4 Documentation to do: (any release a little bit at a time)
5 - DB upgrade to version 5 in bacula-1.27b, DB upgrade to
6 version 6 in 1.31; DB upgrade to version 7 in 1.33/4.
7 - Document running a test version.
8 - Document query file format.
9 - Document static linking
10 - Document problems with Verify and pruning.
11 - Document how to use multiple databases.
12 - For FreeBSD typical /dev/nrsa0 and for mtx /dev/pass1
13 - VXA drives have a "cleaning required"
14 indicator, but Exabyte recommends preventive cleaning after every 75
17 In this context, it should be noted that Exabyte has a command-line
18 vxatool utility available for free download. (The current version is
19 vxatool-3.72.) It can get diagnostic info, read, write and erase tapes,
20 test the drive, unload tapes, change drive settings, flash new firmware,
22 Of particular interest in this context is that vxatool <device> -i will
23 report, among other details, the time since last cleaning in tape motion
24 minutes. This information can be retrieved (and settings changed, for
25 that matter) through the generic-SCSI device even when Bacula has the
26 regular tape device locked. (Needless to say, I don't recommend
27 changing tape settings while a job is running.)
28 - Lookup HP cleaning recommendations.
29 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
30 - Create a man page for each binary (Debian package requirement).
32 Testing to do: (painful)
34 - that ALL console command line options work and are always implemented
35 - blocksize recognition code.
36 - Test if rewind at end of tape waits for tape to rewind.
39 For 1.33 Testing/Documentation:
40 - Newly labeled tapes are chosen before ones already in use.
41 - Document new alias records in Director. SDAddress SDDeviceName, SDPassword.
42 FDPassword, FDAddress, DBAddress, DBPort, DBPassword.
43 - Document new Include/Exclude ...
44 - Add test of exclusion, test multiple Include {} statements.
45 - Add counter variable test.
46 - Document ln -sf /usr/lib/libncurses.so /usr/lib/libtermcap.so
47 and install the esound-dev package for compiling Console on SuSE.
48 This should read 'Document LDFLAGS="-L/usr/lib/termcap" ... '
49 - Add an example of using a FIFO in dirdconf.wml
50 - Add an item to the FAQ about running jobs in different timezones.
51 - Add some examples of job editing codes.
53 - Document Dan's new --with-dir-user, ... options.
55 - Figure out how to use ssh or stunnel to protect Bacula communications.
56 Add Dan's work to manual
58 - Add db check test to regression. Test each function like delete,
60 - Add subsections to the Disaster Recovery index section.
61 - Document Pool keyword for restore.
62 - If you use restore replace=never, the directory attributes for
63 non-existent directories will not be restored properly.
64 - In the Bacula User Guide you write:"Note, one major disadvantage of
65 writing to a NFS mounted volume as I do isthat if the other machine goes
66 down, the OS will wait forever on the fopen()call that Bacula makes. As
67 a consequence, Bacula will completely stall untilthe machine exporting
68 the NSF mounts comes back up. If someone knows a wayaround this, please
69 let me know."I haven't tried using NFS in years, but I think that the
70 "soft" and "intr"remount options may well help you. The only way of
71 being sure would be totry it.See, for example,
72 http://howtos.linux.com/guides/nag2/x-087-2-nfs.mountd.shtml
75 - Fix FreeBSD mt_count problem.
76 - During install, copy any console.conf to bconsole.conf.
77 - Have each daemon save the last_jobs structure when exiting and
78 read it back in when starting up.
79 - "restore jobid=1 select" calls get_storage_xxx, which prints "JobId 1 is
81 - Add FromClient and ToClient keywords on restore command (or
82 BackupClient RestoreClient).
83 - Automatic "update slots" on user configuration directive when a
85 - Allow "delete job jobid=xx jobid=xxx".
86 - Allow "delete job jobid=xxx,yyy,aaa-bbb" i.e. list + ranges.
87 - Implement multiple Volume in "purge jobs volume=".
88 - Implement a JobSet, which groups any number of jobs. If the
89 JobSet is started, all the jobs are started together.
90 Allow Pool, Level, and Schedule overrides.
91 - Enhance cancel to timeout BSOCK packets after a specific delay.
92 - When I restore to Windows the Created, Accessed and Modifiedtimes are
93 those of the time of the restore, not those of the originalfile.
94 The dates you will find in your restore log seem to be the original
96 - Rescue builds incorrect script files on Rufus.
97 - Write a Qmsg() to be used in bnet.c to prevent recursion. Queue the
98 message. If dequeueing toss the messages. Lock while dequeuing so that
99 it cannot be called recursively and set dequeuing flag.
100 - Add all pools in Dir conf to DB also update them to catch changed
101 LabelFormats and such.
102 - Symbolic link a directory to another one, then backup the symbolic
104 - Build console in client-only build.
105 - Restore attributes of directory if replace=never set but directory
107 - Check why Phil's Verify exclude does not work.
108 - Phil says that Windows file sizes mismatch in Verify when they should,
109 and that either the file size or the catalog size was zero.
110 - Fix option 2 of restore -- list where file is backed up -- require Client,
111 then list last 20 backups.
112 - Allow browsing the catalog to see all versions of a file (with
113 stat data on each file).
114 - Finish code passing files=nnn to restore start.
115 - Add level to estimate command.
116 - Check time/dates printed during restore when using Win32 API.
117 - Volume "add"ed to Pool gets recycled in first use. VolBytes=0
118 - Get rid of 0 dates in LastWritten, ...
119 - Fix "llist jobid=xx" where no fileset or client exists.
120 - Release SQLite 2.8.9
121 - If a tape is recycled while it is mounted, Stanislav Tvrudy must do an
122 additional mount to deblock the job.
124 bscan: sql_update.c:65 UPDATE File SET MD5='Ij+5kwN6TFIxK+8l8+/I+A' WHERE FileId=0
125 bscan: bscan.c:1074 Could not add MD5/SHA1 to File record. ERR=sql_update.c:65 Update problem: affected_rows=0
126 - Do scheduling by UTC using gmtime_r() in run_conf, scheduler, and
127 ua_status.!!! Thanks to Alan Brown for this tip.
