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 - Possibly turn on St. Bernard code.
32 - Fix bextract to restore ACLs, or better yet, use common
34 - Do we migrate appendable Volumes?
35 - Remove queue.c code.
36 - Add bconsole option to use stdin/out instead of conio.
37 - Fix ClientRunBefore/AfterJob compatibility.
38 - Fix re-read of last block to check if job has actually written
39 a block, and check if block was written by a different job
40 (i.e. multiple simultaneous jobs writing).
41 - JobStatus and Termination codes.
42 - Some users claim that they must do two prune commands to get a
43 Volume marked as purged.
44 - Print warning message if LANG environment variable does not specify
46 - New dot commands from Arno.
47 .update volume [enabled|disabled|*see below]
48 > However, I could easily imagine an option to "update slots" that says
49 > "enable=yes|no" that would automatically enable or disable all the Volumes
50 > found in the autochanger. This will permit the user to optionally mark all
51 > the Volumes in the magazine disabled prior to taking them offsite, and mark
52 > them all enabled when bringing them back on site. Coupled with the options
53 > to the slots keyword, you can apply the enable/disable to any or all volumes.
54 .show device=xxx lists information from one storage device, including
55 devices (I'm not even sure that information exists in the DIR...)
56 .move eject device=xxx mostly the same as 'unmount xxx' but perhaps with
57 better machine-readable output like "Ok" or "Error busy"
58 .move eject device=xxx toslot=yyy the same as above, but with a new
59 target slot. The catalog should be updated accordingly.
60 .move transfer device=xxx fromslot=yyy toslot=zzz
63 - Get Perl replacement for bregex.c
65 - Given all the problems with FIFOs, I think the solution is to do something a
66 little different, though I will look at the code and see if there is not some
67 simple solution (i.e. some bug that was introduced). What might be a better
68 solution would be to use a FIFO as a sort of "key" to tell Bacula to read and
69 write data to a program rather than the FIFO. For example, suppose you
74 Then, I could imagine if you backup and restore this file with a direct
75 reference as is currently done for fifos, instead, during backup Bacula will
78 /home/kern/my-fifo.backup
80 and read the data that my-fifo.backup writes to stdout. For restore, Bacula
83 /home/kern/my-fifo.restore
85 and send the data backed up to stdout. These programs can either be an
86 executable or a shell script and they need only read/write to stdin/stdout.
88 I think this would give a lot of flexibility to the user without making any
89 significant changes to Bacula.
94 select FilenameId from Filename where Name='';
95 # Get list of all directories referenced in a Backup.
96 select Path.Path from Path,File where File.JobId=nnn and
97 File.FilenameId=(FilenameId-from-above) and File.PathId=Path.PathId
98 order by Path.Path ASC;
100 - Look into using Dart for testing
101 http://public.kitware.com/Dart/HTML/Index.shtml
103 - Look into replacing autotools with cmake
104 http://www.cmake.org/HTML/Index.html
106 === Migration from David ===
107 What I'd like to see:
110 Name = "<poolname>-migrate"
114 Migration Selection Type = LowestUtil | OldestVol | PoolOccupancy |
115 Client | PoolResidence | Volume | JobName | SQLquery
116 Migration Selection Pattern = "regexp"
117 Next Pool = <override>
120 There should be no need for a Level (migration is always Full, since you
121 don't calculate differential/incremental differences for migration),
122 Storage should be determined by the volume types in the pool, and Client
123 is really a selection issue. Migration should always occur to the
124 NextPool defined in the pool definition. If no nextpool is defined, the
125 job should end with a reason of "no place to go". If Next Pool statement
126 is present, we override the check in the pool definition and use the
129 Here's how I'd define Migration Selection Types:
132 Client -- Migrate data from selected client only. Migration Selection
133 Pattern regexp provides pattern to select client names, eg ^FS00* makes
134 all client names starting with FS00 eligible for migration.
136 Jobname -- Migration all jobs matching name. Migration Selection Pattern
137 regexp provides pattern to select jobnames existing in pool.
139 Volume -- Migrate all data on specified volumes. Migration Selection
140 Pattern regexp provides selection criteria for volumes to be migrated.
141 Volumes must exist in pool to be eligible for migration.
145 LowestUtil -- Identify the volume in the pool with the least data on it
146 and empty it. No Migration Selection Pattern required.
148 OldestVol -- Identify the LRU volume with data written, and empty it. No
149 Migration Selection Pattern required.
151 PoolOccupancy -- if pool occupancy exceeds <highmig>, migrate volumes
152 (starting with most full volumes) until pool occupancy drops below
153 <lowmig>. Pool highmig and lowmig values are in pool definition, no
154 Migration Selection Pattern required.
158 SQLQuery -- Migrate all jobuids returned by the supplied SQL query.
159 Migration Selection Pattern contains SQL query to execute; should return
160 a list of 1 or more jobuids to migrate.
162 PoolResidence -- Migrate data sitting in pool for longer than
163 PoolResidence value in pool definition. Migration Selection Pattern
164 optional; if specified, override value in pool definition (value in
168 [ possibly a Python event -- kes ]
170 - Mount on an Autochanger with no tape in the drive causes:
171 Automatically selected Storage: LTO-changer
172 Enter autochanger drive[0]: 0
173 3301 Issuing autochanger "loaded drive 0" command.
174 3302 Autochanger "loaded drive 0", result: nothing loaded.
175 3301 Issuing autochanger "loaded drive 0" command.
176 3302 Autochanger "loaded drive 0", result: nothing loaded.
177 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
178 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
179 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
180 If this is not a blank tape, try unmounting and remounting the Volume.
181 - If Drive 0 is blocked, and drive 1 is set "Autoselect=no", drive 1 will
183 - Autochanger did not change volumes.
184 select * from Storage;
185 +-----------+-------------+-------------+
186 | StorageId | Name | AutoChanger |
187 +-----------+-------------+-------------+
188 | 1 | LTO-changer | 0 |
189 +-----------+-------------+-------------+
190 05-May 03:50 roxie-sd: 3302 Autochanger "loaded drive 0", result is Slot 11.
191 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Warning: Director wanted Volume "LT
192 Current Volume "LT0-002" not acceptable because:
193 1997 Volume "LT0-002" not in catalog.
194 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Error: Autochanger Volume "LT0-002"
195 Setting InChanger to zero in catalog.
196 05-May 03:50 roxie-dir: Tibs.2006-05-05_03.05.02 Error: Unable to get Media record
198 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Error getting Volume i
199 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Job 530 canceled.
200 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: spool.c:249 Fatal appe
201 05-May 03:49 Tibs: Tibs.2006-05-05_03.05.02 Fatal error: c:\cygwin\home\kern\bacula
210 FirstWritten: 2006-05-05 03:11:54
211 LastWritten: 2006-05-05 03:50:23
212 LabelDate: 2005-12-26 16:52:40
223 VolRetention: 31,536,000
235 Note VolStatus is blank!!!!!
242 FirstWritten: 0000-00-00 00:00:00
243 LastWritten: 0000-00-00 00:00:00
244 LabelDate: 2005-12-26 16:52:40
255 VolRetention: 31,536,000
268 Automatically selected Storage: LTO-changer
269 Enter autochanger drive[0]: 0
270 3301 Issuing autochanger "loaded drive 0" command.
271 3302 Autochanger "loaded drive 0", result: nothing loaded.
272 3301 Issuing autochanger "loaded drive 0" command.
273 3302 Autochanger "loaded drive 0", result: nothing loaded.
274 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
275 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
277 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
278 If this is not a blank tape, try unmounting and remounting the Volume.
280 - Add VolumeState (enable, disable, archive)
281 - Add VolumeLock to prevent all but lock holder (SD) from updating
282 the Volume data (with the exception of VolumeState).
283 - The btape fill command does not seem to use the Autochanger
284 - Make Windows installer default to system disk drive.
