6 - !!! Cannot restore two jobs a the same time that were
7 written simultaneously unless they were totally spooled.
8 - Document cleaning up the spool files:
9 db, pid, state, bsr, mail, conmsg, spool
10 - Document the multiple-drive-changer.txt script.
11 - Pruning with Admin job.
12 - Does WildFile match against full name? Doc.
13 - %d and %v only valid on Director, not for ClientRunBefore/After.
14 - During tests with the 260 char fix code, I found one problem:
15 if the system "sees" a long path once, it seems to forget it's
16 working drive (e.g. c:\), which will lead to a problem during
17 the next job (create bootstrap file will fail). Here is the
18 workaround: specify absolute working and pid directory in
19 bacula-fd.conf (e.g. c:\bacula\working instead of
21 - Document techniques for restoring large numbers of files.
22 - Document setting my.cnf to big file usage.
23 - Add example of proper index output to doc. show index from File;
24 - Correct the Include syntax in the m4.xxx files in examples/conf
25 - Document JobStatus and Termination codes.
26 - Fix the error with the "DVI file can't be opened" while
27 building the French PDF.
28 - Document more DVD stuff
36 - Document all the little details of setting up certificates for
37 the Bacula data encryption code.
38 - Document more precisely how to use master keys -- especially
39 for disaster recovery.
42 - Migration from other vendors
46 - Backup conf/exe (all daemons)
47 - Backup up system state
48 - Detect state change of system (verify)
49 - Synthetic Full, Diff, Inc (Virtual, Reconstructed)
51 - Modules for Databases, Exchange, ...
54 - How does restore JobId=nnn work? (Dirk)
55 - What does: restore select fileset="TestSet" client="workplay-fd" pool="Default"
56 > storage="File2" before="2007-02-05 23:05:04" do? (Dirk)
58 - Implement update jobid=
59 - Mention Eric and Marc's work + Marc's doc.
61 - When Pool specifies Storage command override does not work.
62 - Implement wait_for_sysop() message display in wait_for_device(), which
63 now prints warnings too often.
65 - Ensure that each device in an Autochanger has a different
67 - Add Catalog = to Pool resource so that pools will exist
68 in only one catalog -- currently Pools are "global".
69 - Look at sg_logs -a /dev/sg0 for getting soft errors.
70 - btape "test" command with Offline on Unmount = yes
72 This test is essential to Bacula.
74 I'm going to write one record in file 0,
75 two records in file 1,
76 and three records in file 2
78 02-Feb 11:00 btape: ABORTING due to ERROR in dev.c:715
79 dev.c:714 Bad call to rewind. Device "LTO" (/dev/nst0) not open
80 02-Feb 11:00 btape: Fatal Error because: Bacula interrupted by signal 11: Segmentation violation
81 Kaboom! btape, btape got signal 11. Attempting traceback.
83 - Ensure that moving a purged Volume in ua_purge.c to the RecyclePool
85 - Why doesn't @"xxx abc" work in a conf file?
86 - Figure out some way to "automatically" backup conf changes.
87 - Add the OS version back to the Win32 client info.
88 - Restarted jobs have a NULL in the from field.
89 - Modify SD status command to indicate when the SD is writing
90 to a DVD (the device is not open -- see bug #732).
91 - Look at the possibility of adding "SET NAMES UTF8" for MySQL,
92 and possibly changing the blobs into varchar.
93 - Ensure that the SD re-reads the Media record if the JobFiles
94 does not match -- it may have been updated by another job.
95 - Look at moving the Storage directive from the Job to the
96 Pool in the default conf files.
98 - Test Volume compatibility between machine architectures
99 - Encryption documentation
100 - Wrong jobbytes with query 12 (todo)
101 - bacula-1.38.2-ssl.patch
102 - Bare-metal recovery Windows (todo)
106 - Average tape size from Eric
107 SELECT COALESCE(media_avg_size.volavg,0) * count(Media.MediaId) AS volmax, GROUP BY Media.MediaType, Media.PoolId, media_avg_size.volavg
108 count(Media.MediaId) AS volnum,
109 sum(Media.VolBytes) AS voltotal,
110 Media.PoolId AS PoolId,
111 Media.MediaType AS MediaType
113 LEFT JOIN (SELECT avg(Media.VolBytes) AS volavg,
114 Media.MediaType AS MediaType
116 WHERE Media.VolStatus = 'Full'
117 GROUP BY Media.MediaType
118 ) AS media_avg_size ON (Media.MediaType = media_avg_size.MediaType)
119 GROUP BY Media.MediaType, Media.PoolId, media_avg_size.volavg
123 - Add doc for bweb -- especially Installation
125 http://www.orangecrate.com/modules.php?name=News&file=article&sid=501
127 - FD-SD quick disconnect
128 - Despool attributes in separate thread
131 - Check why restore repeatedly sends Rechdrs between
132 each data chunk -- according to James Harper 9Jan07.
133 - Building the in memory restore tree is slow.
136 - Full at least once a month, ...
137 - Cancel Inc if Diff/Full running
138 - More intelligent re-run
139 - New/deleted file backup
141 - Incremental backup -- rsync, Stow
145 - Look at mondo/mindi
146 - Don't restore Solaris Door files:
147 #define S_IFDOOR in st_mode.
148 see: http://docs.sun.com/app/docs/doc/816-5173/6mbb8ae23?a=view#indexterm-360
149 - Make Bacula by default not backup tmpfs, procfs, sysfs, ...
150 - Fix hardlinked immutable files when linking a second file, the
151 immutable flag must be removed prior to trying to link it.
152 - Implement Python event for backing up/restoring a file.
153 - Change dbcheck to tell users to use native tools for fixing
154 broken databases, and to ensure they have the proper indexes.
155 - add udev rules for Bacula devices.
156 - If a job terminates, the DIR connection can close before the
157 Volume info is updated, leaving the File count wrong.
158 - Look at why SIGPIPE during connection can cause seg fault in
159 writing the daemon message, when Dir dropped to bacula:bacula
160 - Look at zlib 32 => 64 problems.
161 - Possibly turn on St. Bernard code.
162 - Fix bextract to restore ACLs, or better yet, use common routines.
163 - Do we migrate appendable Volumes?
164 - Remove queue.c code.
165 - Print warning message if LANG environment variable does not specify
167 - New dot commands from Arno.
168 .show device=xxx lists information from one storage device, including
169 devices (I'm not even sure that information exists in the DIR...)
170 .move eject device=xxx mostly the same as 'unmount xxx' but perhaps with
171 better machine-readable output like "Ok" or "Error busy"
172 .move eject device=xxx toslot=yyy the same as above, but with a new
173 target slot. The catalog should be updated accordingly.
174 .move transfer device=xxx fromslot=yyy toslot=zzz
177 - Article: http://www.heise.de/open/news/meldung/83231
178 - Article: http://www.golem.de/0701/49756.html
179 - Article: http://lwn.net/Articles/209809/
180 - Article: http://www.onlamp.com/pub/a/onlamp/2004/01/09/bacula.html
181 - Article: http://www.linuxdevcenter.com/pub/a/linux/2005/04/07/bacula.html
182 - Article: http://www.osreviews.net/reviews/admin/bacula
183 - Article: http://www.debianhelp.co.uk/baculaweb.htm
185 - It appears to me that you have run into some sort of race
186 condition where two threads want to use the same Volume and they
187 were both given access. Normally that is no problem. However,
188 one thread wanted the particular Volume in drive 0, but it was
189 loaded into drive 1 so it decided to unload it from drive 1 and
190 then loaded it into drive 0, while the second thread went on
191 thinking that the Volume could be used in drive 1 not realizing
192 that in between time, it was loaded in drive 0.
193 I'll look at the code to see if there is some way we can avoid
194 this kind of problem. Probably the best solution is to make the
195 first thread simply start using the Volume in drive 1 rather than
196 transferring it to drive 0.
197 - Fix re-read of last block to check if job has actually written
198 a block, and check if block was written by a different job
199 (i.e. multiple simultaneous jobs writing).
