4 Documentation to do: (any release a little bit at a time)
5 - Document running a test version.
6 - Document query file format.
7 - Document static linking
8 - Document problems with Verify and pruning.
9 - Document how to use multiple databases.
10 - For FreeBSD typical /dev/nrsa0 and for mtx /dev/pass1
11 - VXA drives have a "cleaning required"
12 indicator, but Exabyte recommends preventive cleaning after every 75
14 - Lookup HP cleaning recommendations.
15 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
16 - Create a man page for each binary (Debian package requirement).
18 Testing to do: (painful)
19 - that ALL console command line options work and are always implemented
20 - blocksize recognition code.
21 - Test if rewind at end of tape waits for tape to rewind.
23 - Test not zeroing Autochanger slot when it is wrong.
24 - Figure out how to use ssh or stunnel to protect Bacula communications.
26 For 1.33 Testing/Documentation:
27 - bextract is sending everything to the log file ****FIXME****
28 - Document to start higher priorty jobs before lower ones.
29 - suppress "Do not forget to mount the drive!!!" if error
30 - Document new records in Director. SDAddress SDDeviceName, SDPassword.
31 FDPassword, FDAddress, DBAddress, DBPort, DBPassword.
32 - Document new Include/Exclude ...
33 - Add test of exclusion, test multiple Include {} statements.
34 - Add counter variable test.
35 - Document ln -sf /usr/lib/libncurses.so /usr/lib/libtermcap.so
36 and install the esound-dev package for compiling Console on
40 - Limit the number of block checksum/header BB01, ... errors printed.
41 - Add VerifyJob to "run" summary (yes/mod/no) prompt.
42 - Add device name to "Current Volume not acceptable because ..."
43 - Make sure that Bacula rechecks the tape after the 20 min wait.
44 - Set IO_NOWAIT on Bacula TCP/IP packets.
45 - Try doing a raw partition backup and restore by mounting a
47 - Implement Verify=DiskToCatalog
48 - Implement a RunAfterFailedJob
49 - Report CVS problems to SourceForge.
50 - Implement .consolerc for Console
51 - Is it really important to make Job name the same to find the
52 Full backup to avoid promoting an Incremental job?
53 - Start label, then run job when tape labeled, it should broadcast.
54 - Zap illegal characters in job name for mail files (e.g. /).
56 Yes, it would allow to highly automatic the request for new tapes. If a
57 tape is empty, bacula reads the barcodes (native or simulated), and if
58 an unused tape is found, it runs the label command with all the
61 By the way can bacula automatically "move" an empry/purged volume say
62 in the "short" pool to the "long" pool if this pool runs out of volume
64 - Implement a move Volume from one pool to another.
65 - Either restrict the characters in a name, or fix the problem
66 emailing with names containing / (smtp command line breaks).
67 - Eliminate ua_retention.c (retentioncmd) if possible.
68 - Eliminate orphaned jobs: dbcheck, normal pruning, delete job command.
69 Hm. Well, there are the remaining orphaned job records:
71 | 105 | Llioness Save | 0000-00-00 00:00:00 | B | D | 0 | 0 | f |
72 | 110 | Llioness Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
73 | 115 | Llioness Save | 2003-09-10 02:22:03 | B | I | 0 | 0 | A |
74 | 128 | Catalog Save | 2003-09-11 03:53:32 | B | I | 0 | 0 | C |
75 | 131 | Catalog Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
77 As you can see, three of the five are failures. I already deleted the
78 one restore and one other failure using the by-client option. Deciding
79 what is an orphaned job is a tricky problem though, I agree. All these
80 records have or had 0 files/ 0 bytes, except for the restore. With no
81 files, of course, I don't know of the job ever actually becomes
82 associated with a Volume.
84 (I'm not sure if this is documented anywhere -- what are the meanings of
85 all the possible JobStatus codes?)
87 Looking at my database, it appears to me as though all the "orphaned"
88 jobs fit into one of two categories:
90 1) The Job record has a StartTime but no EndTime, and the job is not
93 2) The Job record has an EndTime, indicating that it completed, but
94 it has no associated JobMedia record.
97 This does suggest an approach. If failed jobs (or jobs that, for some
98 other reason, write no files) are associated with a volume via a
99 JobMedia record, then they should be purged when the associated volume
100 is purged. I see two ways to handle jobs that are NOT associated with a
103 1) purge them automatically whenever any volume is manually purged;
105 2) add an option to the purge command to manually purge all jobs with
106 no associated volume.
108 I think Restore jobs also fall into category 2 above .... so one might
109 want to make that "The Job record has an EndTime,, but no associated
110 JobMedia record, and is not a Restore job."
111 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
112 which would take precidence over the Catalog "JobRetention".
113 - Implement Label Format in Add and Label console commands.