285 - Look at using ioctl(FIOBMAP, ...) on Linux, and
286 DeviceIoControl(..., FSCTL_QUERY_ALLOCATED_RANGES, ...) on
287 Win32 for sparse files.
288 http://www.flexhex.com/docs/articles/sparse-files.phtml
289 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
290 - Directive: at <event> "command"
291 - Command: pycmd "command" generates "command" event. How to
292 attach to a specific job?
293 - Integrate Christopher's St. Bernard code.
294 - run_cmd() returns int should return JobId_t
295 - get_next_jobid_from_list() returns int should return JobId_t
296 - Document export LDFLAGS=-L/usr/lib64
297 - Don't attempt to restore from "Disabled" Volumes.
298 - Network error on Win32 should set Win32 error code.
299 - What happens when you rename a Disk Volume?
300 - Job retention period in a Pool (and hence Volume). The job would
302 - Detect resource deadlock in Migrate when same job wants to read
303 and write the same device.
304 - Queue warning/error messages during restore so that they
305 are reported at the end of the report rather than being
306 hidden in the file listing ...
307 - Look at -D_FORTIFY_SOURCE=2
308 - Add Win32 FileSet definition somewhere
309 - Look at fixing restore status stats in SD.
310 - Make selection of Database used in restore correspond to
312 - Look at using ioctl(FIMAP) and FIGETBSZ for sparse files.
313 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
314 - Implement a mode that says when a hard read error is
315 encountered, read many times (as it currently does), and if the
316 block cannot be read, skip to the next block, and try again. If
317 that fails, skip to the next file and try again, ...
319 create table LevelType (LevelType binary(1), LevelTypeLong tinyblob);
320 insert into LevelType (LevelType,LevelTypeLong) values
324 - Add ACL to restore only to original location.
325 - Show files/second in client status output.
326 - Add a recursive mark command (rmark) to restore.
327 - "Minimum Job Interval = nnn" sets minimum interval between Jobs
328 of the same level and does not permit multiple simultaneous
329 running of that Job (i.e. lets any previous invocation finish
330 before doing Interval testing).
331 - Look at simplifying File exclusions.
332 - New directive "Delete purged Volumes"
333 - new pool XXX with ScratchPoolId = MyScratchPool's PoolId and
334 let it fill itself, and RecyclePoolId = XXX's PoolId so I can
335 see if it become stable and I just have to supervise
337 - If I want to remove this pool, I set RecyclePoolId = MyScratchPool's
338 PoolId, and when it is empty remove it.
339 - Figure out how to recycle Scratch volumes back to the Scratch Pool.
341 - Allow Check Labels to be used with Bacula labels.
342 - "Resuming" a failed backup (lost line for example) by using the
343 failed backup as a sort of "base" job.
345 - Email to the user when the tape is about to need changing x
346 days before it needs changing.
347 - Command to show next tape that will be used for a job even
348 if the job is not scheduled.
349 - From: Arunav Mandal <amandal@trolltech.com>
350 1. When jobs are running and bacula for some reason crashes or if I do a
351 restart it remembers and jobs it was running before it crashed or restarted
352 as of now I loose all jobs if I restart it.
354 2. When spooling and in the midway if client is disconnected for instance a
355 laptop bacula completely discard the spool. It will be nice if it can write
356 that spool to tape so there will be some backups for that client if not all.
358 3. We have around 150 clients machines it will be nice to have a option to
359 upgrade all the client machines bacula version automatically.
361 4. Atleast one connection should be reserved for the bconsole so at heavy load
362 I should connect to the director via bconsole which at sometimes I can't
364 5. Another most important feature that is missing, say at 10am I manually
365 started backup of client abc and it was a full backup since client abc has
366 no backup history and at 10.30am bacula again automatically started backup of
367 client abc as that was in the schedule. So now we have 2 multiple Full
368 backups of the same client and if we again try to start a full backup of
369 client backup abc bacula won't complain. That should be fixed.
371 - Fix bpipe.c so that it does not modify results pointer.
372 ***FIXME*** calling sequence should be changed.
373 - For Windows disaster recovery see http://unattended.sf.net/
374 - regardless of the retention period, Bacula will not prune the
375 last Full, Diff, or Inc File data until a month after the
376 retention period for the last Full backup that was done.
377 - update volume=xxx --- add status=Full
378 - Remove old spool files on startup.
379 - Exclude SD spool/working directory.
380 - Refuse to prune last valid Full backup. Same goes for Catalog.
382 - Make a callback when Rerun failed levels is called.
383 - Give Python program access to Scheduled jobs.
384 - Add setting Volume State via Python.
385 - Python script to save with Python, not save, save with Bacula.
386 - Python script to do backup.
388 - Change the Priority, Client, Storage, JobStatus (error)
389 at the start of a job.
390 - Why is SpoolDirectory = /home/bacula/spool; not reported
391 as an error when writing a DVD?
392 - Make bootstrap file handle multiple MediaTypes (SD)
393 - Remove all old Device resource code in Dir and code to pass it
394 back in SD -- better, rework it to pass back device statistics.
395 - Check locking of resources -- be sure to lock devices where previously
396 resources were locked.
397 - The last part is left in the spool dir.
400 - In restore don't compare byte count on a raw device -- directory
401 entry does not contain bytes.
402 - To mark files as deleted, run essentially a Verify to disk, and
403 when a file is found missing (MarkId != JobId), then create
404 a new File record with FileIndex == -1. This could be done
405 by the FD at the same time as the backup.
409 MA = (last_MA * 3 + rate) / 4
410 rate = (bytes - last_bytes) / (runtime - last_runtime)
411 - Max Vols limit in Pool off by one?
412 - Implement Files/Bytes,... stats for restore job.
413 - Implement Total Bytes Written, ... for restore job.
414 - Despool attributes simultaneously with data in a separate
415 thread, rejoined at end of data spooling.
416 - 7. Implement new Console commands to allow offlining/reserving drives,
417 and possibly manipulating the autochanger (much asked for).
418 - Add start/end date editing in messages (%t %T, %e?) ...
419 - Add ClientDefs similar to JobDefs.
420 - Print more info when bextract -p accepts a bad block.
421 - Fix FD JobType to be set before RunBeforeJob in FD.
422 - Look at adding full Volume and Pool information to a Volume
423 label so that bscan can get *all* the info.
424 - If the user puts "Purge Oldest Volume = yes" or "Recycle Oldest Volume = yes"
425 and there is only one volume in the pool, refuse to do it -- otherwise
426 he fills the Volume, then immediately starts reusing it.
427 - Implement copies and stripes.
428 - Add history file to console.
429 - Each file on tape creates a JobMedia record. Peter has 4 million
430 files spread over 10000 tape files and four tapes. A restore takes
431 16 hours to build the restore list.
432 - Add and option to check if the file size changed during backup.
433 - Make sure SD deletes spool files on error exit.
434 - Delete old spool files when SD starts.
435 - When labeling tapes, if you enter 000026, Bacula uses
436 the tape index rather than the Volume name 000026.
437 - Add offline tape command to Bacula console.
439 Enter MediaId or Volume name: 32
440 Enter new Volume name: DLT-20Dec04
441 Automatically selected Pool: Default
442 Connecting to Storage daemon DLTDrive at 192.168.68.104:9103 ...
443 Sending relabel command from "DLT-28Jun03" to "DLT-20Dec04" ...
444 block.c:552 Write error at 0:0 on device /dev/nst0. ERR=Bad file descriptor.
445 Error writing final EOF to tape. This tape may not be readable.
446 dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
447 askdir.c:219 NULL Volume name. This shouldn't happen!!!
448 3912 Failed to label Volume: ERR=dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
449 Label command failed for Volume DLT-20Dec04.
450 Do not forget to mount the drive!!!
451 - Bug: if a job is manually scheduled to run later, it does not appear
452 in any status report and cannot be cancelled.
