9 - Document cleaning up the spool files:
10 db, pid, state, bsr, mail, conmsg, spool
11 - Document the multiple-drive-changer.txt script.
12 - Pruning with Admin job.
13 - Does WildFile match against full name? Doc.
14 - %d and %v only valid on Director, not for ClientRunBefore/After.
15 - During tests with the 260 char fix code, I found one problem:
16 if the system "sees" a long path once, it seems to forget it's
17 working drive (e.g. c:\), which will lead to a problem during
18 the next job (create bootstrap file will fail). Here is the
19 workaround: specify absolute working and pid directory in
20 bacula-fd.conf (e.g. c:\bacula\working instead of
22 - Document techniques for restoring large numbers of files.
23 - Document setting my.cnf to big file usage.
24 - Add example of proper index output to doc.
26 - Correct the Include syntax in the m4.xxx files in examples/conf
31 - Remove queue.c code.
32 - Add bconsole option to use stdin/out instead of conio.
33 - Fix ClientRunBefore/AfterJob compatibility.
34 - Fix re-read of last block to check if job has actually written
35 a block, and check if block was written by a different job
36 (i.e. multiple simultaneous jobs writing).
37 - JobStatus and Termination codes.
38 - Some users claim that they must do two prune commands to get a
39 Volume marked as purged.
40 - Print warning message if LANG environment variable does not specify
42 - New dot commands from Arno.
43 .update volume [enabled|disabled|*see below]
44 > However, I could easily imagine an option to "update slots" that says
45 > "enable=yes|no" that would automatically enable or disable all the Volumes
46 > found in the autochanger. This will permit the user to optionally mark all
47 > the Volumes in the magazine disabled prior to taking them offsite, and mark
48 > them all enabled when bringing them back on site. Coupled with the options
49 > to the slots keyword, you can apply the enable/disable to any or all volumes.
50 .show device=xxx lists information from one storage device, including
51 devices (I'm not even sure that information exists in the DIR...)
52 .move eject device=xxx mostly the same as 'unmount xxx' but perhaps with
53 better machine-readable output like "Ok" or "Error busy"
54 .move eject device=xxx toslot=yyy the same as above, but with a new
55 target slot. The catalog should be updated accordingly.
56 .move transfer device=xxx fromslot=yyy toslot=zzz
59 - Get Perl replacement for bregex.c
61 - Given all the problems with FIFOs, I think the solution is to do something a
62 little different, though I will look at the code and see if there is not some
63 simple solution (i.e. some bug that was introduced). What might be a better
64 solution would be to use a FIFO as a sort of "key" to tell Bacula to read and
65 write data to a program rather than the FIFO. For example, suppose you
70 Then, I could imagine if you backup and restore this file with a direct
71 reference as is currently done for fifos, instead, during backup Bacula will
74 /home/kern/my-fifo.backup
76 and read the data that my-fifo.backup writes to stdout. For restore, Bacula
79 /home/kern/my-fifo.restore
81 and send the data backed up to stdout. These programs can either be an
82 executable or a shell script and they need only read/write to stdin/stdout.
84 I think this would give a lot of flexibility to the user without making any
85 significant changes to Bacula.
90 select FilenameId from Filename where Name='';
91 # Get list of all directories referenced in a Backup.
92 select Path.Path from Path,File where File.JobId=nnn and
93 File.FilenameId=(FilenameId-from-above) and File.PathId=Path.PathId
94 order by Path.Path ASC;
96 - Look into using Dart for testing
97 http://public.kitware.com/Dart/HTML/Index.shtml
99 - Look into replacing autotools with cmake
100 http://www.cmake.org/HTML/Index.html
102 === Migration from David ===
103 What I'd like to see:
106 Name = "<poolname>-migrate"
110 Migration Selection Type = LowestUtil | OldestVol | PoolOccupancy |
111 Client | PoolResidence | Volume | JobName | SQLquery
112 Migration Selection Pattern = "regexp"
113 Next Pool = <override>
116 There should be no need for a Level (migration is always Full, since you
117 don't calculate differential/incremental differences for migration),
118 Storage should be determined by the volume types in the pool, and Client
119 is really a selection issue. Migration should always occur to the
120 NextPool defined in the pool definition. If no nextpool is defined, the
121 job should end with a reason of "no place to go". If Next Pool statement
122 is present, we override the check in the pool definition and use the
125 Here's how I'd define Migration Selection Types:
128 Client -- Migrate data from selected client only. Migration Selection
129 Pattern regexp provides pattern to select client names, eg ^FS00* makes
130 all client names starting with FS00 eligible for migration.
132 Jobname -- Migration all jobs matching name. Migration Selection Pattern
133 regexp provides pattern to select jobnames existing in pool.
135 Volume -- Migrate all data on specified volumes. Migration Selection
136 Pattern regexp provides selection criteria for volumes to be migrated.
137 Volumes must exist in pool to be eligible for migration.
141 LowestUtil -- Identify the volume in the pool with the least data on it
142 and empty it. No Migration Selection Pattern required.
144 OldestVol -- Identify the LRU volume with data written, and empty it. No
145 Migration Selection Pattern required.
147 PoolOccupancy -- if pool occupancy exceeds <highmig>, migrate volumes
148 (starting with most full volumes) until pool occupancy drops below
149 <lowmig>. Pool highmig and lowmig values are in pool definition, no
150 Migration Selection Pattern required.
154 SQLQuery -- Migrate all jobuids returned by the supplied SQL query.
155 Migration Selection Pattern contains SQL query to execute; should return
156 a list of 1 or more jobuids to migrate.
158 PoolResidence -- Migrate data sitting in pool for longer than
159 PoolResidence value in pool definition. Migration Selection Pattern
160 optional; if specified, override value in pool definition (value in
164 [ possibly a Python event -- kes ]
166 - Mount on an Autochanger with no tape in the drive causes:
167 Automatically selected Storage: LTO-changer
168 Enter autochanger drive[0]: 0
169 3301 Issuing autochanger "loaded drive 0" command.
170 3302 Autochanger "loaded drive 0", result: nothing loaded.
171 3301 Issuing autochanger "loaded drive 0" command.
172 3302 Autochanger "loaded drive 0", result: nothing loaded.
173 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
174 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
175 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
176 If this is not a blank tape, try unmounting and remounting the Volume.
177 - If Drive 0 is blocked, and drive 1 is set "Autoselect=no", drive 1 will
179 - Autochanger did not change volumes.
180 select * from Storage;
181 +-----------+-------------+-------------+
182 | StorageId | Name | AutoChanger |
183 +-----------+-------------+-------------+
184 | 1 | LTO-changer | 0 |
185 +-----------+-------------+-------------+
186 05-May 03:50 roxie-sd: 3302 Autochanger "loaded drive 0", result is Slot 11.
187 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Warning: Director wanted Volume "LT
188 Current Volume "LT0-002" not acceptable because:
189 1997 Volume "LT0-002" not in catalog.
190 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Error: Autochanger Volume "LT0-002"
191 Setting InChanger to zero in catalog.
192 05-May 03:50 roxie-dir: Tibs.2006-05-05_03.05.02 Error: Unable to get Media record
194 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Error getting Volume i
195 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Job 530 canceled.
196 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: spool.c:249 Fatal appe
197 05-May 03:49 Tibs: Tibs.2006-05-05_03.05.02 Fatal error: c:\cygwin\home\kern\bacula
206 FirstWritten: 2006-05-05 03:11:54
207 LastWritten: 2006-05-05 03:50:23
208 LabelDate: 2005-12-26 16:52:40
219 VolRetention: 31,536,000
231 Note VolStatus is blank!!!!!
238 FirstWritten: 0000-00-00 00:00:00
239 LastWritten: 0000-00-00 00:00:00
240 LabelDate: 2005-12-26 16:52:40
251 VolRetention: 31,536,000
264 Automatically selected Storage: LTO-changer
265 Enter autochanger drive[0]: 0
266 3301 Issuing autochanger "loaded drive 0" command.
