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)
17 Testing to do: (painful)
18 - that ALL console command line options work and are always implemented
19 - blocksize recognition code.
20 - Test if rewind at end of tape waits for tape to rewind.
22 - Test not zeroing Autochanger slot when it is wrong.
23 - Figure out how to use ssh or stunnel to protect Bacula communications.
25 For 1.33 Testing/Documentation:
26 - Document to start higher priorty jobs before lower ones.
27 - suppress "Do not forget to mount the drive!!!" if error
28 - Document new records in Director. SDAddress SDDeviceName, SDPassword.
29 FDPassword, FDAddress, DBAddress, DBPort, DBPassword.
30 - Document new Include/Exclude ...
31 - Add test of exclusion, test multiple Include {} statements.
32 - Add counter variable test.
33 - Document ln -sf /usr/lib/libncurses.so /usr/lib/libtermcap.so
34 and install the esound-dev package for compiling Console on
38 - Implement a RunAfterFailedJob
39 - Zap illegal characters in job name for mail files (e.g. /).
41 Yes, it would allow to highly automatic the request for new tapes. If a
42 tape is empty, bacula reads the barcodes (native or simulated), and if
43 an unused tape is found, it runs the label command with all the
46 By the way can bacula automatically "move" an empry/purged volume say
47 in the "short" pool to the "long" pool if this pool runs out of volume
49 - Implement a move Volume from one pool to another.
50 - Either restrict the characters in a name, or fix the problem
51 emailing with names containing / (smtp command line breaks).
52 - Eliminate ua_retention.c (retentioncmd) if possible.
53 - Eliminate orphaned jobs: dbcheck, normal pruning, delete job command.
54 Hm. Well, there are the remaining orphaned job records:
56 | 105 | Llioness Save | 0000-00-00 00:00:00 | B | D | 0 | 0 | f |
57 | 110 | Llioness Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
58 | 115 | Llioness Save | 2003-09-10 02:22:03 | B | I | 0 | 0 | A |
59 | 128 | Catalog Save | 2003-09-11 03:53:32 | B | I | 0 | 0 | C |
60 | 131 | Catalog Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
62 As you can see, three of the five are failures. I already deleted the
63 one restore and one other failure using the by-client option. Deciding
64 what is an orphaned job is a tricky problem though, I agree. All these
65 records have or had 0 files/ 0 bytes, except for the restore. With no
66 files, of course, I don't know of the job ever actually becomes
67 associated with a Volume.
69 (I'm not sure if this is documented anywhere -- what are the meanings of
70 all the possible JobStatus codes?)
72 Looking at my database, it appears to me as though all the "orphaned"
73 jobs fit into one of two categories:
75 1) The Job record has a StartTime but no EndTime, and the job is not
78 2) The Job record has an EndTime, indicating that it completed, but
79 it has no associated JobMedia record.
82 This does suggest an approach. If failed jobs (or jobs that, for some
83 other reason, write no files) are associated with a volume via a
84 JobMedia record, then they should be purged when the associated volume
85 is purged. I see two ways to handle jobs that are NOT associated with a
88 1) purge them automatically whenever any volume is manually purged;
90 2) add an option to the purge command to manually purge all jobs with
93 I think Restore jobs also fall into category 2 above .... so one might
94 want to make that "The Job record has an EndTime,, but no associated
95 JobMedia record, and is not a Restore job."
96 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
97 which would take precidence over the Catalog "JobRetention".
98 - Implement Label Format in Add and Label console commands.
99 - make "btape /tmp" work.
100 - Make sure a rescheduled job is properly reported by status.
101 - Walk through the Pool records rather than the Job records
102 in dird.c to create/update pools.
103 - Figure out a way to move Volumes from one pool to another.
104 - What to do about "list files job=xxx".
105 - Implement delete Job.
106 - Document need to put LabelFormat in quotes.
107 - Implement scan: for every slot it finds, zero the slot of
108 Volume other volume having that slot.
109 - When job rescheduled, status gives is waiting for Client Rufus
110 to connect to Storage File. Dir needs to inform SD that job
112 - Fix get_storage_from_media_type (ua_restore) to use command line
114 - Enhance "update slots" to include a "scan" feature
115 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
116 - Allow a slot or range of slots on the label barcodes command.
117 - Don't print "Warning: Wrong Volume mounted ..." if mounting second volume.
118 - Make Dmsg look at global before calling subroutine.
