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
27 - Document JobStatus and Termination codes.
32 - If a job terminates, the DIR connection can close before the
33 Volume info is updated, leaving the File count wrong.
34 - Look at why SIGPIPE during connection can cause seg fault in
35 writing the daemon message, when Dir dropped to bacula:bacula
36 - Look at zlib 32 => 64 problems.
37 - Ensure that connection to daemon failure always indicates what
38 daemon it was trying to connect to.
39 - Try turning on disk seek code.
40 - Possibly turn on St. Bernard code.
41 - Fix bscan to report the JobType when restoring a job.
42 - Fix bextract to restore ACLs, or better yet, use common
44 - Do we migrate appendable Volumes?
45 - Remove queue.c code.
46 - Add bconsole option to use stdin/out instead of conio.
47 - Fix ClientRunBefore/AfterJob compatibility.
48 - Fix re-read of last block to check if job has actually written
49 a block, and check if block was written by a different job
50 (i.e. multiple simultaneous jobs writing).
51 - Some users claim that they must do two prune commands to get a
52 Volume marked as purged.
53 - Print warning message if LANG environment variable does not specify
55 - New dot commands from Arno.
56 .update volume [enabled|disabled|*see below]
57 > However, I could easily imagine an option to "update slots" that says
58 > "enable=yes|no" that would automatically enable or disable all the Volumes
59 > found in the autochanger. This will permit the user to optionally mark all
60 > the Volumes in the magazine disabled prior to taking them offsite, and mark
61 > them all enabled when bringing them back on site. Coupled with the options
62 > to the slots keyword, you can apply the enable/disable to any or all volumes.
63 .show device=xxx lists information from one storage device, including
64 devices (I'm not even sure that information exists in the DIR...)
65 .move eject device=xxx mostly the same as 'unmount xxx' but perhaps with
66 better machine-readable output like "Ok" or "Error busy"
67 .move eject device=xxx toslot=yyy the same as above, but with a new
68 target slot. The catalog should be updated accordingly.
69 .move transfer device=xxx fromslot=yyy toslot=zzz
72 - Get Perl replacement for bregex.c
73 - Given all the problems with FIFOs, I think the solution is to do something a
74 little different, though I will look at the code and see if there is not some
75 simple solution (i.e. some bug that was introduced). What might be a better
76 solution would be to use a FIFO as a sort of "key" to tell Bacula to read and
77 write data to a program rather than the FIFO. For example, suppose you
82 Then, I could imagine if you backup and restore this file with a direct
83 reference as is currently done for fifos, instead, during backup Bacula will
86 /home/kern/my-fifo.backup
88 and read the data that my-fifo.backup writes to stdout. For restore, Bacula
91 /home/kern/my-fifo.restore
93 and send the data backed up to stdout. These programs can either be an
94 executable or a shell script and they need only read/write to stdin/stdout.
96 I think this would give a lot of flexibility to the user without making any
97 significant changes to Bacula.
102 select FilenameId from Filename where Name='';
103 # Get list of all directories referenced in a Backup.
104 select Path.Path from Path,File where File.JobId=nnn and
105 File.FilenameId=(FilenameId-from-above) and File.PathId=Path.PathId
106 order by Path.Path ASC;
108 - Look into using Dart for testing
109 http://public.kitware.com/Dart/HTML/Index.shtml
111 - Look into replacing autotools with cmake
112 http://www.cmake.org/HTML/Index.html
114 === Migration from David ===
115 What I'd like to see:
118 Name = "<poolname>-migrate"
122 Migration Selection Type = LowestUtil | OldestVol | PoolOccupancy |
123 Client | PoolResidence | Volume | JobName | SQLquery
124 Migration Selection Pattern = "regexp"
125 Next Pool = <override>
128 There should be no need for a Level (migration is always Full, since you
129 don't calculate differential/incremental differences for migration),
130 Storage should be determined by the volume types in the pool, and Client
131 is really a selection issue. Migration should always occur to the
132 NextPool defined in the pool definition. If no nextpool is defined, the
133 job should end with a reason of "no place to go". If Next Pool statement
134 is present, we override the check in the pool definition and use the
137 Here's how I'd define Migration Selection Types:
140 Client -- Migrate data from selected client only. Migration Selection
141 Pattern regexp provides pattern to select client names, eg ^FS00* makes
142 all client names starting with FS00 eligible for migration.
144 Jobname -- Migration all jobs matching name. Migration Selection Pattern
145 regexp provides pattern to select jobnames existing in pool.
147 Volume -- Migrate all data on specified volumes. Migration Selection
148 Pattern regexp provides selection criteria for volumes to be migrated.
149 Volumes must exist in pool to be eligible for migration.
153 LowestUtil -- Identify the volume in the pool with the least data on it
154 and empty it. No Migration Selection Pattern required.
156 OldestVol -- Identify the LRU volume with data written, and empty it. No
157 Migration Selection Pattern required.
159 PoolOccupancy -- if pool occupancy exceeds <highmig>, migrate volumes
160 (starting with most full volumes) until pool occupancy drops below
161 <lowmig>. Pool highmig and lowmig values are in pool definition, no
162 Migration Selection Pattern required.
166 SQLQuery -- Migrate all jobuids returned by the supplied SQL query.
167 Migration Selection Pattern contains SQL query to execute; should return
168 a list of 1 or more jobuids to migrate.
170 PoolResidence -- Migrate data sitting in pool for longer than
171 PoolResidence value in pool definition. Migration Selection Pattern
172 optional; if specified, override value in pool definition (value in
176 [ possibly a Python event -- kes ]
178 - Mount on an Autochanger with no tape in the drive causes:
179 Automatically selected Storage: LTO-changer
180 Enter autochanger drive[0]: 0
181 3301 Issuing autochanger "loaded drive 0" command.
182 3302 Autochanger "loaded drive 0", result: nothing loaded.
183 3301 Issuing autochanger "loaded drive 0" command.
184 3302 Autochanger "loaded drive 0", result: nothing loaded.
185 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
186 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
187 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
188 If this is not a blank tape, try unmounting and remounting the Volume.
189 - If Drive 0 is blocked, and drive 1 is set "Autoselect=no", drive 1 will
191 - Autochanger did not change volumes.
192 select * from Storage;
193 +-----------+-------------+-------------+
194 | StorageId | Name | AutoChanger |
195 +-----------+-------------+-------------+
196 | 1 | LTO-changer | 0 |
197 +-----------+-------------+-------------+
198 05-May 03:50 roxie-sd: 3302 Autochanger "loaded drive 0", result is Slot 11.
199 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Warning: Director wanted Volume "LT
200 Current Volume "LT0-002" not acceptable because:
201 1997 Volume "LT0-002" not in catalog.
202 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Error: Autochanger Volume "LT0-002"
203 Setting InChanger to zero in catalog.
204 05-May 03:50 roxie-dir: Tibs.2006-05-05_03.05.02 Error: Unable to get Media record
206 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Error getting Volume i
207 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Job 530 canceled.
208 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: spool.c:249 Fatal appe
209 05-May 03:49 Tibs: Tibs.2006-05-05_03.05.02 Fatal error: c:\cygwin\home\kern\bacula
218 FirstWritten: 2006-05-05 03:11:54
219 LastWritten: 2006-05-05 03:50:23
220 LabelDate: 2005-12-26 16:52:40
231 VolRetention: 31,536,000
243 Note VolStatus is blank!!!!!
250 FirstWritten: 0000-00-00 00:00:00
251 LastWritten: 0000-00-00 00:00:00
252 LabelDate: 2005-12-26 16:52:40
263 VolRetention: 31,536,000
276 Automatically selected Storage: LTO-changer
277 Enter autochanger drive[0]: 0
278 3301 Issuing autochanger "loaded drive 0" command.
279 3302 Autochanger "loaded drive 0", result: nothing loaded.
280 3301 Issuing autochanger "loaded drive 0" command.
281 3302 Autochanger "loaded drive 0", result: nothing loaded.
