1 <? require_once("inc/header.php"); ?>
12 Bacula Projects Roadmap
13 Status updated 12 January 2007
16 Item 1: Accurate restoration of renamed/deleted files
17 Item 2: Implement a Bacula GUI/management tool.
18 Item 3: Implement Base jobs.
19 Item 4: Implement from-client and to-client on restore command line.
20 Item 5: Implement creation and maintenance of copy pools
21 Item 6: Merge multiple backups (Synthetic Backup or Consolidation).
22 Item 7: Deletion of Disk-Based Bacula Volumes
23 Item 8: Implement a Python interface to the Bacula catalog.
24 Item 9: Archival (removal) of User Files to Tape
25 Item 10: Add Plug-ins to the FileSet Include statements.
26 Item 11: Implement more Python events in Bacula.
27 Item 12: Quick release of FD-SD connection after backup.
28 Item 13: Implement huge exclude list support using hashing.
29 Item 14: Allow skipping execution of Jobs
30 Item 15: Tray monitor window cleanups
31 Item 16: Split documentation
32 Item 17: Automatic promotion of backup levels
33 Item 18: Add an override in Schedule for Pools based on backup types.
34 Item 10: An option to operate on all pools with update vol parameters
35 Item 20: Include JobID in spool file name
36 Item 21: Include timestamp of job launch in "stat clients" output
37 Item 22: Message mailing based on backup types
38 Item 23: Allow inclusion/exclusion of files in a fileset by creation/mod times
39 Item 24: Add a scheduling syntax that permits weekly rotations
40 Item 25: Improve Bacula's tape and drive usage and cleaning management.
41 Item 26: Implement support for stacking arbitrary stream filters, sinks.
42 Item 27: Allow FD to initiate a backup
43 Item 28: Directive/mode to backup only file changes, not entire file
44 Item 29: Automatic disabling of devices
45 Item 30: Incorporation of XACML2/SAML2 parsing
46 Item 31: Clustered file-daemons
47 Item 32: Commercial database support
49 Item 34: Filesystem watch triggered backup.
50 Item 35: Implement multiple numeric backup levels as supported by dump
51 Item 36: Implement a server-side compression feature
52 Item 37: Cause daemons to use a specific IP address to source communications
53 Item 38: Multiple threads in file daemon for the same job
54 Item 39: Restore only file attributes (permissions, ACL, owner, group...)
55 Item 40: Add an item to the restore option where you can select a pool
57 Below, you will find more information on future projects:
59 Item 1: Accurate restoration of renamed/deleted files
60 Date: 28 November 2005
61 Origin: Martin Simmons (martin at lispworks dot com)
62 Status: Robert Nelson will implement this
64 What: When restoring a fileset for a specified date (including "most
65 recent"), Bacula should give you exactly the files and directories
66 that existed at the time of the last backup prior to that date.
68 Currently this only works if the last backup was a Full backup.
69 When the last backup was Incremental/Differential, files and
70 directories that have been renamed or deleted since the last Full
71 backup are not currently restored correctly. Ditto for files with
72 extra/fewer hard links than at the time of the last Full backup.
74 Why: Incremental/Differential would be much more useful if this worked.
76 Notes: Merging of multiple backups into a single one seems to
77 rely on this working, otherwise the merged backups will not be
78 truly equivalent to a Full backup.
80 Kern: notes shortened. This can be done without the need for
81 inodes. It is essentially the same as the current Verify job,
82 but one additional database record must be written, which does
83 not need any database change.
85 Kern: see if we can correct restoration of directories if
86 replace=ifnewer is set. Currently, if the directory does not
87 exist, a "dummy" directory is created, then when all the files
88 are updated, the dummy directory is newer so the real values
91 Item 2: Implement a Bacula GUI/management tool.
96 What: Implement a Bacula console, and management tools
97 probably using Qt3 and C++.
99 Why: Don't we already have a wxWidgets GUI? Yes, but
100 it is written in C++ and changes to the user interface
101 must be hand tailored using C++ code. By developing
102 the user interface using Qt designer, the interface
103 can be very easily updated and most of the new Python
104 code will be automatically created. The user interface
105 changes become very simple, and only the new features
106 must be implement. In addition, the code will be in
107 Python, which will give many more users easy (or easier)
108 access to making additions or modifications.
110 Notes: There is a partial Python-GTK implementation
111 Lucas Di Pentima <lucas at lunix dot com dot ar> but
112 it is no longer being developed.
115 Item 3: Implement Base jobs.
116 Date: 28 October 2005
120 What: A base job is sort of like a Full save except that you
121 will want the FileSet to contain only files that are
122 unlikely to change in the future (i.e. a snapshot of
123 most of your system after installing it). After the
124 base job has been run, when you are doing a Full save,
125 you specify one or more Base jobs to be used. All
126 files that have been backed up in the Base job/jobs but
127 not modified will then be excluded from the backup.
128 During a restore, the Base jobs will be automatically
129 pulled in where necessary.
131 Why: This is something none of the competition does, as far as
132 we know (except perhaps BackupPC, which is a Perl program that
133 saves to disk only). It is big win for the user, it
134 makes Bacula stand out as offering a unique
135 optimization that immediately saves time and money.