119 - Enable trace output at runtime for Win32
120 - Make sure that Volumes are recycled based on "Least recently used"
121 rather than lowest MediaId.
122 - Available volumes for autochangers (see patrick@baanboard.com 3 Sep 03
123 and 4 Sep) scan slots.
124 - Upgrade to cygwin 1.5
126 - Get and test MySQL 4.0
127 - Do a complete audit of all pthreads_mutex, cond, ... to ensure that
128 any that are dynamically initialized are destroyed when no longer used.
129 - Write a mini-readline with history and editing.
130 - Look at how fuser works and /proc/PID/fd that is how Nic found the
131 file descriptor leak in Bacula.
132 - Implement WrapCounters in Counters.
133 - Turn on SIGHUP in dird.c and test.
134 - Use system dependent calls to get more precise info on tape errors.
135 - Add heartbeat from FD to SD if hb interval expires.
136 - Suppress read error on blank tape when doing a label.
137 - Can we dynamically change FileSets?
138 - If pool specified to label command and Label Format is specified,
139 automatically generate the Volume name.
140 - Take a careful look a the Basic recycling algorithm. When Bacula
141 chooses, the order should be:
143 - Look for Recycle or Purged
146 Instead of using lowest media Id, find the least recently used
149 When the tape is mounted and Bacula requests the status
150 - Do everything possible to use it.
152 Define a "available" status, which is the currently mounted
153 Volume and all volumes that are currently in the autochanger.
155 - Why can't SQL do the filename sort for restore?
156 - Is a pool specification really needed for a restore? Yes, and
157 you may want to exclude archive Pools.
158 - Look at libkse (man kse) for FreeBSD threading.
159 - Look into Microsoft Volume Shadowcopy Service VSS for backing
160 up system state components (Active Directory, System Volume, ...)
161 - Add ExhautiveRestoreSearch
162 - Look at the possibility of loading only the necessary
163 data into the restore tree (i.e. do it one directory at a
164 time as the user walks through the tree).
165 - Possibly use the hash code if the user selects all for a restore command.
166 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
167 - Implement some way for the File daemon to contact the Director
168 to start a job or pass its DHCP obtained IP number.
169 - Implement multiple Consoles.
170 - Add Console usr permissions.
171 - Fix "restore all" to bypass building the tree.
172 - Fix restore to list errors if Invalid block found, and if # files
173 restored does not match # expected.
174 - Prohibit backing up archive device (findlib/find_one.c:128)
175 - Implement Release Device in the Job resource to unmount a drive.
176 - Implement Acquire Device in the Job resource to mount a drive,
177 be sure this works with admin jobs so that the user can get
178 prompted to insert the correct tape. Possibly some way to say to
179 run the job but don't save the files.
180 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
181 - Implement FileOptions (see end of this document)
182 - Make things like list where a file is saved case independent for
184 - Edit the Client/Storage name into authentication failure messages.
185 - Implement job in VerifyToCatalog
187 - Implement a PostgreSQL driver.
188 - Bacula needs to propagate SD errors.
189 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
190 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
191 /dev/nst0. ERR=Input/output error.
192 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
193 Last block=5162 Current block=5164.
194 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
196 - Use autochanger to handle multiple devices.
197 - Fix packet too big problem. This is most likely a Windows TCP stack
199 - Add SuSE install doc to list.
200 - Check and rechedk "Invalid block number"
201 - Make bextract release the drive properly between tapes
202 so that an autochanger can be made to work.
203 - User wants to NOT backup up certain big files (email files).
204 - Maybe remove multiple simultaneous devices code in SD.
205 - On Windows with very long path names, it may be impossible to create
206 a file (and thus restore it) because the total length is too long.
207 We must cd into the directory then create the file without the
209 - lstat() is not going to work on Win32 for testing date.
210 - Something is not right in last block of fill command.
211 - Implement a Recycle command
212 - Add FileSet to command line arguments for restore.
213 - Allow multiple Storage specifications (or multiple names on
214 a single Storage specification) in the Job record. Thus a job
215 can be backed up to a number of storage devices.
216 - Add client name to cram-md5 challenge so Director can immediately
217 verify if it is the correct client.
218 - Implement ClientRunBeforeJob and ClientRunAfterJob.
219 - Add JobLevel in FD status (but make sure it is defined).
220 - Audit all UA commands to ensure that we always prompt where possible.
221 - Restrict characters permitted in a Resource name, and don't permit
223 - Check Jmsg in bnet, may not work, must dup bsock.