282 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
283 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
285 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
286 If this is not a blank tape, try unmounting and remounting the Volume.
288 - Add VolumeState (enable, disable, archive)
289 - Add VolumeLock to prevent all but lock holder (SD) from updating
290 the Volume data (with the exception of VolumeState).
291 - The btape fill command does not seem to use the Autochanger
292 - Make Windows installer default to system disk drive.
293 - Look at using ioctl(FIOBMAP, ...) on Linux, and
294 DeviceIoControl(..., FSCTL_QUERY_ALLOCATED_RANGES, ...) on
295 Win32 for sparse files.
296 http://www.flexhex.com/docs/articles/sparse-files.phtml
297 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
298 - Directive: at <event> "command"
299 - Command: pycmd "command" generates "command" event. How to
300 attach to a specific job?
301 - Integrate Christopher's St. Bernard code.
302 - run_cmd() returns int should return JobId_t
303 - get_next_jobid_from_list() returns int should return JobId_t
304 - Document export LDFLAGS=-L/usr/lib64
305 - Don't attempt to restore from "Disabled" Volumes.
306 - Network error on Win32 should set Win32 error code.
307 - What happens when you rename a Disk Volume?
308 - Job retention period in a Pool (and hence Volume). The job would
310 - Detect resource deadlock in Migrate when same job wants to read
311 and write the same device.
312 - Queue warning/error messages during restore so that they
313 are reported at the end of the report rather than being
314 hidden in the file listing ...
315 - Look at -D_FORTIFY_SOURCE=2
316 - Add Win32 FileSet definition somewhere
317 - Look at fixing restore status stats in SD.
318 - Make selection of Database used in restore correspond to
320 - Look at using ioctl(FIMAP) and FIGETBSZ for sparse files.
321 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
322 - Implement a mode that says when a hard read error is
323 encountered, read many times (as it currently does), and if the
324 block cannot be read, skip to the next block, and try again. If
325 that fails, skip to the next file and try again, ...
327 create table LevelType (LevelType binary(1), LevelTypeLong tinyblob);
328 insert into LevelType (LevelType,LevelTypeLong) values
332 - Add ACL to restore only to original location.
333 - Show files/second in client status output.
334 - Add a recursive mark command (rmark) to restore.
335 - "Minimum Job Interval = nnn" sets minimum interval between Jobs
336 of the same level and does not permit multiple simultaneous
337 running of that Job (i.e. lets any previous invocation finish
338 before doing Interval testing).
339 - Look at simplifying File exclusions.
340 - New directive "Delete purged Volumes"
341 - new pool XXX with ScratchPoolId = MyScratchPool's PoolId and
342 let it fill itself, and RecyclePoolId = XXX's PoolId so I can
343 see if it become stable and I just have to supervise
345 - If I want to remove this pool, I set RecyclePoolId = MyScratchPool's
346 PoolId, and when it is empty remove it.
347 - Figure out how to recycle Scratch volumes back to the Scratch Pool.
349 - Allow Check Labels to be used with Bacula labels.
350 - "Resuming" a failed backup (lost line for example) by using the
351 failed backup as a sort of "base" job.
353 - Email to the user when the tape is about to need changing x
354 days before it needs changing.
355 - Command to show next tape that will be used for a job even
356 if the job is not scheduled.
357 - From: Arunav Mandal <amandal@trolltech.com>
358 1. When jobs are running and bacula for some reason crashes or if I do a
359 restart it remembers and jobs it was running before it crashed or restarted
360 as of now I loose all jobs if I restart it.
362 2. When spooling and in the midway if client is disconnected for instance a
363 laptop bacula completely discard the spool. It will be nice if it can write
364 that spool to tape so there will be some backups for that client if not all.
366 3. We have around 150 clients machines it will be nice to have a option to
367 upgrade all the client machines bacula version automatically.
369 4. Atleast one connection should be reserved for the bconsole so at heavy load
370 I should connect to the director via bconsole which at sometimes I can't
372 5. Another most important feature that is missing, say at 10am I manually
373 started backup of client abc and it was a full backup since client abc has
374 no backup history and at 10.30am bacula again automatically started backup of
375 client abc as that was in the schedule. So now we have 2 multiple Full
376 backups of the same client and if we again try to start a full backup of
377 client backup abc bacula won't complain. That should be fixed.
379 - Fix bpipe.c so that it does not modify results pointer.
380 ***FIXME*** calling sequence should be changed.
381 - For Windows disaster recovery see http://unattended.sf.net/
382 - regardless of the retention period, Bacula will not prune the
383 last Full, Diff, or Inc File data until a month after the
384 retention period for the last Full backup that was done.
385 - update volume=xxx --- add status=Full
386 - Remove old spool files on startup.
387 - Exclude SD spool/working directory.
388 - Refuse to prune last valid Full backup. Same goes for Catalog.
390 - Make a callback when Rerun failed levels is called.
391 - Give Python program access to Scheduled jobs.
392 - Add setting Volume State via Python.
393 - Python script to save with Python, not save, save with Bacula.
394 - Python script to do backup.
396 - Change the Priority, Client, Storage, JobStatus (error)
397 at the start of a job.
398 - Why is SpoolDirectory = /home/bacula/spool; not reported
399 as an error when writing a DVD?
400 - Make bootstrap file handle multiple MediaTypes (SD)
401 - Remove all old Device resource code in Dir and code to pass it
402 back in SD -- better, rework it to pass back device statistics.
403 - Check locking of resources -- be sure to lock devices where previously
404 resources were locked.
405 - The last part is left in the spool dir.
408 - In restore don't compare byte count on a raw device -- directory
409 entry does not contain bytes.
410 - To mark files as deleted, run essentially a Verify to disk, and
411 when a file is found missing (MarkId != JobId), then create
412 a new File record with FileIndex == -1. This could be done
413 by the FD at the same time as the backup.
417 MA = (last_MA * 3 + rate) / 4
418 rate = (bytes - last_bytes) / (runtime - last_runtime)
419 - Max Vols limit in Pool off by one?
420 - Implement Files/Bytes,... stats for restore job.
421 - Implement Total Bytes Written, ... for restore job.
422 - Despool attributes simultaneously with data in a separate
423 thread, rejoined at end of data spooling.
424 - 7. Implement new Console commands to allow offlining/reserving drives,
425 and possibly manipulating the autochanger (much asked for).
426 - Add start/end date editing in messages (%t %T, %e?) ...
427 - Add ClientDefs similar to JobDefs.
428 - Print more info when bextract -p accepts a bad block.
429 - Fix FD JobType to be set before RunBeforeJob in FD.
430 - Look at adding full Volume and Pool information to a Volume
431 label so that bscan can get *all* the info.
432 - If the user puts "Purge Oldest Volume = yes" or "Recycle Oldest Volume = yes"
433 and there is only one volume in the pool, refuse to do it -- otherwise
434 he fills the Volume, then immediately starts reusing it.
435 - Implement copies and stripes.
436 - Add history file to console.
437 - Each file on tape creates a JobMedia record. Peter has 4 million
438 files spread over 10000 tape files and four tapes. A restore takes
439 16 hours to build the restore list.
440 - Add and option to check if the file size changed during backup.
441 - Make sure SD deletes spool files on error exit.
442 - Delete old spool files when SD starts.
443 - When labeling tapes, if you enter 000026, Bacula uses
444 the tape index rather than the Volume name 000026.
445 - Add offline tape command to Bacula console.
447 Enter MediaId or Volume name: 32
448 Enter new Volume name: DLT-20Dec04
449 Automatically selected Pool: Default
450 Connecting to Storage daemon DLTDrive at 192.168.68.104:9103 ...
451 Sending relabel command from "DLT-28Jun03" to "DLT-20Dec04" ...
452 block.c:552 Write error at 0:0 on device /dev/nst0. ERR=Bad file descriptor.
453 Error writing final EOF to tape. This tape may not be readable.
454 dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
455 askdir.c:219 NULL Volume name. This shouldn't happen!!!