136 Basically, imagine that you have 100 nearly identical
137 Windows or Linux machine containing the OS and user
138 files. Now for the OS part, a Base job will be backed
139 up once, and rather than making 100 copies of the OS,
140 there will be only one. If one or more of the systems
141 have some files updated, no problem, they will be
142 automatically restored.
144 Notes: Huge savings in tape usage even for a single machine.
145 Will require more resources because the DIR must send
146 FD a list of files/attribs, and the FD must search the
147 list and compare it for each file to be saved.
149 Item 4: Implement from-client and to-client on restore command line.
150 Date: 11 December 2006
151 Origin: Discussion on Bacula-users entitled 'Scripted restores to
152 different clients', December 2006
153 Status: New feature request
155 What: While using bconsole interactively, you can specify the client
156 that a backup job is to be restored for, and then you can
157 specify later a different client to send the restored files
158 back to. However, using the 'restore' command with all options
159 on the command line, this cannot be done, due to the ambiguous
160 'client' parameter. Additionally, this parameter means different
161 things depending on if it's specified on the command line or
162 afterwards, in the Modify Job screens.
164 Why: This feature would enable restore jobs to be more completely
165 automated, for example by a web or GUI front-end.
167 Notes: client can also be implied by specifying the jobid on the command
170 Item 5: Implement creation and maintenance of copy pools
171 Date: 27 November 2005
172 Origin: David Boyes (dboyes at sinenomine dot net)
175 What: I would like Bacula to have the capability to write copies
176 of backed-up data on multiple physical volumes selected
177 from different pools without transferring the data
178 multiple times, and to accept any of the copy volumes
179 as valid for restore.
181 Why: In many cases, businesses are required to keep offsite
182 copies of backup volumes, or just wish for simple
183 protection against a human operator dropping a storage
184 volume and damaging it. The ability to generate multiple
185 volumes in the course of a single backup job allows
186 customers to simple check out one copy and send it
187 offsite, marking it as out of changer or otherwise
188 unavailable. Currently, the library and magazine
189 management capability in Bacula does not make this process
192 Restores would use the copy of the data on the first
193 available volume, in order of copy pool chain definition.
195 This is also a major scalability issue -- as the number of
196 clients increases beyond several thousand, and the volume
197 of data increases, transferring the data multiple times to
198 produce additional copies of the backups will become
199 physically impossible due to transfer speed
200 issues. Generating multiple copies at server side will
201 become the only practical option.
203 How: I suspect that this will require adding a multiplexing
204 SD that appears to be a SD to a specific FD, but 1-n FDs
205 to the specific back end SDs managing the primary and copy
206 pools. Storage pools will also need to acquire parameters
207 to define the pools to be used for copies.
209 Notes: I would commit some of my developers' time if we can agree
210 on the design and behavior.
212 Item 6: Merge multiple backups (Synthetic Backup or Consolidation).
213 Origin: Marc Cousin and Eric Bollengier
214 Date: 15 November 2005
215 Status: Waiting implementation. Depends on first implementing
216 project Item 2 (Migration) which is now done.
218 What: A merged backup is a backup made without connecting to the Client.
219 It would be a Merge of existing backups into a single backup.
220 In effect, it is like a restore but to the backup medium.
222 For instance, say that last Sunday we made a full backup. Then
223 all week long, we created incremental backups, in order to do
224 them fast. Now comes Sunday again, and we need another full.
225 The merged backup makes it possible to do instead an incremental
226 backup (during the night for instance), and then create a merged
227 backup during the day, by using the full and incrementals from
228 the week. The merged backup will be exactly like a full made
229 Sunday night on the tape, but the production interruption on the
230 Client will be minimal, as the Client will only have to send
233 In fact, if it's done correctly, you could merge all the
234 Incrementals into single Incremental, or all the Incrementals
235 and the last Differential into a new Differential, or the Full,
236 last differential and all the Incrementals into a new Full
237 backup. And there is no need to involve the Client.
239 Why: The benefit is that :
240 - the Client just does an incremental ;
241 - the merged backup on tape is just as a single full backup,
242 and can be restored very fast.
244 This is also a way of reducing the backup data since the old
245 data can then be pruned (or not) from the catalog, possibly
246 allowing older volumes to be recycled
248 Item 7: Deletion of Disk-Based Bacula Volumes
250 Origin: Ross Boylan <RossBoylan at stanfordalumni dot org> (edited
254 What: Provide a way for Bacula to automatically remove Volumes
255 from the filesystem, or optionally to truncate them.
256 Obviously, the Volume must be pruned prior removal.
258 Why: This would allow users more control over their Volumes and
259 prevent disk based volumes from consuming too much space.
261 Notes: The following two directives might do the trick:
263 Volume Data Retention = <time period>
264 Remove Volume After = <time period>
266 The migration project should also remove a Volume that is
267 migrated. This might also work for tape Volumes.
269 Item 8: Implement a Python interface to the Bacula catalog.
270 Date: 28 October 2005
274 What: Implement an interface for Python scripts to access
275 the catalog through Bacula.
277 Why: This will permit users to customize Bacula through
280 Item 9: Archival (removal) of User Files to Tape
284 Origin: Ray Pengelly [ray at biomed dot queensu dot ca
287 What: The ability to archive data to storage based on certain parameters
288 such as age, size, or location. Once the data has been written to
289 storage and logged it is then pruned from the originating
290 filesystem. Note! We are talking about user's files and not
293 Why: This would allow fully automatic storage management which becomes
294 useful for large datastores. It would also allow for auto-staging
295 from one media type to another.