200 - Figure out how to configure query.sql. Suggestion to use m4:
201 == changequote.m4 ===
202 changequote(`[',`]')dnl
203 ==== query.sql.in ===
204 :List next 20 volumes to expire
206 Pool.Name AS PoolName,
211 [ FROM_UNIXTIME(UNIX_TIMESTAMP(Media.LastWritten) Media.VolRetention) AS Expire, ])dnl
213 [ media.lastwritten + interval '1 second' * media.volretention as expire, ])dnl
217 ON Media.PoolId=Pool.PoolId
218 WHERE Media.LastWritten>0
222 Command: m4 -DmySQL changequote.m4 query.sql.in >query.sql
224 The problem is that it requires m4, which is not present on all machines
226 - Given all the problems with FIFOs, I think the solution is to do something a
227 little different, though I will look at the code and see if there is not some
228 simple solution (i.e. some bug that was introduced). What might be a better
229 solution would be to use a FIFO as a sort of "key" to tell Bacula to read and
230 write data to a program rather than the FIFO. For example, suppose you
235 Then, I could imagine if you backup and restore this file with a direct
236 reference as is currently done for fifos, instead, during backup Bacula will
239 /home/kern/my-fifo.backup
241 and read the data that my-fifo.backup writes to stdout. For restore, Bacula
244 /home/kern/my-fifo.restore
246 and send the data backed up to stdout. These programs can either be an
247 executable or a shell script and they need only read/write to stdin/stdout.
249 I think this would give a lot of flexibility to the user without making any
250 significant changes to Bacula.
255 select FilenameId from Filename where Name='';
256 # Get list of all directories referenced in a Backup.
257 select Path.Path from Path,File where File.JobId=nnn and
258 File.FilenameId=(FilenameId-from-above) and File.PathId=Path.PathId
259 order by Path.Path ASC;
261 - Look into using Dart for testing
262 http://public.kitware.com/Dart/HTML/Index.shtml
264 - Look into replacing autotools with cmake
265 http://www.cmake.org/HTML/Index.html
267 === Migration from David ===
268 What I'd like to see:
271 Name = "<poolname>-migrate"
275 Migration Selection Type = LowestUtil | OldestVol | PoolOccupancy |
276 Client | PoolResidence | Volume | JobName | SQLquery
277 Migration Selection Pattern = "regexp"
278 Next Pool = <override>
281 There should be no need for a Level (migration is always Full, since you
282 don't calculate differential/incremental differences for migration),
283 Storage should be determined by the volume types in the pool, and Client
284 is really a selection issue. Migration should always occur to the
285 NextPool defined in the pool definition. If no nextpool is defined, the
286 job should end with a reason of "no place to go". If Next Pool statement
287 is present, we override the check in the pool definition and use the
290 Here's how I'd define Migration Selection Types:
293 Client -- Migrate data from selected client only. Migration Selection
294 Pattern regexp provides pattern to select client names, eg ^FS00* makes
295 all client names starting with FS00 eligible for migration.
297 Jobname -- Migration all jobs matching name. Migration Selection Pattern
298 regexp provides pattern to select jobnames existing in pool.
300 Volume -- Migrate all data on specified volumes. Migration Selection
301 Pattern regexp provides selection criteria for volumes to be migrated.
302 Volumes must exist in pool to be eligible for migration.
306 LowestUtil -- Identify the volume in the pool with the least data on it
307 and empty it. No Migration Selection Pattern required.
309 OldestVol -- Identify the LRU volume with data written, and empty it. No
310 Migration Selection Pattern required.
312 PoolOccupancy -- if pool occupancy exceeds <highmig>, migrate volumes
313 (starting with most full volumes) until pool occupancy drops below
314 <lowmig>. Pool highmig and lowmig values are in pool definition, no
315 Migration Selection Pattern required.
319 SQLQuery -- Migrate all jobuids returned by the supplied SQL query.
320 Migration Selection Pattern contains SQL query to execute; should return
321 a list of 1 or more jobuids to migrate.
323 PoolResidence -- Migrate data sitting in pool for longer than
324 PoolResidence value in pool definition. Migration Selection Pattern
325 optional; if specified, override value in pool definition (value in
329 [ possibly a Python event -- kes ]
331 - Mount on an Autochanger with no tape in the drive causes:
332 Automatically selected Storage: LTO-changer
333 Enter autochanger drive[0]: 0
334 3301 Issuing autochanger "loaded drive 0" command.
335 3302 Autochanger "loaded drive 0", result: nothing loaded.
336 3301 Issuing autochanger "loaded drive 0" command.
337 3302 Autochanger "loaded drive 0", result: nothing loaded.
338 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
339 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
340 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
341 If this is not a blank tape, try unmounting and remounting the Volume.
342 - If Drive 0 is blocked, and drive 1 is set "Autoselect=no", drive 1 will
344 - Autochanger did not change volumes.
345 select * from Storage;
346 +-----------+-------------+-------------+
347 | StorageId | Name | AutoChanger |
348 +-----------+-------------+-------------+
349 | 1 | LTO-changer | 0 |
350 +-----------+-------------+-------------+
351 05-May 03:50 roxie-sd: 3302 Autochanger "loaded drive 0", result is Slot 11.
352 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Warning: Director wanted Volume "LT
353 Current Volume "LT0-002" not acceptable because:
354 1997 Volume "LT0-002" not in catalog.
355 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Error: Autochanger Volume "LT0-002"
356 Setting InChanger to zero in catalog.
357 05-May 03:50 roxie-dir: Tibs.2006-05-05_03.05.02 Error: Unable to get Media record
359 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Error getting Volume i
360 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Job 530 canceled.
361 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: spool.c:249 Fatal appe
362 05-May 03:49 Tibs: Tibs.2006-05-05_03.05.02 Fatal error: c:\cygwin\home\kern\bacula
371 FirstWritten: 2006-05-05 03:11:54
372 LastWritten: 2006-05-05 03:50:23
373 LabelDate: 2005-12-26 16:52:40
384 VolRetention: 31,536,000
396 Note VolStatus is blank!!!!!
403 FirstWritten: 0000-00-00 00:00:00
404 LastWritten: 0000-00-00 00:00:00
405 LabelDate: 2005-12-26 16:52:40
416 VolRetention: 31,536,000
429 Automatically selected Storage: LTO-changer
430 Enter autochanger drive[0]: 0
431 3301 Issuing autochanger "loaded drive 0" command.
432 3302 Autochanger "loaded drive 0", result: nothing loaded.
433 3301 Issuing autochanger "loaded drive 0" command.
434 3302 Autochanger "loaded drive 0", result: nothing loaded.
435 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
436 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
438 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
439 If this is not a blank tape, try unmounting and remounting the Volume.
441 - http://www.dwheeler.com/essays/commercial-floss.html
442 - Add VolumeLock to prevent all but lock holder (SD) from updating
443 the Volume data (with the exception of VolumeState).
444 - The btape fill command does not seem to use the Autochanger
445 - Make Windows installer default to system disk drive.
446 - Look at using ioctl(FIOBMAP, ...) on Linux, and
447 DeviceIoControl(..., FSCTL_QUERY_ALLOCATED_RANGES, ...) on
448 Win32 for sparse files.
449 http://www.flexhex.com/docs/articles/sparse-files.phtml
450 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
451 - Directive: at <event> "command"
452 - Command: pycmd "command" generates "command" event. How to
453 attach to a specific job?
454 - Integrate Christopher's St. Bernard code.
455 - run_cmd() returns int should return JobId_t
456 - get_next_jobid_from_list() returns int should return JobId_t
457 - Document export LDFLAGS=-L/usr/lib64
458 - Don't attempt to restore from "Disabled" Volumes.
459 - Network error on Win32 should set Win32 error code.
460 - What happens when you rename a Disk Volume?
461 - Job retention period in a Pool (and hence Volume). The job would
463 - Look at -D_FORTIFY_SOURCE=2
464 - Add Win32 FileSet definition somewhere
465 - Look at fixing restore status stats in SD.
466 - Look at using ioctl(FIMAP) and FIGETBSZ for sparse files.
467 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
468 - Implement a mode that says when a hard read error is
469 encountered, read many times (as it currently does), and if the
470 block cannot be read, skip to the next block, and try again. If
471 that fails, skip to the next file and try again, ...
473 create table LevelType (LevelType binary(1), LevelTypeLong tinyblob);
474 insert into LevelType (LevelType,LevelTypeLong) values
478 - Show files/second in client status output.
479 - Add a recursive mark command (rmark) to restore.
480 - "Minimum Job Interval = nnn" sets minimum interval between Jobs
481 of the same level and does not permit multiple simultaneous
482 running of that Job (i.e. lets any previous invocation finish
483 before doing Interval testing).
484 - Look at simplifying File exclusions.
485 - New directive "Delete purged Volumes"
486 - new pool XXX with ScratchPoolId = MyScratchPool's PoolId and
487 let it fill itself, and RecyclePoolId = XXX's PoolId so I can
488 see if it become stable and I just have to supervise
490 - If I want to remove this pool, I set RecyclePoolId = MyScratchPool's
491 PoolId, and when it is empty remove it.
492 - Figure out how to recycle Scratch volumes back to the Scratch Pool.