267 3302 Autochanger "loaded drive 0", result: nothing loaded.
268 3301 Issuing autochanger "loaded drive 0" command.
269 3302 Autochanger "loaded drive 0", result: nothing loaded.
270 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
271 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
273 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
274 If this is not a blank tape, try unmounting and remounting the Volume.
276 - Add VolumeState (enable, disable, archive)
277 - Add VolumeLock to prevent all but lock holder (SD) from updating
278 the Volume data (with the exception of VolumeState).
279 - The btape fill command does not seem to use the Autochanger
280 - Make Windows installer default to system disk drive.
281 - Look at using ioctl(FIOBMAP, ...) on Linux, and
282 DeviceIoControl(..., FSCTL_QUERY_ALLOCATED_RANGES, ...) on
283 Win32 for sparse files.
284 http://www.flexhex.com/docs/articles/sparse-files.phtml
285 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
286 - Directive: at <event> "command"
287 - Command: pycmd "command" generates "command" event. How to
288 attach to a specific job?
289 - Integrate Christopher's St. Bernard code.
290 - run_cmd() returns int should return JobId_t
291 - get_next_jobid_from_list() returns int should return JobId_t
292 - Document export LDFLAGS=-L/usr/lib64
293 - Don't attempt to restore from "Disabled" Volumes.
294 - Network error on Win32 should set Win32 error code.
295 - What happens when you rename a Disk Volume?
296 - Job retention period in a Pool (and hence Volume). The job would
298 - Detect resource deadlock in Migrate when same job wants to read
299 and write the same device.
300 - Queue warning/error messages during restore so that they
301 are reported at the end of the report rather than being
302 hidden in the file listing ...
303 - Look at -D_FORTIFY_SOURCE=2
304 - Add Win32 FileSet definition somewhere
305 - Look at fixing restore status stats in SD.
306 - Make selection of Database used in restore correspond to
308 - Look at using ioctl(FIMAP) and FIGETBSZ for sparse files.
309 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
310 - Implement a mode that says when a hard read error is
311 encountered, read many times (as it currently does), and if the
312 block cannot be read, skip to the next block, and try again. If
313 that fails, skip to the next file and try again, ...
315 create table LevelType (LevelType binary(1), LevelTypeLong tinyblob);
316 insert into LevelType (LevelType,LevelTypeLong) values
320 - Add ACL to restore only to original location.
321 - Show files/second in client status output.
322 - Add a recursive mark command (rmark) to restore.
323 - "Minimum Job Interval = nnn" sets minimum interval between Jobs
324 of the same level and does not permit multiple simultaneous
325 running of that Job (i.e. lets any previous invocation finish
326 before doing Interval testing).
327 - Look at simplifying File exclusions.
328 - New directive "Delete purged Volumes"
329 - new pool XXX with ScratchPoolId = MyScratchPool's PoolId and
330 let it fill itself, and RecyclePoolId = XXX's PoolId so I can
331 see if it become stable and I just have to supervise
333 - If I want to remove this pool, I set RecyclePoolId = MyScratchPool's
334 PoolId, and when it is empty remove it.
335 - Figure out how to recycle Scratch volumes back to the Scratch Pool.
337 - Allow Check Labels to be used with Bacula labels.
338 - "Resuming" a failed backup (lost line for example) by using the
339 failed backup as a sort of "base" job.
341 - Email to the user when the tape is about to need changing x
342 days before it needs changing.
343 - Command to show next tape that will be used for a job even
344 if the job is not scheduled.
345 - From: Arunav Mandal <amandal@trolltech.com>
346 1. When jobs are running and bacula for some reason crashes or if I do a
347 restart it remembers and jobs it was running before it crashed or restarted
348 as of now I loose all jobs if I restart it.
350 2. When spooling and in the midway if client is disconnected for instance a
351 laptop bacula completely discard the spool. It will be nice if it can write
352 that spool to tape so there will be some backups for that client if not all.
354 3. We have around 150 clients machines it will be nice to have a option to
355 upgrade all the client machines bacula version automatically.
357 4. Atleast one connection should be reserved for the bconsole so at heavy load
358 I should connect to the director via bconsole which at sometimes I can't
360 5. Another most important feature that is missing, say at 10am I manually
361 started backup of client abc and it was a full backup since client abc has
362 no backup history and at 10.30am bacula again automatically started backup of
363 client abc as that was in the schedule. So now we have 2 multiple Full
364 backups of the same client and if we again try to start a full backup of
365 client backup abc bacula won't complain. That should be fixed.
367 - Fix bpipe.c so that it does not modify results pointer.
368 ***FIXME*** calling sequence should be changed.
369 - For Windows disaster recovery see http://unattended.sf.net/
370 - regardless of the retention period, Bacula will not prune the
371 last Full, Diff, or Inc File data until a month after the
372 retention period for the last Full backup that was done.
373 - update volume=xxx --- add status=Full
374 - Remove old spool files on startup.
375 - Exclude SD spool/working directory.
376 - Refuse to prune last valid Full backup. Same goes for Catalog.
378 - Make a callback when Rerun failed levels is called.
379 - Give Python program access to Scheduled jobs.
380 - Add setting Volume State via Python.
381 - Python script to save with Python, not save, save with Bacula.
382 - Python script to do backup.
384 - Change the Priority, Client, Storage, JobStatus (error)
385 at the start of a job.
386 - Why is SpoolDirectory = /home/bacula/spool; not reported
387 as an error when writing a DVD?
388 - Make bootstrap file handle multiple MediaTypes (SD)
389 - Remove all old Device resource code in Dir and code to pass it
390 back in SD -- better, rework it to pass back device statistics.
391 - Check locking of resources -- be sure to lock devices where previously
392 resources were locked.
393 - The last part is left in the spool dir.
396 - In restore don't compare byte count on a raw device -- directory
397 entry does not contain bytes.
398 - To mark files as deleted, run essentially a Verify to disk, and
399 when a file is found missing (MarkId != JobId), then create
400 a new File record with FileIndex == -1. This could be done
401 by the FD at the same time as the backup.
405 MA = (last_MA * 3 + rate) / 4
406 rate = (bytes - last_bytes) / (runtime - last_runtime)
407 - Max Vols limit in Pool off by one?
408 - Implement Files/Bytes,... stats for restore job.
409 - Implement Total Bytes Written, ... for restore job.
410 - Despool attributes simultaneously with data in a separate
411 thread, rejoined at end of data spooling.
412 - 7. Implement new Console commands to allow offlining/reserving drives,
413 and possibly manipulating the autochanger (much asked for).
414 - Add start/end date editing in messages (%t %T, %e?) ...
415 - Add ClientDefs similar to JobDefs.
416 - Print more info when bextract -p accepts a bad block.
417 - Fix FD JobType to be set before RunBeforeJob in FD.
418 - Look at adding full Volume and Pool information to a Volume
419 label so that bscan can get *all* the info.
420 - If the user puts "Purge Oldest Volume = yes" or "Recycle Oldest Volume = yes"
421 and there is only one volume in the pool, refuse to do it -- otherwise
422 he fills the Volume, then immediately starts reusing it.
423 - Implement copies and stripes.
424 - Add history file to console.
425 - Each file on tape creates a JobMedia record. Peter has 4 million
426 files spread over 10000 tape files and four tapes. A restore takes
427 16 hours to build the restore list.
428 - Add and option to check if the file size changed during backup.
429 - Make sure SD deletes spool files on error exit.
430 - Delete old spool files when SD starts.
431 - When labeling tapes, if you enter 000026, Bacula uses
432 the tape index rather than the Volume name 000026.
433 - Add offline tape command to Bacula console.
435 Enter MediaId or Volume name: 32
436 Enter new Volume name: DLT-20Dec04
437 Automatically selected Pool: Default
438 Connecting to Storage daemon DLTDrive at 192.168.68.104:9103 ...
439 Sending relabel command from "DLT-28Jun03" to "DLT-20Dec04" ...
440 block.c:552 Write error at 0:0 on device /dev/nst0. ERR=Bad file descriptor.