297 Example 1) Medical imaging needs to store large amounts of data.
298 They decide to keep data on their servers for 6 months and then put
299 it away for long term storage. The server then finds all files
300 older than 6 months writes them to tape. The files are then removed
303 Example 2) All data that hasn't been accessed in 2 months could be
304 moved from high-cost, fibre-channel disk storage to a low-cost
305 large-capacity SATA disk storage pool which doesn't have as quick of
306 access time. Then after another 6 months (or possibly as one
307 storage pool gets full) data is migrated to Tape.
309 Item 10: Add Plug-ins to the FileSet Include statements.
310 Date: 28 October 2005
312 Status: Partially coded in 1.37 -- much more to do.
314 What: Allow users to specify wild-card and/or regular
315 expressions to be matched in both the Include and
316 Exclude directives in a FileSet. At the same time,
317 allow users to define plug-ins to be called (based on
318 regular expression/wild-card matching).
320 Why: This would give the users the ultimate ability to control
321 how files are backed up/restored. A user could write a
322 plug-in knows how to backup his Oracle database without
323 stopping/starting it, for example.
325 Item 11: Implement more Python events in Bacula.
326 Date: 28 October 2005
330 What: Allow Python scripts to be called at more places
331 within Bacula and provide additional access to Bacula
334 Why: This will permit users to customize Bacula through
342 Also add a way to get a listing of currently running
343 jobs (possibly also scheduled jobs).
346 Item 12: Quick release of FD-SD connection after backup.
347 Origin: Frank Volf (frank at deze dot org)
348 Date: 17 November 2005
351 What: In the Bacula implementation a backup is finished after all data
352 and attributes are successfully written to storage. When using a
353 tape backup it is very annoying that a backup can take a day,
354 simply because the current tape (or whatever) is full and the
355 administrator has not put a new one in. During that time the
356 system cannot be taken off-line, because there is still an open
357 session between the storage daemon and the file daemon on the
360 Although this is a very good strategy for making "safe backups"
361 This can be annoying for e.g. laptops, that must remain
362 connected until the backup is completed.
364 Using a new feature called "migration" it will be possible to
365 spool first to harddisk (using a special 'spool' migration
366 scheme) and then migrate the backup to tape.
368 There is still the problem of getting the attributes committed.
369 If it takes a very long time to do, with the current code, the
370 job has not terminated, and the File daemon is not freed up. The
371 Storage daemon should release the File daemon as soon as all the
372 file data and all the attributes have been sent to it (the SD).
373 Currently the SD waits until everything is on tape and all the
374 attributes are transmitted to the Director before signaling
375 completion to the FD. I don't think I would have any problem
376 changing this. The reason is that even if the FD reports back to
377 the Dir that all is OK, the job will not terminate until the SD
378 has done the same thing -- so in a way keeping the SD-FD link
379 open to the very end is not really very productive ...
381 Why: Makes backup of laptops much faster.
385 Item 13: Implement huge exclude list support using hashing.
386 Date: 28 October 2005
390 What: Allow users to specify very large exclude list (currently
391 more than about 1000 files is too many).
393 Why: This would give the users the ability to exclude all
394 files that are loaded with the OS (e.g. using rpms
395 or debs). If the user can restore the base OS from
396 CDs, there is no need to backup all those files. A
397 complete restore would be to restore the base OS, then
398 do a Bacula restore. By excluding the base OS files, the
399 backup set will be *much* smaller.
402 Item 14: Allow skipping execution of Jobs
403 Date: 29 November 2005
404 Origin: Florian Schnabel <florian.schnabel at docufy dot de>
407 What: An easy option to skip a certain job on a certain date.
408 Why: You could then easily skip tape backups on holidays. Especially
409 if you got no autochanger and can only fit one backup on a tape
410 that would be really handy, other jobs could proceed normally
411 and you won't get errors that way.
414 Item 15: Tray monitor window cleanups
415 Origin: Alan Brown ajb2 at mssl dot ucl dot ac dot uk
418 What: Resizeable and scrollable windows in the tray monitor.
420 Why: With multiple clients, or with many jobs running, the displayed
421 window often ends up larger than the available screen, making
422 the trailing items difficult to read.
425 Item 16: Split documentation
426 Origin: Maxx <maxxatworkat gmail dot com>
430 What: Split documentation in several books
432 Why: Bacula manual has now more than 600 pages, and looking for
433 implementation details is getting complicated. I think
434 it would be good to split the single volume in two or
437 1) Introduction, requirements and tutorial, typically
438 are useful only until first installation time
440 2) Basic installation and configuration, with all the
441 gory details about the directives supported 3)
442 Advanced Bacula: testing, troubleshooting, GUI and
443 ancillary programs, security managements, scripting,
448 Item 17: Automatic promotion of backup levels
449 Date: 19 January 2006
450 Origin: Adam Thornton <athornton@sinenomine.net>
453 What: Amanda has a feature whereby it estimates the space that a
454 differential, incremental, and full backup would take. If the
455 difference in space required between the scheduled level and the next
456 level up is beneath some user-defined critical threshold, the backup
457 level is bumped to the next type. Doing this minimizes the number of
458 volumes necessary during a restore, with a fairly minimal cost in
461 Why: I know at least one (quite sophisticated and smart) user
462 for whom the absence of this feature is a deal-breaker in terms of
463 using Bacula; if we had it it would eliminate the one cool thing
464 Amanda can do and we can't (at least, the one cool thing I know of).
467 Item 18: Add an override in Schedule for Pools based on backup types.
469 Origin: Chad Slater <chad.slater@clickfox.com>
472 What: Adding a FullStorage=BigTapeLibrary in the Schedule resource
473 would help those of us who use different storage devices for different
474 backup levels cope with the "auto-upgrade" of a backup.