128 - Look at updating Volume Jobs so that Max Volume Jobs = 1 will work
129 correctly for multiple simultaneous jobs.
130 - Correct code so that FileSet MD5 is calculated for < and | filename
132 - Mark Volume in error on error from WEOF.
133 - Why does Bacula need the drive open to do "autochanger list" ?
134 - Add a .list all files in the restore tree (probably also a list all files)
135 Do both a long and short form.
136 - Implement the Media record flag that indicates that the Volume does disk
138 - Implement VolAddr, which is used when Volume is addressed like a disk,
139 and form it from VolFile and VolBlock.
140 - Make multiple restore jobs for multiple media types specifying
141 the proper storage type.
142 - Implement MediaType keyword in bsr?
143 - Fix fast block rejection (stored/read_record.c:118). It passes a null
144 pointer (rec) to try_repositioning().
145 - Look at extracting Win data from BackupRead.
146 - Having dashes in filenames apparently creates problems for restore
148 - Add data compare on write/read in btape "test".
149 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
150 which would take precidence over the Catalog "JobRetention".
151 - Implement Label Format in Add and Label console commands.
152 - Possibly up network buffers to 65K. Put on variable.
153 - Put email tape request delays on one or more variables. User wants
154 to cancel the job after a certain time interval. Maximum Mount Wait?
155 Job, Client, Device, Pool, or Volume?
156 Is it possible to make this a directive which is *optional* in multiple
157 resources, like Level? If so, I think I'd make it an optional directive
158 in Job, Client, and Pool, with precedence such that Job overrides Client
159 which in turn overrides Pool.
160 - Print a message when a job starts if the conf file is not current.
161 - Finish work on conio.c
162 - To pass Include 1 or two letter commands
163 I Name Include name - first record
164 B Name Base name - repeat
167 E zzz Exclude expression (wild card)
169 D "reader" Reader program
170 T "writer" Writer program
171 O Options In current commpressed format
172 (compression, signature, onefs, recurse, sparse,
173 replace, verify options, ...)
175 B BaseName Start second option set any letter
178 F Number Number of filenames to follow
183 F Number Number of filenames to follow
187 - Spooling ideas taken from Volker Sauer's and other's emails:
188 > IMHO job spooling should be turned on
194 > where and 2) overrides 1) and 3) is independent.
196 Yes, this is the minimum that I think is necessary.
200 > It should be switched by job, because the job that backs up the machine
201 > with the bacula-sd on doesn't need spooling.
202 > It should be switched by schedule, because for full-backups I don't need
203 > spooling, so I can switch it off (because the network faster then the
206 True, with the exception that if you have enough disk spool space,
207 and you want to run concurrent jobs, spooling can eliminate the block
208 interleaving restore inefficiencies.
210 > And you should be able to turn it of by sd for sd-machines with low disk
211 > capacity or if you just don't need or want this feature.
214 > - definitly the possibility for multipe spool direcories
216 Having multiple directories is no problem -- having different maximum
217 sizes creates specification problems. At some point, I will probably
218 have a common SD pool of spool directories as well as a set of
219 private spool directories for each device. The first implementation
220 will be a set of private spool directories for each device since
221 managing a global pool with a bunch of threads writing into the same
222 directory is *much* more complicated and prone to error.
224 > - the ability to spool parts of a backup (not the whole client)
226 This may change in the future, but for the moment, it will spool
227 either to a job high water mark, or until the directory is full
228 (reaches max spool size or I/O error). It will then write to tape,
229 truncate the spool file, and begin spooling again.
231 > - spooling while writing to tape
233 Not within a job, but yes, if you run concurrent jobs -- each is a
234 different thread. Within a job could be a feature, but *much* later.
236 > - parallel spooling (like parallel jobs/ concurrent jobs) of clients
238 Yes, this is one of my main motivations for doing it (aside from
239 eliminating tape "shoe shine" during incremental backups.
241 > - flushing a backup that only went to disk (like amflush in amanda)
243 This will be a future feature, since spooling is different from backing
244 up to disk. The future feature will be "migration" which will move a job
245 from one backup Volume to another.
247 - New Storage specifications:
248 Passed to SD as a sort of BSR record called Storage Specification
252 MediaType -> Next MediaType
254 Device -> Next Device
255 Write Copy Resource that makes a copy of a resource.
257 Allow multiple Storage specifications
265 Allow Multiple Pool specifications (note, Pool currently
267 Allow Multiple MediaType specifications
268 Allow Multiple Device specifications
269 Perhaps keep this in a single SSR
270 Tie a Volume to a specific device by using a MediaType that
271 is contained in only one device.
272 In SD allow Device to have Multiple MediaTypes
275 - Look at www.nu2.nu/pebuilder as a helper for full windows
278 it seems to be complaining about 12:00pm which should be a valid 12
279 hour time. I changed the time to 11:59am and everything works fine.
280 Also 12:00am works fine. 0:00pm also works (which I don't think
281 should). None of the values 12:00pm - 12:59pm work for that matter.
282 Ideas from Jerry Scharf:
283 First let's point out some big pluses that bacula has for this
285 more importantly it's active. Thank you so much for that
286 even more important, it's not flaky
287 it has an open access catalog, opening many possibilities
288 it's pushing toward heterogeneous systems capability
290 I don't remember an include file directive for config files
291 (not filesets, actual config directives)
292 can you check the configs without starting the daemon?
293 some warnings about possible common mistakes
295 doing the testing and blessing of concurrent backup writes
296 this is absolutely necessary in the enterprise
297 easy user recovery GUI with full access checking
298 Macintosh file client
299 macs are an interesting niche, but I fear a server is a rathole
300 working bare iron recovery for windows
301 much better handling on running config changes
302 thinking through the logic of what happens to jobs in progress
303 the option for inc/diff backups not reset on fileset revision
304 a) use both change and inode update time against base time
305 b) do the full catalog check (expensive but accurate)
306 sizing guide (how much system is needed to back up N systems/files)
307 consultants on using bacula in building a disaster recovery system
309 or how to get at fancy things that one could do with bacula
310 logwatch code for bacula logs (or similar)
311 linux distro inclusion of bacula (brings good and bad, but necessary)
312 win2k/XP server capability (icky but you asked)
313 support for Oracle database ??