441 Error writing final EOF to tape. This tape may not be readable.
442 dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
443 askdir.c:219 NULL Volume name. This shouldn't happen!!!
444 3912 Failed to label Volume: ERR=dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
445 Label command failed for Volume DLT-20Dec04.
446 Do not forget to mount the drive!!!
447 - Bug: if a job is manually scheduled to run later, it does not appear
448 in any status report and cannot be cancelled.
450 ==== Keeping track of deleted files ====
451 My "trick" for keeping track of deletions is the following.
452 Assuming the user turns on this option, after all the files
453 have been backed up, but before the job has terminated, the
454 FD will make a pass through all the files and send their
455 names to the DIR (*exactly* the same as what a Verify job
456 currently does). This will probably be done at the same
457 time the files are being sent to the SD avoiding a second
458 pass. The DIR will then compare that to what is stored in
459 the catalog. Any files in the catalog but not in what the
460 FD sent will receive a catalog File entry that indicates
461 that at that point in time the file was deleted.
463 During a restore, any file initially picked up by some
464 backup (Full, ...) then subsequently having a File entry
465 marked "delete" will be removed from the tree, so will not
466 be restored. If a file with the same name is later OK it
467 will be inserted in the tree -- this already happens. All
468 will be consistent except for possible changes during the
471 Since I'm on the subject, some of you may be wondering what
472 the utility of the in memory tree is if you are going to
473 restore everything (at least it comes up from time to time
474 on the list). Well, it is still *very* useful because it
475 allows only the last item found for a particular filename
476 (full path) to be entered into the tree, and thus if a file
477 is backed up 10 times, only the last copy will be restored.
478 I recently (last Friday) restored a complete directory, and
479 the Full and all the Differential and Incremental backups
480 spanned 3 Volumes. The first Volume was not even mounted
481 because all the files had been updated and hence backed up
482 since the Full backup was made. In this case, the tree
483 saved me a *lot* of time.
485 Make sure this information is stored on the tape too so
486 that it can be restored directly from the tape.
488 Comments from Martin Simmons (I think they are all covered):
489 Ok, that should cover the basics. There are few issues though:
491 - Restore will depend on the catalog. I think it is better to include the
492 extra data in the backup as well, so it can be seen by bscan and bextract.
494 - I'm not sure if it will preserve multiple hard links to the same inode. Or
495 maybe adding or removing links will cause the data to be dumped again?
497 - I'm not sure if it will handle renamed directories. Possibly it will work
498 by dumping the whole tree under a renamed directory?
500 - It remains to be seen how the backup performance of the DIR's will be
501 affected when comparing the catalog for a large filesystem.
505 How about introducing a Type = MgmtPolicy job type? That job type would
506 be responsible for scanning the Bacula environment looking for specific
507 conditions, and submitting the appropriate jobs for implementing said
511 Name = "Migration-Policy"
513 Policy Selection Job Type = Migrate
514 Scope = "<keyword> <operator> <regexp>"
515 Threshold = "<keyword> <operator> <regexp>"
516 Job Template = <template-name>
519 Where <keyword> is any legal job keyword, <operator> is a comparison
520 operator (=,<,>,!=, logical operators AND/OR/NOT) and <regexp> is a
521 appropriate regexp. I could see an argument for Scope and Threshold
522 being SQL queries if we want to support full flexibility. The
523 Migration-Policy job would then get scheduled as frequently as a site
524 felt necessary (suggested default: every 15 minutes).
529 Name = "Migration-Policy"
531 Policy Selection Job Type = Migration
533 Threshold = "Migration Selection Type = LowestUtil"
534 Job Template = "MigrationTemplate"
537 would select all pools for examination and generate a job based on
538 MigrationTemplate to automatically select the volume with the lowest
539 usage and migrate it's contents to the nextpool defined for that pool.
541 This policy abstraction would be really handy for adjusting the behavior
542 of Bacula according to site-selectable criteria (one thing that pops
543 into mind is Amanda's ability to automatically adjust backup levels
544 depending on various criteria).
550 - Add Pool/Storage override regression test.
551 - Add delete JobId to regression.
552 - Add a regression test for dbcheck.
553 - New test to add bscan to four-concurrent-jobs regression,
554 i.e. after the four-concurrent jobs zap the
555 database as is done in the bscan-test, then use bscan to
556 restore the database, do a restore and compare with the
558 - Add restore of specific JobId to regression (item 3
559 on the restore prompt)
560 - Add IPv6 to regression
561 - Add database test to regression. Test each function like delete,
564 - AntiVir can slow down backups on Win32 systems.
565 - Win32 systems with FAT32 can be much slower than NTFS for
566 more than 1000 files per directory.
570 - A HOLD command to stop all jobs from starting.
571 - A PAUSE command to pause all running jobs ==> release the
573 - Media Type = LTO,LTO-2,LTO-3
574 Media Type Read = LTO,LTO2,LTO3
575 Media Type Write = LTO2, LTO3
577 === From Carsten Menke <bootsy52@gmx.net>
579 Following is a list of what I think in the situations where I'm faced with,
580 could be a usefull enhancement to bacula, which I'm certain other users will
581 benefit from as well.
583 1. NextJob/NextJobs Directive within a Job Resource in the form of
584 NextJobs = job1,job2.
587 I currently solved the problem with running multiple jobs each after each
588 by setting the Max Wait Time for a job to 8 hours, and give
589 the jobs different Priorities. However, there scenarios where
590 1 Job is directly depending on another job, so if the former job fails,
591 the job after it needn't to be run
592 while maybe other jobs should run despite of that
595 A Backup Job and a Verify job, if the backup job fails there is no need to run
596 the verify job, as the backup job already failed. However, one may like
597 to backup the Catalog to disk despite of that the main backup job failed.
600 I see that this is related to the Event Handlers which are on the ToDo
601 list, also it is maybe a good idea to check for the return value and
602 execute different actions based on the return value
605 3. offline capability to bconsole
608 Currently I use a script which I execute within the last Job via the
609 RunAfterJob Directive, to release and eject the tape.
610 So I have to call bconsole "release=Storage-Name" and afterwards
611 mt -f /dev/nst0 eject to get the tape out.
613 If I have multiple Storage Devices, than these may not be /dev/nst0 and
614 I have to modify the script or call it with parameters etc.
615 This would actually not be needed, as everything is already defined
616 in bacula-sd.conf and if I can invoke bconsole with the
617 storage name via $1 in the script than I'm done and information is
620 4. %s for Storage Name added to the chars being substituted in "RunAfterJob"
624 For the reason mentioned in 3. to have the ability to call a
625 script with /scripts/foobar %s and in the script use $1
626 to pass the Storage Name to bconsole
628 5. Setting Volume State within a Job Resource
631 Instead of using "Maximum Volume Jobs" in the Pool Resource,
632 I would have the possibilty to define
633 in a Job Resource that after this certain job is run, the Volume State
634 should be set to "Volume State = Used", this give more flexibility (IMHO).
636 6. Localization of Bacula Messages
639 Unfortunatley many,many people I work with don't speak english very well.
640 So if at least the Reporting messages would be localized then they
641 would understand that they have to change the tape,etc. etc.
643 I volunteer to do the german translations, and if I can convince my wife also
644 french and Morre (western african language).
646 7. OK, this is evil, probably bound to security risks and maybe not possible
647 due to the design of bacula.
649 Implementation of Backtics ( `command` ) for shell comand execution to
650 the "Label Format" Directive.
654 Currently I have defined BACULA_DAY_OF_WEEK="day1|day2..." resulting in
655 Label Format = "HolyBackup-${BACULA_DAY_OF_WEEK[${WeekDay}]}". If I could
656 use backticks than I could use "Label Format = HolyBackup-`date +%A` to have
657 the localized name for the day of the week appended to the
658 format string. Then I have the tape labeled automatically with weekday
659 name in the correct language.