476 Why: Assume I add several new device to be backed up, i.e. several
477 hosts with 1TB RAID. To avoid tape switching hassles, incrementals are
478 stored in a disk set on a 2TB RAID. If you add these devices in the
479 middle of the month, the incrementals are upgraded to "full" backups,
480 but they try to use the same storage device as requested in the
481 incremental job, filling up the RAID holding the differentials. If we
482 could override the Storage parameter for full and/or differential
483 backups, then the Full job would use the proper Storage device, which
484 has more capacity (i.e. a 8TB tape library.
486 Item 19: An option to operate on all pools with update vol parameters
487 Origin: Dmitriy Pinchukov <absh@bossdev.kiev.ua>
491 What: When I do update -> Volume parameters -> All Volumes
492 from Pool, then I have to select pools one by one. I'd like
493 console to have an option like "0: All Pools" in the list of
496 Why: I have many pools and therefore unhappy with manually
497 updating each of them using update -> Volume parameters -> All
498 Volumes from Pool -> pool #.
502 Item 20: Include JobID in spool file name ****DONE****
503 Origin: Mark Bergman <mark.bergman@uphs.upenn.edu>
504 Date: Tue Aug 22 17:13:39 EDT 2006
505 Status: Done. (patches/testing/project-include-jobid-in-spool-name.patch)
506 No need to vote for this item.
508 What: Change the name of the spool file to include the JobID
510 Why: JobIDs are the common key used to refer to jobs, yet the
511 spoolfile name doesn't include that information. The date/time
512 stamp is useful (and should be retained).
516 Item 21: Include timestamp of job launch in "stat clients" output
517 Origin: Mark Bergman <mark.bergman@uphs.upenn.edu>
518 Date: Tue Aug 22 17:13:39 EDT 2006
521 What: The "stat clients" command doesn't include any detail on when
522 the active backup jobs were launched.
524 Why: Including the timestamp would make it much easier to decide whether
525 a job is running properly.
527 Notes: It may be helpful to have the output from "stat clients" formatted
528 more like that from "stat dir" (and other commands), in a column
529 format. The per-client information that's currently shown (level,
530 client name, JobId, Volume, pool, device, Files, etc.) is good, but
531 somewhat hard to parse (both programmatically and visually),
532 particularly when there are many active clients.
536 Item 22: Message mailing based on backup types
537 Origin: Evan Kaufman <evan.kaufman@gmail.com>
538 Date: January 6, 2006
541 What: In the "Messages" resource definitions, allowing messages
542 to be mailed based on the type (backup, restore, etc.) and level
543 (full, differential, etc) of job that created the originating
546 Why: It would, for example, allow someone's boss to be emailed
547 automatically only when a Full Backup job runs, so he can
548 retrieve the tapes for offsite storage, even if the IT dept.
549 doesn't (or can't) explicitly notify him. At the same time, his
550 mailbox wouldnt be filled by notifications of Verifies, Restores,
551 or Incremental/Differential Backups (which would likely be kept
554 Notes: One way this could be done is through additional message types, for example:
557 # email the boss only on full system backups
558 Mail = boss@mycompany.com = full, !incremental, !differential, !restore,
560 # email us only when something breaks
561 MailOnError = itdept@mycompany.com = all
565 Item 23: Allow inclusion/exclusion of files in a fileset by creation/mod times
566 Origin: Evan Kaufman <evan.kaufman@gmail.com>
567 Date: January 11, 2006
570 What: In the vein of the Wild and Regex directives in a Fileset's
571 Options, it would be helpful to allow a user to include or exclude
572 files and directories by creation or modification times.
574 You could factor the Exclude=yes|no option in much the same way it
575 affects the Wild and Regex directives. For example, you could exclude
576 all files modified before a certain date:
580 Modified Before = ####
583 Or you could exclude all files created/modified since a certain date:
587 Created Modified Since = ####
590 The format of the time/date could be done several ways, say the number
591 of seconds since the epoch:
592 1137008553 = Jan 11 2006, 1:42:33PM # result of `date +%s`
594 Or a human readable date in a cryptic form:
595 20060111134233 = Jan 11 2006, 1:42:33PM # YYYYMMDDhhmmss
597 Why: I imagine a feature like this could have many uses. It would
598 allow a user to do a full backup while excluding the base operating
599 system files, so if I installed a Linux snapshot from a CD yesterday,
600 I'll *exclude* all files modified *before* today. If I need to
601 recover the system, I use the CD I already have, plus the tape backup.