454 ==== Keeping track of deleted files ====
455 My "trick" for keeping track of deletions is the following.
456 Assuming the user turns on this option, after all the files
457 have been backed up, but before the job has terminated, the
458 FD will make a pass through all the files and send their
459 names to the DIR (*exactly* the same as what a Verify job
460 currently does). This will probably be done at the same
461 time the files are being sent to the SD avoiding a second
462 pass. The DIR will then compare that to what is stored in
463 the catalog. Any files in the catalog but not in what the
464 FD sent will receive a catalog File entry that indicates
465 that at that point in time the file was deleted.
467 During a restore, any file initially picked up by some
468 backup (Full, ...) then subsequently having a File entry
469 marked "delete" will be removed from the tree, so will not
470 be restored. If a file with the same name is later OK it
471 will be inserted in the tree -- this already happens. All
472 will be consistent except for possible changes during the
475 Since I'm on the subject, some of you may be wondering what
476 the utility of the in memory tree is if you are going to
477 restore everything (at least it comes up from time to time
478 on the list). Well, it is still *very* useful because it
479 allows only the last item found for a particular filename
480 (full path) to be entered into the tree, and thus if a file
481 is backed up 10 times, only the last copy will be restored.
482 I recently (last Friday) restored a complete directory, and
483 the Full and all the Differential and Incremental backups
484 spanned 3 Volumes. The first Volume was not even mounted
485 because all the files had been updated and hence backed up
486 since the Full backup was made. In this case, the tree
487 saved me a *lot* of time.
489 Make sure this information is stored on the tape too so
490 that it can be restored directly from the tape.
492 Comments from Martin Simmons (I think they are all covered):
493 Ok, that should cover the basics. There are few issues though:
495 - Restore will depend on the catalog. I think it is better to include the
496 extra data in the backup as well, so it can be seen by bscan and bextract.
498 - I'm not sure if it will preserve multiple hard links to the same inode. Or
499 maybe adding or removing links will cause the data to be dumped again?
501 - I'm not sure if it will handle renamed directories. Possibly it will work
502 by dumping the whole tree under a renamed directory?
504 - It remains to be seen how the backup performance of the DIR's will be
505 affected when comparing the catalog for a large filesystem.
509 How about introducing a Type = MgmtPolicy job type? That job type would
510 be responsible for scanning the Bacula environment looking for specific
511 conditions, and submitting the appropriate jobs for implementing said
515 Name = "Migration-Policy"
517 Policy Selection Job Type = Migrate
518 Scope = "<keyword> <operator> <regexp>"
519 Threshold = "<keyword> <operator> <regexp>"
520 Job Template = <template-name>
523 Where <keyword> is any legal job keyword, <operator> is a comparison
524 operator (=,<,>,!=, logical operators AND/OR/NOT) and <regexp> is a
525 appropriate regexp. I could see an argument for Scope and Threshold
526 being SQL queries if we want to support full flexibility. The
527 Migration-Policy job would then get scheduled as frequently as a site
528 felt necessary (suggested default: every 15 minutes).
533 Name = "Migration-Policy"
535 Policy Selection Job Type = Migration
537 Threshold = "Migration Selection Type = LowestUtil"
538 Job Template = "MigrationTemplate"
541 would select all pools for examination and generate a job based on
542 MigrationTemplate to automatically select the volume with the lowest
543 usage and migrate it's contents to the nextpool defined for that pool.
545 This policy abstraction would be really handy for adjusting the behavior
546 of Bacula according to site-selectable criteria (one thing that pops
547 into mind is Amanda's ability to automatically adjust backup levels
548 depending on various criteria).
554 - Add Pool/Storage override regression test.
555 - Add delete JobId to regression.
556 - Add a regression test for dbcheck.
557 - New test to add bscan to four-concurrent-jobs regression,
558 i.e. after the four-concurrent jobs zap the
559 database as is done in the bscan-test, then use bscan to
560 restore the database, do a restore and compare with the
562 - Add restore of specific JobId to regression (item 3
563 on the restore prompt)
564 - Add IPv6 to regression
565 - Add database test to regression. Test each function like delete,
568 - AntiVir can slow down backups on Win32 systems.
569 - Win32 systems with FAT32 can be much slower than NTFS for
570 more than 1000 files per directory.
574 - A HOLD command to stop all jobs from starting.
575 - A PAUSE command to pause all running jobs ==> release the
577 - Media Type = LTO,LTO-2,LTO-3
578 Media Type Read = LTO,LTO2,LTO3
579 Media Type Write = LTO2, LTO3
581 === From Carsten Menke <bootsy52@gmx.net>
583 Following is a list of what I think in the situations where I'm faced with,
584 could be a usefull enhancement to bacula, which I'm certain other users will
585 benefit from as well.
587 1. NextJob/NextJobs Directive within a Job Resource in the form of
588 NextJobs = job1,job2.
591 I currently solved the problem with running multiple jobs each after each
592 by setting the Max Wait Time for a job to 8 hours, and give
593 the jobs different Priorities. However, there scenarios where
594 1 Job is directly depending on another job, so if the former job fails,
595 the job after it needn't to be run
596 while maybe other jobs should run despite of that
599 A Backup Job and a Verify job, if the backup job fails there is no need to run
600 the verify job, as the backup job already failed. However, one may like
601 to backup the Catalog to disk despite of that the main backup job failed.
604 I see that this is related to the Event Handlers which are on the ToDo
605 list, also it is maybe a good idea to check for the return value and
606 execute different actions based on the return value
609 3. offline capability to bconsole
612 Currently I use a script which I execute within the last Job via the
613 RunAfterJob Directive, to release and eject the tape.
614 So I have to call bconsole "release=Storage-Name" and afterwards
615 mt -f /dev/nst0 eject to get the tape out.
617 If I have multiple Storage Devices, than these may not be /dev/nst0 and
618 I have to modify the script or call it with parameters etc.
619 This would actually not be needed, as everything is already defined
620 in bacula-sd.conf and if I can invoke bconsole with the
621 storage name via $1 in the script than I'm done and information is
624 4. %s for Storage Name added to the chars being substituted in "RunAfterJob"
628 For the reason mentioned in 3. to have the ability to call a
629 script with /scripts/foobar %s and in the script use $1
630 to pass the Storage Name to bconsole
632 5. Setting Volume State within a Job Resource
635 Instead of using "Maximum Volume Jobs" in the Pool Resource,
636 I would have the possibilty to define
637 in a Job Resource that after this certain job is run, the Volume State
638 should be set to "Volume State = Used", this give more flexibility (IMHO).
640 6. Localization of Bacula Messages
643 Unfortunatley many,many people I work with don't speak english very well.
644 So if at least the Reporting messages would be localized then they
645 would understand that they have to change the tape,etc. etc.
647 I volunteer to do the german translations, and if I can convince my wife also
648 french and Morre (western african language).
650 7. OK, this is evil, probably bound to security risks and maybe not possible
651 due to the design of bacula.
653 Implementation of Backtics ( `command` ) for shell comand execution to
654 the "Label Format" Directive.
658 Currently I have defined BACULA_DAY_OF_WEEK="day1|day2..." resulting in
659 Label Format = "HolyBackup-${BACULA_DAY_OF_WEEK[${WeekDay}]}". If I could
660 use backticks than I could use "Label Format = HolyBackup-`date +%A` to have
661 the localized name for the day of the week appended to the
662 format string. Then I have the tape labeled automatically with weekday
663 name in the correct language.
665 - Yes, that is surely the case. I probably should turn those into Warning
666 errors. In addition, you just made me think that it might not be bad to
667 add an option to check the file size after backing up the file and
668 report if it changes. This would be done as an option because it would
671 Kern, good idea. If you do do that, mention in the output: file
672 shrunk, or file expanded, just to make it obvious to the user
673 (without having to the refer to file size), just how the file size
676 Would this option be for all file, or just one file? Or a fileset?
677 - Make output from status use html table tags for nicely
678 presenting in a browser.