661 - Yes, that is surely the case. I probably should turn those into Warning
662 errors. In addition, you just made me think that it might not be bad to
663 add an option to check the file size after backing up the file and
664 report if it changes. This would be done as an option because it would
667 Kern, good idea. If you do do that, mention in the output: file
668 shrunk, or file expanded, just to make it obvious to the user
669 (without having to the refer to file size), just how the file size
672 Would this option be for all file, or just one file? Or a fileset?
673 - Make output from status use html table tags for nicely
674 presenting in a browser.
675 - Can one write tapes faster with 8192 byte block sizes?
676 - Document security problems with the same password for everyone in
677 rpm and Win32 releases.
678 - Browse generations of files.
679 - I've seen an error when my catalog's File table fills up. I
680 then have to recreate the File table with a larger maximum row
681 size. Relevant information is at
682 http://dev.mysql.com/doc/mysql/en/Full_table.html ; I think the
683 "Installing and Configuring MySQL" chapter should talk a bit
684 about this potential problem, and recommend a solution.
685 - For Solaris must use POSIX awk.
686 - Want speed of writing to tape while despooling.
687 - Supported autochanger:
695 Wangtek 6525ES (SCSI-1 QIC drive, 525MB), under Linux 2.4.something,
696 bacula 1.36.0/1 works with blocksize 16k INSIDE bacula-sd.conf.
697 - Add regex from http://www.pcre.org to Bacula for Win32.
698 - Use only shell tools no make in CDROM package.
699 - Include within include does it work?
700 - Implement a Pool of type Cleaning?
701 - Implement VolReadTime and VolWriteTime in SD
702 - Modify Backing up Your Database to include a bootstrap file.
703 - Think about making certain database errors fatal.
704 - Look at correcting the time jump in the scheduler for daylight
705 savings time changes.
706 - Add a "real" timer to network connections.
707 - Promote to Full = Time period
708 - Check dates entered by user for correctness (month/day/... ranges)
709 - Compress restore Volume listing by date and first file.
710 - Look at patches/bacula_db.b2z postgresql that loops during restore.
712 - Perhaps add read/write programs and/or plugins to FileSets.
713 - How to handle backing up portables ...
714 - Add some sort of guaranteed Interval for upgrading jobs.
715 - Can we write the state file after every job terminates? On Win32
716 the system crashes and the state file is not updated.
719 Documentation to do: (any release a little bit at a time)
720 - Doc to do unmount before removing magazine.
721 - Alternative to static linking "ldd prog" save all binaries listed,
722 restore them and point LD_LIBRARY_PATH to them.
723 - Document add "</dev/null >/dev/null 2>&1" to the bacula-fd command line
724 - Document query file format.
725 - Add more documentation for bsr files.
726 - Document problems with Verify and pruning.
727 - Document how to use multiple databases.
728 - VXA drives have a "cleaning required"
729 indicator, but Exabyte recommends preventive cleaning after every 75
732 In this context, it should be noted that Exabyte has a command-line
733 vxatool utility available for free download. (The current version is
734 vxatool-3.72.) It can get diagnostic info, read, write and erase tapes,
735 test the drive, unload tapes, change drive settings, flash new firmware,
737 Of particular interest in this context is that vxatool <device> -i will
738 report, among other details, the time since last cleaning in tape motion
739 minutes. This information can be retrieved (and settings changed, for
740 that matter) through the generic-SCSI device even when Bacula has the
741 regular tape device locked. (Needless to say, I don't recommend
742 changing tape settings while a job is running.)
743 - Lookup HP cleaning recommendations.
744 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
745 - Document doing table repair
748 ===================================
749 - Add macro expansions in JobDefs.
750 Run Before Job = "SomeFile %{Level} %{Client}"
751 Write Bootstrap="/some/dir/%{JobName}_%{Client}.bsr"
752 - Use non-blocking network I/O but if no data is available, use
754 - Use gather write() for network I/O.
755 - Autorestart on crash.
756 - Add bandwidth limiting.
757 - Add acks every once and a while from the SD to keep
758 the line from timing out.
759 - When an error in input occurs and conio beeps, you can back
760 up through the prompt.
761 - Detect fixed tape block mode during positioning by looking at
762 block numbers in btape "test". Possibly adjust in Bacula.
763 - Fix list volumes to output volume retention in some other
764 units, perhaps via a directive.
765 - If opening a tape in read/write mode fails attempt to open
766 it in read-only mode, and mark the tape for read only.
767 - Allow Simultaneous Priorities = yes => run up to Max concurrent jobs even
768 with multiple priorities.
769 - If you use restore replace=never, the directory attributes for
770 non-existent directories will not be restored properly.
772 - see lzma401.zip in others directory for new compression
774 - Minimal autochanger handling in Bacula and in btape.
775 - Look into how tar does not save sockets and the possiblity of
776 not saving them in Bacula (Martin Simmons reported this).
777 - Fix restore jobs so that multiple jobs can run if they
778 are not using the same tape(s).
779 - Allow the user to select JobType for manual pruning/purging.
780 - bscan does not put first of two volumes back with all info in
782 - Implement the FreeBSD nodump flag in chflags.
783 - Figure out how to make named console messages go only to that
784 console and to the non-restricted console (new console class?).
785 - Make restricted console prompt for password if *ask* is set or
786 perhaps if password is undefined.
787 - Implement "from ISO-date/time every x hours/days/weeks/months" in
790 ==== from Marc Schoechlin
791 - the help-command should be more verbose
792 (it should explain the paramters of the different
794 -> its time-comsuming to consult the manual anytime
795 you need a special parameter
796 -> maybe its more easy to maintain this, if the
797 descriptions of that commands are outsourced to
799 - the cd-command should allow complete paths
800 i.e. cd /foo/bar/foo/bar
801 -> if a customer mails me the path to a certain file,
802 its faster to enter the specified directory
803 - if the password is not configured in bconsole.conf
804 you should be asked for it.
805 -> sometimes you like to do restore on a customer-machine
806 which shouldnt know the password for bacula.
807 -> adding the password to the file favours admins
808 to forget to remove the password after usage
810 the protection of that file is less important
811 - long-listed-output of commands should be scrollable
812 like the unix more/less-command does
813 -> if someone runs 200 and more machines, the lists could
814 be a little long and complex
815 - command-output should be shown column by column
816 to reduce scrolling and to increase clarity
818 - lsmark should list the selected files with full
820 - wildcards for selecting and file and directories would be nice
821 - any actions should be interuptable with STRG+C
822 - command-expansion would be pretty cool
824 - When the replace Never option is set, new directory permissions
825 are not restored. See bug 213. To fix this requires creating a
826 list of newly restored directories so that those directory
827 permissions *can* be restored.
828 - Compaction of Disk space by "migrating" Volumes that have pruned
829 Jobs (what criteria? size, #jobs, time).
830 - Add prune all command
831 - Document fact that purge can destroy a part of a restore by purging
832 one volume while others remain valid -- perhaps mark Jobs.
833 - Add multiple-media-types.txt
834 - look at mxt-changer.html
835 - Make ? do a help command (no return needed).
836 - Implement restore directory.
837 - Document streams and how to implement them.
838 - Try not to re-backup a file if a new hard link is added.
839 - Add feature to backup hard links only, but not the data.
840 - Fix stream handling to be simpler.
841 - Add Priority and Bootstrap to Run a Job.
842 - Eliminate Restore "Run Restore Job" prompt by allowing new "run command
844 - Remove View FileSet button from Run a Job dialog.
845 - Handle prompt for restore job at end of Restore command.
846 - Add display of total selected files to Restore window.
847 - Add tree pane to left of window.
848 - Add progress meter.
849 - Max wait time or max run time causes seg fault -- see runtime-bug.txt
850 - Document writing to a CD/DVD with Bacula.
851 - Add a "base" package to the window installer for pthreadsVCE.dll
852 which is needed by all packages.
853 - Add message to user to check for fixed block size when the forward
854 space test fails in btape.