602 Or if, say, a Windows client is hit by a particularly corrosive
603 virus, and I need to *exclude* any files created/modified *since* the
606 Notes: Of course, this feature would work in concert with other
607 in/exclude rules, and wouldnt override them (or each other).
609 Notes: The directives I'd imagine would be along the lines of
610 "[Created] [Modified] [Before|Since] = <date>".
611 So one could compare against 'ctime' and/or 'mtime', but ONLY 'before'
615 Item 24: Add a scheduling syntax that permits weekly rotations
616 Date: 15 December 2006
617 Origin: Gregory Brauer (greg at wildbrain dot com)
620 What: Currently, Bacula only understands how to deal with weeks of the
621 month or weeks of the year in schedules. This makes it impossible
622 to do a true weekly rotation of tapes. There will always be a
623 discontinuity that will require disruptive manual intervention at
624 least monthly or yearly because week boundaries never align with
625 month or year boundaries.
627 A solution would be to add a new syntax that defines (at least)
628 a start timestamp, and repetition period.
630 Why: Rotated backups done at weekly intervals are useful, and Bacula
631 cannot currently do them without extensive hacking.
633 Notes: Here is an example syntax showing a 3-week rotation where full
634 Backups would be performed every week on Saturday, and an
635 incremental would be performed every week on Tuesday. Each
636 set of tapes could be removed from the loader for the following
637 two cycles before coming back and being reused on the third
638 week. Since the execution times are determined by intervals
639 from a given point in time, there will never be any issues with
640 having to adjust to any sort of arbitrary time boundary. In
641 the example provided, I even define the starting schedule
642 as crossing both a year and a month boundary, but the run times
643 would be based on the "Repeat" value and would therefore happen
648 Name = "Week 1 Rotation"
649 #Saturday. Would run Dec 30, Jan 20, Feb 10, etc.
653 Start = 2006-12-30 01:00
657 #Tuesday. Would run Jan 2, Jan 23, Feb 13, etc.
661 Start = 2007-01-02 01:00
668 Name = "Week 2 Rotation"
669 #Saturday. Would run Jan 6, Jan 27, Feb 17, etc.
673 Start = 2007-01-06 01:00
677 #Tuesday. Would run Jan 9, Jan 30, Feb 20, etc.
681 Start = 2007-01-09 01:00
688 Name = "Week 3 Rotation"
689 #Saturday. Would run Jan 13, Feb 3, Feb 24, etc.
693 Start = 2007-01-13 01:00
697 #Tuesday. Would run Jan 16, Feb 6, Feb 27, etc.
701 Start = 2007-01-16 01:00
708 Item 25: Improve Bacula's tape and drive usage and cleaning management.
709 Date: 8 November 2005, November 11, 2005
710 Origin: Adam Thornton <athornton at sinenomine dot net>,
711 Arno Lehmann <al at its-lehmann dot de>
714 What: Make Bacula manage tape life cycle information, tape reuse
715 times and drive cleaning cycles.
717 Why: All three parts of this project are important when operating
719 We need to know which tapes need replacement, and we need to
720 make sure the drives are cleaned when necessary. While many
721 tape libraries and even autoloaders can handle all this
722 automatically, support by Bacula can be helpful for smaller
723 (older) libraries and single drives. Limiting the number of
724 times a tape is used might prevent tape errors when using
725 tapes until the drives can't read it any more. Also, checking
726 drive status during operation can prevent some failures (as I
727 [Arno] had to learn the hard way...)
729 Notes: First, Bacula could (and even does, to some limited extent)
730 record tape and drive usage. For tapes, the number of mounts,
731 the amount of data, and the time the tape has actually been
732 running could be recorded. Data fields for Read and Write
733 time and Number of mounts already exist in the catalog (I'm
734 not sure if VolBytes is the sum of all bytes ever written to
735 that volume by Bacula). This information can be important
736 when determining which media to replace. The ability to mark
737 Volumes as "used up" after a given number of write cycles
738 should also be implemented so that a tape is never actually
739 worn out. For the tape drives known to Bacula, similar
740 information is interesting to determine the device status and
741 expected life time: Time it's been Reading and Writing, number
742 of tape Loads / Unloads / Errors. This information is not yet
743 recorded as far as I [Arno] know. A new volume status would
744 be necessary for the new state, like "Used up" or "Worn out".
745 Volumes with this state could be used for restores, but not
746 for writing. These volumes should be migrated first (assuming
747 migration is implemented) and, once they are no longer needed,
748 could be moved to a Trash pool.
750 The next step would be to implement a drive cleaning setup.
751 Bacula already has knowledge about cleaning tapes. Once it
752 has some information about cleaning cycles (measured in drive
753 run time, number of tapes used, or calender days, for example)
754 it can automatically execute tape cleaning (with an
755 autochanger, obviously) or ask for operator assistance loading
758 The final step would be to implement TAPEALERT checks not only
759 when changing tapes and only sending the information to the
760 administrator, but rather checking after each tape error,
761 checking on a regular basis (for example after each tape
762 file), and also before unloading and after loading a new tape.
763 Then, depending on the drives TAPEALERT state and the known
764 drive cleaning state Bacula could automatically schedule later
765 cleaning, clean immediately, or inform the operator.
767 Implementing this would perhaps require another catalog change
768 and perhaps major changes in SD code and the DIR-SD protocol,
769 so I'd only consider this worth implementing if it would
770 actually be used or even needed by many people.