114 - make "btape /tmp" work.
115 - Make sure a rescheduled job is properly reported by status.
116 - Walk through the Pool records rather than the Job records
117 in dird.c to create/update pools.
118 - What to do about "list files job=xxx".
119 - Implement delete Job.
120 - Document need to put LabelFormat in quotes.
121 - Implement scan: for every slot it finds, zero the slot of
122 Volume other volume having that slot.
123 - When job rescheduled, status gives is waiting for Client Rufus
124 to connect to Storage File. Dir needs to inform SD that job
126 - Fix get_storage_from_media_type (ua_restore) to use command line
128 - Enhance "update slots" to include a "scan" feature
129 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
130 - Allow a slot or range of slots on the label barcodes command.
131 - Don't print "Warning: Wrong Volume mounted ..." if mounting second volume.
132 - Make Dmsg look at global before calling subroutine.
133 - Enable trace output at runtime for Win32
134 - Make sure that Volumes are recycled based on "Least recently used"
135 rather than lowest MediaId.
136 - Available volumes for autochangers (see patrick@baanboard.com 3 Sep 03
137 and 4 Sep) scan slots.
138 - Upgrade to cygwin 1.5
140 - Get and test MySQL 4.0
141 - Do a complete audit of all pthreads_mutex, cond, ... to ensure that
142 any that are dynamically initialized are destroyed when no longer used.
143 - Write a mini-readline with history and editing.
144 - Look at how fuser works and /proc/PID/fd that is how Nic found the
145 file descriptor leak in Bacula.
146 - Implement WrapCounters in Counters.
147 - Turn on SIGHUP in dird.c and test.
148 - Use system dependent calls to get more precise info on tape errors.
149 - Add heartbeat from FD to SD if hb interval expires.
150 - Suppress read error on blank tape when doing a label.
151 - Can we dynamically change FileSets?
152 - If pool specified to label command and Label Format is specified,
153 automatically generate the Volume name.
154 - Take a careful look a the Basic recycling algorithm. When Bacula
155 chooses, the order should be:
157 - Look for Recycle or Purged
160 Instead of using lowest media Id, find the least recently used
163 When the tape is mounted and Bacula requests the status
164 - Do everything possible to use it.
166 Define a "available" status, which is the currently mounted
167 Volume and all volumes that are currently in the autochanger.
169 - Why can't SQL do the filename sort for restore?
170 - Is a pool specification really needed for a restore? Yes, and
171 you may want to exclude archive Pools.
172 - Look at libkse (man kse) for FreeBSD threading.
173 - Look into Microsoft Volume Shadowcopy Service VSS for backing
174 up system state components (Active Directory, System Volume, ...)
175 - Add ExhautiveRestoreSearch
176 - Look at the possibility of loading only the necessary
177 data into the restore tree (i.e. do it one directory at a
178 time as the user walks through the tree).
179 - Possibly use the hash code if the user selects all for a restore command.
180 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
181 - Implement some way for the File daemon to contact the Director
182 to start a job or pass its DHCP obtained IP number.
183 - Implement multiple Consoles.
184 - Add Console usr permissions.
185 - Fix "restore all" to bypass building the tree.
186 - Fix restore to list errors if Invalid block found, and if # files
187 restored does not match # expected.
188 - Prohibit backing up archive device (findlib/find_one.c:128)
189 - Implement Release Device in the Job resource to unmount a drive.
190 - Implement Acquire Device in the Job resource to mount a drive,
191 be sure this works with admin jobs so that the user can get
192 prompted to insert the correct tape. Possibly some way to say to
193 run the job but don't save the files.
194 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
195 - Implement FileOptions (see end of this document)
196 - Make things like list where a file is saved case independent for
198 - Edit the Client/Storage name into authentication failure messages.
199 - Implement job in VerifyToCatalog
201 - Implement a PostgreSQL driver.
202 - Bacula needs to propagate SD errors.
203 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
204 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
205 /dev/nst0. ERR=Input/output error.
206 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
207 Last block=5162 Current block=5164.
208 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
210 - Use autochanger to handle multiple devices.
211 - Fix packet too big problem. This is most likely a Windows TCP stack
213 - Add SuSE install doc to list.
214 - Check and rechedk "Invalid block number"
215 - Make bextract release the drive properly between tapes
216 so that an autochanger can be made to work.
217 - User wants to NOT backup up certain big files (email files).
218 - Maybe remove multiple simultaneous devices code in SD.
219 - On Windows with very long path names, it may be impossible to create
220 a file (and thus restore it) because the total length is too long.
221 We must cd into the directory then create the file without the
223 - lstat() is not going to work on Win32 for testing date.
224 - Something is not right in last block of fill command.
225 - Implement a Recycle command
226 - Add FileSet to command line arguments for restore.