855 - When unmarking a directory check if all files below are unmarked and
856 then remove the + flag -- in the restore tree.
857 - Possibly implement: Action = Unmount Device="TapeDrive1" in Admin jobs.
858 - Setup lrrd graphs: (http://www.linpro.no/projects/lrrd/) Mike Acar.
859 - Revisit the question of multiple Volumes (disk) on a single device.
860 - Add a block copy option to bcopy.
861 - Investigate adding Mac Resource Forks.
862 - Finish work on Gnome restore GUI.
863 - Fix "llist jobid=xx" where no fileset or client exists.
864 - For each job type (Admin, Restore, ...) require only the really necessary
865 fields.- Pass Director resource name as an option to the Console.
866 - Add a "batch" mode to the Console (no unsolicited queries, ...).
867 - Add a .list all files in the restore tree (probably also a list all files)
868 Do both a long and short form.
869 - Allow browsing the catalog to see all versions of a file (with
870 stat data on each file).
871 - Restore attributes of directory if replace=never set but directory
873 - Use SHA1 on authentication if possible.
874 - See comtest-xxx.zip for Windows code to talk to USB.
875 - Add John's appended files:
876 Appended = { /files/server/logs/http/*log }
877 and such files would be treated as follows.On a FULL backup, they would
878 be backed up like any other file.On an INCREMENTAL backup, where a
879 previous INCREMENTAL or FULL was already in thecatalogue and the length
880 of the file wasgreater than the length of the last backup, only thedata
881 added since the last backup will be dumped.On an INCREMENTAL backup, if
882 the length of the file is less than thelength of the file with the same
883 name last backed up, the completefile is dumped.On Windows systems, with
884 creation date of files, we can be evensmarter about this and not count
885 entirely upon the length.On a restore, the full and all incrementals
886 since it will beapplied in sequence to restore the file.
887 - Check new HAVE_WIN32 open bits.
888 - Check if the tape has moved before writing.
889 - Handling removable disks -- see below:
890 - Keep track of tape use time, and report when cleaning is necessary.
891 - Add FromClient and ToClient keywords on restore command (or
892 BackupClient RestoreClient).
893 - Implement a JobSet, which groups any number of jobs. If the
894 JobSet is started, all the jobs are started together.
895 Allow Pool, Level, and Schedule overrides.
896 - Enhance cancel to timeout BSOCK packets after a specific delay.
897 - Do scheduling by UTC using gmtime_r() in run_conf, scheduler, and
898 ua_status.!!! Thanks to Alan Brown for this tip.
899 - Look at updating Volume Jobs so that Max Volume Jobs = 1 will work
900 correctly for multiple simultaneous jobs.
901 - Correct code so that FileSet MD5 is calculated for < and | filename
903 - Implement the Media record flag that indicates that the Volume does disk
905 - Implement VolAddr, which is used when Volume is addressed like a disk,
906 and form it from VolFile and VolBlock.
907 - Make multiple restore jobs for multiple media types specifying
908 the proper storage type.
909 - Fix fast block rejection (stored/read_record.c:118). It passes a null
910 pointer (rec) to try_repositioning().
911 - Look at extracting Win data from BackupRead.
912 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
913 which would take precidence over the Catalog "JobRetention".
914 - Implement Label Format in Add and Label console commands.
915 - Possibly up network buffers to 65K. Put on variable.
916 - Put email tape request delays on one or more variables. User wants
917 to cancel the job after a certain time interval. Maximum Mount Wait?
918 - Job, Client, Device, Pool, or Volume?
919 Is it possible to make this a directive which is *optional* in multiple
920 resources, like Level? If so, I think I'd make it an optional directive
921 in Job, Client, and Pool, with precedence such that Job overrides Client
922 which in turn overrides Pool.
924 - New Storage specifications:
925 - Want to write to multiple storage devices simultaneously
926 - Want to write to multiple storage devices sequentially (in one job)
927 - Want to read/write simultaneously
928 - Key is MediaType -- it must match
930 Passed to SD as a sort of BSR record called Storage Specification
934 MediaType -> Next MediaType
936 Device -> Next Device
938 Allow multiple Storage specifications
946 Allow Multiple Pool specifications (note, Pool currently
948 Allow Multiple MediaType specifications in Dir conf
949 Allow Multiple Device specifications in Dir conf
950 Perhaps keep this in a single SSR
951 Tie a Volume to a specific device by using a MediaType that
952 is contained in only one device.
953 In SD allow Device to have Multiple MediaTypes
955 - Ideas from Jerry Scharf:
956 First let's point out some big pluses that bacula has for this
958 more importantly it's active. Thank you so much for that
959 even more important, it's not flaky
960 it has an open access catalog, opening many possibilities
961 it's pushing toward heterogeneous systems capability
963 Macintosh file client
964 macs are an interesting niche, but I fear a server is a rathole
965 working bare iron recovery for windows
966 the option for inc/diff backups not reset on fileset revision
967 a) use both change and inode update time against base time
968 b) do the full catalog check (expensive but accurate)
969 sizing guide (how much system is needed to back up N systems/files)
970 consultants on using bacula in building a disaster recovery system
972 or how to get at fancy things that one could do with bacula
973 logwatch code for bacula logs (or similar)
974 linux distro inclusion of bacula (brings good and bad, but necessary)
975 win2k/XP server capability (icky but you asked)
976 support for Oracle database ??
978 - Look at adding SQL server and Exchange support for Windows.
979 - Make dev->file and dev->block_num signed integers so that -1 can
980 be an invalid value which happens with BSR.
981 - Create VolAddr for disk files in place of VolFile and VolBlock. This
982 is needed to properly specify ranges.
983 - Add progress of files/bytes to SD and FD.
984 - Print warning message if FileId > 4 billion
985 - do a "messages" before the first prompt in Console
986 - Client does not show busy during Estimate command.
987 - Implement Console mtx commands.
988 - Implement a Mount Command and an Unmount Command where
989 the users could specify a system command to be performed
990 to do the mount, after which Bacula could attempt to
991 read the device. This is for Removeable media such as a CDROM.
992 - Most likely, this mount command would be invoked explicitly
993 by the user using the current Console "mount" and "unmount"
994 commands -- the Storage Daemon would do the right thing
995 depending on the exact nature of the device.
996 - As with tape drives, when Bacula wanted a new removable
997 disk mounted, it would unmount the old one, and send a message
998 to the user, who would then use "mount" as described above
999 once he had actually inserted the disk.
1000 - Implement dump/print label to UA
1001 - Spool to disk only when the tape is full, then when a tape is hung move
1003 - bextract is sending everything to the log file ****FIXME****
1004 - Allow multiple Storage specifications (or multiple names on
1005 a single Storage specification) in the Job record. Thus a job
1006 can be backed up to a number of storage devices.
1007 - Implement some way for the File daemon to contact the Director
1008 to start a job or pass its DHCP obtained IP number.
1009 - Implement a query tape prompt/replace feature for a console
1010 - Copy console @ code to gnome2-console
1011 - Make AES the only encryption algorithm see
1012 http://csrc.nist.gov/CryptoToolkit/aes/). It's
1013 an officially adopted standard, has survived peer
1014 review, and provides keys up to 256 bits.
1015 - Take a careful look at SetACL http://setacl.sourceforge.net
1016 - Make tree walk routines like cd, ls, ... more user friendly
1017 by handling spaces better.
1018 - Make sure that Bacula rechecks the tape after the 20 min wait.
1019 - Set IO_NOWAIT on Bacula TCP/IP packets.
1020 - Try doing a raw partition backup and restore by mounting a
1022 - From Lars Kellers:
1023 Yes, it would allow to highly automatic the request for new tapes. If a
1024 tape is empty, bacula reads the barcodes (native or simulated), and if
1025 an unused tape is found, it runs the label command with all the
1026 necessary parameters.
1028 By the way can bacula automatically "move" an empty/purged volume say
1029 in the "short" pool to the "long" pool if this pool runs out of volume
1031 - What to do about "list files job=xxx".