772 Implementation of these projects could happen in three distinct
773 sub-projects: Measuring Tape and Drive usage, retiring
774 volumes, and handling drive cleaning and TAPEALERTs.
776 Item 26: Implement support for stacking arbitrary stream filters, sinks.
777 Date: 23 November 2006
778 Origin: Landon Fuller <landonf@threerings.net>
779 Status: Planning. Assigned to landonf.
781 What: Implement support for the following:
782 - Stacking arbitrary stream filters (eg, encryption, compression,
783 sparse data handling))
784 - Attaching file sinks to terminate stream filters (ie, write out
785 the resultant data to a file)
786 - Refactor the restoration state machine accordingly
788 Why: The existing stream implementation suffers from the following:
789 - All state (compression, encryption, stream restoration), is
790 global across the entire restore process, for all streams. There are
791 multiple entry and exit points in the restoration state machine, and
792 thus multiple places where state must be allocated, deallocated,
793 initialized, or reinitialized. This results in exceptional complexity
794 for the author of a stream filter.
795 - The developer must enumerate all possible combinations of filters
796 and stream types (ie, win32 data with encryption, without encryption,
797 with encryption AND compression, etc).
799 Notes: This feature request only covers implementing the stream filters/
800 sinks, and refactoring the file daemon's restoration implementation
801 accordingly. If I have extra time, I will also rewrite the backup
802 implementation. My intent in implementing the restoration first is to
803 solve pressing bugs in the restoration handling, and to ensure that
804 the new restore implementation handles existing backups correctly.
806 I do not plan on changing the network or tape data structures to
807 support defining arbitrary stream filters, but supporting that
808 functionality is the ultimate goal.
810 Assistance with either code or testing would be fantastic.
812 Item 27: Allow FD to initiate a backup
813 Origin: Frank Volf (frank at deze dot org)
814 Date: 17 November 2005
817 What: Provide some means, possibly by a restricted console that
818 allows a FD to initiate a backup, and that uses the connection
819 established by the FD to the Director for the backup so that
820 a Director that is firewalled can do the backup.
822 Why: Makes backup of laptops much easier.
824 Item 28: Directive/mode to backup only file changes, not entire file
825 Date: 11 November 2005
826 Origin: Joshua Kugler <joshua dot kugler at uaf dot edu>
827 Marek Bajon <mbajon at bimsplus dot com dot pl>
830 What: Currently when a file changes, the entire file will be backed up in
831 the next incremental or full backup. To save space on the tapes
832 it would be nice to have a mode whereby only the changes to the
833 file would be backed up when it is changed.
835 Why: This would save lots of space when backing up large files such as
836 logs, mbox files, Outlook PST files and the like.
838 Notes: This would require the usage of disk-based volumes as comparing
839 files would not be feasible using a tape drive.
841 Item 29: Automatic disabling of devices
843 Origin: Peter Eriksson <peter at ifm.liu dot se>
846 What: After a configurable amount of fatal errors with a tape drive
847 Bacula should automatically disable further use of a certain
848 tape drive. There should also be "disable"/"enable" commands in
851 Why: On a multi-drive jukebox there is a possibility of tape drives
852 going bad during large backups (needing a cleaning tape run,
853 tapes getting stuck). It would be advantageous if Bacula would
854 automatically disable further use of a problematic tape drive
855 after a configurable amount of errors has occurred.
857 An example: I have a multi-drive jukebox (6 drives, 380+ slots)
858 where tapes occasionally get stuck inside the drive. Bacula will
859 notice that the "mtx-changer" command will fail and then fail
860 any backup jobs trying to use that drive. However, it will still
861 keep on trying to run new jobs using that drive and fail -
862 forever, and thus failing lots and lots of jobs... Since we have
863 many drives Bacula could have just automatically disabled
864 further use of that drive and used one of the other ones
867 Item 30: Incorporation of XACML2/SAML2 parsing
868 Date: 19 January 2006
869 Origin: Adam Thornton <athornton@sinenomine.net>
872 What: XACML is "eXtensible Access Control Markup Language" and
873 "SAML is the "Security Assertion Markup Language"--an XML standard
874 for making statements about identity and authorization. Having these
875 would give us a framework to approach ACLs in a generic manner, and
876 in a way flexible enough to support the four major sorts of ACLs I
877 see as a concern to Bacula at this point, as well as (probably) to
878 deal with new sorts of ACLs that may appear in the future.
880 Why: Bacula is beginning to need to back up systems with ACLs
881 that do not map cleanly onto traditional Unix permissions. I see
882 four sets of ACLs--in general, mutually incompatible with one
883 another--that we're going to need to deal with. These are: NTFS
884 ACLs, POSIX ACLs, NFSv4 ACLS, and AFS ACLS. (Some may question the
885 relevance of AFS; AFS is one of Sine Nomine's core consulting
886 businesses, and having a reputable file-level backup and restore
887 technology for it (as Tivoli is probably going to drop AFS support
888 soon since IBM no longer supports AFS) would be of huge benefit to
889 our customers; we'd most likely create the AFS support at Sine Nomine
890 for inclusion into the Bacula (and perhaps some changes to the
891 OpenAFS volserver) core code.)
893 Now, obviously, Bacula already handles NTFS just fine. However, I
894 think there's a lot of value in implementing a generic ACL model, so
895 that it's easy to support whatever particular instances of ACLs come
896 down the pike: POSIX ACLS (think SELinux) and NFSv4 are the obvious
897 things arriving in the Linux world in a big way in the near future.