456 3912 Failed to label Volume: ERR=dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
457 Label command failed for Volume DLT-20Dec04.
458 Do not forget to mount the drive!!!
459 - Bug: if a job is manually scheduled to run later, it does not appear
460 in any status report and cannot be cancelled.
462 ==== Keeping track of deleted files ====
463 My "trick" for keeping track of deletions is the following.
464 Assuming the user turns on this option, after all the files
465 have been backed up, but before the job has terminated, the
466 FD will make a pass through all the files and send their
467 names to the DIR (*exactly* the same as what a Verify job
468 currently does). This will probably be done at the same
469 time the files are being sent to the SD avoiding a second
470 pass. The DIR will then compare that to what is stored in
471 the catalog. Any files in the catalog but not in what the
472 FD sent will receive a catalog File entry that indicates
473 that at that point in time the file was deleted.
475 During a restore, any file initially picked up by some
476 backup (Full, ...) then subsequently having a File entry
477 marked "delete" will be removed from the tree, so will not
478 be restored. If a file with the same name is later OK it
479 will be inserted in the tree -- this already happens. All
480 will be consistent except for possible changes during the
483 Since I'm on the subject, some of you may be wondering what
484 the utility of the in memory tree is if you are going to
485 restore everything (at least it comes up from time to time
486 on the list). Well, it is still *very* useful because it
487 allows only the last item found for a particular filename
488 (full path) to be entered into the tree, and thus if a file
489 is backed up 10 times, only the last copy will be restored.
490 I recently (last Friday) restored a complete directory, and
491 the Full and all the Differential and Incremental backups
492 spanned 3 Volumes. The first Volume was not even mounted
493 because all the files had been updated and hence backed up
494 since the Full backup was made. In this case, the tree
495 saved me a *lot* of time.
497 Make sure this information is stored on the tape too so
498 that it can be restored directly from the tape.
500 Comments from Martin Simmons (I think they are all covered):
501 Ok, that should cover the basics. There are few issues though:
503 - Restore will depend on the catalog. I think it is better to include the
504 extra data in the backup as well, so it can be seen by bscan and bextract.
506 - I'm not sure if it will preserve multiple hard links to the same inode. Or
507 maybe adding or removing links will cause the data to be dumped again?
509 - I'm not sure if it will handle renamed directories. Possibly it will work
510 by dumping the whole tree under a renamed directory?
512 - It remains to be seen how the backup performance of the DIR's will be
513 affected when comparing the catalog for a large filesystem.
517 How about introducing a Type = MgmtPolicy job type? That job type would
518 be responsible for scanning the Bacula environment looking for specific
519 conditions, and submitting the appropriate jobs for implementing said
523 Name = "Migration-Policy"
525 Policy Selection Job Type = Migrate
526 Scope = "<keyword> <operator> <regexp>"
527 Threshold = "<keyword> <operator> <regexp>"
528 Job Template = <template-name>
531 Where <keyword> is any legal job keyword, <operator> is a comparison
532 operator (=,<,>,!=, logical operators AND/OR/NOT) and <regexp> is a
533 appropriate regexp. I could see an argument for Scope and Threshold
534 being SQL queries if we want to support full flexibility. The
535 Migration-Policy job would then get scheduled as frequently as a site
536 felt necessary (suggested default: every 15 minutes).
541 Name = "Migration-Policy"
543 Policy Selection Job Type = Migration
545 Threshold = "Migration Selection Type = LowestUtil"
546 Job Template = "MigrationTemplate"
549 would select all pools for examination and generate a job based on
550 MigrationTemplate to automatically select the volume with the lowest
551 usage and migrate it's contents to the nextpool defined for that pool.
553 This policy abstraction would be really handy for adjusting the behavior
554 of Bacula according to site-selectable criteria (one thing that pops
555 into mind is Amanda's ability to automatically adjust backup levels
556 depending on various criteria).
562 - Add Pool/Storage override regression test.
563 - Add delete JobId to regression.
564 - Add a regression test for dbcheck.
565 - New test to add bscan to four-concurrent-jobs regression,
566 i.e. after the four-concurrent jobs zap the
567 database as is done in the bscan-test, then use bscan to
568 restore the database, do a restore and compare with the
570 - Add restore of specific JobId to regression (item 3
571 on the restore prompt)
572 - Add IPv6 to regression
573 - Add database test to regression. Test each function like delete,
576 - AntiVir can slow down backups on Win32 systems.
577 - Win32 systems with FAT32 can be much slower than NTFS for
578 more than 1000 files per directory.
582 - A HOLD command to stop all jobs from starting.
583 - A PAUSE command to pause all running jobs ==> release the
585 - Media Type = LTO,LTO-2,LTO-3
586 Media Type Read = LTO,LTO2,LTO3
587 Media Type Write = LTO2, LTO3
589 === From Carsten Menke <bootsy52@gmx.net>
591 Following is a list of what I think in the situations where I'm faced with,
592 could be a usefull enhancement to bacula, which I'm certain other users will
593 benefit from as well.
595 1. NextJob/NextJobs Directive within a Job Resource in the form of
596 NextJobs = job1,job2.
599 I currently solved the problem with running multiple jobs each after each
600 by setting the Max Wait Time for a job to 8 hours, and give
601 the jobs different Priorities. However, there scenarios where
602 1 Job is directly depending on another job, so if the former job fails,
603 the job after it needn't to be run
604 while maybe other jobs should run despite of that
607 A Backup Job and a Verify job, if the backup job fails there is no need to run
608 the verify job, as the backup job already failed. However, one may like
609 to backup the Catalog to disk despite of that the main backup job failed.
612 I see that this is related to the Event Handlers which are on the ToDo
613 list, also it is maybe a good idea to check for the return value and
614 execute different actions based on the return value
617 3. offline capability to bconsole
620 Currently I use a script which I execute within the last Job via the
621 RunAfterJob Directive, to release and eject the tape.
622 So I have to call bconsole "release=Storage-Name" and afterwards
623 mt -f /dev/nst0 eject to get the tape out.
625 If I have multiple Storage Devices, than these may not be /dev/nst0 and
626 I have to modify the script or call it with parameters etc.
627 This would actually not be needed, as everything is already defined
628 in bacula-sd.conf and if I can invoke bconsole with the
629 storage name via $1 in the script than I'm done and information is
632 4. %s for Storage Name added to the chars being substituted in "RunAfterJob"
636 For the reason mentioned in 3. to have the ability to call a
637 script with /scripts/foobar %s and in the script use $1
638 to pass the Storage Name to bconsole
640 5. Setting Volume State within a Job Resource
643 Instead of using "Maximum Volume Jobs" in the Pool Resource,
644 I would have the possibilty to define
645 in a Job Resource that after this certain job is run, the Volume State
646 should be set to "Volume State = Used", this give more flexibility (IMHO).
648 6. Localization of Bacula Messages
651 Unfortunatley many,many people I work with don't speak english very well.
652 So if at least the Reporting messages would be localized then they
653 would understand that they have to change the tape,etc. etc.
655 I volunteer to do the german translations, and if I can convince my wife also
656 french and Morre (western african language).
658 7. OK, this is evil, probably bound to security risks and maybe not possible
659 due to the design of bacula.
661 Implementation of Backtics ( `command` ) for shell comand execution to
662 the "Label Format" Directive.
666 Currently I have defined BACULA_DAY_OF_WEEK="day1|day2..." resulting in
667 Label Format = "HolyBackup-${BACULA_DAY_OF_WEEK[${WeekDay}]}". If I could
668 use backticks than I could use "Label Format = HolyBackup-`date +%A` to have
669 the localized name for the day of the week appended to the
670 format string. Then I have the tape labeled automatically with weekday
671 name in the correct language.