494 - Allow Check Labels to be used with Bacula labels.
495 - "Resuming" a failed backup (lost line for example) by using the
496 failed backup as a sort of "base" job.
498 - Email to the user when the tape is about to need changing x
499 days before it needs changing.
500 - Command to show next tape that will be used for a job even
501 if the job is not scheduled.
502 - From: Arunav Mandal <amandal@trolltech.com>
503 1. When jobs are running and bacula for some reason crashes or if I do a
504 restart it remembers and jobs it was running before it crashed or restarted
505 as of now I loose all jobs if I restart it.
507 2. When spooling and in the midway if client is disconnected for instance a
508 laptop bacula completely discard the spool. It will be nice if it can write
509 that spool to tape so there will be some backups for that client if not all.
511 3. We have around 150 clients machines it will be nice to have a option to
512 upgrade all the client machines bacula version automatically.
514 4. Atleast one connection should be reserved for the bconsole so at heavy load
515 I should connect to the director via bconsole which at sometimes I can't
517 5. Another most important feature that is missing, say at 10am I manually
518 started backup of client abc and it was a full backup since client abc has
519 no backup history and at 10.30am bacula again automatically started backup of
520 client abc as that was in the schedule. So now we have 2 multiple Full
521 backups of the same client and if we again try to start a full backup of
522 client backup abc bacula won't complain. That should be fixed.
524 - Fix bpipe.c so that it does not modify results pointer.
525 ***FIXME*** calling sequence should be changed.
526 - For Windows disaster recovery see http://unattended.sf.net/
527 - regardless of the retention period, Bacula will not prune the
528 last Full, Diff, or Inc File data until a month after the
529 retention period for the last Full backup that was done.
530 - update volume=xxx --- add status=Full
531 - Remove old spool files on startup.
532 - Exclude SD spool/working directory.
533 - Refuse to prune last valid Full backup. Same goes for Catalog.
535 - Make a callback when Rerun failed levels is called.
536 - Give Python program access to Scheduled jobs.
537 - Add setting Volume State via Python.
538 - Python script to save with Python, not save, save with Bacula.
539 - Python script to do backup.
541 - Change the Priority, Client, Storage, JobStatus (error)
542 at the start of a job.
543 - Why is SpoolDirectory = /home/bacula/spool; not reported
544 as an error when writing a DVD?
545 - Make bootstrap file handle multiple MediaTypes (SD)
546 - Remove all old Device resource code in Dir and code to pass it
547 back in SD -- better, rework it to pass back device statistics.
548 - Check locking of resources -- be sure to lock devices where previously
549 resources were locked.
550 - The last part is left in the spool dir.
553 - In restore don't compare byte count on a raw device -- directory
554 entry does not contain bytes.
558 MA = (last_MA * 3 + rate) / 4
559 rate = (bytes - last_bytes) / (runtime - last_runtime)
560 - Max Vols limit in Pool off by one?
561 - Implement Files/Bytes,... stats for restore job.
562 - Implement Total Bytes Written, ... for restore job.
563 - Despool attributes simultaneously with data in a separate
564 thread, rejoined at end of data spooling.
565 - 7. Implement new Console commands to allow offlining/reserving drives,
566 and possibly manipulating the autochanger (much asked for).
567 - Add start/end date editing in messages (%t %T, %e?) ...
568 - Add ClientDefs similar to JobDefs.
569 - Print more info when bextract -p accepts a bad block.
570 - Fix FD JobType to be set before RunBeforeJob in FD.
571 - Look at adding full Volume and Pool information to a Volume
572 label so that bscan can get *all* the info.
573 - If the user puts "Purge Oldest Volume = yes" or "Recycle Oldest Volume = yes"
574 and there is only one volume in the pool, refuse to do it -- otherwise
575 he fills the Volume, then immediately starts reusing it.
576 - Implement copies and stripes.
577 - Add history file to console.
578 - Each file on tape creates a JobMedia record. Peter has 4 million
579 files spread over 10000 tape files and four tapes. A restore takes
580 16 hours to build the restore list.
581 - Add and option to check if the file size changed during backup.
582 - Make sure SD deletes spool files on error exit.
583 - Delete old spool files when SD starts.
584 - When labeling tapes, if you enter 000026, Bacula uses
585 the tape index rather than the Volume name 000026.
586 - Add offline tape command to Bacula console.
588 Enter MediaId or Volume name: 32
589 Enter new Volume name: DLT-20Dec04
590 Automatically selected Pool: Default
591 Connecting to Storage daemon DLTDrive at 192.168.68.104:9103 ...
592 Sending relabel command from "DLT-28Jun03" to "DLT-20Dec04" ...
593 block.c:552 Write error at 0:0 on device /dev/nst0. ERR=Bad file descriptor.
594 Error writing final EOF to tape. This tape may not be readable.
595 dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
596 askdir.c:219 NULL Volume name. This shouldn't happen!!!
597 3912 Failed to label Volume: ERR=dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
598 Label command failed for Volume DLT-20Dec04.
599 Do not forget to mount the drive!!!
600 - Bug: if a job is manually scheduled to run later, it does not appear
601 in any status report and cannot be cancelled.
603 ==== Keeping track of deleted/new files ====
604 - To mark files as deleted, run essentially a Verify to disk, and
605 when a file is found missing (MarkId != JobId), then create
606 a new File record with FileIndex == -1. This could be done
607 by the FD at the same time as the backup.
609 My "trick" for keeping track of deletions is the following.
610 Assuming the user turns on this option, after all the files
611 have been backed up, but before the job has terminated, the
612 FD will make a pass through all the files and send their
613 names to the DIR (*exactly* the same as what a Verify job
614 currently does). This will probably be done at the same
615 time the files are being sent to the SD avoiding a second
616 pass. The DIR will then compare that to what is stored in
617 the catalog. Any files in the catalog but not in what the
618 FD sent will receive a catalog File entry that indicates
619 that at that point in time the file was deleted. This
620 either transmitted to the FD or simultaneously computed in
621 the FD, so that the FD can put a record on the tape that
622 indicates that the file has been deleted at this point.
623 A delete file entry could potentially be one with a FileIndex
624 of 0 or perhaps -1 (need to check if FileIndex is used for
625 some other thing as many of the Bacula fields are "overloaded"
628 During a restore, any file initially picked up by some
629 backup (Full, ...) then subsequently having a File entry
630 marked "delete" will be removed from the tree, so will not
631 be restored. If a file with the same name is later OK it
632 will be inserted in the tree -- this already happens. All
633 will be consistent except for possible changes during the
636 Since I'm on the subject, some of you may be wondering what
637 the utility of the in memory tree is if you are going to
638 restore everything (at least it comes up from time to time
639 on the list). Well, it is still *very* useful because it
640 allows only the last item found for a particular filename
641 (full path) to be entered into the tree, and thus if a file
642 is backed up 10 times, only the last copy will be restored.
643 I recently (last Friday) restored a complete directory, and
644 the Full and all the Differential and Incremental backups
645 spanned 3 Volumes. The first Volume was not even mounted
646 because all the files had been updated and hence backed up
647 since the Full backup was made. In this case, the tree
648 saved me a *lot* of time.
650 Make sure this information is stored on the tape too so
651 that it can be restored directly from the tape.
653 All the code (with the exception of formally generating and
654 saving the delete file entries) already exists in the Verify
655 Catalog command. It explicitly recognizes added/deleted files since
656 the last InitCatalog. It is more or less a "simple" matter of
657 taking that code and adapting it slightly to work for backups.
659 Comments from Martin Simmons (I think they are all covered):
660 Ok, that should cover the basics. There are few issues though:
662 - Restore will depend on the catalog. I think it is better to include the
663 extra data in the backup as well, so it can be seen by bscan and bextract.
665 - I'm not sure if it will preserve multiple hard links to the same inode. Or
666 maybe adding or removing links will cause the data to be dumped again?
668 - I'm not sure if it will handle renamed directories. Possibly it will work
669 by dumping the whole tree under a renamed directory?
671 - It remains to be seen how the backup performance of the DIR's will be
672 affected when comparing the catalog for a large filesystem.