898 XACML, although overcomplicated for our needs, provides this
899 framework, and we should be able to leverage other people's
900 implementations to minimize the amount of work *we* have to do to get
901 a generic ACL framework. Basically, the costs of implementation are
902 high, but they're largely both external to Bacula and already sunk.
905 Item 31: Clustered file-daemons
906 Origin: Alan Brown ajb2 at mssl dot ucl dot ac dot uk
909 What: A "virtual" filedaemon, which is actually a cluster of real ones.
911 Why: In the case of clustered filesystems (SAN setups, GFS, or OCFS2, etc)
912 multiple machines may have access to the same set of filesystems
914 For performance reasons, one may wish to initate backups from
915 several of these machines simultaneously, instead of just using
916 one backup source for the common clustered filesystem.
918 For obvious reasons, normally backups of $A-FD/$PATH and
919 B-FD/$PATH are treated as different backup sets. In this case
920 they are the same communal set.
922 Likewise when restoring, it would be easier to just specify
923 one of the cluster machines and let bacula decide which to use.
925 This can be faked to some extent using DNS round robin entries
926 and a virtual IP address, however it means "status client" will
927 always give bogus answers. Additionally there is no way of
928 spreading the load evenly among the servers.
930 What is required is something similar to the storage daemon
931 autochanger directives, so that Bacula can keep track of
932 operating backups/restores and direct new jobs to a "free"
937 Item 32: Commercial database support
938 Origin: Russell Howe <russell_howe dot wreckage dot org>
942 What: It would be nice for the database backend to support more
943 databases. I'm thinking of SQL Server at the moment, but I guess Oracle,
944 DB2, MaxDB, etc are all candidates. SQL Server would presumably be
945 implemented using FreeTDS or maybe an ODBC library?
947 Why: We only really have one database server, which is MS SQL Server
948 2000. Maintaining a second one for the backup software (we grew out of
949 SQLite, which I liked, but which didn't work so well with our database
950 size). We don't really have a machine with the resources to run
951 postgres, and would rather only maintain a single DBMS. We're stuck with
952 SQL Server because pretty much all the company's custom applications
953 (written by consultants) are locked into SQL Server 2000. I can imagine
954 this scenario is fairly common, and it would be nice to use the existing
955 properly specced database server for storing Bacula's catalog, rather
956 than having to run a second DBMS.
959 Item 33: Archive data
961 Origin: calvin streeting calvin at absentdream dot com
964 What: The abilty to archive to media (dvd/cd) in a uncompressed format
965 for dead filing (archiving not backing up)
967 Why: At my works when jobs are finished and moved off of the main file
968 servers (raid based systems) onto a simple linux file server (ide based
969 system) so users can find old information without contacting the IT
972 So this data dosn't realy change it only gets added to,
973 But it also needs backing up. At the moment it takes
974 about 8 hours to back up our servers (working data) so
975 rather than add more time to existing backups i am trying
976 to implement a system where we backup the acrhive data to
977 cd/dvd these disks would only need to be appended to
978 (burn only new/changed files to new disks for off site
979 storage). basialy understand the differnce between
980 achive data and live data.
982 Notes: Scan the data and email me when it needs burning divide
983 into predifind chunks keep a recored of what is on what
984 disk make me a label (simple php->mysql=>pdf stuff) i
985 could do this bit ability to save data uncompresed so
986 it can be read in any other system (future proof data)
987 save the catalog with the disk as some kind of menu
990 Item 34: Filesystem watch triggered backup.
992 Origin: Jesper Krogh <jesper@krogh.cc>
993 Status: Unimplemented, depends probably on "client initiated backups"
995 What: With inotify and similar filesystem triggeret notification
996 systems is it possible to have the file-daemon to monitor
997 filesystem changes and initiate backup.
999 Why: There are 2 situations where this is nice to have.
1000 1) It is possible to get a much finer-grained backup than
1001 the fixed schedules used now.. A file created and deleted
1002 a few hours later, can automatically be caught.
1004 2) The introduced load on the system will probably be
1005 distributed more even on the system.
1007 Notes: This can be combined with configration that specifies
1008 something like: "at most every 15 minutes or when changes
1011 Kern Notes: I would rather see this implemented by an external program
1012 that monitors the Filesystem changes, then uses the console
1013 to start the appropriate job.
1015 Item 35: Implement multiple numeric backup levels as supported by dump
1017 Origin: Daniel Rich <drich@employees.org>
1019 What: Dump allows specification of backup levels numerically instead of just
1020 "full", "incr", and "diff". In this system, at any given level, all
1021 files are backed up that were were modified since the last backup of a
1022 higher level (with 0 being the highest and 9 being the lowest). A
1023 level 0 is therefore equivalent to a full, level 9 an incremental, and
1024 the levels 1 through 8 are varying levels of differentials. For
1025 bacula's sake, these could be represented as "full", "incr", and
1026 "diff1", "diff2", etc.