315 - Look at adding SQL server and Exchange support for Windows.
316 - Restore: Enter Filename: 'C:/Documents and Settings/Comercial/My
317 Documents/MOP/formulário de registro BELAS ARTES.doc' causes Bacula to
319 - Each DVD-RAM disk would be a volume, just like each tape is
320 a volume. It's a 4.7GB media with random access, but there's nothing about
321 it that I can see that makes it so different than a tape from bacula's
322 perspective. Why couldn't I back up to a bare floppy as a volume (ignoring
324 - Make dev->file and dev->block_num signed integers so that -1 can
325 be an invalid value which happens with BSR.
326 - Create VolAddr for disk files in place of VolFile and VolBlock. This
327 is needed to properly specify ranges.
328 - Print bsmtp output to job report so that problems will be seen.
329 - Pass the number of files to be restored to the FD for reporting
330 - Add progress of files/bytes to SD and FD.
331 - Don't continue Restore if no files selected.
332 - Print warning message if FileId > 4 billion
333 - do a "messages" before the first prompt in Console
334 - Add a date and time stamp at the beginning of every line in the
335 Job report (Volker Sauer).
336 - Client does not show busy during Estimate command.
337 - Implement Console mtx commands.
338 - Add a default DB password to MySQL.
339 GRANT all privileges ON bacula.* TO bacula@localhost IDENTIFIED BY
342 - Implement a Mount Command and an Unmount Command where
343 the users could specify a system command to be performed
344 to do the mount, after which Bacula could attempt to
345 read the device. This is for Removeable media such as a CDROM.
346 - Most likely, this mount command would be invoked explicitly
347 by the user using the current Console "mount" and "unmount"
348 commands -- the Storage Daemon would do the right thing
349 depending on the exact nature of the device.
350 - As with tape drives, when Bacula wanted a new removable
351 disk mounted, it would unmount the old one, and send a message
352 to the user, who would then use "mount" as described above
353 once he had actually inserted the disk.
354 - Implement dump/print label to UA
355 - Implement disk spooling. Two parts: 1. Spool to disk then
356 immediately to tape to speed up tape operations. 2. Spool to
357 disk only when the tape is full, then when a tape is hung move
359 - Scratch Pool where the volumes can be re-assigned to any Pool.
360 - bextract is sending everything to the log file ****FIXME****
361 - Add Progress command that periodically reports the progress of
363 - Restrict characters permitted in a Resource name, and don't permit
365 - Allow multiple Storage specifications (or multiple names on
366 a single Storage specification) in the Job record. Thus a job
367 can be backed up to a number of storage devices.
368 - Implement some way for the File daemon to contact the Director
369 to start a job or pass its DHCP obtained IP number.
370 - Implement multiple Consoles.
371 - Implement a query tape prompt/replace feature for a console
373 Two jobs ready to go, first one blocked waiting for media
374 Cancel 2nd job ("waiting execution" one)
377 - Copy console @ code to gnome2-console
378 - Make AES the only encryption algorithm see
379 http://csrc.nist.gov/CryptoToolkit/aes/). It's
380 an officially adopted standard, has survived peer
381 review, and provides keys up to 256 bits.
382 - Add ctl-c to console to stop current command and discard buffered
384 - Estimate to Tibs never returns.
385 - Think about how space could be freed up on a tape -- perhaps this
386 is a Merge or Compact feature that is needed.
387 - Modify FileSet, did not upgrade the current Increment job, but
388 waited for the next job to be upgraded.
389 - Take a careful look at SetACL http://setacl.sourceforge.net
390 - Implement a where command for the tree telling where a file
392 - Take a careful look at Level for the estimate command, maybe make
393 it a command line option.
394 - Add Volume name to "I cannot write on this volume because"
395 - Make tree walk routines like cd, ls, ... more user friendly
396 by handling spaces better.
397 - Write your PID file and chown root:wheel before drop.
398 - Make sure there is no symlink in a file before creating a
400 - Look at mktemp or mkstemp(3).
401 mktemp and mkstemp create files with predictable names too. That's
402 not the vulnerability. The vulnerability is in creating files without
403 using the O_EXCL flag, which means "only create this file if it doesn't
404 exist, including if the file is a dangling symlink."
406 It is *NOT* enough to do the equivalent of
408 if doesn't exist $filename
409 then create $filename
411 because between the test and the create another process could have
412 gotten the CPU and created the file. You must use atomic functions
413 (those that don't get interrupted by other processes) and O_EXCL is
414 the only way for this particular example.
415 - Automatically create pools, but instead of looking for what
416 in in Job records, walk through the pool resources.
417 - Check and double check tree code, why does it take so long?
418 - Add device name to "Current Volume not acceptable because ..."
419 - Make sure that Bacula rechecks the tape after the 20 min wait.
420 - Set IO_NOWAIT on Bacula TCP/IP packets.
421 - Try doing a raw partition backup and restore by mounting a
424 Yes, it would allow to highly automatic the request for new tapes. If a
425 tape is empty, bacula reads the barcodes (native or simulated), and if
426 an unused tape is found, it runs the label command with all the
427 necessary parameters.
429 By the way can bacula automatically "move" an empty/purged volume say
430 in the "short" pool to the "long" pool if this pool runs out of volume
432 - Eliminate orphaned jobs: dbcheck, normal pruning, delete job command.
433 Hm. Well, there are the remaining orphaned job records:
435 | 105 | Llioness Save | 0000-00-00 00:00:00 | B | D | 0 | 0 | f |
436 | 110 | Llioness Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
437 | 115 | Llioness Save | 2003-09-10 02:22:03 | B | I | 0 | 0 | A |
438 | 128 | Catalog Save | 2003-09-11 03:53:32 | B | I | 0 | 0 | C |
439 | 131 | Catalog Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
441 As you can see, three of the five are failures. I already deleted the
442 one restore and one other failure using the by-client option. Deciding
443 what is an orphaned job is a tricky problem though, I agree. All these
444 records have or had 0 files/ 0 bytes, except for the restore. With no
445 files, of course, I don't know of the job ever actually becomes
446 associated with a Volume.