227 - Allow multiple Storage specifications (or multiple names on
228 a single Storage specification) in the Job record. Thus a job
229 can be backed up to a number of storage devices.
230 - Add client name to cram-md5 challenge so Director can immediately
231 verify if it is the correct client.
232 - Implement ClientRunBeforeJob and ClientRunAfterJob.
233 - Add JobLevel in FD status (but make sure it is defined).
234 - Audit all UA commands to ensure that we always prompt where possible.
235 - Restrict characters permitted in a Resource name, and don't permit
237 - Check Jmsg in bnet, may not work, must dup bsock.
238 - Suppress Job Name in Jmsg for console
239 - Create Pools that are referenced in a Run statement at startup if possible.
240 - Use runbeforejob to unload, then reload a volume previously used,
241 then the next job run gets an error reading the drive.
242 - Make bootstrap filename unique.
243 - Test a second language e.g. french.
244 - Start working on Base jobs.
245 - Make "make binary-release" work from any directory.
246 - Implement UnsavedFiles DB record.
247 - Implement argc/argv for daemon command line scanning using table driven
249 - Implement table driven single argc/argv scanner to pickup all arguments.
250 Much like xxx_conf.c scan table.
251 keyword, handler(store_routine), store_address, code, flags, default.
252 - From Phil Stracchino:
253 It would probably be a per-client option, and would be called
254 something like, say, "Automatically purge obsoleted jobs". What it
255 would do is, when you successfully complete a Differential backup of a
256 client, it would automatically purge all Incremental backups for that
257 client that are rendered redundant by that Differential. Likewise,
258 when a Full backup on a client completed, it would automatically purge
259 all Differential and Incremental jobs obsoleted by that Full backup.
260 This would let people minimize the number of tapes they're keeping on
261 hand without having to master the art of retention times.
262 - Implement new serialize subroutines
263 send(socket, "string", &Vol, "uint32", &i, NULL)
264 - Scratch Pool where the volumes can be re-assigned to any Pool.
265 - Implement a M_SECURITY message class.
266 - Implement forward spacing block/file: position_device(bsr) --
267 just before read_block_from_device();
268 - When doing a Backup send all attributes back to the Director, who
269 would then figure out what files have been deleted.
270 - Currently in mount.c:236 the SD simply creates a Volume. It should have
271 explicit permission to do so. It should also mark the tape in error
272 if there is an error.
273 - Make sure all restore counters are working correctly in the FD.
274 - SD Bytes Read is wrong.
275 - Look at ALL higher level routines that call block.c to be sure
276 they don't expect something in errmsg.
277 - Investigate doing RAW backup of Win32 partition.
278 - Add thread specific data to hold the jcr -- send error messages from
279 low level routines by accessing it and using Jmsg().
280 - Cancel waiting for Client connect in SD if FD goes away.
282 - Add Progress command that periodically reports the progress of
284 - One block was orphaned in the SD probably after cancel.
287 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
288 - Implement timeout in response() when it should come quickly.
289 - Implement console @echo command.
290 - Implement a Slot priority (loaded/not loaded).
291 - Implement "vacation" Incremental only saves.
292 - Implement single pane restore (much like the Gftp panes).
293 - Implement Automatic Mount even in operator wait.
294 - Implement create "FileSet"?
295 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
296 - Implement "scratch pool" where tapes are defined and can be
297 taken by any pool that needs them.
298 - Implement restore "current system", but take all files without
299 doing selection tree -- so that jobs without File records can
301 - Implement disk spooling. Two parts: 1. Spool to disk then
302 immediately to tape to speed up tape operations. 2. Spool to
303 disk only when the tape is full, then when a tape is hung move
305 - Implement a relocatable bacula.spec
306 - Implement dump/print label to UA
307 - Add prefixlinks to where or not where absolute links to FD.
308 - Issue message to mount a new tape before the rewind.
309 - Simplified client job initiation for portables.
310 - If SD cannot open a drive, make it periodically retry.
311 - Implement LabelTemplate (at least first cut).
312 - Add more of the config info to the tape label.
313 - Implement a Mount Command and an Unmount Command where
314 the users could specify a system command to be performed
315 to do the mount, after which Bacula could attempt to
316 read the device. This is for Removeable media such as a CDROM.
317 - Most likely, this mount command would be invoked explicitly
318 by the user using the current Console "mount" and "unmount"
319 commands -- the Storage Daemon would do the right thing
320 depending on the exact nature of the device.
321 - As with tape drives, when Bacula wanted a new removable
322 disk mounted, it would unmount the old one, and send a message
323 to the user, who would then use "mount" as described above
324 once he had actually inserted the disk.
326 - If tape is marked read-only, then try opening it read-only rather than
327 failing, and remember that it cannot be written.
