schedules used now.. A file created and deleted
861 a few hours later, can automatically be caught.
863 2) The introduced load on the system will probably be
864 distributed more even on the system.
866 Notes: This can be combined with configration that specifies
867 something like: "at most every 15 minutes or when changes
870 Kern Notes: I would rather see this implemented by an external program
871 that monitors the Filesystem changes, then uses the console
872 to start the appropriate job.
874 Item 29: Allow inclusion/exclusion of files in a fileset by creation/mod times
875 Origin: Evan Kaufman <evan.kaufman@gmail.com>
876 Date: January 11, 2006
879 What: In the vein of the Wild and Regex directives in a Fileset's
880 Options, it would be helpful to allow a user to include or exclude
881 files and directories by creation or modification times.
883 You could factor the Exclude=yes|no option in much the same way it
884 affects the Wild and Regex directives. For example, you could exclude
885 all files modified before a certain date:
889 Modified Before = ####
892 Or you could exclude all files created/modified since a certain date:
896 Created Modified Since = ####
899 The format of the time/date could be done several ways, say the number
900 of seconds since the epoch:
901 1137008553 = Jan 11 2006, 1:42:33PM # result of `date +%s`
903 Or a human readable date in a cryptic form:
904 20060111134233 = Jan 11 2006, 1:42:33PM # YYYYMMDDhhmmss
906 Why: I imagine a feature like this could have many uses. It would
907 allow a user to do a full backup while excluding the base operating
908 system files, so if I installed a Linux snapshot from a CD yesterday,
909 I'll *exclude* all files modified *before* today. If I need to
910 recover the system, I use the CD I already have, plus the tape backup.
911 Or if, say, a Windows client is hit by a particularly corrosive
912 virus, and I need to *exclude* any files created/modified *since* the
915 Notes: Of course, this feature would work in concert with other
916 in/exclude rules, and wouldnt override them (or each other).
918 Notes: The directives I'd imagine would be along the lines of
919 "[Created] [Modified] [Before|Since] = <date>".
920 So one could compare against 'ctime' and/or 'mtime', but ONLY 'before'
924 Item 30: Tray monitor window cleanups
925 Origin: Alan Brown ajb2 at mssl dot ucl dot ac dot uk
928 What: Resizeable and scrollable windows in the tray monitor.
930 Why: With multiple clients, or with many jobs running, the displayed
931 window often ends up larger than the available screen, making
932 the trailing items difficult to read.
935 Item 31: Implement multiple numeric backup levels as supported by dump
937 Origin: Daniel Rich <drich@employees.org>
939 What: Dump allows specification of backup levels numerically instead of just
940 "full", "incr", and "diff". In this system, at any given level, all
941 files are backed up that were were modified since the last backup of a
942 higher level (with 0 being the highest and 9 being the lowest). A
943 level 0 is therefore equivalent to a full, level 9 an incremental, and
944 the levels 1 through 8 are varying levels of differentials. For
945 bacula's sake, these could be represented as "full", "incr", and
946 "diff1", "diff2", etc.