448 (I'm not sure if this is documented anywhere -- what are the meanings of
449 all the possible JobStatus codes?)
451 Looking at my database, it appears to me as though all the "orphaned"
452 jobs fit into one of two categories:
454 1) The Job record has a StartTime but no EndTime, and the job is not
457 2) The Job record has an EndTime, indicating that it completed, but
458 it has no associated JobMedia record.
461 This does suggest an approach. If failed jobs (or jobs that, for some
462 other reason, write no files) are associated with a volume via a
463 JobMedia record, then they should be purged when the associated volume
464 is purged. I see two ways to handle jobs that are NOT associated with a
467 1) purge them automatically whenever any volume is manually purged;
469 2) add an option to the purge command to manually purge all jobs with
470 no associated volume.
472 I think Restore jobs also fall into category 2 above .... so one might
473 want to make that "The Job record has an EndTime,, but no associated
474 JobMedia record, and is not a Restore job."
475 - make "btape /tmp" work.
476 - Make sure a rescheduled job is properly reported by status.
477 - Walk through the Pool records rather than the Job records
478 in dird.c to create/update pools.
479 - What to do about "list files job=xxx".
480 - When job rescheduled, status gives is waiting for Client Rufus
481 to connect to Storage File. Dir needs to inform SD that job
483 - Make Dmsg look at global before calling subroutine.
484 - Enable trace output at runtime for Win32
485 - Available volumes for autochangers (see patrick@baanboard.com 3 Sep 03
486 and 4 Sep) scan slots.
487 - Get and test MySQL 4.0
488 - Do a complete audit of all pthreads_mutex, cond, ... to ensure that
489 any that are dynamically initialized are destroyed when no longer used.
490 - Look at how fuser works and /proc/PID/fd that is how Nic found the
491 file descriptor leak in Bacula.
492 - Implement WrapCounters in Counters.
493 - Turn on SIGHUP in dird.c and test.
494 - Use system dependent calls to get more precise info on tape errors.
495 - Add heartbeat from FD to SD if hb interval expires.
496 - Suppress read error on blank tape when doing a label.
497 - Can we dynamically change FileSets?
498 - If pool specified to label command and Label Format is specified,
499 automatically generate the Volume name.
500 - Why can't SQL do the filename sort for restore?
501 - Look at libkse (man kse) for FreeBSD threading.
502 - Look into Microsoft Volume Shadowcopy Service VSS for backing
503 up system state components (Active Directory, System Volume, ...)
504 - Add ExhautiveRestoreSearch
505 - Look at the possibility of loading only the necessary
506 data into the restore tree (i.e. do it one directory at a
507 time as the user walks through the tree).
508 - Possibly use the hash code if the user selects all for a restore command.
509 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
510 - Fix "restore all" to bypass building the tree.
511 - Prohibit backing up archive device (findlib/find_one.c:128)
512 - Implement Release Device in the Job resource to unmount a drive.
513 - Implement Acquire Device in the Job resource to mount a drive,
514 be sure this works with admin jobs so that the user can get
515 prompted to insert the correct tape. Possibly some way to say to
516 run the job but don't save the files.
517 - Implement FileOptions (see end of this document)
518 - Make things like list where a file is saved case independent for
521 - Bacula needs to propagate SD errors.
522 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
523 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
524 /dev/nst0. ERR=Input/output error.
525 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
526 Last block=5162 Current block=5164.
527 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
529 - Use autochanger to handle multiple devices.
530 - Add SuSE install doc to list.
531 - Check and rechedk "Invalid block number"
532 - Make bextract release the drive properly between tapes
533 so that an autochanger can be made to work.
534 - User wants to NOT backup up certain big files (email files).
535 - Maybe remove multiple simultaneous devices code in SD.
536 - On Windows with very long path names, it may be impossible to create
537 a file (and thus restore it) because the total length is too long.
538 We must cd into the directory then create the file without the
540 - lstat() is not going to work on Win32 for testing date.
541 - Implement a Recycle command
542 - Add client name to cram-md5 challenge so Director can immediately
543 verify if it is the correct client.
544 - Add JobLevel in FD status (but make sure it is defined).
545 - Audit all UA commands to ensure that we always prompt where possible.
546 - Check Jmsg in bnet, may not work, must dup bsock.
547 - Suppress Job Name in Jmsg for console
548 - Create Pools that are referenced in a Run statement at startup if possible.
549 - Use runbeforejob to unload, then reload a volume previously used,
550 then the next job run gets an error reading the drive.
551 - Make bootstrap filename unique.
552 - Test a second language e.g. french.
553 - Start working on Base jobs.
554 - Make "make binary-release" work from any directory.
555 - Implement UnsavedFiles DB record.
556 - Implement argc/argv for daemon command line scanning using table driven
558 - Implement table driven single argc/argv scanner to pickup all arguments.
559 Much like xxx_conf.c scan table.
560 keyword, handler(store_routine), store_address, code, flags, default.
561 - From Phil Stracchino:
562 It would probably be a per-client option, and would be called
563 something like, say, "Automatically purge obsoleted jobs". What it
564 would do is, when you successfully complete a Differential backup of a
565 client, it would automatically purge all Incremental backups for that
566 client that are rendered redundant by that Differential. Likewise,
567 when a Full backup on a client completed, it would automatically purge
568 all Differential and Incremental jobs obsoleted by that Full backup.
569 This would let people minimize the number of tapes they're keeping on
570 hand without having to master the art of retention times.
571 - Implement a M_SECURITY message class.
572 - When doing a Backup send all attributes back to the Director, who
573 would then figure out what files have been deleted.
574 - Currently in mount.c:236 the SD simply creates a Volume. It should have
575 explicit permission to do so. It should also mark the tape in error
576 if there is an error.
577 - Make sure all restore counters are working correctly in the FD.
578 - SD Bytes Read is wrong.
579 - Look at ALL higher level routines that call block.c to be sure
580 they don't expect something in errmsg.
581 - Investigate doing RAW backup of Win32 partition.