224 - Suppress Job Name in Jmsg for console
225 - Create Pools that are referenced in a Run statement at startup if possible.
226 - Use runbeforejob to unload, then reload a volume previously used,
227 then the next job run gets an error reading the drive.
228 - Make bootstrap filename unique.
229 - Test a second language e.g. french.
230 - Start working on Base jobs.
231 - Make "make binary-release" work from any directory.
232 - Implement UnsavedFiles DB record.
233 - Implement argc/argv for daemon command line scanning using table driven
235 - Implement table driven single argc/argv scanner to pickup all arguments.
236 Much like xxx_conf.c scan table.
237 keyword, handler(store_routine), store_address, code, flags, default.
238 - From Phil Stracchino:
239 It would probably be a per-client option, and would be called
240 something like, say, "Automatically purge obsoleted jobs". What it
241 would do is, when you successfully complete a Differential backup of a
242 client, it would automatically purge all Incremental backups for that
243 client that are rendered redundant by that Differential. Likewise,
244 when a Full backup on a client completed, it would automatically purge
245 all Differential and Incremental jobs obsoleted by that Full backup.
246 This would let people minimize the number of tapes they're keeping on
247 hand without having to master the art of retention times.
248 - Implement new serialize subroutines
249 send(socket, "string", &Vol, "uint32", &i, NULL)
250 - Scratch Pool where the volumes can be re-assigned to any Pool.
251 - Implement a M_SECURITY message class.
252 - Implement forward spacing block/file: position_device(bsr) --
253 just before read_block_from_device();
254 - When doing a Backup send all attributes back to the Director, who
255 would then figure out what files have been deleted.
256 - Currently in mount.c:236 the SD simply creates a Volume. It should have
257 explicit permission to do so. It should also mark the tape in error
258 if there is an error.
259 - Make sure all restore counters are working correctly in the FD.
260 - SD Bytes Read is wrong.
261 - Look at ALL higher level routines that call block.c to be sure
262 they don't expect something in errmsg.
263 - Investigate doing RAW backup of Win32 partition.
264 - Add thread specific data to hold the jcr -- send error messages from
265 low level routines by accessing it and using Jmsg().
266 - Cancel waiting for Client connect in SD if FD goes away.
268 - Add Progress command that periodically reports the progress of
270 - One block was orphaned in the SD probably after cancel.
273 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
274 - Implement timeout in response() when it should come quickly.
275 - Implement console @echo command.
276 - Implement a Slot priority (loaded/not loaded).
277 - Implement "vacation" Incremental only saves.
278 - Implement single pane restore (much like the Gftp panes).
279 - Implement Automatic Mount even in operator wait.
280 - Implement create "FileSet"?
281 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
282 - Implement "scratch pool" where tapes are defined and can be
283 taken by any pool that needs them.
284 - Implement restore "current system", but take all files without
285 doing selection tree -- so that jobs without File records can
287 - Implement disk spooling. Two parts: 1. Spool to disk then
288 immediately to tape to speed up tape operations. 2. Spool to
289 disk only when the tape is full, then when a tape is hung move
291 - Implement a relocatable bacula.spec
292 - Implement dump/print label to UA
293 - Add prefixlinks to where or not where absolute links to FD.
294 - Issue message to mount a new tape before the rewind.
295 - Simplified client job initiation for portables.
296 - If SD cannot open a drive, make it periodically retry.
297 - Implement LabelTemplate (at least first cut).
298 - Add more of the config info to the tape label.
299 - Implement a Mount Command and an Unmount Command where
300 the users could specify a system command to be performed
301 to do the mount, after which Bacula could attempt to
302 read the device. This is for Removeable media such as a CDROM.
303 - Most likely, this mount command would be invoked explicitly
304 by the user using the current Console "mount" and "unmount"
305 commands -- the Storage Daemon would do the right thing
306 depending on the exact nature of the device.
307 - As with tape drives, when Bacula wanted a new removable
308 disk mounted, it would unmount the old one, and send a message
309 to the user, who would then use "mount" as described above
310 once he had actually inserted the disk.
312 - If tape is marked read-only, then try opening it read-only rather than
313 failing, and remember that it cannot be written.
314 - Refine SD waiting output:
315 Device is being positioned
316 > Device is being positioned for append
317 > Device is being positioned to file x
319 - Figure out some way to estimate output size and to avoid splitting
320 a backup across two Volumes -- this could be useful for writing CDROMs
321 where you really prefer not to have it split -- not serious.
