9 - !!! Cannot restore two jobs a the same time that were
10 written simultaneously unless they were totally spooled.
11 - Document cleaning up the spool files:
12 db, pid, state, bsr, mail, conmsg, spool
13 - Document the multiple-drive-changer.txt script.
14 - Pruning with Admin job.
15 - Does WildFile match against full name? Doc.
16 - %d and %v only valid on Director, not for ClientRunBefore/After.
17 - During tests with the 260 char fix code, I found one problem:
18 if the system "sees" a long path once, it seems to forget it's
19 working drive (e.g. c:\), which will lead to a problem during
20 the next job (create bootstrap file will fail). Here is the
21 workaround: specify absolute working and pid directory in
22 bacula-fd.conf (e.g. c:\bacula\working instead of
24 - Document techniques for restoring large numbers of files.
25 - Document setting my.cnf to big file usage.
26 - Add example of proper index output to doc. show index from File;
27 - Correct the Include syntax in the m4.xxx files in examples/conf
28 - Document JobStatus and Termination codes.
29 - Fix the error with the "DVI file can't be opened" while
30 building the French PDF.
31 - Document more DVD stuff
39 - Document all the little details of setting up certificates for
40 the Bacula data encryption code.
41 - Document more precisely how to use master keys -- especially
42 for disaster recovery.
46 - Look at using posix_fadvise(2) for backups -- see bug #751.
47 Possibly add the code at findlib/bfile.c:795
49 /* TCP socket options */
50 #define TCP_NODELAY 1 /* Turn off Nagle's algorithm. */
51 #define TCP_MAXSEG 2 /* Limit MSS */
52 #define TCP_CORK 3 /* Never send partially complete segments */
53 #define TCP_KEEPIDLE 4 /* Start keeplives after this period */
54 #define TCP_KEEPINTVL 5 /* Interval between keepalives */
55 #define TCP_KEEPCNT 6 /* Number of keepalives before death */
56 #define TCP_SYNCNT 7 /* Number of SYN retransmits */
57 #define TCP_LINGER2 8 /* Life time of orphaned FIN-WAIT-2 state */
58 #define TCP_DEFER_ACCEPT 9 /* Wake up listener only when data arrive */
59 #define TCP_WINDOW_CLAMP 10 /* Bound advertised window */
60 #define TCP_INFO 11 /* Information about this connection. */
61 #define TCP_QUICKACK 12 /* Block/reenable quick acks */
62 #define TCP_CONGESTION 13 /* Congestion control algorithm */
64 - Fix bnet_connect() code to set a timer and to use time to
67 - the director seg faulted when I omitted the pool directive from a
68 job resource. I was experimenting and thought it redundant that I had
69 specified Pool, Full Backup Pool. and Differential Backup Pool. but
70 apparently not. This happened when I removed the pool directive and
73 - On restore add Restore Client, Original Client.
74 01-Apr 00:42 rufus-dir: Start Backup JobId 55, Job=kernsave.2007-04-01_00.42.48
75 01-Apr 00:42 rufus-sd: Python SD JobStart: JobId=55 Client=Rufus
76 01-Apr 00:42 rufus-dir: Created new Volume "Full0001" in catalog.
77 01-Apr 00:42 rufus-dir: Using Device "File"
78 01-Apr 00:42 rufus-sd: kernsave.2007-04-01_00.42.48 Warning: Device "File" (/tmp) not configured to autolabel Volumes.
79 01-Apr 00:42 rufus-sd: kernsave.2007-04-01_00.42.48 Warning: Device "File" (/tmp) not configured to autolabel Volumes.
80 01-Apr 00:42 rufus-sd: Please mount Volume "Full0001" on Storage Device "File" (/tmp) for Job kernsave.2007-04-01_00.42.48
81 01-Apr 00:44 rufus-sd: Wrote label to prelabeled Volume "Full0001" on device "File" (/tmp)
83 - Add Where: client:/.... to restore job report.
84 - Ensure that each device in an Autochanger has a different
86 - Add Catalog = to Pool resource so that pools will exist
87 in only one catalog -- currently Pools are "global".
88 - Look at sg_logs -a /dev/sg0 for getting soft errors.
89 - btape "test" command with Offline on Unmount = yes
91 This test is essential to Bacula.
93 I'm going to write one record in file 0,
94 two records in file 1,
95 and three records in file 2
97 02-Feb 11:00 btape: ABORTING due to ERROR in dev.c:715
98 dev.c:714 Bad call to rewind. Device "LTO" (/dev/nst0) not open
99 02-Feb 11:00 btape: Fatal Error because: Bacula interrupted by signal 11: Segmentation violation
100 Kaboom! btape, btape got signal 11. Attempting traceback.
102 - Ensure that moving a purged Volume in ua_purge.c to the RecyclePool
103 does the right thing.
104 - Why doesn't @"xxx abc" work in a conf file?
105 - Figure out some way to "automatically" backup conf changes.
106 - Add the OS version back to the Win32 client info.
107 - Restarted jobs have a NULL in the from field.
108 - Modify SD status command to indicate when the SD is writing
109 to a DVD (the device is not open -- see bug #732).
110 - Look at the possibility of adding "SET NAMES UTF8" for MySQL,
111 and possibly changing the blobs into varchar.
112 - Check if gnome-console works with TLS.
113 - Ensure that the SD re-reads the Media record if the JobFiles
114 does not match -- it may have been updated by another job.
115 - Look at moving the Storage directive from the Job to the
116 Pool in the default conf files.
117 - Test FIFO backup/restore -- make regression
119 - Test Volume compatibility between machine architectures
120 - Encryption documentation
121 - Wrong jobbytes with query 12 (todo)
122 - bacula-1.38.2-ssl.patch
123 - Bare-metal recovery Windows (todo)
130 - Add doc for bweb -- especially Installation
132 http://www.orangecrate.com/modules.php?name=News&file=article&sid=501
134 - FD-SD quick disconnect
135 - Despool attributes in separate thread
138 - Check why restore repeatedly sends Rechdrs between
139 each data chunk -- according to James Harper 9Jan07.
140 - Building the in memory restore tree is slow.
143 - Full at least once a month, ...
144 - Cancel Inc if Diff/Full running
145 - More intelligent re-run
146 - New/deleted file backup
148 - Incremental backup -- rsync, Stow
154 - Look at mondo/mindi
155 - Don't restore Solaris Door files:
156 #define S_IFDOOR in st_mode.
157 see: http://docs.sun.com/app/docs/doc/816-5173/6mbb8ae23?a=view#indexterm-360
158 - Make Bacula by default not backup tmpfs, procfs, sysfs, ...
159 - Fix hardlinked immutable files when linking a second file, the
160 immutable flag must be removed prior to trying to link it.
161 - Implement Python event for backing up/restoring a file.
162 - Change dbcheck to tell users to use native tools for fixing
163 broken databases, and to ensure they have the proper indexes.
164 - add udev rules for Bacula devices.
165 - If a job terminates, the DIR connection can close before the
166 Volume info is updated, leaving the File count wrong.
167 - Look at why SIGPIPE during connection can cause seg fault in
168 writing the daemon message, when Dir dropped to bacula:bacula
169 - Look at zlib 32 => 64 problems.
170 - Possibly turn on St. Bernard code.
171 - Fix bextract to restore ACLs, or better yet, use common routines.
172 - Do we migrate appendable Volumes?
173 - Remove queue.c code.
174 - Print warning message if LANG environment variable does not specify
176 - New dot commands from Arno.
177 .show device=xxx lists information from one storage device, including
178 devices (I'm not even sure that information exists in the DIR...)
179 .move eject device=xxx mostly the same as 'unmount xxx' but perhaps with
180 better machine-readable output like "Ok" or "Error busy"
181 .move eject device=xxx toslot=yyy the same as above, but with a new
182 target slot. The catalog should be updated accordingly.
183 .move transfer device=xxx fromslot=yyy toslot=zzz
186 - Article: http://www.heise.de/open/news/meldung/83231
187 - Article: http://www.golem.de/0701/49756.html
188 - Article: http://lwn.net/Articles/209809/
189 - Article: http://www.onlamp.com/pub/a/onlamp/2004/01/09/bacula.html
190 - Article: http://www.linuxdevcenter.com/pub/a/linux/2005/04/07/bacula.html
191 - Article: http://www.osreviews.net/reviews/admin/bacula
192 - Article: http://www.debianhelp.co.uk/baculaweb.htm
194 - It appears to me that you have run into some sort of race
195 condition where two threads want to use the same Volume and they
196 were both given access. Normally that is no problem. However,
197 one thread wanted the particular Volume in drive 0, but it was
198 loaded into drive 1 so it decided to unload it from drive 1 and
199 then loaded it into drive 0, while the second thread went on
200 thinking that the Volume could be used in drive 1 not realizing
201 that in between time, it was loaded in drive 0.