676 How about introducing a Type = MgmtPolicy job type? That job type would
677 be responsible for scanning the Bacula environment looking for specific
678 conditions, and submitting the appropriate jobs for implementing said
682 Name = "Migration-Policy"
684 Policy Selection Job Type = Migrate
685 Scope = "<keyword> <operator> <regexp>"
686 Threshold = "<keyword> <operator> <regexp>"
687 Job Template = <template-name>
690 Where <keyword> is any legal job keyword, <operator> is a comparison
691 operator (=,<,>,!=, logical operators AND/OR/NOT) and <regexp> is a
692 appropriate regexp. I could see an argument for Scope and Threshold
693 being SQL queries if we want to support full flexibility. The
694 Migration-Policy job would then get scheduled as frequently as a site
695 felt necessary (suggested default: every 15 minutes).
700 Name = "Migration-Policy"
702 Policy Selection Job Type = Migration
704 Threshold = "Migration Selection Type = LowestUtil"
705 Job Template = "MigrationTemplate"
708 would select all pools for examination and generate a job based on
709 MigrationTemplate to automatically select the volume with the lowest
710 usage and migrate it's contents to the nextpool defined for that pool.
712 This policy abstraction would be really handy for adjusting the behavior
713 of Bacula according to site-selectable criteria (one thing that pops
714 into mind is Amanda's ability to automatically adjust backup levels
715 depending on various criteria).
721 - Add Pool/Storage override regression test.
722 - Add delete JobId to regression.
723 - Add a regression test for dbcheck.
724 - New test to add bscan to four-concurrent-jobs regression,
725 i.e. after the four-concurrent jobs zap the
726 database as is done in the bscan-test, then use bscan to
727 restore the database, do a restore and compare with the
729 - Add restore of specific JobId to regression (item 3
730 on the restore prompt)
731 - Add IPv6 to regression
732 - Add database test to regression. Test each function like delete,
735 - AntiVir can slow down backups on Win32 systems.
736 - Win32 systems with FAT32 can be much slower than NTFS for
737 more than 1000 files per directory.
741 - A HOLD command to stop all jobs from starting.
742 - A PAUSE command to pause all running jobs ==> release the
744 - Media Type = LTO,LTO-2,LTO-3
745 Media Type Read = LTO,LTO2,LTO3
746 Media Type Write = LTO2, LTO3
748 === From Carsten Menke <bootsy52@gmx.net>
750 Following is a list of what I think in the situations where I'm faced with,
751 could be a usefull enhancement to bacula, which I'm certain other users will
752 benefit from as well.
754 1. NextJob/NextJobs Directive within a Job Resource in the form of
755 NextJobs = job1,job2.
758 I currently solved the problem with running multiple jobs each after each
759 by setting the Max Wait Time for a job to 8 hours, and give
760 the jobs different Priorities. However, there scenarios where
761 1 Job is directly depending on another job, so if the former job fails,
762 the job after it needn't to be run
763 while maybe other jobs should run despite of that
766 A Backup Job and a Verify job, if the backup job fails there is no need to run
767 the verify job, as the backup job already failed. However, one may like
768 to backup the Catalog to disk despite of that the main backup job failed.
771 I see that this is related to the Event Handlers which are on the ToDo
772 list, also it is maybe a good idea to check for the return value and
773 execute different actions based on the return value
776 3. offline capability to bconsole
779 Currently I use a script which I execute within the last Job via the
780 RunAfterJob Directive, to release and eject the tape.
781 So I have to call bconsole "release=Storage-Name" and afterwards
782 mt -f /dev/nst0 eject to get the tape out.
784 If I have multiple Storage Devices, than these may not be /dev/nst0 and
785 I have to modify the script or call it with parameters etc.
786 This would actually not be needed, as everything is already defined
787 in bacula-sd.conf and if I can invoke bconsole with the
788 storage name via $1 in the script than I'm done and information is
791 4. %s for Storage Name added to the chars being substituted in "RunAfterJob"
795 For the reason mentioned in 3. to have the ability to call a
796 script with /scripts/foobar %s and in the script use $1
797 to pass the Storage Name to bconsole
799 5. Setting Volume State within a Job Resource
802 Instead of using "Maximum Volume Jobs" in the Pool Resource,
803 I would have the possibilty to define
804 in a Job Resource that after this certain job is run, the Volume State
805 should be set to "Volume State = Used", this give more flexibility (IMHO).
807 6. Localization of Bacula Messages
810 Unfortunatley many,many people I work with don't speak english very well.
811 So if at least the Reporting messages would be localized then they
812 would understand that they have to change the tape,etc. etc.
814 I volunteer to do the german translations, and if I can convince my wife also
815 french and Morre (western african language).
817 7. OK, this is evil, probably bound to security risks and maybe not possible
818 due to the design of bacula.
820 Implementation of Backtics ( `command` ) for shell comand execution to
821 the "Label Format" Directive.
825 Currently I have defined BACULA_DAY_OF_WEEK="day1|day2..." resulting in
826 Label Format = "HolyBackup-${BACULA_DAY_OF_WEEK[${WeekDay}]}". If I could
827 use backticks than I could use "Label Format = HolyBackup-`date +%A` to have
828 the localized name for the day of the week appended to the
829 format string. Then I have the tape labeled automatically with weekday
830 name in the correct language.
832 - Yes, that is surely the case. I probably should turn those into Warning
833 errors. In addition, you just made me think that it might not be bad to
834 add an option to check the file size after backing up the file and
835 report if it changes. This would be done as an option because it would
838 Kern, good idea. If you do do that, mention in the output: file
839 shrunk, or file expanded, just to make it obvious to the user
840 (without having to the refer to file size), just how the file size
843 Would this option be for all file, or just one file? Or a fileset?
844 - Make output from status use html table tags for nicely
845 presenting in a browser.
846 - Can one write tapes faster with 8192 byte block sizes?
847 - Document security problems with the same password for everyone in
848 rpm and Win32 releases.
849 - Browse generations of files.
850 - I've seen an error when my catalog's File table fills up. I
851 then have to recreate the File table with a larger maximum row
852 size. Relevant information is at
853 http://dev.mysql.com/doc/mysql/en/Full_table.html ; I think the
854 "Installing and Configuring MySQL" chapter should talk a bit
855 about this potential problem, and recommend a solution.
856 - For Solaris must use POSIX awk.
857 - Want speed of writing to tape while despooling.
858 - Supported autochanger:
866 Wangtek 6525ES (SCSI-1 QIC drive, 525MB), under Linux 2.4.something,
867 bacula 1.36.0/1 works with blocksize 16k INSIDE bacula-sd.conf.
868 - Add regex from http://www.pcre.org to Bacula for Win32.
869 - Use only shell tools no make in CDROM package.
870 - Include within include does it work?
871 - Implement a Pool of type Cleaning?
872 - Implement VolReadTime and VolWriteTime in SD
873 - Modify Backing up Your Database to include a bootstrap file.
874 - Think about making certain database errors fatal.
875 - Look at correcting the time jump in the scheduler for daylight
876 savings time changes.
877 - Add a "real" timer to network connections.
878 - Promote to Full = Time period
879 - Check dates entered by user for correctness (month/day/... ranges)
880 - Compress restore Volume listing by date and first file.
881 - Look at patches/bacula_db.b2z postgresql that loops during restore.
883 - Perhaps add read/write programs and/or plugins to FileSets.
884 - How to handle backing up portables ...
885 - Add some sort of guaranteed Interval for upgrading jobs.
886 - Can we write the state file after every job terminates? On Win32
887 the system crashes and the state file is not updated.
890 Documentation to do: (any release a little bit at a time)
891 - Doc to do unmount before removing magazine.
892 - Alternative to static linking "ldd prog" save all binaries listed,
893 restore them and point LD_LIBRARY_PATH to them.
894 - Document add "</dev/null >/dev/null 2>&1" to the bacula-fd command line
895 - Document query file format.
896 - Add more documentation for bsr files.
897 - Document problems with Verify and pruning.
898 - Document how to use multiple databases.
899 - VXA drives have a "cleaning required"
900 indicator, but Exabyte recommends preventive cleaning after every 75
903 In this context, it should be noted that Exabyte has a command-line
904 vxatool utility available for free download. (The current version is
905 vxatool-3.72.) It can get diagnostic info, read, write and erase tapes,
906 test the drive, unload tapes, change drive settings, flash new firmware,
908 Of particular interest in this context is that vxatool <device> -i will
909 report, among other details, the time since last cleaning in tape motion
910 minutes. This information can be retrieved (and settings changed, for
911 that matter) through the generic-SCSI device even when Bacula has the
912 regular tape device locked. (Needless to say, I don't recommend
913 changing tape settings while a job is running.)
914 - Lookup HP cleaning recommendations.
915 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
916 - Document doing table repair
919 ===================================
920 - Add macro expansions in JobDefs.