322 - Have SD compute MD5 or SHA1 and compare to what FD computes.
323 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
324 actual data on the Volume and compare it.
325 - Implement Bacula plugins -- design API
326 - Make bcopy read through bad tape records.
327 - Program files (i.e. execute a program to read/write files).
328 Pass read date of last backup, size of file last time.
329 - Add Signature type to File DB record.
330 - CD into subdirectory when open()ing files for backup to
331 speed up things. Test with testfind().
332 - Priority job to go to top of list.
333 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
334 - Implement some way for the Console to dynamically create a job.
335 - Restore to a particular time -- e.g. before date, after date.
336 - Solaris -I on tar for include list
337 - Need a verbose mode in restore, perhaps to bsr.
338 - bscan without -v is too quiet -- perhaps show jobs.
339 - Add code to reject whole blocks if not wanted on restore.
340 - Check if we can increase Bacula FD priorty in Win2000
341 - Make sure the MaxVolFiles is fully implemented in SD
342 - Check if both CatalogFiles and UseCatalog are set to SD.
343 - Figure out how to do a bare metal Windows restore
344 - Possibly add email to Watchdog if drive is unmounted too
345 long and a job is waiting on the drive.
346 - Restore program that errs in SD due to no tape, reports
347 OK incorrectly in output.
348 - After unmount, if restore job started, ask to mount.
349 - Convert all %x substitution variables, which are hard to remember
350 and read to %(variable-name). Idea from TMDA.
351 - Remove NextId for SQLite. Optimize.
352 - Move all SQL statements into a single location.
353 - Add UA rc and history files.
354 - put termcap (used by console) in ./configure and
355 allow -with-termcap-dir.
356 - Enhance time and size scanning routines.
357 - Fix Autoprune for Volumes to respect need for full save.
358 - Fix Win32 config file definition name on /install
359 - Compare tape to Client files (attributes, or attributes and data)
360 - Make all database Ids 64 bit.
361 - Write an applet for Linux.
362 - Implement new inter-daemon communications protocol.
363 - Allow console commands to detach or run in background.
364 - Fix status delay on storage daemon during rewind.
365 - Add SD message variables to control operator wait time
366 - Maximum Operator Wait
367 - Minimum Message Interval
368 - Maximum Message Interval
369 - Send Operator message when cannot read tape label.
370 - Verify level=Volume (scan only), level=Data (compare of data to file).
371 Verify level=Catalog, level=InitCatalog
373 - Add keyword search to show command in Console.
374 - Events : tape has more than xxx bytes.
375 - Complete code in Bacula Resources -- this will permit
376 reading a new config file at any time.
377 - Handle ctl-c in Console
378 - Implement script driven addition of File daemon to config files.
379 - Think about how to make Bacula work better with File (non-tape) archives.
380 - Write Unix emulator for Windows.
381 - Put memory utilization in Status output of each daemon
382 if full status requested or if some level of debug on.
383 - Make database type selectable by .conf files i.e. at runtime
384 - Set flag for uname -a. Add to Volume label.
385 - Implement throttled work queue.
386 - Restore files modified after date
387 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
388 - Implement Restore FileSet=
389 - Create a protocol.h and protocol.c where all protocol messages
391 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
392 - Timout a job or terminate if link goes down, or reopen link and query.
393 - Concept of precious tapes (cannot be reused).
394 - Make bcopy copy with a single tape drive.
395 - Permit changing ownership during restore.
398 > My suggestion: Add a feature on the systray menu-icon menu to request
399 > an immediate backup now. This would be useful for laptop users who may
400 > not be on the network when the regular scheduled backup is run.
402 > My wife's suggestion: Add a setting to the win32 client to allow it to
403 > shut down the machine after backup is complete (after, of course,
404 > displaying a "System will shut down in one minute, click here to cancel"
405 > warning dialog). This would be useful for sites that want user
406 > woorkstations to be shut down overnight to save power.
409 - From Terry Manderson <terry@apnic.net>
410 jobset { # new structure
411 name = "monthlyUnixBoxen"
414 jobs = "wakame;durian;soy;wasabi;miso" #new!
424 fileset = "genericUnixSet"
430 fileset = "genericUnixSet"
436 fileset = "UnixDevelBoxSet"
441 - Autolabel should be specified by DIR instead of SD.