673 - Yes, that is surely the case. I probably should turn those into Warning
674 errors. In addition, you just made me think that it might not be bad to
675 add an option to check the file size after backing up the file and
676 report if it changes. This would be done as an option because it would
679 Kern, good idea. If you do do that, mention in the output: file
680 shrunk, or file expanded, just to make it obvious to the user
681 (without having to the refer to file size), just how the file size
684 Would this option be for all file, or just one file? Or a fileset?
685 - Make output from status use html table tags for nicely
686 presenting in a browser.
687 - Can one write tapes faster with 8192 byte block sizes?
688 - Document security problems with the same password for everyone in
689 rpm and Win32 releases.
690 - Browse generations of files.
691 - I've seen an error when my catalog's File table fills up. I
692 then have to recreate the File table with a larger maximum row
693 size. Relevant information is at
694 http://dev.mysql.com/doc/mysql/en/Full_table.html ; I think the
695 "Installing and Configuring MySQL" chapter should talk a bit
696 about this potential problem, and recommend a solution.
697 - For Solaris must use POSIX awk.
698 - Want speed of writing to tape while despooling.
699 - Supported autochanger:
707 Wangtek 6525ES (SCSI-1 QIC drive, 525MB), under Linux 2.4.something,
708 bacula 1.36.0/1 works with blocksize 16k INSIDE bacula-sd.conf.
709 - Add regex from http://www.pcre.org to Bacula for Win32.
710 - Use only shell tools no make in CDROM package.
711 - Include within include does it work?
712 - Implement a Pool of type Cleaning?
713 - Implement VolReadTime and VolWriteTime in SD
714 - Modify Backing up Your Database to include a bootstrap file.
715 - Think about making certain database errors fatal.
716 - Look at correcting the time jump in the scheduler for daylight
717 savings time changes.
718 - Add a "real" timer to network connections.
719 - Promote to Full = Time period
720 - Check dates entered by user for correctness (month/day/... ranges)
721 - Compress restore Volume listing by date and first file.
722 - Look at patches/bacula_db.b2z postgresql that loops during restore.
724 - Perhaps add read/write programs and/or plugins to FileSets.
725 - How to handle backing up portables ...
726 - Add some sort of guaranteed Interval for upgrading jobs.
727 - Can we write the state file after every job terminates? On Win32
728 the system crashes and the state file is not updated.
731 Documentation to do: (any release a little bit at a time)
732 - Doc to do unmount before removing magazine.
733 - Alternative to static linking "ldd prog" save all binaries listed,
734 restore them and point LD_LIBRARY_PATH to them.
735 - Document add "</dev/null >/dev/null 2>&1" to the bacula-fd command line
736 - Document query file format.
737 - Add more documentation for bsr files.
738 - Document problems with Verify and pruning.
739 - Document how to use multiple databases.
740 - VXA drives have a "cleaning required"
741 indicator, but Exabyte recommends preventive cleaning after every 75
744 In this context, it should be noted that Exabyte has a command-line
745 vxatool utility available for free download. (The current version is
746 vxatool-3.72.) It can get diagnostic info, read, write and erase tapes,
747 test the drive, unload tapes, change drive settings, flash new firmware,
749 Of particular interest in this context is that vxatool <device> -i will
750 report, among other details, the time since last cleaning in tape motion
751 minutes. This information can be retrieved (and settings changed, for
752 that matter) through the generic-SCSI device even when Bacula has the
753 regular tape device locked. (Needless to say, I don't recommend
754 changing tape settings while a job is running.)
755 - Lookup HP cleaning recommendations.
756 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
757 - Document doing table repair
760 ===================================
761 - Add macro expansions in JobDefs.
762 Run Before Job = "SomeFile %{Level} %{Client}"
763 Write Bootstrap="/some/dir/%{JobName}_%{Client}.bsr"
764 - Use non-blocking network I/O but if no data is available, use
766 - Use gather write() for network I/O.
767 - Autorestart on crash.
768 - Add bandwidth limiting.
769 - Add acks every once and a while from the SD to keep
770 the line from timing out.
771 - When an error in input occurs and conio beeps, you can back
772 up through the prompt.
773 - Detect fixed tape block mode during positioning by looking at
774 block numbers in btape "test". Possibly adjust in Bacula.
775 - Fix list volumes to output volume retention in some other
776 units, perhaps via a directive.
777 - If opening a tape in read/write mode fails attempt to open
778 it in read-only mode, and mark the tape for read only.
779 - Allow Simultaneous Priorities = yes => run up to Max concurrent jobs even
780 with multiple priorities.
781 - If you use restore replace=never, the directory attributes for
782 non-existent directories will not be restored properly.
784 - see lzma401.zip in others directory for new compression
786 - Minimal autochanger handling in Bacula and in btape.
787 - Look into how tar does not save sockets and the possiblity of
788 not saving them in Bacula (Martin Simmons reported this).
789 - Fix restore jobs so that multiple jobs can run if they
790 are not using the same tape(s).
791 - Allow the user to select JobType for manual pruning/purging.
792 - bscan does not put first of two volumes back with all info in
794 - Implement the FreeBSD nodump flag in chflags.
795 - Figure out how to make named console messages go only to that
796 console and to the non-restricted console (new console class?).
797 - Make restricted console prompt for password if *ask* is set or
798 perhaps if password is undefined.
799 - Implement "from ISO-date/time every x hours/days/weeks/months" in
802 ==== from Marc Schoechlin
803 - the help-command should be more verbose
804 (it should explain the paramters of the different
806 -> its time-comsuming to consult the manual anytime
807 you need a special parameter
808 -> maybe its more easy to maintain this, if the
809 descriptions of that commands are outsourced to
811 - the cd-command should allow complete paths
812 i.e. cd /foo/bar/foo/bar
813 -> if a customer mails me the path to a certain file,
814 its faster to enter the specified directory
815 - if the password is not configured in bconsole.conf
816 you should be asked for it.
817 -> sometimes you like to do restore on a customer-machine
818 which shouldnt know the password for bacula.
819 -> adding the password to the file favours admins
820 to forget to remove the password after usage
822 the protection of that file is less important
823 - long-listed-output of commands should be scrollable
824 like the unix more/less-command does
825 -> if someone runs 200 and more machines, the lists could
826 be a little long and complex
827 - command-output should be shown column by column
828 to reduce scrolling and to increase clarity
830 - lsmark should list the selected files with full
832 - wildcards for selecting and file and directories would be nice
833 - any actions should be interuptable with STRG+C
834 - command-expansion would be pretty cool
836 - When the replace Never option is set, new directory permissions
837 are not restored. See bug 213. To fix this requires creating a
838 list of newly restored directories so that those directory
839 permissions *can* be restored.
840 - Compaction of Disk space by "migrating" Volumes that have pruned
841 Jobs (what criteria? size, #jobs, time).
842 - Add prune all command
843 - Document fact that purge can destroy a part of a restore by purging
844 one volume while others remain valid -- perhaps mark Jobs.
845 - Add multiple-media-types.txt
846 - look at mxt-changer.html
847 - Make ? do a help command (no return needed).
848 - Implement restore directory.
849 - Document streams and how to implement them.
850 - Try not to re-backup a file if a new hard link is added.
851 - Add feature to backup hard links only, but not the data.
852 - Fix stream handling to be simpler.
853 - Add Priority and Bootstrap to Run a Job.
854 - Eliminate Restore "Run Restore Job" prompt by allowing new "run command
856 - Remove View FileSet button from Run a Job dialog.
857 - Handle prompt for restore job at end of Restore command.
858 - Add display of total selected files to Restore window.
859 - Add tree pane to left of window.
860 - Add progress meter.
861 - Max wait time or max run time causes seg fault -- see runtime-bug.txt
862 - Document writing to a CD/DVD with Bacula.
863 - Add a "base" package to the window installer for pthreadsVCE.dll
864 which is needed by all packages.
865 - Add message to user to check for fixed block size when the forward
866 space test fails in btape.
867 - When unmarking a directory check if all files below are unmarked and
868 then remove the + flag -- in the restore tree.