1032 - Get and test MySQL 4.0
1033 - Look at how fuser works and /proc/PID/fd that is how Nic found the
1034 file descriptor leak in Bacula.
1035 - Implement WrapCounters in Counters.
1036 - Add heartbeat from FD to SD if hb interval expires.
1037 - Can we dynamically change FileSets?
1038 - If pool specified to label command and Label Format is specified,
1039 automatically generate the Volume name.
1040 - Why can't SQL do the filename sort for restore?
1041 - Add ExhautiveRestoreSearch
1042 - Look at the possibility of loading only the necessary
1043 data into the restore tree (i.e. do it one directory at a
1044 time as the user walks through the tree).
1045 - Possibly use the hash code if the user selects all for a restore command.
1046 - Fix "restore all" to bypass building the tree.
1047 - Prohibit backing up archive device (findlib/find_one.c:128)
1048 - Implement Release Device in the Job resource to unmount a drive.
1049 - Implement Acquire Device in the Job resource to mount a drive,
1050 be sure this works with admin jobs so that the user can get
1051 prompted to insert the correct tape. Possibly some way to say to
1052 run the job but don't save the files.
1053 - Make things like list where a file is saved case independent for
1056 - Use autochanger to handle multiple devices.
1057 - On Windows with very long path names, it may be impossible to create
1058 a file (and thus restore it) because the total length is too long.
1059 We must cd into the directory then create the file without the
1061 - Implement a Recycle command
1062 - Test a second language e.g. french.
1063 - Start working on Base jobs.
1064 - Implement UnsavedFiles DB record.
1065 - From Phil Stracchino:
1066 It would probably be a per-client option, and would be called
1067 something like, say, "Automatically purge obsoleted jobs". What it
1068 would do is, when you successfully complete a Differential backup of a
1069 client, it would automatically purge all Incremental backups for that
1070 client that are rendered redundant by that Differential. Likewise,
1071 when a Full backup on a client completed, it would automatically purge
1072 all Differential and Incremental jobs obsoleted by that Full backup.
1073 This would let people minimize the number of tapes they're keeping on
1074 hand without having to master the art of retention times.
1075 - When doing a Backup send all attributes back to the Director, who
1076 would then figure out what files have been deleted.
1077 - Currently in mount.c:236 the SD simply creates a Volume. It should have
1078 explicit permission to do so. It should also mark the tape in error
1079 if there is an error.
1080 - Cancel waiting for Client connect in SD if FD goes away.
1082 - Implement timeout in response() when it should come quickly.
1083 - Implement a Slot priority (loaded/not loaded).
1084 - Implement "vacation" Incremental only saves.
1085 - Implement create "FileSet"?
1086 - Add prefixlinks to where or not where absolute links to FD.
1087 - Issue message to mount a new tape before the rewind.
1088 - Simplified client job initiation for portables.
1089 - If SD cannot open a drive, make it periodically retry.
1090 - Add more of the config info to the tape label.
1092 - If tape is marked read-only, then try opening it read-only rather than
1093 failing, and remember that it cannot be written.
1094 - Refine SD waiting output:
1095 Device is being positioned
1096 > Device is being positioned for append
1097 > Device is being positioned to file x
1099 - Figure out some way to estimate output size and to avoid splitting
1100 a backup across two Volumes -- this could be useful for writing CDROMs
1101 where you really prefer not to have it split -- not serious.
1102 - Have SD compute MD5 or SHA1 and compare to what FD computes.
1103 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
1104 actual data on the Volume and compare it.
1105 - Implement Bacula plugins -- design API
1106 - Make bcopy read through bad tape records.
1107 - Program files (i.e. execute a program to read/write files).
1108 Pass read date of last backup, size of file last time.
1109 - Add Signature type to File DB record.
1110 - CD into subdirectory when open()ing files for backup to
1111 speed up things. Test with testfind().
1112 - Priority job to go to top of list.
1113 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
1114 - Implement some way for the Console to dynamically create a job.
1115 - Solaris -I on tar for include list
1116 - Need a verbose mode in restore, perhaps to bsr.
1117 - bscan without -v is too quiet -- perhaps show jobs.
1118 - Add code to reject whole blocks if not wanted on restore.
1119 - Check if we can increase Bacula FD priorty in Win2000
1120 - Make sure the MaxVolFiles is fully implemented in SD
1121 - Check if both CatalogFiles and UseCatalog are set to SD.
1122 - Possibly add email to Watchdog if drive is unmounted too
1123 long and a job is waiting on the drive.
1124 - After unmount, if restore job started, ask to mount.
1125 - Add UA rc and history files.
1126 - put termcap (used by console) in ./configure and
1127 allow -with-termcap-dir.
1128 - Fix Autoprune for Volumes to respect need for full save.
1129 - Compare tape to Client files (attributes, or attributes and data)
1130 - Make all database Ids 64 bit.
1131 - Allow console commands to detach or run in background.
1132 - Fix status delay on storage daemon during rewind.
1133 - Add SD message variables to control operator wait time
1134 - Maximum Operator Wait
1135 - Minimum Message Interval
1136 - Maximum Message Interval
1137 - Send Operator message when cannot read tape label.
1138 - Verify level=Volume (scan only), level=Data (compare of data to file).
1139 Verify level=Catalog, level=InitCatalog
1141 - Add keyword search to show command in Console.
1142 - Events : tape has more than xxx bytes.
1143 - Complete code in Bacula Resources -- this will permit
1144 reading a new config file at any time.
1145 - Handle ctl-c in Console
1146 - Implement script driven addition of File daemon to config files.
1147 - Think about how to make Bacula work better with File (non-tape) archives.
1148 - Write Unix emulator for Windows.
1149 - Put memory utilization in Status output of each daemon
1150 if full status requested or if some level of debug on.
1151 - Make database type selectable by .conf files i.e. at runtime
1152 - Set flag for uname -a. Add to Volume label.
1153 - Restore files modified after date
1154 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
1155 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
1156 - Timout a job or terminate if link goes down, or reopen link and query.
1157 - Concept of precious tapes (cannot be reused).
1158 - Make bcopy copy with a single tape drive.
1159 - Permit changing ownership during restore.
1162 > My suggestion: Add a feature on the systray menu-icon menu to request
1163 > an immediate backup now. This would be useful for laptop users who may
1164 > not be on the network when the regular scheduled backup is run.
1166 > My wife's suggestion: Add a setting to the win32 client to allow it to
1167 > shut down the machine after backup is complete (after, of course,
1168 > displaying a "System will shut down in one minute, click here to cancel"
1169 > warning dialog). This would be useful for sites that want user
1170 > woorkstations to be shut down overnight to save power.
1173 - Autolabel should be specified by DIR instead of SD.
1175 - Add media capacity
1176 - AutoScan (check checksum of tape)
1177 - Format command = "format /dev/nst0"
1181 - Seek resolution (usually corresponds to buffer size)
1182 - EODErrorCode=ENOSPC or code
1183 - Partial Read error code
1184 - Partial write error code
1185 - Nonformatted read error
1186 - Nonformatted write error
1187 - WriteProtected error
1191 - IgnoreCloseErrors=yes
1201 - FD sends unsaved file list to Director at end of job (see
1203 - File daemon should build list of files skipped, and then
1204 at end of save retry and report any errors.
1205 - Write a Storage daemon that uses pipes and
1206 standard Unix programs to write to the tape.
1208 - Need something that monitors the JCR queue and
1209 times out jobs by asking the deamons where they are.
1210 - Enhance Jmsg code to permit buffering and saving to disk.
1211 - device driver = "xxxx" for drives.
1212 - Verify from Volume
1213 - Ensure that /dev/null works
1214 - Need report class for messages. Perhaps
1215 report resource where report=group of messages
1216 - enhance scan_attrib and rename scan_jobtype, and
1217 fill in code for "since" option
1218 - Director needs a time after which the report status is sent
1219 anyway -- or better yet, a retry time for the job.