444 - AutoScan (check checksum of tape)
445 - Format command = "format /dev/nst0"
449 - Seek resolution (usually corresponds to buffer size)
450 - EODErrorCode=ENOSPC or code
451 - Partial Read error code
452 - Partial write error code
453 - Nonformatted read error
454 - Nonformatted write error
455 - WriteProtected error
459 - IgnoreCloseErrors=yes
469 - FD sends unsaved file list to Director at end of job (see
471 - File daemon should build list of files skipped, and then
472 at end of save retry and report any errors.
473 - Write a Storage daemon that uses pipes and
474 standard Unix programs to write to the tape.
476 - Need something that monitors the JCR queue and
477 times out jobs by asking the deamons where they are.
478 - Enhance Jmsg code to permit buffering and saving to disk.
479 - device driver = "xxxx" for drives.
481 - Ensure that /dev/null works
482 - Need report class for messages. Perhaps
483 report resource where report=group of messages
484 - enhance scan_attrib and rename scan_jobtype, and
485 fill in code for "since" option
486 - Director needs a time after which the report status is sent
487 anyway -- or better yet, a retry time for the job.
488 - Don't reschedule a job if previous incarnation is still running.
489 - Some way to automatically backup everything is needed????
490 - Need a structure for pending actions:
492 - termination status (part of buffered msgs?)
494 Read, Write, Clean, Delete
495 - Login to Bacula; Bacula users with different permissions:
496 owner, group, user, quotas
497 - Store info on each file system type (probably in the job header on tape.
498 This could be the output of df; or perhaps some sort of /etc/mtab record.
501 - Design at hierarchial storage for Bacula. Migration and Clone.
502 - Implement FSM (File System Modules).
503 - Audit M_ error codes to ensure they are correct and consistent.
504 - Add variable break characters to lex analyzer.
505 Either a bit mask or a string of chars so that
506 the caller can change the break characters.
507 - Make a single T_BREAK to replace T_COMMA, etc.
508 - Ensure that File daemon and Storage daemon can
509 continue a save if the Director goes down (this
510 is NOT currently the case). Must detect socket error,
511 buffer messages for later.
512 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
513 - Add ability to backup to two Storage devices (two SD sessions) at
514 the same time -- e.g. onsite, offsite.
515 - Add the ability to consolidate old backup sets (basically do a restore
516 to tape and appropriately update the catalog). Compress Volume sets.
517 Might need to spool via file is only one drive is available.
518 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
519 someway to do so with every volume that has less than x% valid
523 Migration: Move a backup from one Volume to another
524 Clone: Copy a backup -- two Volumes
526 Bacula Migration is based on Jobs (apparently Networker is file by file).
528 Migration triggered by:
532 Highwater mark (keep total size)
536 Bacula Projects Roadmap
538 last update 8 May 2003
540 Item 1: Multiple simultaneous Jobs. (done)
541 Done -- Restore part needs better implementation to work correctly
542 Also, it needs considerable testing
544 What: Permit multiple simultaneous jobs in Bacula.
546 Why: An enterprise level solution needs to go fast without the
547 need for the system administrator to carefully tweak
548 timing. Based on the benchmarks, during a full
549 backup, NetWorker typically hit 10 times the bandwidth to
550 the tape compared to Bacula--largely. This is probably due to
551 running parallel jobs and multi-threaded filling of buffers
552 and writing them to tape. This should also make things work
553 better when you have a mix of fast and slow machines backing
556 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
557 implementing this is a matter of some small cleanups and
561 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
562 Deferred -- not necessary yet -- possibly implement with Migration.
564 What: If data is coming into the SD too fast, buffer it to
565 disk if the user has configured this option.
567 Why: This would be nice, especially if it more or less falls out
568 when implementing (1) above. If not, it probably should not
569 be given a high priority because fundamentally the backup time
570 is limited by the tape bandwidth. Even though you may finish a
571 client job quicker by spilling to disk, you still have to
572 eventually get it onto tape. If intermediate disk buffering
573 allows us to improve write bandwidth to tape, it may make
576 Notes: Whether or not this is implemented will depend upon performance
577 testing after item 1 is implemented.
580 Item 3: Write the bscan program -- also write a bcopy program.
583 What: Write a program that reads a Bacula tape and puts all the
584 appropriate data into the catalog. This allows recovery
585 from a tape that is no longer in the database, or it allows
586 re-creation of a database if lost.
588 Why: This is a fundamental robustness and disaster recovery tool
589 which will increase the comfort level of a sysadmin
590 considering adopting Bacula.