869 - Possibly implement: Action = Unmount Device="TapeDrive1" in Admin jobs.
870 - Setup lrrd graphs: (http://www.linpro.no/projects/lrrd/) Mike Acar.
871 - Revisit the question of multiple Volumes (disk) on a single device.
872 - Add a block copy option to bcopy.
873 - Investigate adding Mac Resource Forks.
874 - Finish work on Gnome restore GUI.
875 - Fix "llist jobid=xx" where no fileset or client exists.
876 - For each job type (Admin, Restore, ...) require only the really necessary
877 fields.- Pass Director resource name as an option to the Console.
878 - Add a "batch" mode to the Console (no unsolicited queries, ...).
879 - Add a .list all files in the restore tree (probably also a list all files)
880 Do both a long and short form.
881 - Allow browsing the catalog to see all versions of a file (with
882 stat data on each file).
883 - Restore attributes of directory if replace=never set but directory
885 - Use SHA1 on authentication if possible.
886 - See comtest-xxx.zip for Windows code to talk to USB.
887 - Add John's appended files:
888 Appended = { /files/server/logs/http/*log }
889 and such files would be treated as follows.On a FULL backup, they would
890 be backed up like any other file.On an INCREMENTAL backup, where a
891 previous INCREMENTAL or FULL was already in thecatalogue and the length
892 of the file wasgreater than the length of the last backup, only thedata
893 added since the last backup will be dumped.On an INCREMENTAL backup, if
894 the length of the file is less than thelength of the file with the same
895 name last backed up, the completefile is dumped.On Windows systems, with
896 creation date of files, we can be evensmarter about this and not count
897 entirely upon the length.On a restore, the full and all incrementals
898 since it will beapplied in sequence to restore the file.
899 - Check new HAVE_WIN32 open bits.
900 - Check if the tape has moved before writing.
901 - Handling removable disks -- see below:
902 - Keep track of tape use time, and report when cleaning is necessary.
903 - Add FromClient and ToClient keywords on restore command (or
904 BackupClient RestoreClient).
905 - Implement a JobSet, which groups any number of jobs. If the
906 JobSet is started, all the jobs are started together.
907 Allow Pool, Level, and Schedule overrides.
908 - Enhance cancel to timeout BSOCK packets after a specific delay.
909 - Do scheduling by UTC using gmtime_r() in run_conf, scheduler, and
910 ua_status.!!! Thanks to Alan Brown for this tip.
911 - Look at updating Volume Jobs so that Max Volume Jobs = 1 will work
912 correctly for multiple simultaneous jobs.
913 - Correct code so that FileSet MD5 is calculated for < and | filename
915 - Implement the Media record flag that indicates that the Volume does disk
917 - Implement VolAddr, which is used when Volume is addressed like a disk,
918 and form it from VolFile and VolBlock.
919 - Make multiple restore jobs for multiple media types specifying
920 the proper storage type.
921 - Fix fast block rejection (stored/read_record.c:118). It passes a null
922 pointer (rec) to try_repositioning().
923 - Look at extracting Win data from BackupRead.
924 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
925 which would take precidence over the Catalog "JobRetention".
926 - Implement Label Format in Add and Label console commands.
927 - Possibly up network buffers to 65K. Put on variable.
928 - Put email tape request delays on one or more variables. User wants
929 to cancel the job after a certain time interval. Maximum Mount Wait?
930 - Job, Client, Device, Pool, or Volume?
931 Is it possible to make this a directive which is *optional* in multiple
932 resources, like Level? If so, I think I'd make it an optional directive
933 in Job, Client, and Pool, with precedence such that Job overrides Client
934 which in turn overrides Pool.
936 - New Storage specifications:
937 - Want to write to multiple storage devices simultaneously
938 - Want to write to multiple storage devices sequentially (in one job)
939 - Want to read/write simultaneously
940 - Key is MediaType -- it must match
942 Passed to SD as a sort of BSR record called Storage Specification
946 MediaType -> Next MediaType
948 Device -> Next Device
950 Allow multiple Storage specifications
958 Allow Multiple Pool specifications (note, Pool currently
960 Allow Multiple MediaType specifications in Dir conf
961 Allow Multiple Device specifications in Dir conf
962 Perhaps keep this in a single SSR
963 Tie a Volume to a specific device by using a MediaType that
964 is contained in only one device.
965 In SD allow Device to have Multiple MediaTypes
967 - Ideas from Jerry Scharf:
968 First let's point out some big pluses that bacula has for this
970 more importantly it's active. Thank you so much for that
971 even more important, it's not flaky
972 it has an open access catalog, opening many possibilities
973 it's pushing toward heterogeneous systems capability
975 Macintosh file client
976 macs are an interesting niche, but I fear a server is a rathole
977 working bare iron recovery for windows
978 the option for inc/diff backups not reset on fileset revision
979 a) use both change and inode update time against base time
980 b) do the full catalog check (expensive but accurate)
981 sizing guide (how much system is needed to back up N systems/files)
982 consultants on using bacula in building a disaster recovery system
984 or how to get at fancy things that one could do with bacula
985 logwatch code for bacula logs (or similar)
986 linux distro inclusion of bacula (brings good and bad, but necessary)
987 win2k/XP server capability (icky but you asked)
988 support for Oracle database ??
990 - Look at adding SQL server and Exchange support for Windows.
991 - Make dev->file and dev->block_num signed integers so that -1 can
992 be an invalid value which happens with BSR.
993 - Create VolAddr for disk files in place of VolFile and VolBlock. This
994 is needed to properly specify ranges.
995 - Add progress of files/bytes to SD and FD.
996 - Print warning message if FileId > 4 billion
997 - do a "messages" before the first prompt in Console
998 - Client does not show busy during Estimate command.
999 - Implement Console mtx commands.
1000 - Implement a Mount Command and an Unmount Command where
1001 the users could specify a system command to be performed
1002 to do the mount, after which Bacula could attempt to
1003 read the device. This is for Removeable media such as a CDROM.
1004 - Most likely, this mount command would be invoked explicitly
1005 by the user using the current Console "mount" and "unmount"
1006 commands -- the Storage Daemon would do the right thing
1007 depending on the exact nature of the device.
1008 - As with tape drives, when Bacula wanted a new removable
1009 disk mounted, it would unmount the old one, and send a message
1010 to the user, who would then use "mount" as described above
1011 once he had actually inserted the disk.
1012 - Implement dump/print label to UA
1013 - Spool to disk only when the tape is full, then when a tape is hung move
1015 - bextract is sending everything to the log file ****FIXME****
1016 - Allow multiple Storage specifications (or multiple names on
1017 a single Storage specification) in the Job record. Thus a job
1018 can be backed up to a number of storage devices.
1019 - Implement some way for the File daemon to contact the Director
1020 to start a job or pass its DHCP obtained IP number.
1021 - Implement a query tape prompt/replace feature for a console
1022 - Copy console @ code to gnome2-console
1023 - Make AES the only encryption algorithm see
1024 http://csrc.nist.gov/CryptoToolkit/aes/). It's
1025 an officially adopted standard, has survived peer
1026 review, and provides keys up to 256 bits.
1027 - Take a careful look at SetACL http://setacl.sourceforge.net
1028 - Make tree walk routines like cd, ls, ... more user friendly
1029 by handling spaces better.
1030 - Make sure that Bacula rechecks the tape after the 20 min wait.
1031 - Set IO_NOWAIT on Bacula TCP/IP packets.
1032 - Try doing a raw partition backup and restore by mounting a
1034 - From Lars Kellers:
1035 Yes, it would allow to highly automatic the request for new tapes. If a
1036 tape is empty, bacula reads the barcodes (native or simulated), and if
1037 an unused tape is found, it runs the label command with all the
1038 necessary parameters.
1040 By the way can bacula automatically "move" an empty/purged volume say
1041 in the "short" pool to the "long" pool if this pool runs out of volume
1043 - What to do about "list files job=xxx".