1220 - Don't reschedule a job if previous incarnation is still running.
1221 - Some way to automatically backup everything is needed????
1222 - Need a structure for pending actions:
1224 - termination status (part of buffered msgs?)
1226 Read, Write, Clean, Delete
1227 - Login to Bacula; Bacula users with different permissions:
1228 owner, group, user, quotas
1229 - Store info on each file system type (probably in the job header on tape.
1230 This could be the output of df; or perhaps some sort of /etc/mtab record.
1232 ========= ideas ===============
1233 From: "Jerry K. Schieffer" <jerry@skylinetechnology.com>
1234 To: <kern@sibbald.com>
1235 Subject: RE: [Bacula-users] future large programming jobs
1236 Date: Thu, 26 Feb 2004 11:34:54 -0600
1238 I noticed the subject thread and thought I would offer the following
1239 merely as sources of ideas, i.e. something to think about, not even as
1240 strong as a request. In my former life (before retiring) I often
1241 dealt with backups and storage management issues/products as a
1242 developer and as a consultant. I am currently migrating my personal
1243 network from amanda to bacula specifically because of the ability to
1244 cross media boundaries during storing backups.
1245 Are you familiar with the commercial product called ADSM (I think IBM
1246 now sells it under the Tivoli label)? It has a couple of interesting
1247 ideas that may apply to the following topics.
1249 1. Migration: Consider that when you need to restore a system, there
1250 may be pressure to hurry. If all the information for a single client
1251 can eventually end up on the same media (and in chronological order),
1252 the restore is facillitated by not having to search past information
1253 from other clients. ADSM has the concept of "client affinity" that
1254 may be associated with it's storage pools. It seems to me that this
1255 concept (as an optional feature) might fit in your architecture for
1258 ADSM also has the concept of defining one or more storage pools as
1259 "copy pools" (almost mirrors, but only in the sense of contents).
1260 These pools provide the ability to have duplicte data stored both
1261 onsite and offsite. The copy process can be scheduled to be handled
1262 by their storage manager during periods when there is no backup
1263 activity. Again, the migration process might be a place to consider
1264 implementing something like this.
1267 > It strikes me that it would be very nice to be able to do things
1269 > have the Job(s) backing up the machines run, and once they have all
1270 > completed, start a migration job to copy the data from disks Volumes
1272 > a tape library and then to offsite storage. Maybe this can already
1274 > done with some careful scheduling and Job prioritzation; the events
1275 > mechanism described below would probably make it very easy.
1277 This is the goal. In the first step (before events), you simply
1279 the Migration to tape later.
1281 2. Base jobs: In ADSM, each copy of each stored file is tracked in
1282 the database. Once a file (unique by path and metadata such as dates,
1283 size, ownership, etc.) is in a copy pool, no more copies are made. In
1284 other words, when you start ADSM, it begins like your concept of a
1285 base job. After that it is in the "incremental" mode. You can
1286 configure the number of "generations" of files to be retained, plus a
1287 retention date after which even old generations are purged. The
1288 database tracks the contents of media and projects the percentage of
1289 each volume that is valid. When the valid content of a volume drops
1290 below a configured percentage, the valid data are migrated to another
1291 volume and the old volume is marked as empty. Note, this requires
1292 ADSM to have an idea of the contents of a client, i.e. marking the
1293 database when an existing file was deleted, but this would solve your
1294 issue of restoring a client without restoring deleted files.
1296 This is pretty far from what bacula now does, but if you are going to
1297 rip things up for Base jobs,.....
1298 Also, the benefits of this are huge for very large shops, especially
1299 with media robots, but are a pain for shops with manual media
1303 > Base jobs sound pretty useful, but I'm not dying for them.
1305 Nobody is dying for them, but when you see what it does, you will die
1308 3. Restoring deleted files: Since I think my comments in (2) above
1309 have low probability of implementation, I'll also suggest that you
1310 could approach the issue of deleted files by a mechanism of having the
1311 fd report to the dir, a list of all files on the client for every
1312 backup job. The dir could note in the database entry for each file
1313 the date that the file was seen. Then if a restore as of date X takes
1314 place, only files that exist from before X until after X would be
1315 restored. Probably the major cost here is the extra date container in
1316 each row of the files table.
1318 Thanks for "listening". I hope some of this helps. If you want to
1319 contact me, please send me an email - I read some but not all of the
1320 mailing list traffic and might miss a reply there.
1322 Please accept my compliments for bacula. It is doing a great job for
1323 me!! I sympathize with you in the need to wrestle with excelence in
1324 execution vs. excelence in feature inclusion.
1329 ==============================
1332 - Design at hierarchial storage for Bacula. Migration and Clone.
1333 - Implement FSM (File System Modules).
1334 - Audit M_ error codes to ensure they are correct and consistent.
1335 - Add variable break characters to lex analyzer.
1336 Either a bit mask or a string of chars so that
1337 the caller can change the break characters.
1338 - Make a single T_BREAK to replace T_COMMA, etc.
1339 - Ensure that File daemon and Storage daemon can
1340 continue a save if the Director goes down (this
1341 is NOT currently the case). Must detect socket error,
1342 buffer messages for later.
1343 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
1344 - Add ability to backup to two Storage devices (two SD sessions) at
1345 the same time -- e.g. onsite, offsite.
1346 - Add the ability to consolidate old backup sets (basically do a restore
1347 to tape and appropriately update the catalog). Compress Volume sets.
1348 Might need to spool via file is only one drive is available.
1349 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
1350 someway to do so with every volume that has less than x% valid
1354 Migration: Move a backup from one Volume to another
1355 Clone: Copy a backup -- two Volumes
1357 Bacula Migration is based on Jobs (apparently Networker is file by file).
1359 Migration triggered by:
1363 Highwater mark (keep total size)
1368 ======================================================
1370 It is somewhat like a Full save becomes an incremental since
1371 the Base job (or jobs) plus other non-base files.
1373 - A Base backup is same as Full backup, just different type.
1374 - New BaseFiles table that contains:
1376 BaseJobId - Base JobId referenced for this FileId (needed ???)
1377 JobId - JobId currently running
1378 FileId - File not backed up, exists in Base Job
1379 FileIndex - FileIndex from Base Job.
1380 i.e. for each base file that exists but is not saved because
1381 it has not changed, the File daemon sends the JobId, BaseId,
1382 FileId, FileIndex back to the Director who creates the DB entry.
1383 - To initiate a Base save, the Director sends the FD
1384 the FileId, and full filename for each file in the Base.
1385 - When the FD finds a Base file, he requests the Director to
1386 send him the full File entry (stat packet plus MD5), or
1387 conversely, the FD sends it to the Director and the Director
1388 says yes or no. This can be quite rapid if the FileId is kept
1389 by the FD for each Base Filename.
1390 - It is probably better to have the comparison done by the FD
1391 despite the fact that the File entry must be sent across the
1393 - An alternative would be to send the FD the whole File entry
1394 from the start. The disadvantage is that it requires a lot of
1395 space. The advantage is that it requires less communications
1397 - The Job record must be updated to indicate that one or more
1399 - At end of Job, FD returns:
1400 1. Count of base files/bytes not written to tape (i.e. matches)
1401 2. Count of base file that were saved i.e. had changed.
1402 - No tape record would be written for a Base file that matches, in the
1403 same way that no tape record is written for Incremental jobs where
1404 the file is not saved because it is unchanged.
1405 - On a restore, all the Base file records must explicitly be
1406 found from the BaseFile tape. I.e. for each Full save that is marked
1407 to have one or more Base Jobs, search the BaseFile for all occurrences
1409 - An optimization might be to make the BaseFile have:
1415 This would avoid the need to explicitly fetch each File record for
1416 the Base job. The Base Job record will be fetched to get the
1417 VolSessionId and VolSessionTime.
1418 =========================================================
1421 ==========================================================
1423 For each Incremental job that is run, there may be files that
1424 were found but not saved because they were locked (this applies
1425 only to Windows). Such a system could send back to the Director
1426 a list of Unsaved files.