582 - Add thread specific data to hold the jcr -- send error messages from
583 low level routines by accessing it and using Jmsg().
584 - Cancel waiting for Client connect in SD if FD goes away.
586 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
587 - Implement timeout in response() when it should come quickly.
588 - Implement console @echo command.
589 - Implement a Slot priority (loaded/not loaded).
590 - Implement "vacation" Incremental only saves.
591 - Implement single pane restore (much like the Gftp panes).
592 - Implement Automatic Mount even in operator wait.
593 - Implement create "FileSet"?
594 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
595 - Implement "scratch pool" where tapes are defined and can be
596 taken by any pool that needs them.
597 - Implement restore "current system", but take all files without
598 doing selection tree -- so that jobs without File records can
600 - Add prefixlinks to where or not where absolute links to FD.
601 - Issue message to mount a new tape before the rewind.
602 - Simplified client job initiation for portables.
603 - If SD cannot open a drive, make it periodically retry.
604 - Add more of the config info to the tape label.
606 - If tape is marked read-only, then try opening it read-only rather than
607 failing, and remember that it cannot be written.
608 - Refine SD waiting output:
609 Device is being positioned
610 > Device is being positioned for append
611 > Device is being positioned to file x
613 - Figure out some way to estimate output size and to avoid splitting
614 a backup across two Volumes -- this could be useful for writing CDROMs
615 where you really prefer not to have it split -- not serious.
616 - Have SD compute MD5 or SHA1 and compare to what FD computes.
617 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
618 actual data on the Volume and compare it.
619 - Implement Bacula plugins -- design API
620 - Make bcopy read through bad tape records.
621 - Program files (i.e. execute a program to read/write files).
622 Pass read date of last backup, size of file last time.
623 - Add Signature type to File DB record.
624 - CD into subdirectory when open()ing files for backup to
625 speed up things. Test with testfind().
626 - Priority job to go to top of list.
627 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
628 - Implement some way for the Console to dynamically create a job.
629 - Restore to a particular time -- e.g. before date, after date.
630 - Solaris -I on tar for include list
631 - Need a verbose mode in restore, perhaps to bsr.
632 - bscan without -v is too quiet -- perhaps show jobs.
633 - Add code to reject whole blocks if not wanted on restore.
634 - Check if we can increase Bacula FD priorty in Win2000
635 - Make sure the MaxVolFiles is fully implemented in SD
636 - Check if both CatalogFiles and UseCatalog are set to SD.
637 - Figure out how to do a bare metal Windows restore
638 - Possibly add email to Watchdog if drive is unmounted too
639 long and a job is waiting on the drive.
640 - Restore program that errs in SD due to no tape, reports
641 OK incorrectly in output.
642 - After unmount, if restore job started, ask to mount.
643 - Convert all %x substitution variables, which are hard to remember
644 and read to %(variable-name). Idea from TMDA.
645 - Remove NextId for SQLite. Optimize.
646 - Move all SQL statements into a single location.
647 - Add UA rc and history files.
648 - put termcap (used by console) in ./configure and
649 allow -with-termcap-dir.
650 - Fix Autoprune for Volumes to respect need for full save.
651 - Fix Win32 config file definition name on /install
652 - Compare tape to Client files (attributes, or attributes and data)
653 - Make all database Ids 64 bit.
654 - Write an applet for Linux.
655 - Allow console commands to detach or run in background.
656 - Fix status delay on storage daemon during rewind.
657 - Add SD message variables to control operator wait time
658 - Maximum Operator Wait
659 - Minimum Message Interval
660 - Maximum Message Interval
661 - Send Operator message when cannot read tape label.
662 - Verify level=Volume (scan only), level=Data (compare of data to file).
663 Verify level=Catalog, level=InitCatalog
665 - Add keyword search to show command in Console.
666 - Events : tape has more than xxx bytes.
667 - Complete code in Bacula Resources -- this will permit
668 reading a new config file at any time.
669 - Handle ctl-c in Console
670 - Implement script driven addition of File daemon to config files.
671 - Think about how to make Bacula work better with File (non-tape) archives.
672 - Write Unix emulator for Windows.
673 - Put memory utilization in Status output of each daemon
674 if full status requested or if some level of debug on.
675 - Make database type selectable by .conf files i.e. at runtime
676 - Set flag for uname -a. Add to Volume label.
677 - Implement throttled work queue.
678 - Restore files modified after date
679 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
680 - Implement Restore FileSet=
681 - Create a protocol.h and protocol.c where all protocol messages
683 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
684 - Timout a job or terminate if link goes down, or reopen link and query.
685 - Concept of precious tapes (cannot be reused).
686 - Make bcopy copy with a single tape drive.
687 - Permit changing ownership during restore.
690 > My suggestion: Add a feature on the systray menu-icon menu to request
691 > an immediate backup now. This would be useful for laptop users who may
692 > not be on the network when the regular scheduled backup is run.
694 > My wife's suggestion: Add a setting to the win32 client to allow it to
695 > shut down the machine after backup is complete (after, of course,
696 > displaying a "System will shut down in one minute, click here to cancel"
697 > warning dialog). This would be useful for sites that want user
698 > woorkstations to be shut down overnight to save power.
701 - From Terry Manderson <terry@apnic.net>
702 jobdefs { # new structure
703 name = "monthlyUnixBoxen"
715 jobdefs = "genericUnixSet"
721 jobdefs = "genericUnixSet"
727 jobdefs = "UnixDevelBoxSet"
732 - Autolabel should be specified by DIR instead of SD.
735 - AutoScan (check checksum of tape)
736 - Format command = "format /dev/nst0"
740 - Seek resolution (usually corresponds to buffer size)
741 - EODErrorCode=ENOSPC or code
742 - Partial Read error code
743 - Partial write error code
744 - Nonformatted read error
745 - Nonformatted write error
746 - WriteProtected error
750 - IgnoreCloseErrors=yes
760 - FD sends unsaved file list to Director at end of job (see
762 - File daemon should build list of files skipped, and then
763 at end of save retry and report any errors.
764 - Write a Storage daemon that uses pipes and
765 standard Unix programs to write to the tape.