1044 - Get and test MySQL 4.0
1045 - Look at how fuser works and /proc/PID/fd that is how Nic found the
1046 file descriptor leak in Bacula.
1047 - Implement WrapCounters in Counters.
1048 - Add heartbeat from FD to SD if hb interval expires.
1049 - Can we dynamically change FileSets?
1050 - If pool specified to label command and Label Format is specified,
1051 automatically generate the Volume name.
1052 - Why can't SQL do the filename sort for restore?
1053 - Add ExhautiveRestoreSearch
1054 - Look at the possibility of loading only the necessary
1055 data into the restore tree (i.e. do it one directory at a
1056 time as the user walks through the tree).
1057 - Possibly use the hash code if the user selects all for a restore command.
1058 - Fix "restore all" to bypass building the tree.
1059 - Prohibit backing up archive device (findlib/find_one.c:128)
1060 - Implement Release Device in the Job resource to unmount a drive.
1061 - Implement Acquire Device in the Job resource to mount a drive,
1062 be sure this works with admin jobs so that the user can get
1063 prompted to insert the correct tape. Possibly some way to say to
1064 run the job but don't save the files.
1065 - Make things like list where a file is saved case independent for
1068 - Use autochanger to handle multiple devices.
1069 - On Windows with very long path names, it may be impossible to create
1070 a file (and thus restore it) because the total length is too long.
1071 We must cd into the directory then create the file without the
1073 - Implement a Recycle command
1074 - Test a second language e.g. french.
1075 - Start working on Base jobs.
1076 - Implement UnsavedFiles DB record.
1077 - From Phil Stracchino:
1078 It would probably be a per-client option, and would be called
1079 something like, say, "Automatically purge obsoleted jobs". What it
1080 would do is, when you successfully complete a Differential backup of a
1081 client, it would automatically purge all Incremental backups for that
1082 client that are rendered redundant by that Differential. Likewise,
1083 when a Full backup on a client completed, it would automatically purge
1084 all Differential and Incremental jobs obsoleted by that Full backup.
1085 This would let people minimize the number of tapes they're keeping on
1086 hand without having to master the art of retention times.
1087 - When doing a Backup send all attributes back to the Director, who
1088 would then figure out what files have been deleted.
1089 - Currently in mount.c:236 the SD simply creates a Volume. It should have
1090 explicit permission to do so. It should also mark the tape in error
1091 if there is an error.
1092 - Cancel waiting for Client connect in SD if FD goes away.
1094 - Implement timeout in response() when it should come quickly.
1095 - Implement a Slot priority (loaded/not loaded).
1096 - Implement "vacation" Incremental only saves.
1097 - Implement create "FileSet"?
1098 - Add prefixlinks to where or not where absolute links to FD.
1099 - Issue message to mount a new tape before the rewind.
1100 - Simplified client job initiation for portables.
1101 - If SD cannot open a drive, make it periodically retry.
1102 - Add more of the config info to the tape label.
1104 - If tape is marked read-only, then try opening it read-only rather than
1105 failing, and remember that it cannot be written.
1106 - Refine SD waiting output:
1107 Device is being positioned
1108 > Device is being positioned for append
1109 > Device is being positioned to file x
1111 - Figure out some way to estimate output size and to avoid splitting
1112 a backup across two Volumes -- this could be useful for writing CDROMs
1113 where you really prefer not to have it split -- not serious.
1114 - Have SD compute MD5 or SHA1 and compare to what FD computes.
1115 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
1116 actual data on the Volume and compare it.
1117 - Implement Bacula plugins -- design API
1118 - Make bcopy read through bad tape records.
1119 - Program files (i.e. execute a program to read/write files).
1120 Pass read date of last backup, size of file last time.
1121 - Add Signature type to File DB record.
1122 - CD into subdirectory when open()ing files for backup to
1123 speed up things. Test with testfind().
1124 - Priority job to go to top of list.
1125 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
1126 - Implement some way for the Console to dynamically create a job.
1127 - Solaris -I on tar for include list
1128 - Need a verbose mode in restore, perhaps to bsr.
1129 - bscan without -v is too quiet -- perhaps show jobs.
1130 - Add code to reject whole blocks if not wanted on restore.
1131 - Check if we can increase Bacula FD priorty in Win2000
1132 - Make sure the MaxVolFiles is fully implemented in SD
1133 - Check if both CatalogFiles and UseCatalog are set to SD.
1134 - Possibly add email to Watchdog if drive is unmounted too
1135 long and a job is waiting on the drive.
1136 - After unmount, if restore job started, ask to mount.
1137 - Add UA rc and history files.
1138 - put termcap (used by console) in ./configure and
1139 allow -with-termcap-dir.
1140 - Fix Autoprune for Volumes to respect need for full save.
1141 - Compare tape to Client files (attributes, or attributes and data)
1142 - Make all database Ids 64 bit.
1143 - Allow console commands to detach or run in background.
1144 - Fix status delay on storage daemon during rewind.
1145 - Add SD message variables to control operator wait time
1146 - Maximum Operator Wait
1147 - Minimum Message Interval
1148 - Maximum Message Interval
1149 - Send Operator message when cannot read tape label.
1150 - Verify level=Volume (scan only), level=Data (compare of data to file).
1151 Verify level=Catalog, level=InitCatalog
1153 - Add keyword search to show command in Console.
1154 - Events : tape has more than xxx bytes.
1155 - Complete code in Bacula Resources -- this will permit
1156 reading a new config file at any time.
1157 - Handle ctl-c in Console
1158 - Implement script driven addition of File daemon to config files.
1159 - Think about how to make Bacula work better with File (non-tape) archives.
1160 - Write Unix emulator for Windows.
1161 - Put memory utilization in Status output of each daemon
1162 if full status requested or if some level of debug on.
1163 - Make database type selectable by .conf files i.e. at runtime
1164 - Set flag for uname -a. Add to Volume label.
1165 - Restore files modified after date
1166 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
1167 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
1168 - Timout a job or terminate if link goes down, or reopen link and query.
1169 - Concept of precious tapes (cannot be reused).
1170 - Make bcopy copy with a single tape drive.
1171 - Permit changing ownership during restore.
1174 > My suggestion: Add a feature on the systray menu-icon menu to request
1175 > an immediate backup now. This would be useful for laptop users who may
1176 > not be on the network when the regular scheduled backup is run.
1178 > My wife's suggestion: Add a setting to the win32 client to allow it to
1179 > shut down the machine after backup is complete (after, of course,
1180 > displaying a "System will shut down in one minute, click here to cancel"
1181 > warning dialog). This would be useful for sites that want user
1182 > woorkstations to be shut down overnight to save power.
1185 - Autolabel should be specified by DIR instead of SD.
1187 - Add media capacity
1188 - AutoScan (check checksum of tape)
1189 - Format command = "format /dev/nst0"
1193 - Seek resolution (usually corresponds to buffer size)
1194 - EODErrorCode=ENOSPC or code
1195 - Partial Read error code
1196 - Partial write error code
1197 - Nonformatted read error
1198 - Nonformatted write error
1199 - WriteProtected error
1203 - IgnoreCloseErrors=yes
1213 - FD sends unsaved file list to Director at end of job (see
1215 - File daemon should build list of files skipped, and then
1216 at end of save retry and report any errors.
1217 - Write a Storage daemon that uses pipes and
1218 standard Unix programs to write to the tape.
1220 - Need something that monitors the JCR queue and
1221 times out jobs by asking the deamons where they are.
1222 - Enhance Jmsg code to permit buffering and saving to disk.
1223 - device driver = "xxxx" for drives.
1224 - Verify from Volume
1225 - Ensure that /dev/null works
1226 - Need report class for messages. Perhaps
1227 report resource where report=group of messages
1228 - enhance scan_attrib and rename scan_jobtype, and
1229 fill in code for "since" option
1230 - Director needs a time after which the report status is sent
1231 anyway -- or better yet, a retry time for the job.
1232 - Don't reschedule a job if previous incarnation is still running.