679 - Can one write tapes faster with 8192 byte block sizes?
680 - Document security problems with the same password for everyone in
681 rpm and Win32 releases.
682 - Browse generations of files.
683 - I've seen an error when my catalog's File table fills up. I
684 then have to recreate the File table with a larger maximum row
685 size. Relevant information is at
686 http://dev.mysql.com/doc/mysql/en/Full_table.html ; I think the
687 "Installing and Configuring MySQL" chapter should talk a bit
688 about this potential problem, and recommend a solution.
689 - For Solaris must use POSIX awk.
690 - Want speed of writing to tape while despooling.
691 - Supported autochanger:
699 Wangtek 6525ES (SCSI-1 QIC drive, 525MB), under Linux 2.4.something,
700 bacula 1.36.0/1 works with blocksize 16k INSIDE bacula-sd.conf.
701 - Add regex from http://www.pcre.org to Bacula for Win32.
702 - Use only shell tools no make in CDROM package.
703 - Include within include does it work?
704 - Implement a Pool of type Cleaning?
705 - Implement VolReadTime and VolWriteTime in SD
706 - Modify Backing up Your Database to include a bootstrap file.
707 - Think about making certain database errors fatal.
708 - Look at correcting the time jump in the scheduler for daylight
709 savings time changes.
710 - Add a "real" timer to network connections.
711 - Promote to Full = Time period
712 - Check dates entered by user for correctness (month/day/... ranges)
713 - Compress restore Volume listing by date and first file.
714 - Look at patches/bacula_db.b2z postgresql that loops during restore.
716 - Perhaps add read/write programs and/or plugins to FileSets.
717 - How to handle backing up portables ...
718 - Add some sort of guaranteed Interval for upgrading jobs.
719 - Can we write the state file after every job terminates? On Win32
720 the system crashes and the state file is not updated.
723 Documentation to do: (any release a little bit at a time)
724 - Doc to do unmount before removing magazine.
725 - Alternative to static linking "ldd prog" save all binaries listed,
726 restore them and point LD_LIBRARY_PATH to them.
727 - Document add "</dev/null >/dev/null 2>&1" to the bacula-fd command line
728 - Document query file format.
729 - Add more documentation for bsr files.
730 - Document problems with Verify and pruning.
731 - Document how to use multiple databases.
732 - VXA drives have a "cleaning required"
733 indicator, but Exabyte recommends preventive cleaning after every 75
736 In this context, it should be noted that Exabyte has a command-line
737 vxatool utility available for free download. (The current version is
738 vxatool-3.72.) It can get diagnostic info, read, write and erase tapes,
739 test the drive, unload tapes, change drive settings, flash new firmware,
741 Of particular interest in this context is that vxatool <device> -i will
742 report, among other details, the time since last cleaning in tape motion
743 minutes. This information can be retrieved (and settings changed, for
744 that matter) through the generic-SCSI device even when Bacula has the
745 regular tape device locked. (Needless to say, I don't recommend
746 changing tape settings while a job is running.)
747 - Lookup HP cleaning recommendations.
748 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
749 - Document doing table repair
752 ===================================
753 - Add macro expansions in JobDefs.
754 Run Before Job = "SomeFile %{Level} %{Client}"
755 Write Bootstrap="/some/dir/%{JobName}_%{Client}.bsr"
756 - Use non-blocking network I/O but if no data is available, use
758 - Use gather write() for network I/O.
759 - Autorestart on crash.
760 - Add bandwidth limiting.
761 - Add acks every once and a while from the SD to keep
762 the line from timing out.
763 - When an error in input occurs and conio beeps, you can back
764 up through the prompt.
765 - Detect fixed tape block mode during positioning by looking at
766 block numbers in btape "test". Possibly adjust in Bacula.
767 - Fix list volumes to output volume retention in some other
768 units, perhaps via a directive.
769 - If opening a tape in read/write mode fails attempt to open
770 it in read-only mode, and mark the tape for read only.
771 - Allow Simultaneous Priorities = yes => run up to Max concurrent jobs even
772 with multiple priorities.
773 - If you use restore replace=never, the directory attributes for
774 non-existent directories will not be restored properly.
776 - see lzma401.zip in others directory for new compression
778 - Minimal autochanger handling in Bacula and in btape.
779 - Look into how tar does not save sockets and the possiblity of
780 not saving them in Bacula (Martin Simmons reported this).
781 - Fix restore jobs so that multiple jobs can run if they
782 are not using the same tape(s).
783 - Allow the user to select JobType for manual pruning/purging.
784 - bscan does not put first of two volumes back with all info in
786 - Implement the FreeBSD nodump flag in chflags.
787 - Figure out how to make named console messages go only to that
788 console and to the non-restricted console (new console class?).
789 - Make restricted console prompt for password if *ask* is set or
790 perhaps if password is undefined.
791 - Implement "from ISO-date/time every x hours/days/weeks/months" in
794 ==== from Marc Schoechlin
795 - the help-command should be more verbose
796 (it should explain the paramters of the different
798 -> its time-comsuming to consult the manual anytime
799 you need a special parameter
800 -> maybe its more easy to maintain this, if the
801 descriptions of that commands are outsourced to
803 - the cd-command should allow complete paths
804 i.e. cd /foo/bar/foo/bar
805 -> if a customer mails me the path to a certain file,
806 its faster to enter the specified directory
807 - if the password is not configured in bconsole.conf
808 you should be asked for it.
809 -> sometimes you like to do restore on a customer-machine
810 which shouldnt know the password for bacula.
811 -> adding the password to the file favours admins
812 to forget to remove the password after usage
814 the protection of that file is less important
815 - long-listed-output of commands should be scrollable
816 like the unix more/less-command does
817 -> if someone runs 200 and more machines, the lists could
818 be a little long and complex
819 - command-output should be shown column by column
820 to reduce scrolling and to increase clarity
822 - lsmark should list the selected files with full
824 - wildcards for selecting and file and directories would be nice
825 - any actions should be interuptable with STRG+C
826 - command-expansion would be pretty cool
828 - When the replace Never option is set, new directory permissions
829 are not restored. See bug 213. To fix this requires creating a
830 list of newly restored directories so that those directory
831 permissions *can* be restored.
832 - Compaction of Disk space by "migrating" Volumes that have pruned
833 Jobs (what criteria? size, #jobs, time).
834 - Add prune all command
835 - Document fact that purge can destroy a part of a restore by purging
836 one volume while others remain valid -- perhaps mark Jobs.
837 - Add multiple-media-types.txt
838 - look at mxt-changer.html
839 - Make ? do a help command (no return needed).
840 - Implement restore directory.
841 - Document streams and how to implement them.
842 - Try not to re-backup a file if a new hard link is added.
843 - Add feature to backup hard links only, but not the data.
844 - Fix stream handling to be simpler.
845 - Add Priority and Bootstrap to Run a Job.
846 - Eliminate Restore "Run Restore Job" prompt by allowing new "run command
848 - Remove View FileSet button from Run a Job dialog.
849 - Handle prompt for restore job at end of Restore command.
850 - Add display of total selected files to Restore window.
851 - Add tree pane to left of window.
852 - Add progress meter.
853 - Max wait time or max run time causes seg fault -- see runtime-bug.txt
854 - Document writing to a CD/DVD with Bacula.
855 - Add a "base" package to the window installer for pthreadsVCE.dll
856 which is needed by all packages.
857 - Add message to user to check for fixed block size when the forward
858 space test fails in btape.
859 - When unmarking a directory check if all files below are unmarked and
860 then remove the + flag -- in the restore tree.
861 - Possibly implement: Action = Unmount Device="TapeDrive1" in Admin jobs.
862 - Setup lrrd graphs: (http://www.linpro.no/projects/lrrd/) Mike Acar.
863 - Revisit the question of multiple Volumes (disk) on a single device.