202 I'll look at the code to see if there is some way we can avoid
203 this kind of problem. Probably the best solution is to make the
204 first thread simply start using the Volume in drive 1 rather than
205 transferring it to drive 0.
206 - Fix re-read of last block to check if job has actually written
207 a block, and check if block was written by a different job
208 (i.e. multiple simultaneous jobs writing).
209 - Figure out how to configure query.sql. Suggestion to use m4:
210 == changequote.m4 ===
211 changequote(`[',`]')dnl
212 ==== query.sql.in ===
213 :List next 20 volumes to expire
215 Pool.Name AS PoolName,
220 [ FROM_UNIXTIME(UNIX_TIMESTAMP(Media.LastWritten) Media.VolRetention) AS Expire, ])dnl
222 [ media.lastwritten + interval '1 second' * media.volretention as expire, ])dnl
226 ON Media.PoolId=Pool.PoolId
227 WHERE Media.LastWritten>0
231 Command: m4 -DmySQL changequote.m4 query.sql.in >query.sql
233 The problem is that it requires m4, which is not present on all machines
235 - Given all the problems with FIFOs, I think the solution is to do something a
236 little different, though I will look at the code and see if there is not some
237 simple solution (i.e. some bug that was introduced). What might be a better
238 solution would be to use a FIFO as a sort of "key" to tell Bacula to read and
239 write data to a program rather than the FIFO. For example, suppose you
244 Then, I could imagine if you backup and restore this file with a direct
245 reference as is currently done for fifos, instead, during backup Bacula will
248 /home/kern/my-fifo.backup
250 and read the data that my-fifo.backup writes to stdout. For restore, Bacula
253 /home/kern/my-fifo.restore
255 and send the data backed up to stdout. These programs can either be an
256 executable or a shell script and they need only read/write to stdin/stdout.
258 I think this would give a lot of flexibility to the user without making any
259 significant changes to Bacula.
264 select FilenameId from Filename where Name='';
265 # Get list of all directories referenced in a Backup.
266 select Path.Path from Path,File where File.JobId=nnn and
267 File.FilenameId=(FilenameId-from-above) and File.PathId=Path.PathId
268 order by Path.Path ASC;
270 - Look into using Dart for testing
271 http://public.kitware.com/Dart/HTML/Index.shtml
273 - Look into replacing autotools with cmake
274 http://www.cmake.org/HTML/Index.html
276 === Migration from David ===
277 What I'd like to see:
280 Name = "<poolname>-migrate"
284 Migration Selection Type = LowestUtil | OldestVol | PoolOccupancy |
285 Client | PoolResidence | Volume | JobName | SQLquery
286 Migration Selection Pattern = "regexp"
287 Next Pool = <override>
290 There should be no need for a Level (migration is always Full, since you
291 don't calculate differential/incremental differences for migration),
292 Storage should be determined by the volume types in the pool, and Client
293 is really a selection issue. Migration should always occur to the
294 NextPool defined in the pool definition. If no nextpool is defined, the
295 job should end with a reason of "no place to go". If Next Pool statement
296 is present, we override the check in the pool definition and use the
299 Here's how I'd define Migration Selection Types:
302 Client -- Migrate data from selected client only. Migration Selection
303 Pattern regexp provides pattern to select client names, eg ^FS00* makes
304 all client names starting with FS00 eligible for migration.
306 Jobname -- Migration all jobs matching name. Migration Selection Pattern
307 regexp provides pattern to select jobnames existing in pool.
309 Volume -- Migrate all data on specified volumes. Migration Selection
310 Pattern regexp provides selection criteria for volumes to be migrated.
311 Volumes must exist in pool to be eligible for migration.
315 LowestUtil -- Identify the volume in the pool with the least data on it
316 and empty it. No Migration Selection Pattern required.
318 OldestVol -- Identify the LRU volume with data written, and empty it. No
319 Migration Selection Pattern required.
321 PoolOccupancy -- if pool occupancy exceeds <highmig>, migrate volumes
322 (starting with most full volumes) until pool occupancy drops below
323 <lowmig>. Pool highmig and lowmig values are in pool definition, no
324 Migration Selection Pattern required.
328 SQLQuery -- Migrate all jobuids returned by the supplied SQL query.
329 Migration Selection Pattern contains SQL query to execute; should return
330 a list of 1 or more jobuids to migrate.
332 PoolResidence -- Migrate data sitting in pool for longer than
333 PoolResidence value in pool definition. Migration Selection Pattern
334 optional; if specified, override value in pool definition (value in
338 [ possibly a Python event -- kes ]
340 - Mount on an Autochanger with no tape in the drive causes:
341 Automatically selected Storage: LTO-changer
342 Enter autochanger drive[0]: 0
343 3301 Issuing autochanger "loaded drive 0" command.
344 3302 Autochanger "loaded drive 0", result: nothing loaded.
345 3301 Issuing autochanger "loaded drive 0" command.
346 3302 Autochanger "loaded drive 0", result: nothing loaded.
347 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
348 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
349 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
350 If this is not a blank tape, try unmounting and remounting the Volume.
351 - If Drive 0 is blocked, and drive 1 is set "Autoselect=no", drive 1 will
353 - Autochanger did not change volumes.
354 select * from Storage;
355 +-----------+-------------+-------------+
356 | StorageId | Name | AutoChanger |
357 +-----------+-------------+-------------+
358 | 1 | LTO-changer | 0 |
359 +-----------+-------------+-------------+
360 05-May 03:50 roxie-sd: 3302 Autochanger "loaded drive 0", result is Slot 11.
361 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Warning: Director wanted Volume "LT
362 Current Volume "LT0-002" not acceptable because:
363 1997 Volume "LT0-002" not in catalog.
364 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Error: Autochanger Volume "LT0-002"
365 Setting InChanger to zero in catalog.
366 05-May 03:50 roxie-dir: Tibs.2006-05-05_03.05.02 Error: Unable to get Media record
368 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Error getting Volume i
369 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: Job 530 canceled.
370 05-May 03:50 roxie-sd: Tibs.2006-05-05_03.05.02 Fatal error: spool.c:249 Fatal appe
371 05-May 03:49 Tibs: Tibs.2006-05-05_03.05.02 Fatal error: c:\cygwin\home\kern\bacula
380 FirstWritten: 2006-05-05 03:11:54
381 LastWritten: 2006-05-05 03:50:23
382 LabelDate: 2005-12-26 16:52:40
393 VolRetention: 31,536,000
405 Note VolStatus is blank!!!!!
412 FirstWritten: 0000-00-00 00:00:00
413 LastWritten: 0000-00-00 00:00:00
414 LabelDate: 2005-12-26 16:52:40
425 VolRetention: 31,536,000
438 Automatically selected Storage: LTO-changer
439 Enter autochanger drive[0]: 0
440 3301 Issuing autochanger "loaded drive 0" command.
441 3302 Autochanger "loaded drive 0", result: nothing loaded.
442 3301 Issuing autochanger "loaded drive 0" command.
443 3302 Autochanger "loaded drive 0", result: nothing loaded.
444 3902 Cannot mount Volume on Storage Device "LTO-Drive1" (/dev/nst0) because:
445 Couldn't rewind device "LTO-Drive1" (/dev/nst0): ERR=dev.c:678 Rewind error on "LTO-Drive1" (/dev/nst0). ERR=No medium found.
447 3905 Device "LTO-Drive1" (/dev/nst0) open but no Bacula volume is mounted.
448 If this is not a blank tape, try unmounting and remounting the Volume.
450 - http://www.dwheeler.com/essays/commercial-floss.html
451 - Add VolumeLock to prevent all but lock holder (SD) from updating
452 the Volume data (with the exception of VolumeState).
453 - The btape fill command does not seem to use the Autochanger
454 - Make Windows installer default to system disk drive.
455 - Look at using ioctl(FIOBMAP, ...) on Linux, and
456 DeviceIoControl(..., FSCTL_QUERY_ALLOCATED_RANGES, ...) on
457 Win32 for sparse files.
458 http://www.flexhex.com/docs/articles/sparse-files.phtml
459 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
460 - Directive: at <event> "command"
461 - Command: pycmd "command" generates "command" event. How to
462 attach to a specific job?
463 - Integrate Christopher's St. Bernard code.