948 Why: Support of multiple backup levels would provide for more advanced backup
949 rotation schemes such as "Towers of Hanoi". This would allow better
950 flexibility in performing backups, and can lead to shorter recover
953 Notes: Legato Networker supports a similar system with full, incr, and 1-9 as
956 Item 32: Automatic promotion of backup levels
957 Date: 19 January 2006
958 Origin: Adam Thornton <athornton@sinenomine.net>
961 What: Amanda has a feature whereby it estimates the space that a
962 differential, incremental, and full backup would take. If the
963 difference in space required between the scheduled level and the next
964 level up is beneath some user-defined critical threshold, the backup
965 level is bumped to the next type. Doing this minimizes the number of
966 volumes necessary during a restore, with a fairly minimal cost in
969 Why: I know at least one (quite sophisticated and smart) user
970 for whom the absence of this feature is a deal-breaker in terms of
971 using Bacula; if we had it it would eliminate the one cool thing
972 Amanda can do and we can't (at least, the one cool thing I know of).
974 Item 33: Clustered file-daemons
975 Origin: Alan Brown ajb2 at mssl dot ucl dot ac dot uk
978 What: A "virtual" filedaemon, which is actually a cluster of real ones.
980 Why: In the case of clustered filesystems (SAN setups, GFS, or OCFS2, etc)
981 multiple machines may have access to the same set of filesystems
983 For performance reasons, one may wish to initate backups from
984 several of these machines simultaneously, instead of just using
985 one backup source for the common clustered filesystem.
987 For obvious reasons, normally backups of $A-FD/$PATH and
988 B-FD/$PATH are treated as different backup sets. In this case
989 they are the same communal set.
991 Likewise when restoring, it would be easier to just specify
992 one of the cluster machines and let bacula decide which to use.
994 This can be faked to some extent using DNS round robin entries
995 and a virtual IP address, however it means "status client" will
996 always give bogus answers. Additionally there is no way of
997 spreading the load evenly among the servers.
999 What is required is something similar to the storage daemon
1000 autochanger directives, so that Bacula can keep track of
1001 operating backups/restores and direct new jobs to a "free"
1006 Item 34: Commercial database support
1007 Origin: Russell Howe <russell_howe dot wreckage dot org>
1011 What: It would be nice for the database backend to support more
1012 databases. I'm thinking of SQL Server at the moment, but I guess Oracle,
1013 DB2, MaxDB, etc are all candidates. SQL Server would presumably be
1014 implemented using FreeTDS or maybe an ODBC library?
1016 Why: We only really have one database server, which is MS SQL Server
1017 2000. Maintaining a second one for the backup software (we grew out of
1018 SQLite, which I liked, but which didn't work so well with our database
1019 size). We don't really have a machine with the resources to run
1020 postgres, and would rather only maintain a single DBMS. We're stuck with
1021 SQL Server because pretty much all the company's custom applications
1022 (written by consultants) are locked into SQL Server 2000. I can imagine
1023 this scenario is fairly common, and it would be nice to use the existing
1024 properly specced database server for storing Bacula's catalog, rather
1025 than having to run a second DBMS.
1027 Item 35: Automatic disabling of devices
1029 Origin: Peter Eriksson <peter at ifm.liu dot se>
1032 What: After a configurable amount of fatal errors with a tape drive
1033 Bacula should automatically disable further use of a certain
1034 tape drive. There should also be "disable"/"enable" commands in
1035 the "bconsole" tool.
1037 Why: On a multi-drive jukebox there is a possibility of tape drives
1038 going bad during large backups (needing a cleaning tape run,
1039 tapes getting stuck). It would be advantageous if Bacula would
1040 automatically disable further use of a problematic tape drive
1041 after a configurable amount of errors has occurred.
1043 An example: I have a multi-drive jukebox (6 drives, 380+ slots)
1044 where tapes occasionally get stuck inside the drive. Bacula will
1045 notice that the "mtx-changer" command will fail and then fail
1046 any backup jobs trying to use that drive. However, it will still
1047 keep on trying to run new jobs using that drive and fail -
1048 forever, and thus failing lots and lots of jobs... Since we have
1049 many drives Bacula could have just automatically disabled
1050 further use of that drive and used one of the other ones
1053 Item 36: An option to operate on all pools with update vol parameters
1054 Origin: Dmitriy Pinchukov <absh@bossdev.kiev.ua>
1055 Date: 16 August 2006
1058 What: When I do update -> Volume parameters -> All Volumes
1059 from Pool, then I have to select pools one by one. I'd like
1060 console to have an option like "0: All Pools" in the list of
1063 Why: I have many pools and therefore unhappy with manually
1064 updating each of them using update -> Volume parameters -> All
1065 Volumes from Pool -> pool #.
1067 Item 37: Add an item to the restore option where you can select a pool
1068 Origin: kshatriyak at gmail dot com
1072 What: In the restore option (Select the most recent backup for a
1073 client) it would be useful to add an option where you can limit
1074 the selection to a certain pool.