767 - Need something that monitors the JCR queue and
768 times out jobs by asking the deamons where they are.
769 - Enhance Jmsg code to permit buffering and saving to disk.
770 - device driver = "xxxx" for drives.
772 - Ensure that /dev/null works
773 - Need report class for messages. Perhaps
774 report resource where report=group of messages
775 - enhance scan_attrib and rename scan_jobtype, and
776 fill in code for "since" option
777 - Director needs a time after which the report status is sent
778 anyway -- or better yet, a retry time for the job.
779 - Don't reschedule a job if previous incarnation is still running.
780 - Some way to automatically backup everything is needed????
781 - Need a structure for pending actions:
783 - termination status (part of buffered msgs?)
785 Read, Write, Clean, Delete
786 - Login to Bacula; Bacula users with different permissions:
787 owner, group, user, quotas
788 - Store info on each file system type (probably in the job header on tape.
789 This could be the output of df; or perhaps some sort of /etc/mtab record.
792 - Design at hierarchial storage for Bacula. Migration and Clone.
793 - Implement FSM (File System Modules).
794 - Audit M_ error codes to ensure they are correct and consistent.
795 - Add variable break characters to lex analyzer.
796 Either a bit mask or a string of chars so that
797 the caller can change the break characters.
798 - Make a single T_BREAK to replace T_COMMA, etc.
799 - Ensure that File daemon and Storage daemon can
800 continue a save if the Director goes down (this
801 is NOT currently the case). Must detect socket error,
802 buffer messages for later.
803 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
804 - Add ability to backup to two Storage devices (two SD sessions) at
805 the same time -- e.g. onsite, offsite.
806 - Add the ability to consolidate old backup sets (basically do a restore
807 to tape and appropriately update the catalog). Compress Volume sets.
808 Might need to spool via file is only one drive is available.
809 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
810 someway to do so with every volume that has less than x% valid
814 Migration: Move a backup from one Volume to another
815 Clone: Copy a backup -- two Volumes
817 Bacula Migration is based on Jobs (apparently Networker is file by file).
819 Migration triggered by:
823 Highwater mark (keep total size)
828 ======================================================
830 It is somewhat like a Full save becomes an incremental since
831 the Base job (or jobs) plus other non-base files.
833 - A Base backup is same as Full backup, just different type.
834 - New BaseFiles table that contains:
836 BaseJobId - Base JobId referenced for this FileId (needed ???)
837 JobId - JobId currently running
838 FileId - File not backed up, exists in Base Job
839 FileIndex - FileIndex from Base Job.
840 i.e. for each base file that exists but is not saved because
841 it has not changed, the File daemon sends the JobId, BaseId,
842 FileId, FileIndex back to the Director who creates the DB entry.
843 - To initiate a Base save, the Director sends the FD
844 the FileId, and full filename for each file in the Base.
845 - When the FD finds a Base file, he requests the Director to
846 send him the full File entry (stat packet plus MD5), or
847 conversely, the FD sends it to the Director and the Director
848 says yes or no. This can be quite rapid if the FileId is kept
849 by the FD for each Base Filename.
850 - It is probably better to have the comparison done by the FD
851 despite the fact that the File entry must be sent across the
853 - An alternative would be to send the FD the whole File entry
854 from the start. The disadvantage is that it requires a lot of
855 space. The advantage is that it requires less communications
857 - The Job record must be updated to indicate that one or more
859 - At end of Job, FD returns:
860 1. Count of base files/bytes not written to tape (i.e. matches)
861 2. Count of base file that were saved i.e. had changed.
862 - No tape record would be written for a Base file that matches, in the
863 same way that no tape record is written for Incremental jobs where
864 the file is not saved because it is unchanged.
865 - On a restore, all the Base file records must explicitly be
866 found from the BaseFile tape. I.e. for each Full save that is marked
867 to have one or more Base Jobs, search the BaseFile for all occurrences
869 - An optimization might be to make the BaseFile have:
875 This would avoid the need to explicitly fetch each File record for
876 the Base job. The Base Job record will be fetched to get the
877 VolSessionId and VolSessionTime.
878 =========================================================
882 =============================================================
884 Request For Comments For File Backup Options
887 Subject: File Backup Options
890 A few days ago, a Bacula user who is backing up to file volumes and
891 using compression asked if it was possible to suppress compressing
892 all .gz files since it was a waste of CPU time. Although Bacula
893 currently permits using different options (compression, ...) on
894 a directory by directory basis, it cannot do it on a file by
895 file basis, which is clearly what was desired.
897 Proposed Implementation:
898 To solve this problem, I propose the following:
900 - Add a new Director resource type called Options.
902 - The Options resource will have records for all
903 options that can currently be specified on the Include record
904 (in a FileSet). Examples below.
906 - The Options resource will permit an exclude option as well
907 as a number of additional options.
909 - The heart of the Options resource is the ability to
910 supply any number of Match records which specify POSIX
911 regular expressions. These Match regular expressions are
912 applied to the fully qualified filename (path and all). If
913 one matches, then the Options will be used.
915 - When an Match specification matches an included file, the
916 options specified in the Options resource will override
917 the default options specified on the Include record.
919 - Include records will be modified to permit referencing one or
920 more Options resources. The Options will be used
921 in the order listed on the Include record and the first
922 one that matches will be applied.
924 - Options (or specifications) currently supplied on the Include
925 record will be deprecated (i.e. removed in a later version a
926 year or so from now).
928 - The Exclude record will be deprecated as the same functionality
929 can be obtained by using an Exclude = yes in the Options.
932 The following records can appear in the Options resource. An
933 asterisk preceding the name indicates a feature not currently
936 - Regexp "xxx" - Match regular expression
937 - Wild "xxx" - Do a wild card match
940 - Compression= (GZIP, ...)
941 - Signature= (MD5, SHA1, ...)
943 - OneFs= (yes/no) - remain on one filesystem
944 - Recurse= (yes/no) - recurse into subdirectories
945 - Sparse= (yes/no) - do sparse file backup
946 - *Exclude= (yes/no) - exclude file from being saved
947 - *Reader= (filename) - external read (backup) program
948 - *Plugin= (filename) - read/write plugin module
950 - Include= (yes/no) - Include the file matched no additional
951 patterns are applied.