464 - run_cmd() returns int should return JobId_t
465 - get_next_jobid_from_list() returns int should return JobId_t
466 - Document export LDFLAGS=-L/usr/lib64
467 - Don't attempt to restore from "Disabled" Volumes.
468 - Network error on Win32 should set Win32 error code.
469 - What happens when you rename a Disk Volume?
470 - Job retention period in a Pool (and hence Volume). The job would
472 - Look at -D_FORTIFY_SOURCE=2
473 - Add Win32 FileSet definition somewhere
474 - Look at fixing restore status stats in SD.
475 - Look at using ioctl(FIMAP) and FIGETBSZ for sparse files.
476 http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/fibmap.html
477 - Implement a mode that says when a hard read error is
478 encountered, read many times (as it currently does), and if the
479 block cannot be read, skip to the next block, and try again. If
480 that fails, skip to the next file and try again, ...
482 create table LevelType (LevelType binary(1), LevelTypeLong tinyblob);
483 insert into LevelType (LevelType,LevelTypeLong) values
487 - Show files/second in client status output.
488 - Add a recursive mark command (rmark) to restore.
489 - "Minimum Job Interval = nnn" sets minimum interval between Jobs
490 of the same level and does not permit multiple simultaneous
491 running of that Job (i.e. lets any previous invocation finish
492 before doing Interval testing).
493 - Look at simplifying File exclusions.
494 - New directive "Delete purged Volumes"
495 - new pool XXX with ScratchPoolId = MyScratchPool's PoolId and
496 let it fill itself, and RecyclePoolId = XXX's PoolId so I can
497 see if it become stable and I just have to supervise
499 - If I want to remove this pool, I set RecyclePoolId = MyScratchPool's
500 PoolId, and when it is empty remove it.
501 - Figure out how to recycle Scratch volumes back to the Scratch Pool.
503 - Allow Check Labels to be used with Bacula labels.
504 - "Resuming" a failed backup (lost line for example) by using the
505 failed backup as a sort of "base" job.
507 - Email to the user when the tape is about to need changing x
508 days before it needs changing.
509 - Command to show next tape that will be used for a job even
510 if the job is not scheduled.
511 - From: Arunav Mandal <amandal@trolltech.com>
512 1. When jobs are running and bacula for some reason crashes or if I do a
513 restart it remembers and jobs it was running before it crashed or restarted
514 as of now I loose all jobs if I restart it.
516 2. When spooling and in the midway if client is disconnected for instance a
517 laptop bacula completely discard the spool. It will be nice if it can write
518 that spool to tape so there will be some backups for that client if not all.
520 3. We have around 150 clients machines it will be nice to have a option to
521 upgrade all the client machines bacula version automatically.
523 4. Atleast one connection should be reserved for the bconsole so at heavy load
524 I should connect to the director via bconsole which at sometimes I can't
526 5. Another most important feature that is missing, say at 10am I manually
527 started backup of client abc and it was a full backup since client abc has
528 no backup history and at 10.30am bacula again automatically started backup of
529 client abc as that was in the schedule. So now we have 2 multiple Full
530 backups of the same client and if we again try to start a full backup of
531 client backup abc bacula won't complain. That should be fixed.
533 - Fix bpipe.c so that it does not modify results pointer.
534 ***FIXME*** calling sequence should be changed.
535 - For Windows disaster recovery see http://unattended.sf.net/
536 - regardless of the retention period, Bacula will not prune the
537 last Full, Diff, or Inc File data until a month after the
538 retention period for the last Full backup that was done.
539 - update volume=xxx --- add status=Full
540 - Remove old spool files on startup.
541 - Exclude SD spool/working directory.
542 - Refuse to prune last valid Full backup. Same goes for Catalog.
544 - Make a callback when Rerun failed levels is called.
545 - Give Python program access to Scheduled jobs.
546 - Add setting Volume State via Python.
547 - Python script to save with Python, not save, save with Bacula.
548 - Python script to do backup.
550 - Change the Priority, Client, Storage, JobStatus (error)
551 at the start of a job.
552 - Why is SpoolDirectory = /home/bacula/spool; not reported
553 as an error when writing a DVD?
554 - Make bootstrap file handle multiple MediaTypes (SD)
555 - Remove all old Device resource code in Dir and code to pass it
556 back in SD -- better, rework it to pass back device statistics.
557 - Check locking of resources -- be sure to lock devices where previously
558 resources were locked.
559 - The last part is left in the spool dir.
562 - In restore don't compare byte count on a raw device -- directory
563 entry does not contain bytes.
567 MA = (last_MA * 3 + rate) / 4
568 rate = (bytes - last_bytes) / (runtime - last_runtime)
569 - Max Vols limit in Pool off by one?
570 - Implement Files/Bytes,... stats for restore job.
571 - Implement Total Bytes Written, ... for restore job.
572 - Despool attributes simultaneously with data in a separate
573 thread, rejoined at end of data spooling.
574 - 7. Implement new Console commands to allow offlining/reserving drives,
575 and possibly manipulating the autochanger (much asked for).
576 - Add start/end date editing in messages (%t %T, %e?) ...
577 - Add ClientDefs similar to JobDefs.
578 - Print more info when bextract -p accepts a bad block.
579 - Fix FD JobType to be set before RunBeforeJob in FD.
580 - Look at adding full Volume and Pool information to a Volume
581 label so that bscan can get *all* the info.
582 - If the user puts "Purge Oldest Volume = yes" or "Recycle Oldest Volume = yes"
583 and there is only one volume in the pool, refuse to do it -- otherwise
584 he fills the Volume, then immediately starts reusing it.
585 - Implement copies and stripes.
586 - Add history file to console.
587 - Each file on tape creates a JobMedia record. Peter has 4 million
588 files spread over 10000 tape files and four tapes. A restore takes
589 16 hours to build the restore list.
590 - Add and option to check if the file size changed during backup.
591 - Make sure SD deletes spool files on error exit.
592 - Delete old spool files when SD starts.
593 - When labeling tapes, if you enter 000026, Bacula uses
594 the tape index rather than the Volume name 000026.
595 - Add offline tape command to Bacula console.
597 Enter MediaId or Volume name: 32
598 Enter new Volume name: DLT-20Dec04
599 Automatically selected Pool: Default
600 Connecting to Storage daemon DLTDrive at 192.168.68.104:9103 ...
601 Sending relabel command from "DLT-28Jun03" to "DLT-20Dec04" ...
602 block.c:552 Write error at 0:0 on device /dev/nst0. ERR=Bad file descriptor.
603 Error writing final EOF to tape. This tape may not be readable.
604 dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
605 askdir.c:219 NULL Volume name. This shouldn't happen!!!
606 3912 Failed to label Volume: ERR=dev.c:1207 ioctl MTWEOF error on /dev/nst0. ERR=Permission denied.
607 Label command failed for Volume DLT-20Dec04.
608 Do not forget to mount the drive!!!
609 - Bug: if a job is manually scheduled to run later, it does not appear
610 in any status report and cannot be cancelled.
612 ==== Keeping track of deleted/new files ====
613 - To mark files as deleted, run essentially a Verify to disk, and
614 when a file is found missing (MarkId != JobId), then create
615 a new File record with FileIndex == -1. This could be done
616 by the FD at the same time as the backup.
618 My "trick" for keeping track of deletions is the following.
619 Assuming the user turns on this option, after all the files
620 have been backed up, but before the job has terminated, the
621 FD will make a pass through all the files and send their
622 names to the DIR (*exactly* the same as what a Verify job
623 currently does). This will probably be done at the same
624 time the files are being sent to the SD avoiding a second
625 pass. The DIR will then compare that to what is stored in
626 the catalog. Any files in the catalog but not in what the
627 FD sent will receive a catalog File entry that indicates
628 that at that point in time the file was deleted. This
629 either transmitted to the FD or simultaneously computed in
630 the FD, so that the FD can put a record on the tape that
631 indicates that the file has been deleted at this point.
632 A delete file entry could potentially be one with a FileIndex
633 of 0 or perhaps -1 (need to check if FileIndex is used for
634 some other thing as many of the Bacula fields are "overloaded"
637 During a restore, any file initially picked up by some
638 backup (Full, ...) then subsequently having a File entry
639 marked "delete" will be removed from the tree, so will not
640 be restored. If a file with the same name is later OK it
641 will be inserted in the tree -- this already happens. All
642 will be consistent except for possible changes during the
645 Since I'm on the subject, some of you may be wondering what
646 the utility of the in memory tree is if you are going to
647 restore everything (at least it comes up from time to time
648 on the list). Well, it is still *very* useful because it
649 allows only the last item found for a particular filename
650 (full path) to be entered into the tree, and thus if a file
651 is backed up 10 times, only the last copy will be restored.