921 Run Before Job = "SomeFile %{Level} %{Client}"
922 Write Bootstrap="/some/dir/%{JobName}_%{Client}.bsr"
923 - Use non-blocking network I/O but if no data is available, use
925 - Use gather write() for network I/O.
926 - Autorestart on crash.
927 - Add bandwidth limiting.
928 - Add acks every once and a while from the SD to keep
929 the line from timing out.
930 - When an error in input occurs and conio beeps, you can back
931 up through the prompt.
932 - Detect fixed tape block mode during positioning by looking at
933 block numbers in btape "test". Possibly adjust in Bacula.
934 - Fix list volumes to output volume retention in some other
935 units, perhaps via a directive.
936 - Allow Simultaneous Priorities = yes => run up to Max concurrent jobs even
937 with multiple priorities.
938 - If you use restore replace=never, the directory attributes for
939 non-existent directories will not be restored properly.
941 - see lzma401.zip in others directory for new compression
943 - Allow the user to select JobType for manual pruning/purging.
944 - bscan does not put first of two volumes back with all info in
946 - Implement the FreeBSD nodump flag in chflags.
947 - Figure out how to make named console messages go only to that
948 console and to the non-restricted console (new console class?).
949 - Make restricted console prompt for password if *ask* is set or
950 perhaps if password is undefined.
951 - Implement "from ISO-date/time every x hours/days/weeks/months" in
954 ==== from Marc Schoechlin
955 - the help-command should be more verbose
956 (it should explain the paramters of the different
958 -> its time-comsuming to consult the manual anytime
959 you need a special parameter
960 -> maybe its more easy to maintain this, if the
961 descriptions of that commands are outsourced to
963 - the cd-command should allow complete paths
964 i.e. cd /foo/bar/foo/bar
965 -> if a customer mails me the path to a certain file,
966 its faster to enter the specified directory
967 - if the password is not configured in bconsole.conf
968 you should be asked for it.
969 -> sometimes you like to do restore on a customer-machine
970 which shouldnt know the password for bacula.
971 -> adding the password to the file favours admins
972 to forget to remove the password after usage
974 the protection of that file is less important
975 - long-listed-output of commands should be scrollable
976 like the unix more/less-command does
977 -> if someone runs 200 and more machines, the lists could
978 be a little long and complex
979 - command-output should be shown column by column
980 to reduce scrolling and to increase clarity
982 - lsmark should list the selected files with full
984 - wildcards for selecting and file and directories would be nice
985 - any actions should be interuptable with STRG+C
986 - command-expansion would be pretty cool
988 - When the replace Never option is set, new directory permissions
989 are not restored. See bug 213. To fix this requires creating a
990 list of newly restored directories so that those directory
991 permissions *can* be restored.
992 - Add prune all command
993 - Document fact that purge can destroy a part of a restore by purging
994 one volume while others remain valid -- perhaps mark Jobs.
995 - Add multiple-media-types.txt
996 - look at mxt-changer.html
997 - Make ? do a help command (no return needed).
998 - Implement restore directory.
999 - Document streams and how to implement them.
1000 - Try not to re-backup a file if a new hard link is added.
1001 - Add feature to backup hard links only, but not the data.
1002 - Fix stream handling to be simpler.
1003 - Add Priority and Bootstrap to Run a Job.
1004 - Eliminate Restore "Run Restore Job" prompt by allowing new "run command
1006 - Remove View FileSet button from Run a Job dialog.
1007 - Handle prompt for restore job at end of Restore command.
1008 - Add display of total selected files to Restore window.
1009 - Add tree pane to left of window.
1010 - Add progress meter.
1011 - Max wait time or max run time causes seg fault -- see runtime-bug.txt
1012 - Add message to user to check for fixed block size when the forward
1013 space test fails in btape.
1014 - When unmarking a directory check if all files below are unmarked and
1015 then remove the + flag -- in the restore tree.
1016 - Possibly implement: Action = Unmount Device="TapeDrive1" in Admin jobs.
1017 - Setup lrrd graphs: (http://www.linpro.no/projects/lrrd/) Mike Acar.
1018 - Revisit the question of multiple Volumes (disk) on a single device.
1019 - Add a block copy option to bcopy.
1020 - Finish work on Gnome restore GUI.
1021 - Fix "llist jobid=xx" where no fileset or client exists.
1022 - For each job type (Admin, Restore, ...) require only the really necessary
1023 fields.- Pass Director resource name as an option to the Console.
1024 - Add a "batch" mode to the Console (no unsolicited queries, ...).
1025 - Add a .list all files in the restore tree (probably also a list all files)
1026 Do both a long and short form.
1027 - Allow browsing the catalog to see all versions of a file (with
1028 stat data on each file).
1029 - Restore attributes of directory if replace=never set but directory
1031 - Use SHA1 on authentication if possible.
1032 - See comtest-xxx.zip for Windows code to talk to USB.
1033 - Add John's appended files:
1034 Appended = { /files/server/logs/http/*log }
1035 and such files would be treated as follows.On a FULL backup, they would
1036 be backed up like any other file.On an INCREMENTAL backup, where a
1037 previous INCREMENTAL or FULL was already in thecatalogue and the length
1038 of the file wasgreater than the length of the last backup, only thedata
1039 added since the last backup will be dumped.On an INCREMENTAL backup, if
1040 the length of the file is less than thelength of the file with the same
1041 name last backed up, the completefile is dumped.On Windows systems, with
1042 creation date of files, we can be evensmarter about this and not count
1043 entirely upon the length.On a restore, the full and all incrementals
1044 since it will beapplied in sequence to restore the file.
1045 - Check new HAVE_WIN32 open bits.
1046 - Check if the tape has moved before writing.
1047 - Handling removable disks -- see below:
1048 - Keep track of tape use time, and report when cleaning is necessary.
1049 - Add FromClient and ToClient keywords on restore command (or
1050 BackupClient RestoreClient).
1051 - Implement a JobSet, which groups any number of jobs. If the
1052 JobSet is started, all the jobs are started together.
1053 Allow Pool, Level, and Schedule overrides.
1054 - Enhance cancel to timeout BSOCK packets after a specific delay.
1055 - Do scheduling by UTC using gmtime_r() in run_conf, scheduler, and
1056 ua_status.!!! Thanks to Alan Brown for this tip.
1057 - Look at updating Volume Jobs so that Max Volume Jobs = 1 will work
1058 correctly for multiple simultaneous jobs.
1059 - Correct code so that FileSet MD5 is calculated for < and | filename
1061 - Implement the Media record flag that indicates that the Volume does disk
1063 - Implement VolAddr, which is used when Volume is addressed like a disk,
1064 and form it from VolFile and VolBlock.
1065 - Make multiple restore jobs for multiple media types specifying
1066 the proper storage type.
1067 - Fix fast block rejection (stored/read_record.c:118). It passes a null
1068 pointer (rec) to try_repositioning().
1069 - Look at extracting Win data from BackupRead.
1070 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
1071 which would take precidence over the Catalog "JobRetention".
1072 - Implement Label Format in Add and Label console commands.
1073 - Possibly up network buffers to 65K. Put on variable.
1074 - Put email tape request delays on one or more variables. User wants
1075 to cancel the job after a certain time interval. Maximum Mount Wait?
1076 - Job, Client, Device, Pool, or Volume?
1077 Is it possible to make this a directive which is *optional* in multiple
1078 resources, like Level? If so, I think I'd make it an optional directive
1079 in Job, Client, and Pool, with precedence such that Job overrides Client
1080 which in turn overrides Pool.
1082 - New Storage specifications:
1083 - Want to write to multiple storage devices simultaneously
1084 - Want to write to multiple storage devices sequentially (in one job)
1085 - Want to read/write simultaneously
1086 - Key is MediaType -- it must match
1088 Passed to SD as a sort of BSR record called Storage Specification
1092 MediaType -> Next MediaType
1094 Device -> Next Device
1096 Allow multiple Storage specifications
1104 Allow Multiple Pool specifications (note, Pool currently
1106 Allow Multiple MediaType specifications in Dir conf
1107 Allow Multiple Device specifications in Dir conf
1108 Perhaps keep this in a single SSR
1109 Tie a Volume to a specific device by using a MediaType that
1110 is contained in only one device.