1233 - Some way to automatically backup everything is needed????
1234 - Need a structure for pending actions:
1236 - termination status (part of buffered msgs?)
1238 Read, Write, Clean, Delete
1239 - Login to Bacula; Bacula users with different permissions:
1240 owner, group, user, quotas
1241 - Store info on each file system type (probably in the job header on tape.
1242 This could be the output of df; or perhaps some sort of /etc/mtab record.
1244 ========= ideas ===============
1245 From: "Jerry K. Schieffer" <jerry@skylinetechnology.com>
1246 To: <kern@sibbald.com>
1247 Subject: RE: [Bacula-users] future large programming jobs
1248 Date: Thu, 26 Feb 2004 11:34:54 -0600
1250 I noticed the subject thread and thought I would offer the following
1251 merely as sources of ideas, i.e. something to think about, not even as
1252 strong as a request. In my former life (before retiring) I often
1253 dealt with backups and storage management issues/products as a
1254 developer and as a consultant. I am currently migrating my personal
1255 network from amanda to bacula specifically because of the ability to
1256 cross media boundaries during storing backups.
1257 Are you familiar with the commercial product called ADSM (I think IBM
1258 now sells it under the Tivoli label)? It has a couple of interesting
1259 ideas that may apply to the following topics.
1261 1. Migration: Consider that when you need to restore a system, there
1262 may be pressure to hurry. If all the information for a single client
1263 can eventually end up on the same media (and in chronological order),
1264 the restore is facillitated by not having to search past information
1265 from other clients. ADSM has the concept of "client affinity" that
1266 may be associated with it's storage pools. It seems to me that this
1267 concept (as an optional feature) might fit in your architecture for
1270 ADSM also has the concept of defining one or more storage pools as
1271 "copy pools" (almost mirrors, but only in the sense of contents).
1272 These pools provide the ability to have duplicte data stored both
1273 onsite and offsite. The copy process can be scheduled to be handled
1274 by their storage manager during periods when there is no backup
1275 activity. Again, the migration process might be a place to consider
1276 implementing something like this.
1279 > It strikes me that it would be very nice to be able to do things
1281 > have the Job(s) backing up the machines run, and once they have all
1282 > completed, start a migration job to copy the data from disks Volumes
1284 > a tape library and then to offsite storage. Maybe this can already
1286 > done with some careful scheduling and Job prioritzation; the events
1287 > mechanism described below would probably make it very easy.
1289 This is the goal. In the first step (before events), you simply
1291 the Migration to tape later.
1293 2. Base jobs: In ADSM, each copy of each stored file is tracked in
1294 the database. Once a file (unique by path and metadata such as dates,
1295 size, ownership, etc.) is in a copy pool, no more copies are made. In
1296 other words, when you start ADSM, it begins like your concept of a
1297 base job. After that it is in the "incremental" mode. You can
1298 configure the number of "generations" of files to be retained, plus a
1299 retention date after which even old generations are purged. The
1300 database tracks the contents of media and projects the percentage of
1301 each volume that is valid. When the valid content of a volume drops
1302 below a configured percentage, the valid data are migrated to another
1303 volume and the old volume is marked as empty. Note, this requires
1304 ADSM to have an idea of the contents of a client, i.e. marking the
1305 database when an existing file was deleted, but this would solve your
1306 issue of restoring a client without restoring deleted files.
1308 This is pretty far from what bacula now does, but if you are going to
1309 rip things up for Base jobs,.....
1310 Also, the benefits of this are huge for very large shops, especially
1311 with media robots, but are a pain for shops with manual media
1315 > Base jobs sound pretty useful, but I'm not dying for them.
1317 Nobody is dying for them, but when you see what it does, you will die
1320 3. Restoring deleted files: Since I think my comments in (2) above
1321 have low probability of implementation, I'll also suggest that you
1322 could approach the issue of deleted files by a mechanism of having the
1323 fd report to the dir, a list of all files on the client for every
1324 backup job. The dir could note in the database entry for each file
1325 the date that the file was seen. Then if a restore as of date X takes
1326 place, only files that exist from before X until after X would be
1327 restored. Probably the major cost here is the extra date container in
1328 each row of the files table.
1330 Thanks for "listening". I hope some of this helps. If you want to
1331 contact me, please send me an email - I read some but not all of the
1332 mailing list traffic and might miss a reply there.
1334 Please accept my compliments for bacula. It is doing a great job for
1335 me!! I sympathize with you in the need to wrestle with excelence in
1336 execution vs. excelence in feature inclusion.
1341 ==============================
1344 - Design at hierarchial storage for Bacula. Migration and Clone.
1345 - Implement FSM (File System Modules).
1346 - Audit M_ error codes to ensure they are correct and consistent.
1347 - Add variable break characters to lex analyzer.
1348 Either a bit mask or a string of chars so that
1349 the caller can change the break characters.
1350 - Make a single T_BREAK to replace T_COMMA, etc.
1351 - Ensure that File daemon and Storage daemon can
1352 continue a save if the Director goes down (this
1353 is NOT currently the case). Must detect socket error,
1354 buffer messages for later.
1355 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
1356 - Add ability to backup to two Storage devices (two SD sessions) at
1357 the same time -- e.g. onsite, offsite.
1358 - Add the ability to consolidate old backup sets (basically do a restore
1359 to tape and appropriately update the catalog). Compress Volume sets.
1360 Might need to spool via file is only one drive is available.
1361 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
1362 someway to do so with every volume that has less than x% valid
1366 Migration: Move a backup from one Volume to another
1367 Clone: Copy a backup -- two Volumes
1369 Bacula Migration is based on Jobs (apparently Networker is file by file).
1371 Migration triggered by:
1375 Highwater mark (keep total size)
1380 ======================================================
1382 It is somewhat like a Full save becomes an incremental since
1383 the Base job (or jobs) plus other non-base files.
1385 - A Base backup is same as Full backup, just different type.
1386 - New BaseFiles table that contains:
1388 BaseJobId - Base JobId referenced for this FileId (needed ???)
1389 JobId - JobId currently running
1390 FileId - File not backed up, exists in Base Job
1391 FileIndex - FileIndex from Base Job.
1392 i.e. for each base file that exists but is not saved because
1393 it has not changed, the File daemon sends the JobId, BaseId,
1394 FileId, FileIndex back to the Director who creates the DB entry.
1395 - To initiate a Base save, the Director sends the FD
1396 the FileId, and full filename for each file in the Base.
1397 - When the FD finds a Base file, he requests the Director to
1398 send him the full File entry (stat packet plus MD5), or
1399 conversely, the FD sends it to the Director and the Director
1400 says yes or no. This can be quite rapid if the FileId is kept
1401 by the FD for each Base Filename.
1402 - It is probably better to have the comparison done by the FD
1403 despite the fact that the File entry must be sent across the
1405 - An alternative would be to send the FD the whole File entry
1406 from the start. The disadvantage is that it requires a lot of
1407 space. The advantage is that it requires less communications
1409 - The Job record must be updated to indicate that one or more
1411 - At end of Job, FD returns:
1412 1. Count of base files/bytes not written to tape (i.e. matches)
1413 2. Count of base file that were saved i.e. had changed.
1414 - No tape record would be written for a Base file that matches, in the
1415 same way that no tape record is written for Incremental jobs where
1416 the file is not saved because it is unchanged.
1417 - On a restore, all the Base file records must explicitly be
1418 found from the BaseFile tape. I.e. for each Full save that is marked
1419 to have one or more Base Jobs, search the BaseFile for all occurrences
1421 - An optimization might be to make the BaseFile have:
1427 This would avoid the need to explicitly fetch each File record for
1428 the Base job. The Base Job record will be fetched to get the
1429 VolSessionId and VolSessionTime.
1430 =========================================================
1433 ==========================================================
1435 For each Incremental job that is run, there may be files that
1436 were found but not saved because they were locked (this applies
1437 only to Windows). Such a system could send back to the Director
1438 a list of Unsaved files.