328 - Refine SD waiting output:
329 Device is being positioned
330 > Device is being positioned for append
331 > Device is being positioned to file x
333 - Figure out some way to estimate output size and to avoid splitting
334 a backup across two Volumes -- this could be useful for writing CDROMs
335 where you really prefer not to have it split -- not serious.
336 - Have SD compute MD5 or SHA1 and compare to what FD computes.
337 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
338 actual data on the Volume and compare it.
339 - Implement Bacula plugins -- design API
340 - Make bcopy read through bad tape records.
341 - Program files (i.e. execute a program to read/write files).
342 Pass read date of last backup, size of file last time.
343 - Add Signature type to File DB record.
344 - CD into subdirectory when open()ing files for backup to
345 speed up things. Test with testfind().
346 - Priority job to go to top of list.
347 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
348 - Implement some way for the Console to dynamically create a job.
349 - Restore to a particular time -- e.g. before date, after date.
350 - Solaris -I on tar for include list
351 - Need a verbose mode in restore, perhaps to bsr.
352 - bscan without -v is too quiet -- perhaps show jobs.
353 - Add code to reject whole blocks if not wanted on restore.
354 - Check if we can increase Bacula FD priorty in Win2000
355 - Make sure the MaxVolFiles is fully implemented in SD
356 - Check if both CatalogFiles and UseCatalog are set to SD.
357 - Figure out how to do a bare metal Windows restore
358 - Possibly add email to Watchdog if drive is unmounted too
359 long and a job is waiting on the drive.
360 - Restore program that errs in SD due to no tape, reports
361 OK incorrectly in output.
362 - After unmount, if restore job started, ask to mount.
363 - Convert all %x substitution variables, which are hard to remember
364 and read to %(variable-name). Idea from TMDA.
365 - Remove NextId for SQLite. Optimize.
366 - Move all SQL statements into a single location.
367 - Add UA rc and history files.
368 - put termcap (used by console) in ./configure and
369 allow -with-termcap-dir.
370 - Enhance time and size scanning routines.
371 - Fix Autoprune for Volumes to respect need for full save.
372 - Fix Win32 config file definition name on /install
373 - Compare tape to Client files (attributes, or attributes and data)
374 - Make all database Ids 64 bit.
375 - Write an applet for Linux.
376 - Implement new inter-daemon communications protocol.
377 - Allow console commands to detach or run in background.
378 - Fix status delay on storage daemon during rewind.
379 - Add SD message variables to control operator wait time
380 - Maximum Operator Wait
381 - Minimum Message Interval
382 - Maximum Message Interval
383 - Send Operator message when cannot read tape label.
384 - Verify level=Volume (scan only), level=Data (compare of data to file).
385 Verify level=Catalog, level=InitCatalog
387 - Add keyword search to show command in Console.
388 - Events : tape has more than xxx bytes.
389 - Complete code in Bacula Resources -- this will permit
390 reading a new config file at any time.
391 - Handle ctl-c in Console
392 - Implement script driven addition of File daemon to config files.
393 - Think about how to make Bacula work better with File (non-tape) archives.
394 - Write Unix emulator for Windows.
395 - Put memory utilization in Status output of each daemon
396 if full status requested or if some level of debug on.
397 - Make database type selectable by .conf files i.e. at runtime
398 - Set flag for uname -a. Add to Volume label.
399 - Implement throttled work queue.
400 - Restore files modified after date
401 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
402 - Implement Restore FileSet=
403 - Create a protocol.h and protocol.c where all protocol messages
405 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
406 - Timout a job or terminate if link goes down, or reopen link and query.
407 - Concept of precious tapes (cannot be reused).
408 - Make bcopy copy with a single tape drive.
409 - Permit changing ownership during restore.
412 > My suggestion: Add a feature on the systray menu-icon menu to request
413 > an immediate backup now. This would be useful for laptop users who may
414 > not be on the network when the regular scheduled backup is run.
416 > My wife's suggestion: Add a setting to the win32 client to allow it to
417 > shut down the machine after backup is complete (after, of course,
418 > displaying a "System will shut down in one minute, click here to cancel"
419 > warning dialog). This would be useful for sites that want user
420 > woorkstations to be shut down overnight to save power.
423 - From Terry Manderson <terry@apnic.net>
424 jobset { # new structure
425 name = "monthlyUnixBoxen"
428 jobs = "wakame;durian;soy;wasabi;miso" #new!
438 fileset = "genericUnixSet"
444 fileset = "genericUnixSet"
450 fileset = "UnixDevelBoxSet"
455 - Autolabel should be specified by DIR instead of SD.