954 - verify= (ipnougsamc5) - verify options
957 - replace= (always/ifnewer/ifolder/never) - replace options currently
959 - *Writer= (filename) - external write (restore) program
963 Currently options specifying compression, MD5 signatures, recursion,
964 ... of a FileSet are supplied on the Include record. These will now
965 all be collected into a Options resource, which will be
966 specified in the Include in place of the options. Multiple Options
967 may be specified. Since the Options may contain regular expressions
968 that are applied to the full filename, this will give the ability
969 to specify backup options on a file by file basis to whatever level
978 Include = compression=GZIP signature=MD5 {
990 Wild = /*.?*/ # matches all files.
995 That's a lot more to do the same thing, but it gives the ability to
996 apply options on a file by file basis. For example, suppose you
997 want to compress all files but not any file with extensions .gz or .Z.
998 In that case, you will need to group two sets of options using
999 the Options resource. Files may be anywhere except in an
1000 option set??? All OptionSets apply to all files in the order
1001 the OptionSets were specified. To have files included with
1002 different option sets without using wild-cards, use two or more
1003 Includes -- each one is handled in turn using only the files and
1004 optionsets specified in the include.
1011 # Note multiple Matches are ORed
1012 Wild = "*.gz" # matches .gz files
1013 Wild = "*.Z" # matches .Z files
1018 Wild = "*.?*" # matches all files
1024 Now, since the no Compression option is specified in the
1025 first group of Options, *.gz or *.Z file will have an MD5 signature computed,
1026 but will not be compressed. For all other files, the *.gz *.Z will not
1027 match, so the second group of options will be used which will include GZIP
1031 - Is it necessary to provide some means of ANDing regular expressions
1032 and negation? (not currently planned)
1034 e.g. Wild = /*.gz/ && !/big.gz/
1036 - I see that Networker has a "null" module which, if specified, does not
1037 backup the file, but does make an record of the file in the catalog
1038 so that the catalog will reflect an exact picture of the filesystem.
1039 The result is that the file can be "seen" when "browsing" the save
1040 sets, but it cannot be restored.
1042 Is this really useful? Should it be implemented in Bacula?
1045 After implementing the above, the user will be able to specify
1046 on a file by file basis (using regular expressions) what options are
1047 applied for the backup.
1050 =============================================
1052 ==========================================================
1054 For each Incremental job that is run, there may be files that
1055 were found but not saved because they were locked (this applies
1056 only to Windows). Such a system could send back to the Director
1057 a list of Unsaved files.
1059 - New UnSavedFiles table that contains:
1063 - Then in the next Incremental job, the list of Unsaved Files will be
1064 feed to the FD, who will ensure that they are explicitly chosen even
1065 if standard date/time check would not have selected them.
1066 =============================================================
1069 Done: (see kernsdone for more)
1072 - John's Full save failed with 1.32c FD and 1.31 Dir no FD status,
1073 and no error message.
1074 - Add fd and st as Console keywords.
1075 - Recycling volume with a Slot requires an operator intervention:
1076 rufus-dir: Start Backup JobId 18, Job=kernsave.2003-11-01_21.23.52
1077 rufus-dir: Pruned 1 Job on Volume Vol01 from catalog.
1078 rufus-dir: There are no Jobs associated with Volume Vol01. Marking it purged.
1079 rufus-dir: Recycled volume "Vol01"
1080 rufus-sd: Please mount Volume "Vol01" on Storage Device "DDS-4" for Job kernsave.2003-11-01_21.23.52
1081 Use "mount" command to release Job.
1082 - Implement Dan's bacula script (email of 26 Oct).
1083 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
1084 - Edit the Client/Storage name into authentication failure messages.
1085 - Fix packet too big problem. This is most likely a Windows TCP stack
1087 - Implement ClientRunBeforeJob and ClientRunAfterJob.
1088 - Implement forward spacing block/file: position_device(bsr) --
1089 just before read_block_from_device();
1092 - Change console to bconsole.
1093 - Change smtp to bsmtp.
1094 - Fix time difference problem between Bacula and Client
1095 so that everything is in GMT.
1096 - Fix TimeZone problem!
1097 - Mount a tape that is not right for the job (wrong # files on tape)
1098 Bacula asks for another tape, fix problems with first tape and
1099 say "mount". All works OK, but status shows:
1100 Device /dev/nst0 open but no Bacula volume is mounted.
1101 Total Bytes=1,153,820,213 Blocks=17,888 Bytes/block=64,502
1102 Positioned at File=9 Block=3,951
1103 Full Backup job Rufus.2003-10-26_16.45.31 using Volume "DLT-24Oct03" on device /dev/nst0
1104 Files=21,003 Bytes=253,954,408 Bytes/sec=2,919,016
1105 FDReadSeqNo=192,134 in_msg=129830 out_msg=5 fd=7
1106 - Upgrade to cygwin 1.5
1107 - Optimize fsf not to read.
1108 - Use ioctl() fsf if it exists. Figure out where we are from
1109 the mt_status command. Use slow fsf only if other does not work.
1110 - Enhance "update slots" to include a "scan" feature
1111 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
1112 - Allow a slot or range of slots on the label barcodes command.
1113 - Finish implementation of Verify=DiskToCatalog
1114 - Make sure that Volumes are recycled based on "Least recently used"
1115 rather than lowest MediaId.
1116 - Add flag to write only one EOF mark on the tape.
1117 - Implement autochanger testing in btape "test" command.
1118 - Implement lmark to list everyfile marked.
1119 - Make mark/unmark report how many files marked/unmarked.
1120 - Keep last 5 or 10 completed jobs and show them in a similar list.
1121 - Make a Running Jobs: output similar to current Scheduled Jobs:
1122 - Change "create_media_record in bscan to use Archive instead of Full.
1123 - Have some way to estimate the restore size or have it printed.