864 - Add a block copy option to bcopy.
865 - Investigate adding Mac Resource Forks.
866 - Finish work on Gnome restore GUI.
867 - Fix "llist jobid=xx" where no fileset or client exists.
868 - For each job type (Admin, Restore, ...) require only the really necessary
869 fields.- Pass Director resource name as an option to the Console.
870 - Add a "batch" mode to the Console (no unsolicited queries, ...).
871 - Add a .list all files in the restore tree (probably also a list all files)
872 Do both a long and short form.
873 - Allow browsing the catalog to see all versions of a file (with
874 stat data on each file).
875 - Restore attributes of directory if replace=never set but directory
877 - Use SHA1 on authentication if possible.
878 - See comtest-xxx.zip for Windows code to talk to USB.
879 - Add John's appended files:
880 Appended = { /files/server/logs/http/*log }
881 and such files would be treated as follows.On a FULL backup, they would
882 be backed up like any other file.On an INCREMENTAL backup, where a
883 previous INCREMENTAL or FULL was already in thecatalogue and the length
884 of the file wasgreater than the length of the last backup, only thedata
885 added since the last backup will be dumped.On an INCREMENTAL backup, if
886 the length of the file is less than thelength of the file with the same
887 name last backed up, the completefile is dumped.On Windows systems, with
888 creation date of files, we can be evensmarter about this and not count
889 entirely upon the length.On a restore, the full and all incrementals
890 since it will beapplied in sequence to restore the file.
891 - Check new HAVE_WIN32 open bits.
892 - Check if the tape has moved before writing.
893 - Handling removable disks -- see below:
894 - Keep track of tape use time, and report when cleaning is necessary.
895 - Add FromClient and ToClient keywords on restore command (or
896 BackupClient RestoreClient).
897 - Implement a JobSet, which groups any number of jobs. If the
898 JobSet is started, all the jobs are started together.
899 Allow Pool, Level, and Schedule overrides.
900 - Enhance cancel to timeout BSOCK packets after a specific delay.
901 - Do scheduling by UTC using gmtime_r() in run_conf, scheduler, and
902 ua_status.!!! Thanks to Alan Brown for this tip.
903 - Look at updating Volume Jobs so that Max Volume Jobs = 1 will work
904 correctly for multiple simultaneous jobs.
905 - Correct code so that FileSet MD5 is calculated for < and | filename
907 - Implement the Media record flag that indicates that the Volume does disk
909 - Implement VolAddr, which is used when Volume is addressed like a disk,
910 and form it from VolFile and VolBlock.
911 - Make multiple restore jobs for multiple media types specifying
912 the proper storage type.
913 - Fix fast block rejection (stored/read_record.c:118). It passes a null
914 pointer (rec) to try_repositioning().
915 - Look at extracting Win data from BackupRead.
916 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
917 which would take precidence over the Catalog "JobRetention".
918 - Implement Label Format in Add and Label console commands.
919 - Possibly up network buffers to 65K. Put on variable.
920 - Put email tape request delays on one or more variables. User wants
921 to cancel the job after a certain time interval. Maximum Mount Wait?
922 - Job, Client, Device, Pool, or Volume?
923 Is it possible to make this a directive which is *optional* in multiple
924 resources, like Level? If so, I think I'd make it an optional directive
925 in Job, Client, and Pool, with precedence such that Job overrides Client
926 which in turn overrides Pool.
928 - New Storage specifications:
929 - Want to write to multiple storage devices simultaneously
930 - Want to write to multiple storage devices sequentially (in one job)
931 - Want to read/write simultaneously
932 - Key is MediaType -- it must match
934 Passed to SD as a sort of BSR record called Storage Specification
938 MediaType -> Next MediaType
940 Device -> Next Device
942 Allow multiple Storage specifications
950 Allow Multiple Pool specifications (note, Pool currently
952 Allow Multiple MediaType specifications in Dir conf
953 Allow Multiple Device specifications in Dir conf
954 Perhaps keep this in a single SSR
955 Tie a Volume to a specific device by using a MediaType that
956 is contained in only one device.
957 In SD allow Device to have Multiple MediaTypes
959 - Ideas from Jerry Scharf:
960 First let's point out some big pluses that bacula has for this
962 more importantly it's active. Thank you so much for that
963 even more important, it's not flaky
964 it has an open access catalog, opening many possibilities
965 it's pushing toward heterogeneous systems capability
967 Macintosh file client
968 macs are an interesting niche, but I fear a server is a rathole
969 working bare iron recovery for windows
970 the option for inc/diff backups not reset on fileset revision
971 a) use both change and inode update time against base time
972 b) do the full catalog check (expensive but accurate)
973 sizing guide (how much system is needed to back up N systems/files)
974 consultants on using bacula in building a disaster recovery system
976 or how to get at fancy things that one could do with bacula
977 logwatch code for bacula logs (or similar)
978 linux distro inclusion of bacula (brings good and bad, but necessary)
979 win2k/XP server capability (icky but you asked)
980 support for Oracle database ??
982 - Look at adding SQL server and Exchange support for Windows.
983 - Make dev->file and dev->block_num signed integers so that -1 can
984 be an invalid value which happens with BSR.
985 - Create VolAddr for disk files in place of VolFile and VolBlock. This
986 is needed to properly specify ranges.
987 - Add progress of files/bytes to SD and FD.
988 - Print warning message if FileId > 4 billion
989 - do a "messages" before the first prompt in Console
990 - Client does not show busy during Estimate command.
991 - Implement Console mtx commands.
992 - Implement a Mount Command and an Unmount Command where
993 the users could specify a system command to be performed
994 to do the mount, after which Bacula could attempt to
995 read the device. This is for Removeable media such as a CDROM.
996 - Most likely, this mount command would be invoked explicitly
997 by the user using the current Console "mount" and "unmount"
998 commands -- the Storage Daemon would do the right thing
999 depending on the exact nature of the device.
1000 - As with tape drives, when Bacula wanted a new removable
1001 disk mounted, it would unmount the old one, and send a message
1002 to the user, who would then use "mount" as described above
1003 once he had actually inserted the disk.
1004 - Implement dump/print label to UA
1005 - Spool to disk only when the tape is full, then when a tape is hung move
1007 - bextract is sending everything to the log file ****FIXME****
1008 - Allow multiple Storage specifications (or multiple names on
1009 a single Storage specification) in the Job record. Thus a job
1010 can be backed up to a number of storage devices.
1011 - Implement some way for the File daemon to contact the Director
1012 to start a job or pass its DHCP obtained IP number.
1013 - Implement a query tape prompt/replace feature for a console
1014 - Copy console @ code to gnome2-console
1015 - Make AES the only encryption algorithm see
1016 http://csrc.nist.gov/CryptoToolkit/aes/). It's
1017 an officially adopted standard, has survived peer
1018 review, and provides keys up to 256 bits.
1019 - Take a careful look at SetACL http://setacl.sourceforge.net
1020 - Make tree walk routines like cd, ls, ... more user friendly
1021 by handling spaces better.
1022 - Make sure that Bacula rechecks the tape after the 20 min wait.
1023 - Set IO_NOWAIT on Bacula TCP/IP packets.
1024 - Try doing a raw partition backup and restore by mounting a
1026 - From Lars Kellers:
1027 Yes, it would allow to highly automatic the request for new tapes. If a
1028 tape is empty, bacula reads the barcodes (native or simulated), and if
1029 an unused tape is found, it runs the label command with all the
1030 necessary parameters.
1032 By the way can bacula automatically "move" an empty/purged volume say
1033 in the "short" pool to the "long" pool if this pool runs out of volume
1035 - What to do about "list files job=xxx".
1036 - Get and test MySQL 4.0
1037 - Look at how fuser works and /proc/PID/fd that is how Nic found the
1038 file descriptor leak in Bacula.
1039 - Implement WrapCounters in Counters.