458 - AutoScan (check checksum of tape)
459 - Format command = "format /dev/nst0"
463 - Seek resolution (usually corresponds to buffer size)
464 - EODErrorCode=ENOSPC or code
465 - Partial Read error code
466 - Partial write error code
467 - Nonformatted read error
468 - Nonformatted write error
469 - WriteProtected error
473 - IgnoreCloseErrors=yes
483 - FD sends unsaved file list to Director at end of job (see
485 - File daemon should build list of files skipped, and then
486 at end of save retry and report any errors.
487 - Write a Storage daemon that uses pipes and
488 standard Unix programs to write to the tape.
490 - Need something that monitors the JCR queue and
491 times out jobs by asking the deamons where they are.
492 - Enhance Jmsg code to permit buffering and saving to disk.
493 - device driver = "xxxx" for drives.
495 - Ensure that /dev/null works
496 - Need report class for messages. Perhaps
497 report resource where report=group of messages
498 - enhance scan_attrib and rename scan_jobtype, and
499 fill in code for "since" option
500 - Director needs a time after which the report status is sent
501 anyway -- or better yet, a retry time for the job.
502 - Don't reschedule a job if previous incarnation is still running.
503 - Some way to automatically backup everything is needed????
504 - Need a structure for pending actions:
506 - termination status (part of buffered msgs?)
508 Read, Write, Clean, Delete
509 - Login to Bacula; Bacula users with different permissions:
510 owner, group, user, quotas
511 - Store info on each file system type (probably in the job header on tape.
512 This could be the output of df; or perhaps some sort of /etc/mtab record.
515 - Design at hierarchial storage for Bacula. Migration and Clone.
516 - Implement FSM (File System Modules).
517 - Audit M_ error codes to ensure they are correct and consistent.
518 - Add variable break characters to lex analyzer.
519 Either a bit mask or a string of chars so that
520 the caller can change the break characters.
521 - Make a single T_BREAK to replace T_COMMA, etc.
522 - Ensure that File daemon and Storage daemon can
523 continue a save if the Director goes down (this
524 is NOT currently the case). Must detect socket error,
525 buffer messages for later.
526 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
527 - Add ability to backup to two Storage devices (two SD sessions) at
528 the same time -- e.g. onsite, offsite.
529 - Add the ability to consolidate old backup sets (basically do a restore
530 to tape and appropriately update the catalog). Compress Volume sets.
531 Might need to spool via file is only one drive is available.
532 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
533 someway to do so with every volume that has less than x% valid
537 Migration: Move a backup from one Volume to another
538 Clone: Copy a backup -- two Volumes
540 Bacula Migration is based on Jobs (apparently Networker is file by file).
542 Migration triggered by:
546 Highwater mark (keep total size)
550 Bacula Projects Roadmap
552 last update 8 May 2003
554 Item 1: Multiple simultaneous Jobs. (done)
555 Done -- Restore part needs better implementation to work correctly
556 Also, it needs considerable testing
558 What: Permit multiple simultaneous jobs in Bacula.
560 Why: An enterprise level solution needs to go fast without the
561 need for the system administrator to carefully tweak
562 timing. Based on the benchmarks, during a full
563 backup, NetWorker typically hit 10 times the bandwidth to
564 the tape compared to Bacula--largely. This is probably due to
565 running parallel jobs and multi-threaded filling of buffers
566 and writing them to tape. This should also make things work
567 better when you have a mix of fast and slow machines backing
570 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
571 implementing this is a matter of some small cleanups and
575 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
576 Deferred -- not necessary yet -- possibly implement with Migration.
578 What: If data is coming into the SD too fast, buffer it to
579 disk if the user has configured this option.
581 Why: This would be nice, especially if it more or less falls out
582 when implementing (1) above. If not, it probably should not
583 be given a high priority because fundamentally the backup time
584 is limited by the tape bandwidth. Even though you may finish a
585 client job quicker by spilling to disk, you still have to
586 eventually get it onto tape. If intermediate disk buffering
587 allows us to improve write bandwidth to tape, it may make
590 Notes: Whether or not this is implemented will depend upon performance
591 testing after item 1 is implemented.
594 Item 3: Write the bscan program -- also write a bcopy program.
597 What: Write a program that reads a Bacula tape and puts all the
598 appropriate data into the catalog. This allows recovery
599 from a tape that is no longer in the database, or it allows
600 re-creation of a database if lost.
602 Why: This is a fundamental robustness and disaster recovery tool
603 which will increase the comfort level of a sysadmin
604 considering adopting Bacula.
606 Notes: A skeleton of this program already exists, but much work
607 needs to be done. Implementing this will also make apparent
608 any deficiencies in the current Bacula tape format.
611 Item 4: Implement Base jobs.
613 What: A base job is sort of like a Full save except that you
614 will want the FileSet to contain only files that are unlikely
615 to change in the future (i.e. a snapshot of most of your
616 system after installing it). After the base job has been run,
617 when you are doing a Full save, you can specify to exclude
618 all files saved by the base job that have not been modified.