1028 Why: Support of multiple backup levels would provide for more advanced backup
1029 rotation schemes such as "Towers of Hanoi". This would allow better
1030 flexibility in performing backups, and can lead to shorter recover
1033 Notes: Legato Networker supports a similar system with full, incr, and 1-9 as
1036 Item 36: Implement a server-side compression feature
1037 Date: 18 December 2006
1038 Origin: Vadim A. Umanski , e-mail umanski@ext.ru
1040 What: The ability to compress backup data on server receiving data
1041 instead of doing that on client sending data.
1042 Why: The need is practical. I've got some machines that can send
1043 data to the network 4 or 5 times faster than compressing
1044 them (I've measured that). They're using fast enough SCSI/FC
1045 disk subsystems but rather slow CPUs (ex. UltraSPARC II).
1046 And the backup server has got a quite fast CPUs (ex. Dual P4
1047 Xeons) and quite a low load. When you have 20, 50 or 100 GB
1048 of raw data - running a job 4 to 5 times faster - that
1049 really matters. On the other hand, the data can be
1050 compressed 50% or better - so losing twice more space for
1051 disk backup is not good at all. And the network is all mine
1052 (I have a dedicated management/provisioning network) and I
1053 can get as high bandwidth as I need - 100Mbps, 1000Mbps...
1054 That's why the server-side compression feature is needed!
1057 Item 37: Cause daemons to use a specific IP address to source communications
1058 Origin: Bill Moran <wmoran@collaborativefusion.com>
1061 What: Cause Bacula daemons (dir, fd, sd) to always use the ip address
1062 specified in the [DIR|DF|SD]Addr directive as the source IP
1063 for initiating communication.
1064 Why: On complex networks, as well as extremely secure networks, it's
1065 not unusual to have multiple possible routes through the network.
1066 Often, each of these routes is secured by different policies
1067 (effectively, firewalls allow or deny different traffic depending
1068 on the source address)
1069 Unfortunately, it can sometimes be difficult or impossible to
1070 represent this in a system routing table, as the result is
1071 excessive subnetting that quickly exhausts available IP space.
1072 The best available workaround is to provide multiple IPs to
1073 a single machine that are all on the same subnet. In order
1074 for this to work properly, applications must support the ability
1075 to bind outgoing connections to a specified address, otherwise
1076 the operating system will always choose the first IP that
1077 matches the required route.
1078 Notes: Many other programs support this. For example, the following
1079 can be configured in BIND:
1080 query-source address 10.0.0.1;
1081 transfer-source 10.0.0.2;
1082 Which means queries from this server will always come from
1083 10.0.0.1 and zone transfers will always originate from
1086 Item 38: Multiple threads in file daemon for the same job
1087 Date: 27 November 2005
1088 Origin: Ove Risberg (Ove.Risberg at octocode dot com)
1091 What: I want the file daemon to start multiple threads for a backup
1092 job so the fastest possible backup can be made.
1094 The file daemon could parse the FileSet information and start
1095 one thread for each File entry located on a separate
1098 A confiuration option in the job section should be used to
1099 enable or disable this feature. The confgutration option could
1100 specify the maximum number of threads in the file daemon.
1102 If the theads could spool the data to separate spool files
1103 the restore process will not be much slower.
1105 Why: Multiple concurrent backups of a large fileserver with many
1106 disks and controllers will be much faster.
1108 Item 39: Restore only file attributes (permissions, ACL, owner, group...)
1109 Origin: Eric Bollengier
1113 What: The goal of this project is to be able to restore only rights
1114 and attributes of files without crushing them.
1116 Why: Who have never had to repair a chmod -R 777, or a wild update
1117 of recursive right under Windows? At this time, you must have
1118 enough space to restore data, dump attributes (easy with acl,
1119 more complex with unix/windows rights) and apply them to your
1120 broken tree. With this options, it will be very easy to compare
1121 right or ACL over the time.
1123 Notes: If the file is here, we skip restore and we change rights.
1124 If the file isn't here, we can create an empty one and apply
1125 rights or do nothing.
1127 Item 40: Add an item to the restore option where you can select a pool
1128 Origin: kshatriyak at gmail dot com
1132 What: In the restore option (Select the most recent backup for a
1133 client) it would be useful to add an option where you can limit
1134 the selection to a certain pool.
1136 Why: When using cloned jobs, most of the time you have 2 pools - a
1137 disk pool and a tape pool. People who have 2 pools would like to
1138 select the most recent backup from disk, not from tape (tape
1139 would be only needed in emergency). However, the most recent
1140 backup (which may just differ a second from the disk backup) may
1141 be on tape and would be selected. The problem becomes bigger if
1142 you have a full and differential - the most "recent" full backup
1143 may be on disk, while the most recent differential may be on tape
1144 (though the differential on disk may differ even only a second or
1145 so). Bacula will complain that the backups reside on different
1146 media then. For now the only solution now when restoring things
1147 when you have 2 pools is to manually search for the right
1148 job-id's and enter them by hand, which is a bit fault tolerant.
1150 ============= Empty Feature Request form ===========
1151 Item n: One line summary ...
1152 Date: Date submitted
1153 Origin: Name and email of originator.
1156 What: More detailed explanation ...
1158 Why: Why it is important ...
1160 Notes: Additional notes or features (omit if not used)
1161 ============== End Feature Request form ==============
1170 <? require_once("inc/footer.php"); ?>