652 I recently (last Friday) restored a complete directory, and
653 the Full and all the Differential and Incremental backups
654 spanned 3 Volumes. The first Volume was not even mounted
655 because all the files had been updated and hence backed up
656 since the Full backup was made. In this case, the tree
657 saved me a *lot* of time.
659 Make sure this information is stored on the tape too so
660 that it can be restored directly from the tape.
662 All the code (with the exception of formally generating and
663 saving the delete file entries) already exists in the Verify
664 Catalog command. It explicitly recognizes added/deleted files since
665 the last InitCatalog. It is more or less a "simple" matter of
666 taking that code and adapting it slightly to work for backups.
668 Comments from Martin Simmons (I think they are all covered):
669 Ok, that should cover the basics. There are few issues though:
671 - Restore will depend on the catalog. I think it is better to include the
672 extra data in the backup as well, so it can be seen by bscan and bextract.
674 - I'm not sure if it will preserve multiple hard links to the same inode. Or
675 maybe adding or removing links will cause the data to be dumped again?
677 - I'm not sure if it will handle renamed directories. Possibly it will work
678 by dumping the whole tree under a renamed directory?
680 - It remains to be seen how the backup performance of the DIR's will be
681 affected when comparing the catalog for a large filesystem.
685 How about introducing a Type = MgmtPolicy job type? That job type would
686 be responsible for scanning the Bacula environment looking for specific
687 conditions, and submitting the appropriate jobs for implementing said
691 Name = "Migration-Policy"
693 Policy Selection Job Type = Migrate
694 Scope = "<keyword> <operator> <regexp>"
695 Threshold = "<keyword> <operator> <regexp>"
696 Job Template = <template-name>
699 Where <keyword> is any legal job keyword, <operator> is a comparison
700 operator (=,<,>,!=, logical operators AND/OR/NOT) and <regexp> is a
701 appropriate regexp. I could see an argument for Scope and Threshold
702 being SQL queries if we want to support full flexibility. The
703 Migration-Policy job would then get scheduled as frequently as a site
704 felt necessary (suggested default: every 15 minutes).
709 Name = "Migration-Policy"
711 Policy Selection Job Type = Migration
713 Threshold = "Migration Selection Type = LowestUtil"
714 Job Template = "MigrationTemplate"
717 would select all pools for examination and generate a job based on
718 MigrationTemplate to automatically select the volume with the lowest
719 usage and migrate it's contents to the nextpool defined for that pool.
721 This policy abstraction would be really handy for adjusting the behavior
722 of Bacula according to site-selectable criteria (one thing that pops
723 into mind is Amanda's ability to automatically adjust backup levels
724 depending on various criteria).
730 - Add Pool/Storage override regression test.
731 - Add delete JobId to regression.
732 - Add a regression test for dbcheck.
733 - New test to add bscan to four-concurrent-jobs regression,
734 i.e. after the four-concurrent jobs zap the
735 database as is done in the bscan-test, then use bscan to
736 restore the database, do a restore and compare with the
738 - Add restore of specific JobId to regression (item 3
739 on the restore prompt)
740 - Add IPv6 to regression
741 - Add database test to regression. Test each function like delete,
744 - AntiVir can slow down backups on Win32 systems.
745 - Win32 systems with FAT32 can be much slower than NTFS for
746 more than 1000 files per directory.
750 - A HOLD command to stop all jobs from starting.
751 - A PAUSE command to pause all running jobs ==> release the
753 - Media Type = LTO,LTO-2,LTO-3
754 Media Type Read = LTO,LTO2,LTO3
755 Media Type Write = LTO2, LTO3
757 === From Carsten Menke <bootsy52@gmx.net>
759 Following is a list of what I think in the situations where I'm faced with,
760 could be a usefull enhancement to bacula, which I'm certain other users will
761 benefit from as well.
763 1. NextJob/NextJobs Directive within a Job Resource in the form of
764 NextJobs = job1,job2.
767 I currently solved the problem with running multiple jobs each after each
768 by setting the Max Wait Time for a job to 8 hours, and give
769 the jobs different Priorities. However, there scenarios where
770 1 Job is directly depending on another job, so if the former job fails,
771 the job after it needn't to be run
772 while maybe other jobs should run despite of that
775 A Backup Job and a Verify job, if the backup job fails there is no need to run
776 the verify job, as the backup job already failed. However, one may like
777 to backup the Catalog to disk despite of that the main backup job failed.
780 I see that this is related to the Event Handlers which are on the ToDo
781 list, also it is maybe a good idea to check for the return value and
782 execute different actions based on the return value
785 3. offline capability to bconsole
788 Currently I use a script which I execute within the last Job via the
789 RunAfterJob Directive, to release and eject the tape.
790 So I have to call bconsole "release=Storage-Name" and afterwards
791 mt -f /dev/nst0 eject to get the tape out.
793 If I have multiple Storage Devices, than these may not be /dev/nst0 and
794 I have to modify the script or call it with parameters etc.
795 This would actually not be needed, as everything is already defined
796 in bacula-sd.conf and if I can invoke bconsole with the
797 storage name via $1 in the script than I'm done and information is
800 4. %s for Storage Name added to the chars being substituted in "RunAfterJob"
804 For the reason mentioned in 3. to have the ability to call a
805 script with /scripts/foobar %s and in the script use $1
806 to pass the Storage Name to bconsole
808 5. Setting Volume State within a Job Resource
811 Instead of using "Maximum Volume Jobs" in the Pool Resource,
812 I would have the possibilty to define
813 in a Job Resource that after this certain job is run, the Volume State
814 should be set to "Volume State = Used", this give more flexibility (IMHO).
816 6. Localization of Bacula Messages
819 Unfortunatley many,many people I work with don't speak english very well.
820 So if at least the Reporting messages would be localized then they
821 would understand that they have to change the tape,etc. etc.
823 I volunteer to do the german translations, and if I can convince my wife also
824 french and Morre (western african language).
826 7. OK, this is evil, probably bound to security risks and maybe not possible
827 due to the design of bacula.
829 Implementation of Backtics ( `command` ) for shell comand execution to
830 the "Label Format" Directive.
834 Currently I have defined BACULA_DAY_OF_WEEK="day1|day2..." resulting in
835 Label Format = "HolyBackup-${BACULA_DAY_OF_WEEK[${WeekDay}]}". If I could
836 use backticks than I could use "Label Format = HolyBackup-`date +%A` to have
837 the localized name for the day of the week appended to the
838 format string. Then I have the tape labeled automatically with weekday
839 name in the correct language.
841 - Yes, that is surely the case. I probably should turn those into Warning
842 errors. In addition, you just made me think that it might not be bad to
843 add an option to check the file size after backing up the file and
844 report if it changes. This would be done as an option because it would
847 Kern, good idea. If you do do that, mention in the output: file
848 shrunk, or file expanded, just to make it obvious to the user
849 (without having to the refer to file size), just how the file size
852 Would this option be for all file, or just one file? Or a fileset?
853 - Make output from status use html table tags for nicely
854 presenting in a browser.
855 - Can one write tapes faster with 8192 byte block sizes?
856 - Document security problems with the same password for everyone in
857 rpm and Win32 releases.
858 - Browse generations of files.
859 - I've seen an error when my catalog's File table fills up. I
860 then have to recreate the File table with a larger maximum row
861 size. Relevant information is at
862 http://dev.mysql.com/doc/mysql/en/Full_table.html ; I think the
863 "Installing and Configuring MySQL" chapter should talk a bit
864 about this potential problem, and recommend a solution.
865 - For Solaris must use POSIX awk.
866 - Want speed of writing to tape while despooling.
867 - Supported autochanger:
875 Wangtek 6525ES (SCSI-1 QIC drive, 525MB), under Linux 2.4.something,
876 bacula 1.36.0/1 works with blocksize 16k INSIDE bacula-sd.conf.
877 - Add regex from http://www.pcre.org to Bacula for Win32.
878 - Use only shell tools no make in CDROM package.
879 - Include within include does it work?
880 - Implement a Pool of type Cleaning?
881 - Implement VolReadTime and VolWriteTime in SD
882 - Modify Backing up Your Database to include a bootstrap file.
883 - Think about making certain database errors fatal.
884 - Look at correcting the time jump in the scheduler for daylight
885 savings time changes.
886 - Add a "real" timer to network connections.