620 Why: This is something none of the competition does, as far as we know
621 (except BackupPC, which is a Perl program that saves to disk
622 only). It is big win for the user, it makes Bacula stand out
623 as offering a unique optimization that immediately saves time
626 Notes: Big savings in tape usage. Will require more resources because
627 the e. DIR must send FD a list of files/attribs, and the FD must
628 search the list and compare it for each file to be saved.
631 Item 5: Implement Label templates
634 What: This is a mechanism whereby Bacula can automatically create
635 a tape label for new tapes according to a detailed specification
636 provided by the user.
638 Why: It is a major convenience item for folks who use automated label
641 Notes: Bacula already has a working form of automatic tape label
642 creation, but it is very crude. The design for the complete
643 tape labeling project is already documented in the manual.
646 Item 6: Write a regression script.
647 Done -- Continue to expand its testing.
649 What: This is an automatic script that runs and tests as many features
650 of Bacula as possible. The output is compared to previous
651 versions of Bacula and any differences are reported.
653 Why: This is an enormous help in preventing introduction of new
654 errors in parts of the program that already work correctly.
656 Notes: This probably should be ranked higher, it's something the typical
657 user doesn't see. Depending on how it's implemented, it may
658 make sense to defer it until the archival tape format and
659 user interface mature.
662 Item 7: GUI for interactive restore
663 Item 8: GUI for interactive backup
665 What: The current interactive restore is implemented with a tty
666 interface. It would be much nicer to be able to "see" the
667 list of files backed up in typical GUI tree format.
668 The same mechanism could also be used for creating
669 ad-hoc backup FileSets (item 8).
671 Why: Ease of use -- especially for the end user.
673 Notes: Rather than implementing in Gtk, we probably should go directly
674 for a Browser implementation, even if doing so meant the
675 capability wouldn't be available until much later. Not only
676 is there the question of Windows sites, most
677 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
678 without installing lots of stuff admins are very wary about.
679 Real sysadmins will always use the command line anyway, and
680 the user who's doing an interactive restore or backup of his
681 own files will in most cases be on a Windows machine running
685 Item 9: Add SSL to daemon communications.
687 What: This provides for secure communications between the daemons.
689 Why: This would allow doing backup across the Internet without
690 privacy concerns (or with much less concern).
692 Notes: The vast majority of near term potential users will be backing up
693 a single site over a LAN and, correctly or not, they probably
694 won't be concerned with security, at least not enough to go to
695 the trouble to set up keys, etc. to screw things down. We suspect
696 that many users genuinely interested in multi-site backup
697 already run some form of VPN software in their internetwork
698 connections, and are willing to delegate security to that layer.
701 Item 10: Define definitive tape format.
704 What: Define that definitive tape format that will not change
705 for the next millennium.
707 Why: Stability, security.
709 Notes: See notes for item 11 below.
712 Item 11: New daemon communication protocol.
714 What: The current daemon to daemon protocol is basically an ASCII
715 printf() and sending the buffer. On the receiving end, the
716 buffer is sscanf()ed to unpack it. The new scheme would
717 be a binary format that allows quick packing and unpacking
718 of any data type with named fields.
720 Why: Using binary packing would be faster. Named fields will permit
721 error checking to ensure that what is sent is what the
722 receiver really wants.
724 Notes: These are internal improvements in the interest of the
725 long-term stability and evolution of the program. On the one
726 hand, the sooner they're done, the less code we have to rip
727 up when the time comes to install them. On the other hand, they
728 don't bring an immediately perceptible benefit to potential
729 users. Item 10 and possibly item 11 should be deferred until Bacula
730 is well established with a growing user community more or
731 less happy with the feature set. At that time, it will make a
732 good "next generation" upgrade in the interest of data
738 ======================================================
740 It is somewhat like a Full save becomes an incremental since
741 the Base job (or jobs) plus other non-base files.
743 - New BaseFile table that contains:
744 JobId, BaseJobId, FileId (from Base).
745 i.e. for each base file that exists but is not saved because
746 it has not changed, the File daemon sends the JobId, BaseId,
747 and FileId back to the Director who creates the DB entry.
748 - To initiate a Base save, the Director sends the FD
749 the FileId, and full filename for each file in the Base.
750 - When the FD finds a Base file, he requests the Director to
751 send him the full File entry (stat packet plus MD5), or
752 conversely, the FD sends it to the Director and the Director
753 says yes or no. This can be quite rapid if the FileId is kept
754 by the FD for each Base Filename.
755 - It is probably better to have the comparison done by the FD
756 despite the fact that the File entry must be sent across the
758 - An alternative would be to send the FD the whole File entry
759 from the start. The disadvantage is that it requires a lot of
760 space. The advantage is that it requires less communications
762 - The Job record must be updated to indicate that one or more
764 - At end of Job, FD returns:
765 1. Count of base files/bytes not written to tape (i.e. matches)
766 2. Count of base file that were saved i.e. had changed.