1076 Why: When using cloned jobs, most of the time you have 2 pools - a
1077 disk pool and a tape pool. People who have 2 pools would like to
1078 select the most recent backup from disk, not from tape (tape
1079 would be only needed in emergency). However, the most recent
1080 backup (which may just differ a second from the disk backup) may
1081 be on tape and would be selected. The problem becomes bigger if
1082 you have a full and differential - the most "recent" full backup
1083 may be on disk, while the most recent differential may be on tape
1084 (though the differential on disk may differ even only a second or
1085 so). Bacula will complain that the backups reside on different
1086 media then. For now the only solution now when restoring things
1087 when you have 2 pools is to manually search for the right
1088 job-id's and enter them by hand, which is a bit fault tolerant.
1090 Item 38: Include timestamp of job launch in "stat clients" output
1091 Origin: Mark Bergman <mark.bergman@uphs.upenn.edu>
1092 Date: Tue Aug 22 17:13:39 EDT 2006
1095 What: The "stat clients" command doesn't include any detail on when
1096 the active backup jobs were launched.
1098 Why: Including the timestamp would make it much easier to decide whether
1099 a job is running properly.
1101 Notes: It may be helpful to have the output from "stat clients" formatted
1102 more like that from "stat dir" (and other commands), in a column
1103 format. The per-client information that's currently shown (level,
1104 client name, JobId, Volume, pool, device, Files, etc.) is good, but
1105 somewhat hard to parse (both programmatically and visually),
1106 particularly when there are many active clients.
1109 Item 39: Message mailing based on backup types
1110 Origin: Evan Kaufman <evan.kaufman@gmail.com>
1111 Date: January 6, 2006
1114 What: In the "Messages" resource definitions, allowing messages
1115 to be mailed based on the type (backup, restore, etc.) and level
1116 (full, differential, etc) of job that created the originating
1119 Why: It would, for example, allow someone's boss to be emailed
1120 automatically only when a Full Backup job runs, so he can
1121 retrieve the tapes for offsite storage, even if the IT dept.
1122 doesn't (or can't) explicitly notify him. At the same time, his
1123 mailbox wouldnt be filled by notifications of Verifies, Restores,
1124 or Incremental/Differential Backups (which would likely be kept
1127 Notes: One way this could be done is through additional message types, for example:
1130 # email the boss only on full system backups
1131 Mail = boss@mycompany.com = full, !incremental, !differential, !restore,
1133 # email us only when something breaks
1134 MailOnError = itdept@mycompany.com = all
1138 Item 40: Include JobID in spool file name ****DONE****
1139 Origin: Mark Bergman <mark.bergman@uphs.upenn.edu>
1140 Date: Tue Aug 22 17:13:39 EDT 2006
1141 Status: Done. (patches/testing/project-include-jobid-in-spool-name.patch)
1142 No need to vote for this item.
1144 What: Change the name of the spool file to include the JobID
1146 Why: JobIDs are the common key used to refer to jobs, yet the
1147 spoolfile name doesn't include that information. The date/time
1148 stamp is useful (and should be retained).
1150 ============= New Freature Requests after vote of 26 Jan 2007 ========
1151 Item 41: Enable to relocate files and directories when restoring
1153 Origin: Eric Bollengier <eric@eb.homelinux.org>
1156 What: The where= option is not powerful enough. It will be
1157 a great feature if bacula can restore a file in the
1158 same directory, but with a different name, or in
1159 an other directory without recreating the full path.
1161 Why: When i want to restore a production environment to a
1162 development environment, i just want change the first
1163 directory. ie restore /prod/data/file.dat to /rect/data/file.dat.