887 - Promote to Full = Time period
888 - Check dates entered by user for correctness (month/day/... ranges)
889 - Compress restore Volume listing by date and first file.
890 - Look at patches/bacula_db.b2z postgresql that loops during restore.
892 - Perhaps add read/write programs and/or plugins to FileSets.
893 - How to handle backing up portables ...
894 - Add some sort of guaranteed Interval for upgrading jobs.
895 - Can we write the state file after every job terminates? On Win32
896 the system crashes and the state file is not updated.
899 Documentation to do: (any release a little bit at a time)
900 - Doc to do unmount before removing magazine.
901 - Alternative to static linking "ldd prog" save all binaries listed,
902 restore them and point LD_LIBRARY_PATH to them.
903 - Document add "</dev/null >/dev/null 2>&1" to the bacula-fd command line
904 - Document query file format.
905 - Add more documentation for bsr files.
906 - Document problems with Verify and pruning.
907 - Document how to use multiple databases.
908 - VXA drives have a "cleaning required"
909 indicator, but Exabyte recommends preventive cleaning after every 75
912 In this context, it should be noted that Exabyte has a command-line
913 vxatool utility available for free download. (The current version is
914 vxatool-3.72.) It can get diagnostic info, read, write and erase tapes,
915 test the drive, unload tapes, change drive settings, flash new firmware,
917 Of particular interest in this context is that vxatool <device> -i will
918 report, among other details, the time since last cleaning in tape motion
919 minutes. This information can be retrieved (and settings changed, for
920 that matter) through the generic-SCSI device even when Bacula has the
921 regular tape device locked. (Needless to say, I don't recommend
922 changing tape settings while a job is running.)
923 - Lookup HP cleaning recommendations.
924 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
925 - Document doing table repair
928 ===================================
929 - Add macro expansions in JobDefs.
930 Run Before Job = "SomeFile %{Level} %{Client}"
931 Write Bootstrap="/some/dir/%{JobName}_%{Client}.bsr"
932 - Use non-blocking network I/O but if no data is available, use
934 - Use gather write() for network I/O.
935 - Autorestart on crash.
936 - Add bandwidth limiting.
937 - Add acks every once and a while from the SD to keep
938 the line from timing out.
939 - When an error in input occurs and conio beeps, you can back
940 up through the prompt.
941 - Detect fixed tape block mode during positioning by looking at
942 block numbers in btape "test". Possibly adjust in Bacula.
943 - Fix list volumes to output volume retention in some other
944 units, perhaps via a directive.
945 - Allow Simultaneous Priorities = yes => run up to Max concurrent jobs even
946 with multiple priorities.
947 - If you use restore replace=never, the directory attributes for
948 non-existent directories will not be restored properly.
950 - see lzma401.zip in others directory for new compression
952 - Allow the user to select JobType for manual pruning/purging.
953 - bscan does not put first of two volumes back with all info in
955 - Implement the FreeBSD nodump flag in chflags.
956 - Figure out how to make named console messages go only to that
957 console and to the non-restricted console (new console class?).
958 - Make restricted console prompt for password if *ask* is set or
959 perhaps if password is undefined.
960 - Implement "from ISO-date/time every x hours/days/weeks/months" in
963 ==== from Marc Schoechlin
964 - the help-command should be more verbose
965 (it should explain the paramters of the different
967 -> its time-comsuming to consult the manual anytime
968 you need a special parameter
969 -> maybe its more easy to maintain this, if the
970 descriptions of that commands are outsourced to
972 - the cd-command should allow complete paths
973 i.e. cd /foo/bar/foo/bar
974 -> if a customer mails me the path to a certain file,
975 its faster to enter the specified directory
976 - if the password is not configured in bconsole.conf
977 you should be asked for it.
978 -> sometimes you like to do restore on a customer-machine
979 which shouldnt know the password for bacula.
980 -> adding the password to the file favours admins
981 to forget to remove the password after usage
983 the protection of that file is less important
984 - long-listed-output of commands should be scrollable
985 like the unix more/less-command does
986 -> if someone runs 200 and more machines, the lists could
987 be a little long and complex
988 - command-output should be shown column by column
989 to reduce scrolling and to increase clarity
991 - lsmark should list the selected files with full
993 - wildcards for selecting and file and directories would be nice
994 - any actions should be interuptable with STRG+C
995 - command-expansion would be pretty cool
997 - When the replace Never option is set, new directory permissions
998 are not restored. See bug 213. To fix this requires creating a
999 list of newly restored directories so that those directory
1000 permissions *can* be restored.
1001 - Add prune all command
1002 - Document fact that purge can destroy a part of a restore by purging
1003 one volume while others remain valid -- perhaps mark Jobs.
1004 - Add multiple-media-types.txt
1005 - look at mxt-changer.html
1006 - Make ? do a help command (no return needed).
1007 - Implement restore directory.
1008 - Document streams and how to implement them.
1009 - Try not to re-backup a file if a new hard link is added.
1010 - Add feature to backup hard links only, but not the data.
1011 - Fix stream handling to be simpler.
1012 - Add Priority and Bootstrap to Run a Job.
1013 - Eliminate Restore "Run Restore Job" prompt by allowing new "run command
1015 - Remove View FileSet button from Run a Job dialog.
1016 - Handle prompt for restore job at end of Restore command.
1017 - Add display of total selected files to Restore window.
1018 - Add tree pane to left of window.
1019 - Add progress meter.
1020 - Max wait time or max run time causes seg fault -- see runtime-bug.txt
1021 - Add message to user to check for fixed block size when the forward
1022 space test fails in btape.
1023 - When unmarking a directory check if all files below are unmarked and
1024 then remove the + flag -- in the restore tree.
1025 - Possibly implement: Action = Unmount Device="TapeDrive1" in Admin jobs.
1026 - Setup lrrd graphs: (http://www.linpro.no/projects/lrrd/) Mike Acar.
1027 - Revisit the question of multiple Volumes (disk) on a single device.
1028 - Add a block copy option to bcopy.
1029 - Finish work on Gnome restore GUI.
1030 - Fix "llist jobid=xx" where no fileset or client exists.
1031 - For each job type (Admin, Restore, ...) require only the really necessary
1032 fields.- Pass Director resource name as an option to the Console.
1033 - Add a "batch" mode to the Console (no unsolicited queries, ...).
1034 - Add a .list all files in the restore tree (probably also a list all files)
1035 Do both a long and short form.
1036 - Allow browsing the catalog to see all versions of a file (with
1037 stat data on each file).
1038 - Restore attributes of directory if replace=never set but directory
1040 - Use SHA1 on authentication if possible.
1041 - See comtest-xxx.zip for Windows code to talk to USB.
1042 - Add John's appended files:
1043 Appended = { /files/server/logs/http/*log }
1044 and such files would be treated as follows.On a FULL backup, they would
1045 be backed up like any other file.On an INCREMENTAL backup, where a
1046 previous INCREMENTAL or FULL was already in thecatalogue and the length
1047 of the file wasgreater than the length of the last backup, only thedata
1048 added since the last backup will be dumped.On an INCREMENTAL backup, if
1049 the length of the file is less than thelength of the file with the same
1050 name last backed up, the completefile is dumped.On Windows systems, with
1051 creation date of files, we can be evensmarter about this and not count
1052 entirely upon the length.On a restore, the full and all incrementals
1053 since it will beapplied in sequence to restore the file.
1054 - Check new HAVE_WIN32 open bits.
1055 - Check if the tape has moved before writing.
1056 - Handling removable disks -- see below:
1057 - Keep track of tape use time, and report when cleaning is necessary.
1058 - Add FromClient and ToClient keywords on restore command (or
1059 BackupClient RestoreClient).
1060 - Implement a JobSet, which groups any number of jobs. If the
1061 JobSet is started, all the jobs are started together.
1062 Allow Pool, Level, and Schedule overrides.
1063 - Enhance cancel to timeout BSOCK packets after a specific delay.
1064 - Do scheduling by UTC using gmtime_r() in run_conf, scheduler, and
1065 ua_status.!!! Thanks to Alan Brown for this tip.
1066 - Look at updating Volume Jobs so that Max Volume Jobs = 1 will work
1067 correctly for multiple simultaneous jobs.
1068 - Correct code so that FileSet MD5 is calculated for < and | filename
1070 - Implement the Media record flag that indicates that the Volume does disk
1072 - Implement VolAddr, which is used when Volume is addressed like a disk,
1073 and form it from VolFile and VolBlock.
1074 - Make multiple restore jobs for multiple media types specifying
1075 the proper storage type.