767 - No tape record would be written for a Base file that matches, in the
768 same way that no tape record is written for Incremental jobs where
769 the file is not saved because it is unchanged.
770 - On a restore, all the Base file records must explicitly be
771 found from the BaseFile tape. I.e. for each Full save that is marked
772 to have one or more Base Jobs, search the BaseFile for all occurrences
774 - An optimization might be to make the BaseFile have:
780 This would avoid the need to explicitly fetch each File record for
781 the Base job. The Base Job record will be fetched to get the
782 VolSessionId and VolSessionTime.
783 =========================================================
787 =============================================================
789 Request For Comments For File Backup Options
792 Subject: File Backup Options
795 A few days ago, a Bacula user who is backing up to file volumes and
796 using compression asked if it was possible to suppress compressing
797 all .gz files since it was a waste of CPU time. Although Bacula
798 currently permits using different options (compression, ...) on
799 a directory by directory basis, it cannot do it on a file by
800 file basis, which is clearly what was desired.
802 Proposed Implementation:
803 To solve this problem, I propose the following:
805 - Add a new Director resource type called Options.
807 - The Options resource will have records for all
808 options that can currently be specified on the Include record
809 (in a FileSet). Examples below.
811 - The Options resource will permit an exclude option as well
812 as a number of additional options.
814 - The heart of the Options resource is the ability to
815 supply any number of Match records which specify POSIX
816 regular expressions. These Match regular expressions are
817 applied to the fully qualified filename (path and all). If
818 one matches, then the Options will be used.
820 - When an Match specification matches an included file, the
821 options specified in the Options resource will override
822 the default options specified on the Include record.
824 - Include records will be modified to permit referencing one or
825 more Options resources. The Options will be used
826 in the order listed on the Include record and the first
827 one that matches will be applied.
829 - Options (or specifications) currently supplied on the Include
830 record will be deprecated (i.e. removed in a later version a
831 year or so from now).
833 - The Exclude record will be deprecated as the same functionality
834 can be obtained by using an Exclude = yes in the Options.
837 The following records can appear in the Options resource. An
838 asterisk preceding the name indicates a feature not currently
842 - Compression= (GZIP, ...)
843 - Signature= (MD5, SHA1, ...)
845 - OneFs= (yes/no) - remain on one filesystem
846 - Recurse= (yes/no) - recurse into subdirectories
847 - Sparse= (yes/no) - do sparse file backup
848 - *Exclude= (yes/no) - exclude file from being saved
849 - *Reader= (filename) - external read (backup) program
850 - *Plugin= (filename) - read/write plugin module
853 - verify= (ipnougsamc5) - verify options
856 - replace= (always/ifnewer/ifolder/never) - replace options currently
858 - *Writer= (filename) - external write (restore) program
862 Currently options specifying compression, MD5 signatures, recursion,
863 ... of a FileSet are supplied on the Include record. These will now
864 all be collected into a Options resource, which will be
865 specified in the Include in place of the options. Multiple Options
866 may be specified. Since the Options may contain regular expressions
867 that are applied to the full filename, this will give the ability
868 to specify backup options on a file by file basis to whatever level
877 Include = compression=GZIP signature=MD5 {
889 Match = /*.?*/ # matches all files.
894 That's a lot more to do the same thing, but it gives the ability to
895 apply options on a file by file basis. For example, suppose you
896 want to compress all files but not any file with extensions .gz or .Z.
897 In that case, you will need to group two sets of options using
898 the Options resource as follows:
906 # Note multiple Matches are ORed
907 Match = /*.gz/ # matches .gz files */
908 Match = /*.Z/ # matches .Z files */
913 Match = /*.?*/ # matches all files
919 Now, since the no Compression option is specified in the
920 first group of Options, *.gz or *.Z file will have an MD5 signature computed,
921 but will not be compressed. For all other files, the *.gz *.Z will not
922 match, so the second group of options will be used which will include GZIP
926 - Is it necessary to provide some means of ANDing regular expressions
927 and negation? (not currently planned)
929 e.g. Match = /*.gz/ && !/big.gz/
931 - I see that Networker has a "null" module which, if specified, does not
932 backup the file, but does make an record of the file in the catalog
933 so that the catalog will reflect an exact picture of the filesystem.
934 The result is that the file can be "seen" when "browsing" the save
935 sets, but it cannot be restored.
937 Is this really useful? Should it be implemented in Bacula?
940 After implementing the above, the user will be able to specify
941 on a file by file basis (using regular expressions) what options are
942 applied for the backup.