1124 - Volume problems occurs if you have valid volume, written, then it is
1125 truncated. You get 12-Nov-2003 11:48 rufus-sd: kernsave.2003-11-12_11.48.09 Warning: mount.c:228 Volume on /tmp is not a Bacula labeled Volume, because:
1126 block.c:640 Read zero bytes on device /tmp.
1127 - Make sure that 64 bit I/O packets are used on Cygwin.
1128 - Add to supported autochangers
1130 Auto-Changer : QUALSTAR TLS-4210
1131 Manufufactur : Qualstar
1132 Tapes : 12 (AIT1: 36GB, AIT2: 50GB all uncompressed)
1133 Drives : 2xAIT2 (installed in the Qualstar: SONY SDX-500C AIT2)
1134 - Document estimate command in tree.
1135 - Document lsmark command in tree.
1136 - Setup a standard job that builds a bootstrap file and saves
1137 it with the catalog database.
1138 - See if a restore job can add a file to the tape (prohibit this).
1139 - Restrict characters permitted in a name.
1140 - In restore, provide option for limiting to a particular Pool.
1141 - In restore, list FileSets that only have different base names --
1142 i.e. any FileSet with the same name should be treated as the same.
1143 - Make Scheduler sort jobs by StartTime, Priority.
1144 - Make sure smtp and any other useful program is executable by the world
1145 in case Bacula is not running as root.
1146 - Look at Dan's field width problems in PostgreSQL.
1147 - Look at effect of removing GROUP BYs.
1148 - In restore take all filesets with same base name.
1149 - From Alan Brown <ajb2@mssl.ucl.ac.uk>
1150 BTW, there's a make install bug in 1.33 - with --enable-gnome,
1151 gnome-console is built, but the binary and .conf are not being installed.
1152 - Permit Bacula and apcupsd donations (not done for apcupsd).
1153 - Fix Ctl-C crashing the Console (readline?).
1154 - Look at code in recycle_oldes_purged_volume() recycle.c. Why not
1155 let SQL do ORDER BY LastWritten ASC?
1156 - Look at find_next_volume() algorithm. Currently, it selects:
1157 +---------+------------+---------------------+-----------+
1158 | MediaId | VolumeName | LastWritten | VolBytes |
1159 +---------+------------+---------------------+-----------+
1160 | 3 | Test13 | 0000-00-00 00:00:00 | 1 |
1161 | 4 | Test14 | 0000-00-00 00:00:00 | 1 |
1162 | 1 | test11 | 2003-12-03 18:39:55 | 4,004,926 |
1163 | 2 | test12 | 2004-01-04 15:25:56 | 2,078,691 |
1164 +---------+------------+---------------------+-----------+
1165 but perhaps it should fill already used Volumes first, and use
1166 Append volumes before Purged, or Recycled, ...
1167 - Possibly remove the "|| ap == NULL" on lines 123 and 207 of lib/var.c,
1168 which creates compile problems on alpha systems.
1169 var.c:123: no match for `va_list & == long int'
1170 - Check "restore" 3 (JobId), then it asks for Storage resource. Does
1171 it verify that the correct volume is chosen?
1172 - Make Bacula "poll a drive".
1173 - Notes for final checking of Nic's code:
1174 Could I get you to double check the switch () statements in the
1175 job_check_maxwaittime and job_check_maxruntime functions in
1177 - Define week of year for scheduler. W01, W02, ...
1178 Week 01 of a year is per definition the first week that has the
1179 Thursday in this year, which is equivalent to the week that contains the
1180 fourth day of January. In other words, the first week of a new year is
1181 the week that has the majority of its days in the new year. Week 01
1182 might also contain days from the previous year and the week before week
1183 01 of a year is the last week (52 or 53) of the previous year even if it
1184 contains days from the new year. A week starts with Monday (day 1) and
1185 ends with Sunday (day 7). For example, the first week of the year 1997
1186 lasts from 1996-12-30 to 1997-01-05 and can be written in standard
1189 The week notation can also be extended by a number indicating the day
1190 of the week. For example, the day 1996-12-31, which is the Tuesday (day
1191 2) of the first week of 1997, can also be written as
1192 1997-W01-2 or 1997W012
1193 - Either restrict the characters in a name, or fix the problem
1194 emailing with names containing / (smtp command line breaks).
1195 - Implement .consolerc for Console
1196 - Implement scan: for every slot it finds, zero the slot of
1197 Volume other volume having that slot.
1198 - Make restore job check if all the files are actually restored.
1199 - Look at 2Gb limit for SQLite.
1200 - Fix get_storage_from_media_type (ua_restore) to use command line
1202 - Don't print "Warning: Wrong Volume mounted ..." if mounting second volume.
1203 - Write a mini-readline with history and editing.
1204 - Take a careful look a the Basic recycling algorithm. When Bacula
1205 chooses, the order should be:
1207 - Look for Recycle or Purged
1210 Instead of using lowest media Id, find the least recently used
1213 When the tape is mounted and Bacula requests the status
1214 - Do everything possible to use it.
1216 Define a "available" status, which is the currently mounted
1217 Volume and all volumes that are currently in the autochanger.
1218 - Is a pool specification really needed for a restore? Yes, and
1219 you may want to exclude archive Pools.
1220 - Implement a PostgreSQL driver.
1221 - Fix restore to list errors if Invalid block found, and if # files
1222 restored does not match # expected.
1223 - Something is not right in last block of fill command.
1224 - Add FileSet to command line arguments for restore.
1225 - Enhance time and size scanning routines.
1226 - Add Console usr permissions -- do by adding filters for
1227 jobs, clients, storage, ...
1228 - Put max network buffer size on a directive.
1229 - Why does "mark cygwin" take so long!!!!!!!!
1230 - Implement alist processing for ACLs from Console.
1231 - When a file is set for restore, walk back up the chain of
1232 directories, setting them to be restored.
1233 - Figure out a way to set restore on a directory without recursively
1234 decending. (recurse off?).
1235 - Fix restore to only pull in last Differential and later Incrementals.
1236 - Implement 3 Pools for a Job:
1239 Full Backup Pool = xxx
1240 Incremental Backup Pool = yyy
1241 Differential Backup Pool = zzz
1243 - Look at ASSERT() at 384 src/lib/bnet.c