1076 - Fix fast block rejection (stored/read_record.c:118). It passes a null
1077 pointer (rec) to try_repositioning().
1078 - Look at extracting Win data from BackupRead.
1079 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
1080 which would take precidence over the Catalog "JobRetention".
1081 - Implement Label Format in Add and Label console commands.
1082 - Possibly up network buffers to 65K. Put on variable.
1083 - Put email tape request delays on one or more variables. User wants
1084 to cancel the job after a certain time interval. Maximum Mount Wait?
1085 - Job, Client, Device, Pool, or Volume?
1086 Is it possible to make this a directive which is *optional* in multiple
1087 resources, like Level? If so, I think I'd make it an optional directive
1088 in Job, Client, and Pool, with precedence such that Job overrides Client
1089 which in turn overrides Pool.
1091 - New Storage specifications:
1092 - Want to write to multiple storage devices simultaneously
1093 - Want to write to multiple storage devices sequentially (in one job)
1094 - Want to read/write simultaneously
1095 - Key is MediaType -- it must match
1097 Passed to SD as a sort of BSR record called Storage Specification
1101 MediaType -> Next MediaType
1103 Device -> Next Device
1105 Allow multiple Storage specifications
1113 Allow Multiple Pool specifications (note, Pool currently
1115 Allow Multiple MediaType specifications in Dir conf
1116 Allow Multiple Device specifications in Dir conf
1117 Perhaps keep this in a single SSR
1118 Tie a Volume to a specific device by using a MediaType that
1119 is contained in only one device.
1120 In SD allow Device to have Multiple MediaTypes
1122 - Ideas from Jerry Scharf:
1123 First let's point out some big pluses that bacula has for this
1125 more importantly it's active. Thank you so much for that
1126 even more important, it's not flaky
1127 it has an open access catalog, opening many possibilities
1128 it's pushing toward heterogeneous systems capability
1130 Macintosh file client
1131 macs are an interesting niche, but I fear a server is a rathole
1132 working bare iron recovery for windows
1133 the option for inc/diff backups not reset on fileset revision
1134 a) use both change and inode update time against base time
1135 b) do the full catalog check (expensive but accurate)
1136 sizing guide (how much system is needed to back up N systems/files)
1137 consultants on using bacula in building a disaster recovery system
1138 an integration guide
1139 or how to get at fancy things that one could do with bacula
1140 logwatch code for bacula logs (or similar)
1141 linux distro inclusion of bacula (brings good and bad, but necessary)
1142 win2k/XP server capability (icky but you asked)
1143 support for Oracle database ??
1145 - Look at adding SQL server and Exchange support for Windows.
1146 - Make dev->file and dev->block_num signed integers so that -1 can
1147 be an invalid value which happens with BSR.
1148 - Create VolAddr for disk files in place of VolFile and VolBlock. This
1149 is needed to properly specify ranges.
1150 - Add progress of files/bytes to SD and FD.
1151 - Print warning message if FileId > 4 billion
1152 - do a "messages" before the first prompt in Console
1153 - Client does not show busy during Estimate command.
1154 - Implement Console mtx commands.
1155 - Implement a Mount Command and an Unmount Command where
1156 the users could specify a system command to be performed
1157 to do the mount, after which Bacula could attempt to
1158 read the device. This is for Removeable media such as a CDROM.
1159 - Most likely, this mount command would be invoked explicitly
1160 by the user using the current Console "mount" and "unmount"
1161 commands -- the Storage Daemon would do the right thing
1162 depending on the exact nature of the device.
1163 - As with tape drives, when Bacula wanted a new removable
1164 disk mounted, it would unmount the old one, and send a message
1165 to the user, who would then use "mount" as described above
1166 once he had actually inserted the disk.
1167 - Implement dump/print label to UA
1168 - Spool to disk only when the tape is full, then when a tape is hung move
1170 - bextract is sending everything to the log file ****FIXME****
1171 - Allow multiple Storage specifications (or multiple names on
1172 a single Storage specification) in the Job record. Thus a job
1173 can be backed up to a number of storage devices.
1174 - Implement some way for the File daemon to contact the Director
1175 to start a job or pass its DHCP obtained IP number.
1176 - Implement a query tape prompt/replace feature for a console
1177 - Copy console @ code to gnome2-console
1178 - Make tree walk routines like cd, ls, ... more user friendly
1179 by handling spaces better.
1180 - Make sure that Bacula rechecks the tape after the 20 min wait.
1181 - Set IO_NOWAIT on Bacula TCP/IP packets.
1182 - Try doing a raw partition backup and restore by mounting a
1184 - From Lars Kellers:
1185 Yes, it would allow to highly automatic the request for new tapes. If a
1186 tape is empty, bacula reads the barcodes (native or simulated), and if
1187 an unused tape is found, it runs the label command with all the
1188 necessary parameters.
1190 By the way can bacula automatically "move" an empty/purged volume say
1191 in the "short" pool to the "long" pool if this pool runs out of volume
1193 - What to do about "list files job=xxx".
1194 - Look at how fuser works and /proc/PID/fd that is how Nic found the
1195 file descriptor leak in Bacula.
1196 - Implement WrapCounters in Counters.
1197 - Add heartbeat from FD to SD if hb interval expires.
1198 - Can we dynamically change FileSets?
1199 - If pool specified to label command and Label Format is specified,
1200 automatically generate the Volume name.
1201 - Why can't SQL do the filename sort for restore?
1202 - Add ExhautiveRestoreSearch
1203 - Look at the possibility of loading only the necessary
1204 data into the restore tree (i.e. do it one directory at a
1205 time as the user walks through the tree).
1206 - Possibly use the hash code if the user selects all for a restore command.
1207 - Fix "restore all" to bypass building the tree.
1208 - Prohibit backing up archive device (findlib/find_one.c:128)
1209 - Implement Release Device in the Job resource to unmount a drive.
1210 - Implement Acquire Device in the Job resource to mount a drive,
1211 be sure this works with admin jobs so that the user can get
1212 prompted to insert the correct tape. Possibly some way to say to
1213 run the job but don't save the files.
1214 - Make things like list where a file is saved case independent for
1216 - Use autochanger to handle multiple devices.
1217 - Implement a Recycle command
1218 - Start working on Base jobs.
1219 - Implement UnsavedFiles DB record.
1220 - From Phil Stracchino:
1221 It would probably be a per-client option, and would be called
1222 something like, say, "Automatically purge obsoleted jobs". What it
1223 would do is, when you successfully complete a Differential backup of a
1224 client, it would automatically purge all Incremental backups for that
1225 client that are rendered redundant by that Differential. Likewise,
1226 when a Full backup on a client completed, it would automatically purge
1227 all Differential and Incremental jobs obsoleted by that Full backup.
1228 This would let people minimize the number of tapes they're keeping on
1229 hand without having to master the art of retention times.
1230 - When doing a Backup send all attributes back to the Director, who
1231 would then figure out what files have been deleted.
1232 - Currently in mount.c:236 the SD simply creates a Volume. It should have
1233 explicit permission to do so. It should also mark the tape in error
1234 if there is an error.
1235 - Cancel waiting for Client connect in SD if FD goes away.
1237 - Implement timeout in response() when it should come quickly.
1238 - Implement a Slot priority (loaded/not loaded).
1239 - Implement "vacation" Incremental only saves.
1240 - Implement create "FileSet"?
1241 - Add prefixlinks to where or not where absolute links to FD.
1242 - Issue message to mount a new tape before the rewind.
1243 - Simplified client job initiation for portables.
1244 - If SD cannot open a drive, make it periodically retry.
1245 - Add more of the config info to the tape label.
1247 - Refine SD waiting output:
1248 Device is being positioned
1249 > Device is being positioned for append
1250 > Device is being positioned to file x
1252 - Figure out some way to estimate output size and to avoid splitting
1253 a backup across two Volumes -- this could be useful for writing CDROMs
1254 where you really prefer not to have it split -- not serious.
1255 - Have SD compute MD5 or SHA1 and compare to what FD computes.
1256 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
1257 actual data on the Volume and compare it.
1258 - Implement Bacula plugins -- design API
1259 - Make bcopy read through bad tape records.
1260 - Program files (i.e. execute a program to read/write files).
1261 Pass read date of last backup, size of file last time.
1262 - Add Signature type to File DB record.
1263 - CD into subdirectory when open()ing files for backup to
1264 speed up things. Test with testfind().
1265 - Priority job to go to top of list.
1266 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
1267 - Implement some way for the Console to dynamically create a job.