945 =============================================
947 ==========================================================
949 For each Incremental job that is run, there may be files that
950 were found but not saved because they were locked (this applies
951 only to Windows). Such a system could send back to the Director
952 a list of Unsaved files.
954 - New UnSavedFiles table that contains:
958 - Then in the next Incremental job, the list of Unsaved Files will be
959 feed to the FD, who will ensure that they are explicitly chosen even
960 if standard date/time check would not have selected them.
961 =============================================================
964 Done: (see kernsdone for more)
965 - Implement new alist in FileSet scanning.
966 - bls should continue reading even if it finds Win32 data on the tape.
967 The error should be Warning rather the Error.
968 - Add user configurable timeout for connecting to SD.
969 - Unsaved Flag in Job record (use JobMissingFiles).
970 - Base Flag in Job record.
971 - Configure mtx-changer to have correct path to mtx.
972 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
973 - Make some way so that if a machine is skipped because it is not up
974 that Bacula will continue retrying for a specified period of time --
976 - Implement all command line args on run.
977 - Implement command line "restore" args.
978 - Implement "restore current select=no"
979 - Restore file modified before date
980 - Restore -- do nothing but show what would happen
981 - Add estimate to Console commands
982 - Use read_record.c in SD code.
983 - Fix read_record to handle multiple sessions.
984 - Tip from Steve Allam
985 mt -f /dev/nst0 defblksize 0
986 - Document "status" in the console.
987 - Document driving console from shell script.
988 - Write JobMedia records with max file size is reached on tape.
989 - Handle the case of multiple JobMedia records pending (i.e. the
990 thread is slow and multiple situations requiring a JobMedia
992 - Do performance analysis on the restore tree routines.
993 - Fix maximum file size (block.c) to generate JobMedia records.
994 - Make the default file size 1GB on the tape.
995 - Implement forward spacing between files.
996 - Add Machine type (Linux/Windows) to Status report for daemons.
998 - Use repositioning at the beginning of the tape.
999 - Do full check the command line args in update (e.g. VolStatus ...).
1000 - Specify list of files to restore
1001 - Implement ClientRunBeforeJob and ClientRunAfterJob.
1002 - Make | and < work on FD side.
1003 - Check to see if "blocked" is set during restore.
1004 - Figure out what is interrupting sql command in console.
1005 - Make new job print warning User Unmounted Tape.
1006 - Test recycling and purging (code changed in db_find_next_volume and
1008 - Document SDConnectTimeout (in FD).
1009 - Add restore by filename test.
1010 - Document restore by files.
1011 - Make variable expansion work correctly.
1012 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
1013 - Copy static programs into install directory.
1014 - Think about changing Storage resource Device record to be
1016 - Add RunBeforeJob and RunAfterJob to the Client program.
1017 - Need return status on read_cb() from read_records(). Need multiple
1018 records -- one per Job, maybe a JCR or some other structure with
1019 a block and a record.
1020 - LabelFormat on tape volume apparently creates the db record but
1021 never actually labels the volume.
1022 - Recycling a volume when two jobs are using it is going to break. Fixed.
1023 - Document list nextvol and new format status dir.
1024 - Client files in Win32 with Unix eol conventions doesn't work.
1025 - Either fix or document that fill command in btape can be
1026 compressed enormously by the hardware - a 36GB tape wrote 750GB!
1027 - Add multiple character duration qualifiers.
1028 - Require some modifer.
1029 - Restrict characters permitted in a Resource name, and don't permit
1031 - Figure out some way to ignore or get past checksum errors in
1033 - The SD spooling file gets created even if it is not used.
1034 - Look at Cleaning tape in ua_label.c for media create/update
1035 - Add regression testing to the manual
1036 - End time: in job output of rescheduled job is time of first run.
1037 - Document list nextvol and status output.
1038 - Separate Dir heartbeat in FD from the SD heartbeat.
1039 - Fix sparse file handeling so that it always reads a multiple
1040 of 512. Currently, it subtracts 8 bytes (for faddr).
1041 Kludged with #ifdef for FreeBSD.
1042 - Document that Volume pruning can delete last Full backup and
1043 hence you will not have a valid backup.
1044 - Clarify the fact that having the Bacula cygwin1.dll loaded
1045 is not the same as having cygwin installed.
1046 - Document that it is safe to use the drive when the lights stop flashing.
1047 - Document all the status codes JobLevel, JobType, JobStatus.
1048 - Add GUI interface to manual
1049 - Combine the 3 places that search run records for the next
1050 job. Use find_job_pool() modified in ua_output.c
1051 - Test connect timeouts.
1052 - Fix FreeBSD build with tcp_wrapper -- should not have -lnsl
1053 - Implement fast block rejection.
1054 - I want to restore by file to some date.
1056 - Figure out a way to move Volumes from one pool to another.