1040 - Add heartbeat from FD to SD if hb interval expires.
1041 - Can we dynamically change FileSets?
1042 - If pool specified to label command and Label Format is specified,
1043 automatically generate the Volume name.
1044 - Why can't SQL do the filename sort for restore?
1045 - Add ExhautiveRestoreSearch
1046 - Look at the possibility of loading only the necessary
1047 data into the restore tree (i.e. do it one directory at a
1048 time as the user walks through the tree).
1049 - Possibly use the hash code if the user selects all for a restore command.
1050 - Fix "restore all" to bypass building the tree.
1051 - Prohibit backing up archive device (findlib/find_one.c:128)
1052 - Implement Release Device in the Job resource to unmount a drive.
1053 - Implement Acquire Device in the Job resource to mount a drive,
1054 be sure this works with admin jobs so that the user can get
1055 prompted to insert the correct tape. Possibly some way to say to
1056 run the job but don't save the files.
1057 - Make things like list where a file is saved case independent for
1060 - Use autochanger to handle multiple devices.
1061 - On Windows with very long path names, it may be impossible to create
1062 a file (and thus restore it) because the total length is too long.
1063 We must cd into the directory then create the file without the
1065 - Implement a Recycle command
1066 - Test a second language e.g. french.
1067 - Start working on Base jobs.
1068 - Implement UnsavedFiles DB record.
1069 - From Phil Stracchino:
1070 It would probably be a per-client option, and would be called
1071 something like, say, "Automatically purge obsoleted jobs". What it
1072 would do is, when you successfully complete a Differential backup of a
1073 client, it would automatically purge all Incremental backups for that
1074 client that are rendered redundant by that Differential. Likewise,
1075 when a Full backup on a client completed, it would automatically purge
1076 all Differential and Incremental jobs obsoleted by that Full backup.
1077 This would let people minimize the number of tapes they're keeping on
1078 hand without having to master the art of retention times.
1079 - When doing a Backup send all attributes back to the Director, who
1080 would then figure out what files have been deleted.
1081 - Currently in mount.c:236 the SD simply creates a Volume. It should have
1082 explicit permission to do so. It should also mark the tape in error
1083 if there is an error.
1084 - Cancel waiting for Client connect in SD if FD goes away.
1086 - Implement timeout in response() when it should come quickly.
1087 - Implement a Slot priority (loaded/not loaded).
1088 - Implement "vacation" Incremental only saves.
1089 - Implement create "FileSet"?
1090 - Add prefixlinks to where or not where absolute links to FD.
1091 - Issue message to mount a new tape before the rewind.
1092 - Simplified client job initiation for portables.
1093 - If SD cannot open a drive, make it periodically retry.
1094 - Add more of the config info to the tape label.
1096 - If tape is marked read-only, then try opening it read-only rather than
1097 failing, and remember that it cannot be written.
1098 - Refine SD waiting output:
1099 Device is being positioned
1100 > Device is being positioned for append
1101 > Device is being positioned to file x
1103 - Figure out some way to estimate output size and to avoid splitting
1104 a backup across two Volumes -- this could be useful for writing CDROMs
1105 where you really prefer not to have it split -- not serious.
1106 - Have SD compute MD5 or SHA1 and compare to what FD computes.
1107 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
1108 actual data on the Volume and compare it.
1109 - Implement Bacula plugins -- design API
1110 - Make bcopy read through bad tape records.
1111 - Program files (i.e. execute a program to read/write files).
1112 Pass read date of last backup, size of file last time.
1113 - Add Signature type to File DB record.
1114 - CD into subdirectory when open()ing files for backup to
1115 speed up things. Test with testfind().
1116 - Priority job to go to top of list.
1117 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
1118 - Implement some way for the Console to dynamically create a job.
1119 - Solaris -I on tar for include list
1120 - Need a verbose mode in restore, perhaps to bsr.
1121 - bscan without -v is too quiet -- perhaps show jobs.
1122 - Add code to reject whole blocks if not wanted on restore.
1123 - Check if we can increase Bacula FD priorty in Win2000
1124 - Make sure the MaxVolFiles is fully implemented in SD
1125 - Check if both CatalogFiles and UseCatalog are set to SD.
1126 - Possibly add email to Watchdog if drive is unmounted too
1127 long and a job is waiting on the drive.
1128 - After unmount, if restore job started, ask to mount.
1129 - Add UA rc and history files.
1130 - put termcap (used by console) in ./configure and
1131 allow -with-termcap-dir.
1132 - Fix Autoprune for Volumes to respect need for full save.
1133 - Compare tape to Client files (attributes, or attributes and data)
1134 - Make all database Ids 64 bit.
1135 - Allow console commands to detach or run in background.
1136 - Fix status delay on storage daemon during rewind.
1137 - Add SD message variables to control operator wait time
1138 - Maximum Operator Wait
1139 - Minimum Message Interval
1140 - Maximum Message Interval
1141 - Send Operator message when cannot read tape label.
1142 - Verify level=Volume (scan only), level=Data (compare of data to file).
1143 Verify level=Catalog, level=InitCatalog
1145 - Add keyword search to show command in Console.
1146 - Events : tape has more than xxx bytes.
1147 - Complete code in Bacula Resources -- this will permit
1148 reading a new config file at any time.
1149 - Handle ctl-c in Console
1150 - Implement script driven addition of File daemon to config files.
1151 - Think about how to make Bacula work better with File (non-tape) archives.
1152 - Write Unix emulator for Windows.
1153 - Put memory utilization in Status output of each daemon
1154 if full status requested or if some level of debug on.
1155 - Make database type selectable by .conf files i.e. at runtime
1156 - Set flag for uname -a. Add to Volume label.
1157 - Restore files modified after date
1158 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
1159 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
1160 - Timout a job or terminate if link goes down, or reopen link and query.
1161 - Concept of precious tapes (cannot be reused).
1162 - Make bcopy copy with a single tape drive.
1163 - Permit changing ownership during restore.
1166 > My suggestion: Add a feature on the systray menu-icon menu to request
1167 > an immediate backup now. This would be useful for laptop users who may
1168 > not be on the network when the regular scheduled backup is run.
1170 > My wife's suggestion: Add a setting to the win32 client to allow it to
1171 > shut down the machine after backup is complete (after, of course,
1172 > displaying a "System will shut down in one minute, click here to cancel"
1173 > warning dialog). This would be useful for sites that want user
1174 > woorkstations to be shut down overnight to save power.
1177 - Autolabel should be specified by DIR instead of SD.
1179 - Add media capacity
1180 - AutoScan (check checksum of tape)
1181 - Format command = "format /dev/nst0"
1185 - Seek resolution (usually corresponds to buffer size)
1186 - EODErrorCode=ENOSPC or code
1187 - Partial Read error code
1188 - Partial write error code
1189 - Nonformatted read error
1190 - Nonformatted write error
1191 - WriteProtected error
1195 - IgnoreCloseErrors=yes
1205 - FD sends unsaved file list to Director at end of job (see
1207 - File daemon should build list of files skipped, and then
1208 at end of save retry and report any errors.
1209 - Write a Storage daemon that uses pipes and
1210 standard Unix programs to write to the tape.
1212 - Need something that monitors the JCR queue and
1213 times out jobs by asking the deamons where they are.
1214 - Enhance Jmsg code to permit buffering and saving to disk.
1215 - device driver = "xxxx" for drives.
1216 - Verify from Volume
1217 - Ensure that /dev/null works
1218 - Need report class for messages. Perhaps
1219 report resource where report=group of messages
1220 - enhance scan_attrib and rename scan_jobtype, and
1221 fill in code for "since" option
1222 - Director needs a time after which the report status is sent
1223 anyway -- or better yet, a retry time for the job.
1224 - Don't reschedule a job if previous incarnation is still running.
1225 - Some way to automatically backup everything is needed????
1226 - Need a structure for pending actions:
1228 - termination status (part of buffered msgs?)