1111 In SD allow Device to have Multiple MediaTypes
1113 - Ideas from Jerry Scharf:
1114 First let's point out some big pluses that bacula has for this
1116 more importantly it's active. Thank you so much for that
1117 even more important, it's not flaky
1118 it has an open access catalog, opening many possibilities
1119 it's pushing toward heterogeneous systems capability
1121 Macintosh file client
1122 macs are an interesting niche, but I fear a server is a rathole
1123 working bare iron recovery for windows
1124 the option for inc/diff backups not reset on fileset revision
1125 a) use both change and inode update time against base time
1126 b) do the full catalog check (expensive but accurate)
1127 sizing guide (how much system is needed to back up N systems/files)
1128 consultants on using bacula in building a disaster recovery system
1129 an integration guide
1130 or how to get at fancy things that one could do with bacula
1131 logwatch code for bacula logs (or similar)
1132 linux distro inclusion of bacula (brings good and bad, but necessary)
1133 win2k/XP server capability (icky but you asked)
1134 support for Oracle database ??
1136 - Look at adding SQL server and Exchange support for Windows.
1137 - Make dev->file and dev->block_num signed integers so that -1 can
1138 be an invalid value which happens with BSR.
1139 - Create VolAddr for disk files in place of VolFile and VolBlock. This
1140 is needed to properly specify ranges.
1141 - Add progress of files/bytes to SD and FD.
1142 - Print warning message if FileId > 4 billion
1143 - do a "messages" before the first prompt in Console
1144 - Client does not show busy during Estimate command.
1145 - Implement Console mtx commands.
1146 - Implement a Mount Command and an Unmount Command where
1147 the users could specify a system command to be performed
1148 to do the mount, after which Bacula could attempt to
1149 read the device. This is for Removeable media such as a CDROM.
1150 - Most likely, this mount command would be invoked explicitly
1151 by the user using the current Console "mount" and "unmount"
1152 commands -- the Storage Daemon would do the right thing
1153 depending on the exact nature of the device.
1154 - As with tape drives, when Bacula wanted a new removable
1155 disk mounted, it would unmount the old one, and send a message
1156 to the user, who would then use "mount" as described above
1157 once he had actually inserted the disk.
1158 - Implement dump/print label to UA
1159 - Spool to disk only when the tape is full, then when a tape is hung move
1161 - bextract is sending everything to the log file ****FIXME****
1162 - Allow multiple Storage specifications (or multiple names on
1163 a single Storage specification) in the Job record. Thus a job
1164 can be backed up to a number of storage devices.
1165 - Implement some way for the File daemon to contact the Director
1166 to start a job or pass its DHCP obtained IP number.
1167 - Implement a query tape prompt/replace feature for a console
1168 - Copy console @ code to gnome2-console
1169 - Make tree walk routines like cd, ls, ... more user friendly
1170 by handling spaces better.
1171 - Make sure that Bacula rechecks the tape after the 20 min wait.
1172 - Set IO_NOWAIT on Bacula TCP/IP packets.
1173 - Try doing a raw partition backup and restore by mounting a
1175 - From Lars Kellers:
1176 Yes, it would allow to highly automatic the request for new tapes. If a
1177 tape is empty, bacula reads the barcodes (native or simulated), and if
1178 an unused tape is found, it runs the label command with all the
1179 necessary parameters.
1181 By the way can bacula automatically "move" an empty/purged volume say
1182 in the "short" pool to the "long" pool if this pool runs out of volume
1184 - What to do about "list files job=xxx".
1185 - Look at how fuser works and /proc/PID/fd that is how Nic found the
1186 file descriptor leak in Bacula.
1187 - Implement WrapCounters in Counters.
1188 - Add heartbeat from FD to SD if hb interval expires.
1189 - Can we dynamically change FileSets?
1190 - If pool specified to label command and Label Format is specified,
1191 automatically generate the Volume name.
1192 - Why can't SQL do the filename sort for restore?
1193 - Add ExhautiveRestoreSearch
1194 - Look at the possibility of loading only the necessary
1195 data into the restore tree (i.e. do it one directory at a
1196 time as the user walks through the tree).
1197 - Possibly use the hash code if the user selects all for a restore command.
1198 - Fix "restore all" to bypass building the tree.
1199 - Prohibit backing up archive device (findlib/find_one.c:128)
1200 - Implement Release Device in the Job resource to unmount a drive.
1201 - Implement Acquire Device in the Job resource to mount a drive,
1202 be sure this works with admin jobs so that the user can get
1203 prompted to insert the correct tape. Possibly some way to say to
1204 run the job but don't save the files.
1205 - Make things like list where a file is saved case independent for
1207 - Use autochanger to handle multiple devices.
1208 - Implement a Recycle command
1209 - Start working on Base jobs.
1210 - Implement UnsavedFiles DB record.
1211 - From Phil Stracchino:
1212 It would probably be a per-client option, and would be called
1213 something like, say, "Automatically purge obsoleted jobs". What it
1214 would do is, when you successfully complete a Differential backup of a
1215 client, it would automatically purge all Incremental backups for that
1216 client that are rendered redundant by that Differential. Likewise,
1217 when a Full backup on a client completed, it would automatically purge
1218 all Differential and Incremental jobs obsoleted by that Full backup.
1219 This would let people minimize the number of tapes they're keeping on
1220 hand without having to master the art of retention times.
1221 - When doing a Backup send all attributes back to the Director, who
1222 would then figure out what files have been deleted.
1223 - Currently in mount.c:236 the SD simply creates a Volume. It should have
1224 explicit permission to do so. It should also mark the tape in error
1225 if there is an error.
1226 - Cancel waiting for Client connect in SD if FD goes away.
1228 - Implement timeout in response() when it should come quickly.
1229 - Implement a Slot priority (loaded/not loaded).
1230 - Implement "vacation" Incremental only saves.
1231 - Implement create "FileSet"?
1232 - Add prefixlinks to where or not where absolute links to FD.
1233 - Issue message to mount a new tape before the rewind.
1234 - Simplified client job initiation for portables.
1235 - If SD cannot open a drive, make it periodically retry.
1236 - Add more of the config info to the tape label.
1238 - Refine SD waiting output:
1239 Device is being positioned
1240 > Device is being positioned for append
1241 > Device is being positioned to file x
1243 - Figure out some way to estimate output size and to avoid splitting
1244 a backup across two Volumes -- this could be useful for writing CDROMs
1245 where you really prefer not to have it split -- not serious.
1246 - Have SD compute MD5 or SHA1 and compare to what FD computes.
1247 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
1248 actual data on the Volume and compare it.
1249 - Implement Bacula plugins -- design API
1250 - Make bcopy read through bad tape records.
1251 - Program files (i.e. execute a program to read/write files).
1252 Pass read date of last backup, size of file last time.
1253 - Add Signature type to File DB record.
1254 - CD into subdirectory when open()ing files for backup to
1255 speed up things. Test with testfind().
1256 - Priority job to go to top of list.
1257 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
1258 - Implement some way for the Console to dynamically create a job.
1259 - Solaris -I on tar for include list
1260 - Need a verbose mode in restore, perhaps to bsr.
1261 - bscan without -v is too quiet -- perhaps show jobs.
1262 - Add code to reject whole blocks if not wanted on restore.
1263 - Check if we can increase Bacula FD priorty in Win2000
1264 - Make sure the MaxVolFiles is fully implemented in SD
1265 - Check if both CatalogFiles and UseCatalog are set to SD.
1266 - Possibly add email to Watchdog if drive is unmounted too
1267 long and a job is waiting on the drive.
1268 - After unmount, if restore job started, ask to mount.
1269 - Add UA rc and history files.
1270 - put termcap (used by console) in ./configure and
1271 allow -with-termcap-dir.
1272 - Fix Autoprune for Volumes to respect need for full save.
1273 - Compare tape to Client files (attributes, or attributes and data)
1274 - Make all database Ids 64 bit.
1275 - Allow console commands to detach or run in background.
1276 - Add SD message variables to control operator wait time
1277 - Maximum Operator Wait
1278 - Minimum Message Interval
1279 - Maximum Message Interval
1280 - Send Operator message when cannot read tape label.
1281 - Verify level=Volume (scan only), level=Data (compare of data to file).
1282 Verify level=Catalog, level=InitCatalog
1284 - Add keyword search to show command in Console.
1285 - Events : tape has more than xxx bytes.
1286 - Complete code in Bacula Resources -- this will permit
1287 reading a new config file at any time.
1288 - Handle ctl-c in Console
1289 - Implement script driven addition of File daemon to config files.
1290 - Think about how to make Bacula work better with File (non-tape) archives.
1291 - Write Unix emulator for Windows.
1292 - Put memory utilization in Status output of each daemon
1293 if full status requested or if some level of debug on.
1294 - Make database type selectable by .conf files i.e. at runtime
1295 - Set flag for uname -a. Add to Volume label.