1440 - New UnSavedFiles table that contains:
1444 - Then in the next Incremental job, the list of Unsaved Files will be
1445 feed to the FD, who will ensure that they are explicitly chosen even
1446 if standard date/time check would not have selected them.
1447 =============================================================
1451 Multiple drive autochanger data: see Alan Brown
1452 mtx -f xxx unloadStorage Element 1 is Already Full(drive 0 was empty)
1453 Unloading Data Transfer Element into Storage Element 1...source Element
1454 Address 480 is Empty
1456 (drive 0 was empty and so was slot 1)
1457 > mtx -f xxx load 15 0
1458 no response, just returns to the command prompt when complete.
1459 > mtx -f xxx status Storage Changer /dev/changer:2 Drives, 60 Slots ( 2 Import/Export )
1460 Data Transfer Element 0:Full (Storage Element 15 Loaded):VolumeTag = HX001
1461 Data Transfer Element 1:Empty
1462 Storage Element 1:Empty
1463 Storage Element 2:Full :VolumeTag=HX002
1464 Storage Element 3:Full :VolumeTag=HX003
1465 Storage Element 4:Full :VolumeTag=HX004
1466 Storage Element 5:Full :VolumeTag=HX005
1467 Storage Element 6:Full :VolumeTag=HX006
1468 Storage Element 7:Full :VolumeTag=HX007
1469 Storage Element 8:Full :VolumeTag=HX008
1470 Storage Element 9:Full :VolumeTag=HX009
1471 Storage Element 10:Full :VolumeTag=HX010
1472 Storage Element 11:Empty
1473 Storage Element 12:Empty
1474 Storage Element 13:Empty
1475 Storage Element 14:Empty
1476 Storage Element 15:Empty
1477 Storage Element 16:Empty....
1478 Storage Element 28:Empty
1479 Storage Element 29:Full :VolumeTag=CLNU01L1
1480 Storage Element 30:Empty....
1481 Storage Element 57:Empty
1482 Storage Element 58:Full :VolumeTag=NEX261L2
1483 Storage Element 59 IMPORT/EXPORT:Empty
1484 Storage Element 60 IMPORT/EXPORT:Empty
1486 Unloading Data Transfer Element into Storage Element 15...done
1488 (just to verify it remembers where it came from, however it can be
1489 overrriden with mtx unload {slotnumber} to go to any storage slot.)
1491 There needs to be a table of drive # to devices somewhere - If there are
1492 multiple changers or drives there may not be a 1:1 correspondance between
1493 changer drive number and system device name - and depending on the way the
1494 drives are hooked up to scsi busses, they may not be linearly numbered
1495 from an offset point either.something like
1497 Autochanger drives = 2
1498 Autochanger drive 0 = /dev/nst1
1499 Autochanger drive 1 = /dev/nst2
1500 IMHO, it would be _safest_ to use explicit mtx unload commands at all
1501 times, not just for multidrive changers. For a 1 drive changer, that's
1507 MTX's manpage (1.2.15):
1508 unload [<slotnum>] [ <drivenum> ]
1509 Unloads media from drive <drivenum> into slot
1510 <slotnum>. If <drivenum> is omitted, defaults to
1511 drive 0 (as do all commands). If <slotnum> is
1512 omitted, defaults to the slot that the drive was
1513 loaded from. Note that there's currently no way
1514 to say 'unload drive 1's media to the slot it
1515 came from', other than to explicitly use that
1516 slot number as the destination.AB
1522 undef# camcontrol devlist
1523 <WANGTEK 51000 SCSI M74H 12B3> at scbus0 target 2 lun 0 (pass0,sa0)
1524 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 0 (pass1,sa1)
1525 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 1 (pass2)
1527 tapeinfo -f /dev/sg0 with a bad tape in drive 1:
1528 [kern@rufus mtx-1.2.17kes]$ ./tapeinfo -f /dev/sg0
1529 Product Type: Tape Drive
1531 Product ID: 'C5713A '
1533 Attached Changer: No
1534 TapeAlert[3]: Hard Error: Uncorrectable read/write error.
1535 TapeAlert[20]: Clean Now: The tape drive neads cleaning NOW.
1542 Medium Type: Not Loaded
1545 DataCompEnabled: yes
1546 DataCompCapable: yes
1547 DataDeCompEnabled: yes
1554 Handling removable disks
1556 From: Karl Cunningham <karlc@keckec.com>
1558 My backups are only to hard disk these days, in removable bays. This is my
1559 idea of how a backup to hard disk would work more smoothly. Some of these
1560 things Bacula does already, but I mention them for completeness. If others
1561 have better ways to do this, I'd like to hear about it.
1563 1. Accommodate several disks, rotated similar to how tapes are. Identified
1564 by partition volume ID or perhaps by the name of a subdirectory.
1565 2. Abort & notify the admin if the wrong disk is in the bay.
1566 3. Write backups to different subdirectories for each machine to be backed
1568 4. Volumes (files) get created as needed in the proper subdirectory, one
1570 5. When a disk is recycled, remove or zero all old backup files. This is
1571 important as the disk being recycled may be close to full. This may be
1572 better done manually since the backup files for many machines may be
1573 scattered in many subdirectories.
1578 - Make sure that all do_prompt() calls in Dir check for
1579 -1 (error) and -2 (cancel) returns.
1580 - Fix foreach_jcr() to have free_jcr() inside next().
1581 jcr=jcr_walk_start();
1582 for ( ; jcr; (jcr=jcr_walk_next(jcr)) )
1585 - A Volume taken from Scratch should take on the retention period
1587 - Correct doc for Maximum Changer Wait (and others) accepting only
1589 - Implement status that shows why a job is being held in reserve, or
1590 rather why none of the drives are suitable.
1591 - Implement a way to disable a drive (so you can use the second
1592 drive of an autochanger, and the first one will not be used or
1594 - Make sure Maximum Volumes is respected in Pools when adding
1595 Volumes (e.g. when pulling a Scratch volume).
1596 - Keep same dcr when switching device ...
1597 - Implement code that makes the Dir aware that a drive is an
1598 autochanger (so the user doesn't need to use the Autochanger = yes
1600 - Make catalog respect ACL.
1601 - Add recycle count to Media record.
1602 - Add initial write date to Media record.
1603 - Fix store_yesno to be store_bitmask.
1604 --- create_file.c.orig Fri Jul 8 12:13:05 2005
1605 +++ create_file.c Fri Jul 8 12:13:07 2005
1607 attr->ofname, be.strerror());
1610 + } else if(S_ISSOCK(attr->statp.st_mode)) {
1611 + Dmsg1(200, "Skipping socket: %s\n", attr->ofname);
1613 Dmsg1(200, "Restore node: %s\n", attr->ofname);
1614 if (mknod(attr->ofname, attr->statp.st_mode, attr->statp.st_rdev) != 0 && errno != EEXIST) {
1615 - Add true/false to conf same as yes/no
1616 - Reserve blocks other restore jobs when first cannot connect to SD.
1617 - Fix Maximum Changer Wait, Maximum Open Wait, Maximum Rewind Wait to
1618 accept time qualifiers.
1619 - Does ClientRunAfterJob fail the job on a bad return code?
1620 - Make hardlink code at line 240 of find_one.c use binary search.
1621 - Add ACL error messages in src/filed/acl.c.
1622 - Make authentication failures single threaded.
1623 - Make Dir and SD authentication errors single threaded.
1625 - Fix catreq.c digestbuf at line 411 in src/dird/catreq.c
1626 - Make base64.c (bin_to_base64) take a buffer length
1627 argument to avoid overruns.
1628 and verify that other buffers cannot overrun.
1629 - Implement VolumeState as discussed with Arno.
1630 - Add LocationId to update volume
1638 - Add Comment to Media record
1639 - Fix auth compatibility with 1.38
1640 - Update dbcheck to include Log table
1641 - Update llist to include new fields.