1230 Read, Write, Clean, Delete
1231 - Login to Bacula; Bacula users with different permissions:
1232 owner, group, user, quotas
1233 - Store info on each file system type (probably in the job header on tape.
1234 This could be the output of df; or perhaps some sort of /etc/mtab record.
1236 ========= ideas ===============
1237 From: "Jerry K. Schieffer" <jerry@skylinetechnology.com>
1238 To: <kern@sibbald.com>
1239 Subject: RE: [Bacula-users] future large programming jobs
1240 Date: Thu, 26 Feb 2004 11:34:54 -0600
1242 I noticed the subject thread and thought I would offer the following
1243 merely as sources of ideas, i.e. something to think about, not even as
1244 strong as a request. In my former life (before retiring) I often
1245 dealt with backups and storage management issues/products as a
1246 developer and as a consultant. I am currently migrating my personal
1247 network from amanda to bacula specifically because of the ability to
1248 cross media boundaries during storing backups.
1249 Are you familiar with the commercial product called ADSM (I think IBM
1250 now sells it under the Tivoli label)? It has a couple of interesting
1251 ideas that may apply to the following topics.
1253 1. Migration: Consider that when you need to restore a system, there
1254 may be pressure to hurry. If all the information for a single client
1255 can eventually end up on the same media (and in chronological order),
1256 the restore is facillitated by not having to search past information
1257 from other clients. ADSM has the concept of "client affinity" that
1258 may be associated with it's storage pools. It seems to me that this
1259 concept (as an optional feature) might fit in your architecture for
1262 ADSM also has the concept of defining one or more storage pools as
1263 "copy pools" (almost mirrors, but only in the sense of contents).
1264 These pools provide the ability to have duplicte data stored both
1265 onsite and offsite. The copy process can be scheduled to be handled
1266 by their storage manager during periods when there is no backup
1267 activity. Again, the migration process might be a place to consider
1268 implementing something like this.
1271 > It strikes me that it would be very nice to be able to do things
1273 > have the Job(s) backing up the machines run, and once they have all
1274 > completed, start a migration job to copy the data from disks Volumes
1276 > a tape library and then to offsite storage. Maybe this can already
1278 > done with some careful scheduling and Job prioritzation; the events
1279 > mechanism described below would probably make it very easy.
1281 This is the goal. In the first step (before events), you simply
1283 the Migration to tape later.
1285 2. Base jobs: In ADSM, each copy of each stored file is tracked in
1286 the database. Once a file (unique by path and metadata such as dates,
1287 size, ownership, etc.) is in a copy pool, no more copies are made. In
1288 other words, when you start ADSM, it begins like your concept of a
1289 base job. After that it is in the "incremental" mode. You can
1290 configure the number of "generations" of files to be retained, plus a
1291 retention date after which even old generations are purged. The
1292 database tracks the contents of media and projects the percentage of
1293 each volume that is valid. When the valid content of a volume drops
1294 below a configured percentage, the valid data are migrated to another
1295 volume and the old volume is marked as empty. Note, this requires
1296 ADSM to have an idea of the contents of a client, i.e. marking the
1297 database when an existing file was deleted, but this would solve your
1298 issue of restoring a client without restoring deleted files.
1300 This is pretty far from what bacula now does, but if you are going to
1301 rip things up for Base jobs,.....
1302 Also, the benefits of this are huge for very large shops, especially
1303 with media robots, but are a pain for shops with manual media
1307 > Base jobs sound pretty useful, but I'm not dying for them.
1309 Nobody is dying for them, but when you see what it does, you will die
1312 3. Restoring deleted files: Since I think my comments in (2) above
1313 have low probability of implementation, I'll also suggest that you
1314 could approach the issue of deleted files by a mechanism of having the
1315 fd report to the dir, a list of all files on the client for every
1316 backup job. The dir could note in the database entry for each file
1317 the date that the file was seen. Then if a restore as of date X takes
1318 place, only files that exist from before X until after X would be
1319 restored. Probably the major cost here is the extra date container in
1320 each row of the files table.
1322 Thanks for "listening". I hope some of this helps. If you want to
1323 contact me, please send me an email - I read some but not all of the
1324 mailing list traffic and might miss a reply there.
1326 Please accept my compliments for bacula. It is doing a great job for
1327 me!! I sympathize with you in the need to wrestle with excelence in
1328 execution vs. excelence in feature inclusion.
1333 ==============================
1336 - Design at hierarchial storage for Bacula. Migration and Clone.
1337 - Implement FSM (File System Modules).
1338 - Audit M_ error codes to ensure they are correct and consistent.
1339 - Add variable break characters to lex analyzer.
1340 Either a bit mask or a string of chars so that
1341 the caller can change the break characters.
1342 - Make a single T_BREAK to replace T_COMMA, etc.
1343 - Ensure that File daemon and Storage daemon can
1344 continue a save if the Director goes down (this
1345 is NOT currently the case). Must detect socket error,
1346 buffer messages for later.
1347 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
1348 - Add ability to backup to two Storage devices (two SD sessions) at
1349 the same time -- e.g. onsite, offsite.
1350 - Add the ability to consolidate old backup sets (basically do a restore
1351 to tape and appropriately update the catalog). Compress Volume sets.
1352 Might need to spool via file is only one drive is available.
1353 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
1354 someway to do so with every volume that has less than x% valid
1358 Migration: Move a backup from one Volume to another
1359 Clone: Copy a backup -- two Volumes
1361 Bacula Migration is based on Jobs (apparently Networker is file by file).
1363 Migration triggered by:
1367 Highwater mark (keep total size)
1372 ======================================================
1374 It is somewhat like a Full save becomes an incremental since
1375 the Base job (or jobs) plus other non-base files.
1377 - A Base backup is same as Full backup, just different type.
1378 - New BaseFiles table that contains:
1380 BaseJobId - Base JobId referenced for this FileId (needed ???)
1381 JobId - JobId currently running
1382 FileId - File not backed up, exists in Base Job
1383 FileIndex - FileIndex from Base Job.
1384 i.e. for each base file that exists but is not saved because
1385 it has not changed, the File daemon sends the JobId, BaseId,
1386 FileId, FileIndex back to the Director who creates the DB entry.
1387 - To initiate a Base save, the Director sends the FD
1388 the FileId, and full filename for each file in the Base.
1389 - When the FD finds a Base file, he requests the Director to
1390 send him the full File entry (stat packet plus MD5), or
1391 conversely, the FD sends it to the Director and the Director
1392 says yes or no. This can be quite rapid if the FileId is kept
1393 by the FD for each Base Filename.
1394 - It is probably better to have the comparison done by the FD
1395 despite the fact that the File entry must be sent across the
1397 - An alternative would be to send the FD the whole File entry
1398 from the start. The disadvantage is that it requires a lot of
1399 space. The advantage is that it requires less communications
1401 - The Job record must be updated to indicate that one or more
1403 - At end of Job, FD returns:
1404 1. Count of base files/bytes not written to tape (i.e. matches)
1405 2. Count of base file that were saved i.e. had changed.
1406 - No tape record would be written for a Base file that matches, in the
1407 same way that no tape record is written for Incremental jobs where
1408 the file is not saved because it is unchanged.
1409 - On a restore, all the Base file records must explicitly be
1410 found from the BaseFile tape. I.e. for each Full save that is marked
1411 to have one or more Base Jobs, search the BaseFile for all occurrences
1413 - An optimization might be to make the BaseFile have:
1419 This would avoid the need to explicitly fetch each File record for
1420 the Base job. The Base Job record will be fetched to get the
1421 VolSessionId and VolSessionTime.
1422 =========================================================
1425 ==========================================================
1427 For each Incremental job that is run, there may be files that
1428 were found but not saved because they were locked (this applies
1429 only to Windows). Such a system could send back to the Director
1430 a list of Unsaved files.