1296 - Restore files modified after date
1297 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
1298 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
1299 - Timout a job or terminate if link goes down, or reopen link and query.
1300 - Concept of precious tapes (cannot be reused).
1301 - Make bcopy copy with a single tape drive.
1302 - Permit changing ownership during restore.
1305 > My suggestion: Add a feature on the systray menu-icon menu to request
1306 > an immediate backup now. This would be useful for laptop users who may
1307 > not be on the network when the regular scheduled backup is run.
1309 > My wife's suggestion: Add a setting to the win32 client to allow it to
1310 > shut down the machine after backup is complete (after, of course,
1311 > displaying a "System will shut down in one minute, click here to cancel"
1312 > warning dialog). This would be useful for sites that want user
1313 > woorkstations to be shut down overnight to save power.
1316 - Autolabel should be specified by DIR instead of SD.
1318 - Add media capacity
1319 - AutoScan (check checksum of tape)
1320 - Format command = "format /dev/nst0"
1324 - Seek resolution (usually corresponds to buffer size)
1325 - EODErrorCode=ENOSPC or code
1326 - Partial Read error code
1327 - Partial write error code
1328 - Nonformatted read error
1329 - Nonformatted write error
1330 - WriteProtected error
1334 - IgnoreCloseErrors=yes
1344 - FD sends unsaved file list to Director at end of job (see
1346 - File daemon should build list of files skipped, and then
1347 at end of save retry and report any errors.
1348 - Write a Storage daemon that uses pipes and
1349 standard Unix programs to write to the tape.
1351 - Need something that monitors the JCR queue and
1352 times out jobs by asking the deamons where they are.
1353 - Enhance Jmsg code to permit buffering and saving to disk.
1354 - device driver = "xxxx" for drives.
1355 - Verify from Volume
1356 - Ensure that /dev/null works
1357 - Need report class for messages. Perhaps
1358 report resource where report=group of messages
1359 - enhance scan_attrib and rename scan_jobtype, and
1360 fill in code for "since" option
1361 - Director needs a time after which the report status is sent
1362 anyway -- or better yet, a retry time for the job.
1363 - Don't reschedule a job if previous incarnation is still running.
1364 - Some way to automatically backup everything is needed????
1365 - Need a structure for pending actions:
1367 - termination status (part of buffered msgs?)
1369 Read, Write, Clean, Delete
1370 - Login to Bacula; Bacula users with different permissions:
1371 owner, group, user, quotas
1372 - Store info on each file system type (probably in the job header on tape.
1373 This could be the output of df; or perhaps some sort of /etc/mtab record.
1375 ========= ideas ===============
1376 From: "Jerry K. Schieffer" <jerry@skylinetechnology.com>
1377 To: <kern@sibbald.com>
1378 Subject: RE: [Bacula-users] future large programming jobs
1379 Date: Thu, 26 Feb 2004 11:34:54 -0600
1381 I noticed the subject thread and thought I would offer the following
1382 merely as sources of ideas, i.e. something to think about, not even as
1383 strong as a request. In my former life (before retiring) I often
1384 dealt with backups and storage management issues/products as a
1385 developer and as a consultant. I am currently migrating my personal
1386 network from amanda to bacula specifically because of the ability to
1387 cross media boundaries during storing backups.
1388 Are you familiar with the commercial product called ADSM (I think IBM
1389 now sells it under the Tivoli label)? It has a couple of interesting
1390 ideas that may apply to the following topics.
1392 1. Migration: Consider that when you need to restore a system, there
1393 may be pressure to hurry. If all the information for a single client
1394 can eventually end up on the same media (and in chronological order),
1395 the restore is facillitated by not having to search past information
1396 from other clients. ADSM has the concept of "client affinity" that
1397 may be associated with it's storage pools. It seems to me that this
1398 concept (as an optional feature) might fit in your architecture for
1401 ADSM also has the concept of defining one or more storage pools as
1402 "copy pools" (almost mirrors, but only in the sense of contents).
1403 These pools provide the ability to have duplicte data stored both
1404 onsite and offsite. The copy process can be scheduled to be handled
1405 by their storage manager during periods when there is no backup
1406 activity. Again, the migration process might be a place to consider
1407 implementing something like this.
1410 > It strikes me that it would be very nice to be able to do things
1412 > have the Job(s) backing up the machines run, and once they have all
1413 > completed, start a migration job to copy the data from disks Volumes
1415 > a tape library and then to offsite storage. Maybe this can already
1417 > done with some careful scheduling and Job prioritzation; the events
1418 > mechanism described below would probably make it very easy.
1420 This is the goal. In the first step (before events), you simply
1422 the Migration to tape later.
1424 2. Base jobs: In ADSM, each copy of each stored file is tracked in
1425 the database. Once a file (unique by path and metadata such as dates,
1426 size, ownership, etc.) is in a copy pool, no more copies are made. In
1427 other words, when you start ADSM, it begins like your concept of a
1428 base job. After that it is in the "incremental" mode. You can
1429 configure the number of "generations" of files to be retained, plus a
1430 retention date after which even old generations are purged. The
1431 database tracks the contents of media and projects the percentage of
1432 each volume that is valid. When the valid content of a volume drops
1433 below a configured percentage, the valid data are migrated to another
1434 volume and the old volume is marked as empty. Note, this requires
1435 ADSM to have an idea of the contents of a client, i.e. marking the
1436 database when an existing file was deleted, but this would solve your
1437 issue of restoring a client without restoring deleted files.
1439 This is pretty far from what bacula now does, but if you are going to
1440 rip things up for Base jobs,.....
1441 Also, the benefits of this are huge for very large shops, especially
1442 with media robots, but are a pain for shops with manual media
1446 > Base jobs sound pretty useful, but I'm not dying for them.
1448 Nobody is dying for them, but when you see what it does, you will die
1451 3. Restoring deleted files: Since I think my comments in (2) above
1452 have low probability of implementation, I'll also suggest that you
1453 could approach the issue of deleted files by a mechanism of having the
1454 fd report to the dir, a list of all files on the client for every
1455 backup job. The dir could note in the database entry for each file
1456 the date that the file was seen. Then if a restore as of date X takes
1457 place, only files that exist from before X until after X would be
1458 restored. Probably the major cost here is the extra date container in
1459 each row of the files table.
1461 Thanks for "listening". I hope some of this helps. If you want to
1462 contact me, please send me an email - I read some but not all of the
1463 mailing list traffic and might miss a reply there.
1465 Please accept my compliments for bacula. It is doing a great job for
1466 me!! I sympathize with you in the need to wrestle with excelence in
1467 execution vs. excelence in feature inclusion.
1472 ==============================
1475 - Design at hierarchial storage for Bacula. Migration and Clone.
1476 - Implement FSM (File System Modules).
1477 - Audit M_ error codes to ensure they are correct and consistent.
1478 - Add variable break characters to lex analyzer.
1479 Either a bit mask or a string of chars so that
1480 the caller can change the break characters.
1481 - Make a single T_BREAK to replace T_COMMA, etc.
1482 - Ensure that File daemon and Storage daemon can
1483 continue a save if the Director goes down (this
1484 is NOT currently the case). Must detect socket error,
1485 buffer messages for later.
1486 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
1487 - Add ability to backup to two Storage devices (two SD sessions) at
1488 the same time -- e.g. onsite, offsite.
1489 - Add the ability to consolidate old backup sets (basically do a restore
1490 to tape and appropriately update the catalog). Compress Volume sets.
1491 Might need to spool via file is only one drive is available.
1492 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
1493 someway to do so with every volume that has less than x% valid
1497 Migration: Move a backup from one Volume to another
1498 Clone: Copy a backup -- two Volumes
1501 ======================================================
1503 It is somewhat like a Full save becomes an incremental since
1504 the Base job (or jobs) plus other non-base files.
1506 - A Base backup is same as Full backup, just different type.
1507 - New BaseFiles table that contains:
1509 BaseJobId - Base JobId referenced for this FileId (needed ???)
1510 JobId - JobId currently running
1511 FileId - File not backed up, exists in Base Job
1512 FileIndex - FileIndex from Base Job.
1513 i.e. for each base file that exists but is not saved because
1514 it has not changed, the File daemon sends the JobId, BaseId,
1515 FileId, FileIndex back to the Director who creates the DB entry.
1516 - To initiate a Base save, the Director sends the FD
1517 the FileId, and full filename for each file in the Base.
1518 - When the FD finds a Base file, he requests the Director to
1519 send him the full File entry (stat packet plus MD5), or
1520 conversely, the FD sends it to the Director and the Director
1521 says yes or no. This can be quite rapid if the FileId is kept
1522 by the FD for each Base Filename.