1268 - Solaris -I on tar for include list
1269 - Need a verbose mode in restore, perhaps to bsr.
1270 - bscan without -v is too quiet -- perhaps show jobs.
1271 - Add code to reject whole blocks if not wanted on restore.
1272 - Check if we can increase Bacula FD priorty in Win2000
1273 - Make sure the MaxVolFiles is fully implemented in SD
1274 - Check if both CatalogFiles and UseCatalog are set to SD.
1275 - Possibly add email to Watchdog if drive is unmounted too
1276 long and a job is waiting on the drive.
1277 - After unmount, if restore job started, ask to mount.
1278 - Add UA rc and history files.
1279 - put termcap (used by console) in ./configure and
1280 allow -with-termcap-dir.
1281 - Fix Autoprune for Volumes to respect need for full save.
1282 - Compare tape to Client files (attributes, or attributes and data)
1283 - Make all database Ids 64 bit.
1284 - Allow console commands to detach or run in background.
1285 - Add SD message variables to control operator wait time
1286 - Maximum Operator Wait
1287 - Minimum Message Interval
1288 - Maximum Message Interval
1289 - Send Operator message when cannot read tape label.
1290 - Verify level=Volume (scan only), level=Data (compare of data to file).
1291 Verify level=Catalog, level=InitCatalog
1293 - Add keyword search to show command in Console.
1294 - Events : tape has more than xxx bytes.
1295 - Complete code in Bacula Resources -- this will permit
1296 reading a new config file at any time.
1297 - Handle ctl-c in Console
1298 - Implement script driven addition of File daemon to config files.
1299 - Think about how to make Bacula work better with File (non-tape) archives.
1300 - Write Unix emulator for Windows.
1301 - Put memory utilization in Status output of each daemon
1302 if full status requested or if some level of debug on.
1303 - Make database type selectable by .conf files i.e. at runtime
1304 - Set flag for uname -a. Add to Volume label.
1305 - Restore files modified after date
1306 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
1307 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
1308 - Timout a job or terminate if link goes down, or reopen link and query.
1309 - Concept of precious tapes (cannot be reused).
1310 - Make bcopy copy with a single tape drive.
1311 - Permit changing ownership during restore.
1314 > My suggestion: Add a feature on the systray menu-icon menu to request
1315 > an immediate backup now. This would be useful for laptop users who may
1316 > not be on the network when the regular scheduled backup is run.
1318 > My wife's suggestion: Add a setting to the win32 client to allow it to
1319 > shut down the machine after backup is complete (after, of course,
1320 > displaying a "System will shut down in one minute, click here to cancel"
1321 > warning dialog). This would be useful for sites that want user
1322 > woorkstations to be shut down overnight to save power.
1325 - Autolabel should be specified by DIR instead of SD.
1327 - Add media capacity
1328 - AutoScan (check checksum of tape)
1329 - Format command = "format /dev/nst0"
1333 - Seek resolution (usually corresponds to buffer size)
1334 - EODErrorCode=ENOSPC or code
1335 - Partial Read error code
1336 - Partial write error code
1337 - Nonformatted read error
1338 - Nonformatted write error
1339 - WriteProtected error
1343 - IgnoreCloseErrors=yes
1353 - FD sends unsaved file list to Director at end of job (see
1355 - File daemon should build list of files skipped, and then
1356 at end of save retry and report any errors.
1357 - Write a Storage daemon that uses pipes and
1358 standard Unix programs to write to the tape.
1360 - Need something that monitors the JCR queue and
1361 times out jobs by asking the deamons where they are.
1362 - Enhance Jmsg code to permit buffering and saving to disk.
1363 - device driver = "xxxx" for drives.
1364 - Verify from Volume
1365 - Ensure that /dev/null works
1366 - Need report class for messages. Perhaps
1367 report resource where report=group of messages
1368 - enhance scan_attrib and rename scan_jobtype, and
1369 fill in code for "since" option
1370 - Director needs a time after which the report status is sent
1371 anyway -- or better yet, a retry time for the job.
1372 - Don't reschedule a job if previous incarnation is still running.
1373 - Some way to automatically backup everything is needed????
1374 - Need a structure for pending actions:
1376 - termination status (part of buffered msgs?)
1378 Read, Write, Clean, Delete
1379 - Login to Bacula; Bacula users with different permissions:
1380 owner, group, user, quotas
1381 - Store info on each file system type (probably in the job header on tape.
1382 This could be the output of df; or perhaps some sort of /etc/mtab record.
1384 ========= ideas ===============
1385 From: "Jerry K. Schieffer" <jerry@skylinetechnology.com>
1386 To: <kern@sibbald.com>
1387 Subject: RE: [Bacula-users] future large programming jobs
1388 Date: Thu, 26 Feb 2004 11:34:54 -0600
1390 I noticed the subject thread and thought I would offer the following
1391 merely as sources of ideas, i.e. something to think about, not even as
1392 strong as a request. In my former life (before retiring) I often
1393 dealt with backups and storage management issues/products as a
1394 developer and as a consultant. I am currently migrating my personal
1395 network from amanda to bacula specifically because of the ability to
1396 cross media boundaries during storing backups.
1397 Are you familiar with the commercial product called ADSM (I think IBM
1398 now sells it under the Tivoli label)? It has a couple of interesting
1399 ideas that may apply to the following topics.
1401 1. Migration: Consider that when you need to restore a system, there
1402 may be pressure to hurry. If all the information for a single client
1403 can eventually end up on the same media (and in chronological order),
1404 the restore is facillitated by not having to search past information
1405 from other clients. ADSM has the concept of "client affinity" that
1406 may be associated with it's storage pools. It seems to me that this
1407 concept (as an optional feature) might fit in your architecture for
1410 ADSM also has the concept of defining one or more storage pools as
1411 "copy pools" (almost mirrors, but only in the sense of contents).
1412 These pools provide the ability to have duplicte data stored both
1413 onsite and offsite. The copy process can be scheduled to be handled
1414 by their storage manager during periods when there is no backup
1415 activity. Again, the migration process might be a place to consider
1416 implementing something like this.
1419 > It strikes me that it would be very nice to be able to do things
1421 > have the Job(s) backing up the machines run, and once they have all
1422 > completed, start a migration job to copy the data from disks Volumes
1424 > a tape library and then to offsite storage. Maybe this can already
1426 > done with some careful scheduling and Job prioritzation; the events
1427 > mechanism described below would probably make it very easy.
1429 This is the goal. In the first step (before events), you simply
1431 the Migration to tape later.
1433 2. Base jobs: In ADSM, each copy of each stored file is tracked in
1434 the database. Once a file (unique by path and metadata such as dates,
1435 size, ownership, etc.) is in a copy pool, no more copies are made. In
1436 other words, when you start ADSM, it begins like your concept of a
1437 base job. After that it is in the "incremental" mode. You can
1438 configure the number of "generations" of files to be retained, plus a
1439 retention date after which even old generations are purged. The
1440 database tracks the contents of media and projects the percentage of
1441 each volume that is valid. When the valid content of a volume drops
1442 below a configured percentage, the valid data are migrated to another
1443 volume and the old volume is marked as empty. Note, this requires
1444 ADSM to have an idea of the contents of a client, i.e. marking the
1445 database when an existing file was deleted, but this would solve your
1446 issue of restoring a client without restoring deleted files.
1448 This is pretty far from what bacula now does, but if you are going to
1449 rip things up for Base jobs,.....
1450 Also, the benefits of this are huge for very large shops, especially
1451 with media robots, but are a pain for shops with manual media
1455 > Base jobs sound pretty useful, but I'm not dying for them.
1457 Nobody is dying for them, but when you see what it does, you will die
1460 3. Restoring deleted files: Since I think my comments in (2) above
1461 have low probability of implementation, I'll also suggest that you
1462 could approach the issue of deleted files by a mechanism of having the
1463 fd report to the dir, a list of all files on the client for every
1464 backup job. The dir could note in the database entry for each file
1465 the date that the file was seen. Then if a restore as of date X takes
1466 place, only files that exist from before X until after X would be
1467 restored. Probably the major cost here is the extra date container in
1468 each row of the files table.
1470 Thanks for "listening". I hope some of this helps. If you want to
1471 contact me, please send me an email - I read some but not all of the
1472 mailing list traffic and might miss a reply there.
1474 Please accept my compliments for bacula. It is doing a great job for
1475 me!! I sympathize with you in the need to wrestle with excelence in
1476 execution vs. excelence in feature inclusion.
1481 ==============================
1484 - Design at hierarchial storage for Bacula. Migration and Clone.