1432 - New UnSavedFiles table that contains:
1436 - Then in the next Incremental job, the list of Unsaved Files will be
1437 feed to the FD, who will ensure that they are explicitly chosen even
1438 if standard date/time check would not have selected them.
1439 =============================================================
1443 Multiple drive autochanger data: see Alan Brown
1444 mtx -f xxx unloadStorage Element 1 is Already Full(drive 0 was empty)
1445 Unloading Data Transfer Element into Storage Element 1...source Element
1446 Address 480 is Empty
1448 (drive 0 was empty and so was slot 1)
1449 > mtx -f xxx load 15 0
1450 no response, just returns to the command prompt when complete.
1451 > mtx -f xxx status Storage Changer /dev/changer:2 Drives, 60 Slots ( 2 Import/Export )
1452 Data Transfer Element 0:Full (Storage Element 15 Loaded):VolumeTag = HX001
1453 Data Transfer Element 1:Empty
1454 Storage Element 1:Empty
1455 Storage Element 2:Full :VolumeTag=HX002
1456 Storage Element 3:Full :VolumeTag=HX003
1457 Storage Element 4:Full :VolumeTag=HX004
1458 Storage Element 5:Full :VolumeTag=HX005
1459 Storage Element 6:Full :VolumeTag=HX006
1460 Storage Element 7:Full :VolumeTag=HX007
1461 Storage Element 8:Full :VolumeTag=HX008
1462 Storage Element 9:Full :VolumeTag=HX009
1463 Storage Element 10:Full :VolumeTag=HX010
1464 Storage Element 11:Empty
1465 Storage Element 12:Empty
1466 Storage Element 13:Empty
1467 Storage Element 14:Empty
1468 Storage Element 15:Empty
1469 Storage Element 16:Empty....
1470 Storage Element 28:Empty
1471 Storage Element 29:Full :VolumeTag=CLNU01L1
1472 Storage Element 30:Empty....
1473 Storage Element 57:Empty
1474 Storage Element 58:Full :VolumeTag=NEX261L2
1475 Storage Element 59 IMPORT/EXPORT:Empty
1476 Storage Element 60 IMPORT/EXPORT:Empty
1478 Unloading Data Transfer Element into Storage Element 15...done
1480 (just to verify it remembers where it came from, however it can be
1481 overrriden with mtx unload {slotnumber} to go to any storage slot.)
1483 There needs to be a table of drive # to devices somewhere - If there are
1484 multiple changers or drives there may not be a 1:1 correspondance between
1485 changer drive number and system device name - and depending on the way the
1486 drives are hooked up to scsi busses, they may not be linearly numbered
1487 from an offset point either.something like
1489 Autochanger drives = 2
1490 Autochanger drive 0 = /dev/nst1
1491 Autochanger drive 1 = /dev/nst2
1492 IMHO, it would be _safest_ to use explicit mtx unload commands at all
1493 times, not just for multidrive changers. For a 1 drive changer, that's
1499 MTX's manpage (1.2.15):
1500 unload [<slotnum>] [ <drivenum> ]
1501 Unloads media from drive <drivenum> into slot
1502 <slotnum>. If <drivenum> is omitted, defaults to
1503 drive 0 (as do all commands). If <slotnum> is
1504 omitted, defaults to the slot that the drive was
1505 loaded from. Note that there's currently no way
1506 to say 'unload drive 1's media to the slot it
1507 came from', other than to explicitly use that
1508 slot number as the destination.AB
1514 undef# camcontrol devlist
1515 <WANGTEK 51000 SCSI M74H 12B3> at scbus0 target 2 lun 0 (pass0,sa0)
1516 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 0 (pass1,sa1)
1517 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 1 (pass2)
1519 tapeinfo -f /dev/sg0 with a bad tape in drive 1:
1520 [kern@rufus mtx-1.2.17kes]$ ./tapeinfo -f /dev/sg0
1521 Product Type: Tape Drive
1523 Product ID: 'C5713A '
1525 Attached Changer: No
1526 TapeAlert[3]: Hard Error: Uncorrectable read/write error.
1527 TapeAlert[20]: Clean Now: The tape drive neads cleaning NOW.
1534 Medium Type: Not Loaded
1537 DataCompEnabled: yes
1538 DataCompCapable: yes
1539 DataDeCompEnabled: yes
1546 Handling removable disks
1548 From: Karl Cunningham <karlc@keckec.com>
1550 My backups are only to hard disk these days, in removable bays. This is my
1551 idea of how a backup to hard disk would work more smoothly. Some of these
1552 things Bacula does already, but I mention them for completeness. If others
1553 have better ways to do this, I'd like to hear about it.
1555 1. Accommodate several disks, rotated similar to how tapes are. Identified
1556 by partition volume ID or perhaps by the name of a subdirectory.
1557 2. Abort & notify the admin if the wrong disk is in the bay.
1558 3. Write backups to different subdirectories for each machine to be backed
1560 4. Volumes (files) get created as needed in the proper subdirectory, one
1562 5. When a disk is recycled, remove or zero all old backup files. This is
1563 important as the disk being recycled may be close to full. This may be
1564 better done manually since the backup files for many machines may be
1565 scattered in many subdirectories.
1570 - Make sure that all do_prompt() calls in Dir check for
1571 -1 (error) and -2 (cancel) returns.
1572 - Fix foreach_jcr() to have free_jcr() inside next().
1573 jcr=jcr_walk_start();
1574 for ( ; jcr; (jcr=jcr_walk_next(jcr)) )
1577 - A Volume taken from Scratch should take on the retention period
1579 - Correct doc for Maximum Changer Wait (and others) accepting only
1581 - Implement status that shows why a job is being held in reserve, or
1582 rather why none of the drives are suitable.
1583 - Implement a way to disable a drive (so you can use the second
1584 drive of an autochanger, and the first one will not be used or
1586 - Make sure Maximum Volumes is respected in Pools when adding
1587 Volumes (e.g. when pulling a Scratch volume).
1588 - Keep same dcr when switching device ...
1589 - Implement code that makes the Dir aware that a drive is an
1590 autochanger (so the user doesn't need to use the Autochanger = yes
1592 - Make catalog respect ACL.
1593 - Add recycle count to Media record.
1594 - Add initial write date to Media record.
1595 - Fix store_yesno to be store_bitmask.
1596 --- create_file.c.orig Fri Jul 8 12:13:05 2005
1597 +++ create_file.c Fri Jul 8 12:13:07 2005
1599 attr->ofname, be.strerror());
1602 + } else if(S_ISSOCK(attr->statp.st_mode)) {
1603 + Dmsg1(200, "Skipping socket: %s\n", attr->ofname);
1605 Dmsg1(200, "Restore node: %s\n", attr->ofname);
1606 if (mknod(attr->ofname, attr->statp.st_mode, attr->statp.st_rdev) != 0 && errno != EEXIST) {
1607 - Add true/false to conf same as yes/no
1608 - Reserve blocks other restore jobs when first cannot connect to SD.
1609 - Fix Maximum Changer Wait, Maximum Open Wait, Maximum Rewind Wait to
1610 accept time qualifiers.
1611 - Does ClientRunAfterJob fail the job on a bad return code?
1612 - Make hardlink code at line 240 of find_one.c use binary search.
1613 - Add ACL error messages in src/filed/acl.c.
1614 - Make authentication failures single threaded.
1615 - Make Dir and SD authentication errors single threaded.
1617 - Fix catreq.c digestbuf at line 411 in src/dird/catreq.c
1618 - Make base64.c (bin_to_base64) take a buffer length
1619 argument to avoid overruns.
1620 and verify that other buffers cannot overrun.
1621 - Implement VolumeState as discussed with Arno.
1622 - Add LocationId to update volume
1630 - Add Comment to Media record
1631 - Fix auth compatibility with 1.38
1632 - Update dbcheck to include Log table
1633 - Update llist to include new fields.