1523 - It is probably better to have the comparison done by the FD
1524 despite the fact that the File entry must be sent across the
1526 - An alternative would be to send the FD the whole File entry
1527 from the start. The disadvantage is that it requires a lot of
1528 space. The advantage is that it requires less communications
1530 - The Job record must be updated to indicate that one or more
1532 - At end of Job, FD returns:
1533 1. Count of base files/bytes not written to tape (i.e. matches)
1534 2. Count of base file that were saved i.e. had changed.
1535 - No tape record would be written for a Base file that matches, in the
1536 same way that no tape record is written for Incremental jobs where
1537 the file is not saved because it is unchanged.
1538 - On a restore, all the Base file records must explicitly be
1539 found from the BaseFile tape. I.e. for each Full save that is marked
1540 to have one or more Base Jobs, search the BaseFile for all occurrences
1542 - An optimization might be to make the BaseFile have:
1548 This would avoid the need to explicitly fetch each File record for
1549 the Base job. The Base Job record will be fetched to get the
1550 VolSessionId and VolSessionTime.
1551 =========================================================
1554 ==========================================================
1556 For each Incremental job that is run, there may be files that
1557 were found but not saved because they were locked (this applies
1558 only to Windows). Such a system could send back to the Director
1559 a list of Unsaved files.
1561 - New UnSavedFiles table that contains:
1565 - Then in the next Incremental job, the list of Unsaved Files will be
1566 feed to the FD, who will ensure that they are explicitly chosen even
1567 if standard date/time check would not have selected them.
1568 =============================================================
1572 Multiple drive autochanger data: see Alan Brown
1573 mtx -f xxx unloadStorage Element 1 is Already Full(drive 0 was empty)
1574 Unloading Data Transfer Element into Storage Element 1...source Element
1575 Address 480 is Empty
1577 (drive 0 was empty and so was slot 1)
1578 > mtx -f xxx load 15 0
1579 no response, just returns to the command prompt when complete.
1580 > mtx -f xxx status Storage Changer /dev/changer:2 Drives, 60 Slots ( 2 Import/Export )
1581 Data Transfer Element 0:Full (Storage Element 15 Loaded):VolumeTag = HX001
1582 Data Transfer Element 1:Empty
1583 Storage Element 1:Empty
1584 Storage Element 2:Full :VolumeTag=HX002
1585 Storage Element 3:Full :VolumeTag=HX003
1586 Storage Element 4:Full :VolumeTag=HX004
1587 Storage Element 5:Full :VolumeTag=HX005
1588 Storage Element 6:Full :VolumeTag=HX006
1589 Storage Element 7:Full :VolumeTag=HX007
1590 Storage Element 8:Full :VolumeTag=HX008
1591 Storage Element 9:Full :VolumeTag=HX009
1592 Storage Element 10:Full :VolumeTag=HX010
1593 Storage Element 11:Empty
1594 Storage Element 12:Empty
1595 Storage Element 13:Empty
1596 Storage Element 14:Empty
1597 Storage Element 15:Empty
1598 Storage Element 16:Empty....
1599 Storage Element 28:Empty
1600 Storage Element 29:Full :VolumeTag=CLNU01L1
1601 Storage Element 30:Empty....
1602 Storage Element 57:Empty
1603 Storage Element 58:Full :VolumeTag=NEX261L2
1604 Storage Element 59 IMPORT/EXPORT:Empty
1605 Storage Element 60 IMPORT/EXPORT:Empty
1607 Unloading Data Transfer Element into Storage Element 15...done
1609 (just to verify it remembers where it came from, however it can be
1610 overrriden with mtx unload {slotnumber} to go to any storage slot.)
1612 There needs to be a table of drive # to devices somewhere - If there are
1613 multiple changers or drives there may not be a 1:1 correspondance between
1614 changer drive number and system device name - and depending on the way the
1615 drives are hooked up to scsi busses, they may not be linearly numbered
1616 from an offset point either.something like
1618 Autochanger drives = 2
1619 Autochanger drive 0 = /dev/nst1
1620 Autochanger drive 1 = /dev/nst2
1621 IMHO, it would be _safest_ to use explicit mtx unload commands at all
1622 times, not just for multidrive changers. For a 1 drive changer, that's
1628 MTX's manpage (1.2.15):
1629 unload [<slotnum>] [ <drivenum> ]
1630 Unloads media from drive <drivenum> into slot
1631 <slotnum>. If <drivenum> is omitted, defaults to
1632 drive 0 (as do all commands). If <slotnum> is
1633 omitted, defaults to the slot that the drive was
1634 loaded from. Note that there's currently no way
1635 to say 'unload drive 1's media to the slot it
1636 came from', other than to explicitly use that
1637 slot number as the destination.AB
1643 undef# camcontrol devlist
1644 <WANGTEK 51000 SCSI M74H 12B3> at scbus0 target 2 lun 0 (pass0,sa0)
1645 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 0 (pass1,sa1)
1646 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 1 (pass2)
1648 tapeinfo -f /dev/sg0 with a bad tape in drive 1:
1649 [kern@rufus mtx-1.2.17kes]$ ./tapeinfo -f /dev/sg0
1650 Product Type: Tape Drive
1652 Product ID: 'C5713A '
1654 Attached Changer: No
1655 TapeAlert[3]: Hard Error: Uncorrectable read/write error.
1656 TapeAlert[20]: Clean Now: The tape drive neads cleaning NOW.
1663 Medium Type: Not Loaded
1666 DataCompEnabled: yes
1667 DataCompCapable: yes
1668 DataDeCompEnabled: yes
1675 Handling removable disks
1677 From: Karl Cunningham <karlc@keckec.com>
1679 My backups are only to hard disk these days, in removable bays. This is my
1680 idea of how a backup to hard disk would work more smoothly. Some of these
1681 things Bacula does already, but I mention them for completeness. If others
1682 have better ways to do this, I'd like to hear about it.
1684 1. Accommodate several disks, rotated similar to how tapes are. Identified
1685 by partition volume ID or perhaps by the name of a subdirectory.
1686 2. Abort & notify the admin if the wrong disk is in the bay.
1687 3. Write backups to different subdirectories for each machine to be backed
1689 4. Volumes (files) get created as needed in the proper subdirectory, one
1691 5. When a disk is recycled, remove or zero all old backup files. This is
1692 important as the disk being recycled may be close to full. This may be
1693 better done manually since the backup files for many machines may be
1694 scattered in many subdirectories.
1699 - Why the heck doesn't bacula drop root priviledges before connecting to
1701 - Look at using posix_fadvise(2) for backups -- see bug #751.
1702 Possibly add the code at findlib/bfile.c:795
1703 /* TCP socket options */
1704 #define TCP_KEEPIDLE 4 /* Start keeplives after this period */
1705 - Fix bnet_connect() code to set a timer and to use time to
1707 - Implement 4th argument to make_catalog_backup that passes hostname.
1708 - Test FIFO backup/restore -- make regression
1709 - Please mount volume "xxx" on Storage device ... should also list
1710 Pool and MediaType in case user needs to create a new volume.
1711 - On restore add Restore Client, Original Client.
1712 01-Apr 00:42 rufus-dir: Start Backup JobId 55, Job=kernsave.2007-04-01_00.42.48
1713 01-Apr 00:42 rufus-sd: Python SD JobStart: JobId=55 Client=Rufus
1714 01-Apr 00:42 rufus-dir: Created new Volume "Full0001" in catalog.
1715 01-Apr 00:42 rufus-dir: Using Device "File"
1716 01-Apr 00:42 rufus-sd: kernsave.2007-04-01_00.42.48 Warning: Device "File" (/tmp) not configured to autolabel Volumes.
1717 01-Apr 00:42 rufus-sd: kernsave.2007-04-01_00.42.48 Warning: Device "File" (/tmp) not configured to autolabel Volumes.
1718 01-Apr 00:42 rufus-sd: Please mount Volume "Full0001" on Storage Device "File" (/tmp) for Job kernsave.2007-04-01_00.42.48
1719 01-Apr 00:44 rufus-sd: Wrote label to prelabeled Volume "Full0001" on device "File" (/tmp)
1720 - Check if gnome-console works with TLS.
1721 - the director seg faulted when I omitted the pool directive from a
1722 job resource. I was experimenting and thought it redundant that I had
1723 specified Pool, Full Backup Pool. and Differential Backup Pool. but
1724 apparently not. This happened when I removed the pool directive and
1725 started the director.
1726 - Add Where: client:/.... to restore job report.