1428 - New UnSavedFiles table that contains:
1432 - Then in the next Incremental job, the list of Unsaved Files will be
1433 feed to the FD, who will ensure that they are explicitly chosen even
1434 if standard date/time check would not have selected them.
1435 =============================================================
1439 Multiple drive autochanger data: see Alan Brown
1440 mtx -f xxx unloadStorage Element 1 is Already Full(drive 0 was empty)
1441 Unloading Data Transfer Element into Storage Element 1...source Element
1442 Address 480 is Empty
1444 (drive 0 was empty and so was slot 1)
1445 > mtx -f xxx load 15 0
1446 no response, just returns to the command prompt when complete.
1447 > mtx -f xxx status Storage Changer /dev/changer:2 Drives, 60 Slots ( 2 Import/Export )
1448 Data Transfer Element 0:Full (Storage Element 15 Loaded):VolumeTag = HX001
1449 Data Transfer Element 1:Empty
1450 Storage Element 1:Empty
1451 Storage Element 2:Full :VolumeTag=HX002
1452 Storage Element 3:Full :VolumeTag=HX003
1453 Storage Element 4:Full :VolumeTag=HX004
1454 Storage Element 5:Full :VolumeTag=HX005
1455 Storage Element 6:Full :VolumeTag=HX006
1456 Storage Element 7:Full :VolumeTag=HX007
1457 Storage Element 8:Full :VolumeTag=HX008
1458 Storage Element 9:Full :VolumeTag=HX009
1459 Storage Element 10:Full :VolumeTag=HX010
1460 Storage Element 11:Empty
1461 Storage Element 12:Empty
1462 Storage Element 13:Empty
1463 Storage Element 14:Empty
1464 Storage Element 15:Empty
1465 Storage Element 16:Empty....
1466 Storage Element 28:Empty
1467 Storage Element 29:Full :VolumeTag=CLNU01L1
1468 Storage Element 30:Empty....
1469 Storage Element 57:Empty
1470 Storage Element 58:Full :VolumeTag=NEX261L2
1471 Storage Element 59 IMPORT/EXPORT:Empty
1472 Storage Element 60 IMPORT/EXPORT:Empty
1474 Unloading Data Transfer Element into Storage Element 15...done
1476 (just to verify it remembers where it came from, however it can be
1477 overrriden with mtx unload {slotnumber} to go to any storage slot.)
1479 There needs to be a table of drive # to devices somewhere - If there are
1480 multiple changers or drives there may not be a 1:1 correspondance between
1481 changer drive number and system device name - and depending on the way the
1482 drives are hooked up to scsi busses, they may not be linearly numbered
1483 from an offset point either.something like
1485 Autochanger drives = 2
1486 Autochanger drive 0 = /dev/nst1
1487 Autochanger drive 1 = /dev/nst2
1488 IMHO, it would be _safest_ to use explicit mtx unload commands at all
1489 times, not just for multidrive changers. For a 1 drive changer, that's
1495 MTX's manpage (1.2.15):
1496 unload [<slotnum>] [ <drivenum> ]
1497 Unloads media from drive <drivenum> into slot
1498 <slotnum>. If <drivenum> is omitted, defaults to
1499 drive 0 (as do all commands). If <slotnum> is
1500 omitted, defaults to the slot that the drive was
1501 loaded from. Note that there's currently no way
1502 to say 'unload drive 1's media to the slot it
1503 came from', other than to explicitly use that
1504 slot number as the destination.AB
1510 undef# camcontrol devlist
1511 <WANGTEK 51000 SCSI M74H 12B3> at scbus0 target 2 lun 0 (pass0,sa0)
1512 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 0 (pass1,sa1)
1513 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 1 (pass2)
1515 tapeinfo -f /dev/sg0 with a bad tape in drive 1:
1516 [kern@rufus mtx-1.2.17kes]$ ./tapeinfo -f /dev/sg0
1517 Product Type: Tape Drive
1519 Product ID: 'C5713A '
1521 Attached Changer: No
1522 TapeAlert[3]: Hard Error: Uncorrectable read/write error.
1523 TapeAlert[20]: Clean Now: The tape drive neads cleaning NOW.
1530 Medium Type: Not Loaded
1533 DataCompEnabled: yes
1534 DataCompCapable: yes
1535 DataDeCompEnabled: yes
1542 Handling removable disks
1544 From: Karl Cunningham <karlc@keckec.com>
1546 My backups are only to hard disk these days, in removable bays. This is my
1547 idea of how a backup to hard disk would work more smoothly. Some of these
1548 things Bacula does already, but I mention them for completeness. If others
1549 have better ways to do this, I'd like to hear about it.
1551 1. Accommodate several disks, rotated similar to how tapes are. Identified
1552 by partition volume ID or perhaps by the name of a subdirectory.
1553 2. Abort & notify the admin if the wrong disk is in the bay.
1554 3. Write backups to different subdirectories for each machine to be backed
1556 4. Volumes (files) get created as needed in the proper subdirectory, one
1558 5. When a disk is recycled, remove or zero all old backup files. This is
1559 important as the disk being recycled may be close to full. This may be
1560 better done manually since the backup files for many machines may be
1561 scattered in many subdirectories.
1566 - Make sure that all do_prompt() calls in Dir check for
1567 -1 (error) and -2 (cancel) returns.
1568 - Fix foreach_jcr() to have free_jcr() inside next().
1569 jcr=jcr_walk_start();
1570 for ( ; jcr; (jcr=jcr_walk_next(jcr)) )
1573 - A Volume taken from Scratch should take on the retention period
1575 - Correct doc for Maximum Changer Wait (and others) accepting only
1577 - Implement status that shows why a job is being held in reserve, or
1578 rather why none of the drives are suitable.
1579 - Implement a way to disable a drive (so you can use the second
1580 drive of an autochanger, and the first one will not be used or
1582 - Make sure Maximum Volumes is respected in Pools when adding
1583 Volumes (e.g. when pulling a Scratch volume).
1584 - Keep same dcr when switching device ...
1585 - Implement code that makes the Dir aware that a drive is an
1586 autochanger (so the user doesn't need to use the Autochanger = yes
1588 - Make catalog respect ACL.
1589 - Add recycle count to Media record.
1590 - Add initial write date to Media record.
1591 - Fix store_yesno to be store_bitmask.
1592 --- create_file.c.orig Fri Jul 8 12:13:05 2005
1593 +++ create_file.c Fri Jul 8 12:13:07 2005
1595 attr->ofname, be.strerror());
1598 + } else if(S_ISSOCK(attr->statp.st_mode)) {
1599 + Dmsg1(200, "Skipping socket: %s\n", attr->ofname);
1601 Dmsg1(200, "Restore node: %s\n", attr->ofname);
1602 if (mknod(attr->ofname, attr->statp.st_mode, attr->statp.st_rdev) != 0 && errno != EEXIST) {
1603 - Add true/false to conf same as yes/no
1604 - Reserve blocks other restore jobs when first cannot connect to SD.
1605 - Fix Maximum Changer Wait, Maximum Open Wait, Maximum Rewind Wait to
1606 accept time qualifiers.
1607 - Does ClientRunAfterJob fail the job on a bad return code?
1608 - Make hardlink code at line 240 of find_one.c use binary search.
1609 - Add ACL error messages in src/filed/acl.c.
1610 - Make authentication failures single threaded.
1611 - Make Dir and SD authentication errors single threaded.
1613 - Fix catreq.c digestbuf at line 411 in src/dird/catreq.c
1614 - Make base64.c (bin_to_base64) take a buffer length
1615 argument to avoid overruns.
1616 and verify that other buffers cannot overrun.
1617 - Implement VolumeState as discussed with Arno.
1618 - Add LocationId to update volume
1626 - Add Comment to Media record
1627 - Fix auth compatibility with 1.38
1628 - Update dbcheck to include Log table
1629 - Update llist to include new fields.