1164 At this time, i have to move by hand files. You must have a big
1165 dump space to restore and move data after.
1167 When i use Linux or SAN snapshot, i mount them to /mnt/snap_xxx
1168 so, when a restore a file, i have to move by hand
1169 from /mnt/snap_xxx/file to /xxx/file. I can't replace a file
1172 When a user ask me to restore a file in its personal folder,
1173 (without replace the existing one), i can't restore from
1174 my_file.txt to my_file.txt.old witch is very practical.
1177 Notes: I think we can enhance the where= option very easily by
1178 allowing regexp expression.
1180 Since, many users think that regexp are not user friendly, i think
1181 that bat, bconsole or brestore must provide a simple way to
1182 configure where= option (i think to something like in
1183 openoffice "search and replace").
1185 Ie, if user uses where=/tmp/bacula-restore, we keep the old
1188 If user uses something like where=s!/prod!/test!, files will
1189 be restored from /prod/xxx to /test/xxx.
1191 If user uses something like where=s/$/.old/, files will
1192 be restored from /prod/xxx.txt to /prod/xxx.txt.old.
1194 If user uses something like where=s/txt$/old.txt/, files will
1195 be restored from /prod/xxx.txt to /prod/xxx.old.txt
1197 if user uses something like where=s/([a-z]+)$/old.$1/, files will
1198 be restored from /prod/xxx.ext to /prod/xxx.old.ext
1200 Item n: Implement Catalog directive for Pool resource in Director
1202 Origin: Alan Davis adavis@ruckus.com
1206 What: The current behavior is for the director to create all pools
1207 found in the configuration file in all catalogs. Add a
1208 Catalog directive to the Pool resource to specify which
1209 catalog to use for each pool definition.
1211 Why: This allows different catalogs to have different pool
1212 attributes and eliminates the side-effect of adding
1213 pools to catalogs that don't need/use them.
1218 Item n: Implement NDMP protocol support
1223 What: Network Data Management Protocol is implemented by a number of
1224 NAS filer vendors to enable backups using third-party
1227 Why: This would allow NAS filer backups in Bacula without incurring
1228 the overhead of NFS or SBM/CIFS.
1230 Notes: Further information is available:
1232 http://www.ndmp.org/wp/wp.shtml
1233 http://www.traakan.com/ndmjob/index.html
1235 There are currently no viable open-source NDMP
1236 implementations. There is a reference SDK and example
1237 app available from ndmp.org but it has problems
1238 compiling on recent Linux and Solaris OS'. The ndmjob
1239 reference implementation from Traakan is known to
1240 compile on Solaris 10.
1242 Notes (Kern): I am not at all in favor of this until NDMP becomes
1243 an Open Standard or until there are Open Source libraries
1244 that interface to it.
1246 Item n: make changing "spooldata=yes|no" possible for
1247 manual/interactive jobs
1249 Origin: Marc Schiffbauer <marc@schiffbauer.net>
1251 Date: 12 April 2007)
1255 What: Make it possible to modify the spooldata option
1256 for a job when being run from within the console.
1257 Currently it is possible to modify the backup level
1258 and the spooldata setting in a Schedule resource.
1259 It is also possible to modify the backup level when using
1260 the "run" command in the console.
1261 But it is currently not possible to to the same
1262 with "spooldata=yes|no" like:
1264 run job=MyJob level=incremental spooldata=yes
1266 Why: In some situations it would be handy to be able to switch
1267 spooldata on or off for interactive/manual jobs based on
1268 which data the admin expects or how fast the LAN/WAN
1269 connection currently is.
1273 ============= Empty Feature Request form ===========
1274 Item n: One line summary ...
1275 Date: Date submitted
1276 Origin: Name and email of originator.
1279 What: More detailed explanation ...
1281 Why: Why it is important ...
1283 Notes: Additional notes or features (omit if not used)
1284 ============== End Feature Request form ==============
1292 <? require_once("inc/footer.php"); ?>