1485 - Implement FSM (File System Modules).
1486 - Audit M_ error codes to ensure they are correct and consistent.
1487 - Add variable break characters to lex analyzer.
1488 Either a bit mask or a string of chars so that
1489 the caller can change the break characters.
1490 - Make a single T_BREAK to replace T_COMMA, etc.
1491 - Ensure that File daemon and Storage daemon can
1492 continue a save if the Director goes down (this
1493 is NOT currently the case). Must detect socket error,
1494 buffer messages for later.
1495 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
1496 - Add ability to backup to two Storage devices (two SD sessions) at
1497 the same time -- e.g. onsite, offsite.
1498 - Add the ability to consolidate old backup sets (basically do a restore
1499 to tape and appropriately update the catalog). Compress Volume sets.
1500 Might need to spool via file is only one drive is available.
1501 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
1502 someway to do so with every volume that has less than x% valid
1506 Migration: Move a backup from one Volume to another
1507 Clone: Copy a backup -- two Volumes
1510 ======================================================
1512 It is somewhat like a Full save becomes an incremental since
1513 the Base job (or jobs) plus other non-base files.
1515 - A Base backup is same as Full backup, just different type.
1516 - New BaseFiles table that contains:
1518 BaseJobId - Base JobId referenced for this FileId (needed ???)
1519 JobId - JobId currently running
1520 FileId - File not backed up, exists in Base Job
1521 FileIndex - FileIndex from Base Job.
1522 i.e. for each base file that exists but is not saved because
1523 it has not changed, the File daemon sends the JobId, BaseId,
1524 FileId, FileIndex back to the Director who creates the DB entry.
1525 - To initiate a Base save, the Director sends the FD
1526 the FileId, and full filename for each file in the Base.
1527 - When the FD finds a Base file, he requests the Director to
1528 send him the full File entry (stat packet plus MD5), or
1529 conversely, the FD sends it to the Director and the Director
1530 says yes or no. This can be quite rapid if the FileId is kept
1531 by the FD for each Base Filename.
1532 - It is probably better to have the comparison done by the FD
1533 despite the fact that the File entry must be sent across the
1535 - An alternative would be to send the FD the whole File entry
1536 from the start. The disadvantage is that it requires a lot of
1537 space. The advantage is that it requires less communications
1539 - The Job record must be updated to indicate that one or more
1541 - At end of Job, FD returns:
1542 1. Count of base files/bytes not written to tape (i.e. matches)
1543 2. Count of base file that were saved i.e. had changed.
1544 - No tape record would be written for a Base file that matches, in the
1545 same way that no tape record is written for Incremental jobs where
1546 the file is not saved because it is unchanged.
1547 - On a restore, all the Base file records must explicitly be
1548 found from the BaseFile tape. I.e. for each Full save that is marked
1549 to have one or more Base Jobs, search the BaseFile for all occurrences
1551 - An optimization might be to make the BaseFile have:
1557 This would avoid the need to explicitly fetch each File record for
1558 the Base job. The Base Job record will be fetched to get the
1559 VolSessionId and VolSessionTime.
1560 =========================================================
1563 ==========================================================
1565 For each Incremental job that is run, there may be files that
1566 were found but not saved because they were locked (this applies
1567 only to Windows). Such a system could send back to the Director
1568 a list of Unsaved files.
1570 - New UnSavedFiles table that contains:
1574 - Then in the next Incremental job, the list of Unsaved Files will be
1575 feed to the FD, who will ensure that they are explicitly chosen even
1576 if standard date/time check would not have selected them.
1577 =============================================================
1581 Multiple drive autochanger data: see Alan Brown
1582 mtx -f xxx unloadStorage Element 1 is Already Full(drive 0 was empty)
1583 Unloading Data Transfer Element into Storage Element 1...source Element
1584 Address 480 is Empty
1586 (drive 0 was empty and so was slot 1)
1587 > mtx -f xxx load 15 0
1588 no response, just returns to the command prompt when complete.
1589 > mtx -f xxx status Storage Changer /dev/changer:2 Drives, 60 Slots ( 2 Import/Export )
1590 Data Transfer Element 0:Full (Storage Element 15 Loaded):VolumeTag = HX001
1591 Data Transfer Element 1:Empty
1592 Storage Element 1:Empty
1593 Storage Element 2:Full :VolumeTag=HX002
1594 Storage Element 3:Full :VolumeTag=HX003
1595 Storage Element 4:Full :VolumeTag=HX004
1596 Storage Element 5:Full :VolumeTag=HX005
1597 Storage Element 6:Full :VolumeTag=HX006
1598 Storage Element 7:Full :VolumeTag=HX007
1599 Storage Element 8:Full :VolumeTag=HX008
1600 Storage Element 9:Full :VolumeTag=HX009
1601 Storage Element 10:Full :VolumeTag=HX010
1602 Storage Element 11:Empty
1603 Storage Element 12:Empty
1604 Storage Element 13:Empty
1605 Storage Element 14:Empty
1606 Storage Element 15:Empty
1607 Storage Element 16:Empty....
1608 Storage Element 28:Empty
1609 Storage Element 29:Full :VolumeTag=CLNU01L1
1610 Storage Element 30:Empty....
1611 Storage Element 57:Empty
1612 Storage Element 58:Full :VolumeTag=NEX261L2
1613 Storage Element 59 IMPORT/EXPORT:Empty
1614 Storage Element 60 IMPORT/EXPORT:Empty
1616 Unloading Data Transfer Element into Storage Element 15...done
1618 (just to verify it remembers where it came from, however it can be
1619 overrriden with mtx unload {slotnumber} to go to any storage slot.)
1621 There needs to be a table of drive # to devices somewhere - If there are
1622 multiple changers or drives there may not be a 1:1 correspondance between
1623 changer drive number and system device name - and depending on the way the
1624 drives are hooked up to scsi busses, they may not be linearly numbered
1625 from an offset point either.something like
1627 Autochanger drives = 2
1628 Autochanger drive 0 = /dev/nst1
1629 Autochanger drive 1 = /dev/nst2
1630 IMHO, it would be _safest_ to use explicit mtx unload commands at all
1631 times, not just for multidrive changers. For a 1 drive changer, that's
1637 MTX's manpage (1.2.15):
1638 unload [<slotnum>] [ <drivenum> ]
1639 Unloads media from drive <drivenum> into slot
1640 <slotnum>. If <drivenum> is omitted, defaults to
1641 drive 0 (as do all commands). If <slotnum> is
1642 omitted, defaults to the slot that the drive was
1643 loaded from. Note that there's currently no way
1644 to say 'unload drive 1's media to the slot it
1645 came from', other than to explicitly use that
1646 slot number as the destination.AB
1652 undef# camcontrol devlist
1653 <WANGTEK 51000 SCSI M74H 12B3> at scbus0 target 2 lun 0 (pass0,sa0)
1654 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 0 (pass1,sa1)
1655 <ARCHIVE 4586XX 28887-XXX 4BGD> at scbus0 target 4 lun 1 (pass2)
1657 tapeinfo -f /dev/sg0 with a bad tape in drive 1:
1658 [kern@rufus mtx-1.2.17kes]$ ./tapeinfo -f /dev/sg0
1659 Product Type: Tape Drive
1661 Product ID: 'C5713A '
1663 Attached Changer: No
1664 TapeAlert[3]: Hard Error: Uncorrectable read/write error.
1665 TapeAlert[20]: Clean Now: The tape drive neads cleaning NOW.
1672 Medium Type: Not Loaded
1675 DataCompEnabled: yes
1676 DataCompCapable: yes
1677 DataDeCompEnabled: yes
1684 Handling removable disks
1686 From: Karl Cunningham <karlc@keckec.com>
1688 My backups are only to hard disk these days, in removable bays. This is my
1689 idea of how a backup to hard disk would work more smoothly. Some of these
1690 things Bacula does already, but I mention them for completeness. If others
1691 have better ways to do this, I'd like to hear about it.
1693 1. Accommodate several disks, rotated similar to how tapes are. Identified
1694 by partition volume ID or perhaps by the name of a subdirectory.
1695 2. Abort & notify the admin if the wrong disk is in the bay.
1696 3. Write backups to different subdirectories for each machine to be backed
1698 4. Volumes (files) get created as needed in the proper subdirectory, one
1700 5. When a disk is recycled, remove or zero all old backup files. This is
1701 important as the disk being recycled may be close to full. This may be
1702 better done manually since the backup files for many machines may be
1703 scattered in many subdirectories.
1708 - Why the heck doesn't bacula drop root priviledges before connecting to