592 Notes: A skeleton of this program already exists, but much work
593 needs to be done. Implementing this will also make apparent
594 any deficiencies in the current Bacula tape format.
597 Item 4: Implement Base jobs.
599 What: A base job is sort of like a Full save except that you
600 will want the FileSet to contain only files that are unlikely
601 to change in the future (i.e. a snapshot of most of your
602 system after installing it). After the base job has been run,
603 when you are doing a Full save, you can specify to exclude
604 all files saved by the base job that have not been modified.
606 Why: This is something none of the competition does, as far as we know
607 (except BackupPC, which is a Perl program that saves to disk
608 only). It is big win for the user, it makes Bacula stand out
609 as offering a unique optimization that immediately saves time
612 Notes: Big savings in tape usage. Will require more resources because
613 the e. DIR must send FD a list of files/attribs, and the FD must
614 search the list and compare it for each file to be saved.
617 Item 5: Implement Label templates
620 What: This is a mechanism whereby Bacula can automatically create
621 a tape label for new tapes according to a detailed specification
622 provided by the user.
624 Why: It is a major convenience item for folks who use automated label
627 Notes: Bacula already has a working form of automatic tape label
628 creation, but it is very crude. The design for the complete
629 tape labeling project is already documented in the manual.
632 Item 6: Write a regression script.
633 Done -- Continue to expand its testing.
635 What: This is an automatic script that runs and tests as many features
636 of Bacula as possible. The output is compared to previous
637 versions of Bacula and any differences are reported.
639 Why: This is an enormous help in preventing introduction of new
640 errors in parts of the program that already work correctly.
642 Notes: This probably should be ranked higher, it's something the typical
643 user doesn't see. Depending on how it's implemented, it may
644 make sense to defer it until the archival tape format and
645 user interface mature.
648 Item 7: GUI for interactive restore
649 Item 8: GUI for interactive backup
651 What: The current interactive restore is implemented with a tty
652 interface. It would be much nicer to be able to "see" the
653 list of files backed up in typical GUI tree format.
654 The same mechanism could also be used for creating
655 ad-hoc backup FileSets (item 8).
657 Why: Ease of use -- especially for the end user.
659 Notes: Rather than implementing in Gtk, we probably should go directly
660 for a Browser implementation, even if doing so meant the
661 capability wouldn't be available until much later. Not only
662 is there the question of Windows sites, most
663 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
664 without installing lots of stuff admins are very wary about.
665 Real sysadmins will always use the command line anyway, and
666 the user who's doing an interactive restore or backup of his
667 own files will in most cases be on a Windows machine running
671 Item 9: Add SSL to daemon communications.
673 What: This provides for secure communications between the daemons.
675 Why: This would allow doing backup across the Internet without
676 privacy concerns (or with much less concern).
678 Notes: The vast majority of near term potential users will be backing up
679 a single site over a LAN and, correctly or not, they probably
680 won't be concerned with security, at least not enough to go to
681 the trouble to set up keys, etc. to screw things down. We suspect
682 that many users genuinely interested in multi-site backup
683 already run some form of VPN software in their internetwork
684 connections, and are willing to delegate security to that layer.
687 Item 10: Define definitive tape format.
690 What: Define that definitive tape format that will not change
691 for the next millennium.
693 Why: Stability, security.
695 Notes: See notes for item 11 below.
698 Item 11: New daemon communication protocol.
700 What: The current daemon to daemon protocol is basically an ASCII
701 printf() and sending the buffer. On the receiving end, the
702 buffer is sscanf()ed to unpack it. The new scheme would
703 be a binary format that allows quick packing and unpacking
704 of any data type with named fields.
706 Why: Using binary packing would be faster. Named fields will permit
707 error checking to ensure that what is sent is what the
708 receiver really wants.
710 Notes: These are internal improvements in the interest of the
711 long-term stability and evolution of the program. On the one
712 hand, the sooner they're done, the less code we have to rip
713 up when the time comes to install them. On the other hand, they
714 don't bring an immediately perceptible benefit to potential
715 users. Item 10 and possibly item 11 should be deferred until Bacula
716 is well established with a growing user community more or
717 less happy with the feature set. At that time, it will make a
718 good "next generation" upgrade in the interest of data
724 ======================================================
726 It is somewhat like a Full save becomes an incremental since
727 the Base job (or jobs) plus other non-base files.
729 - New BaseFile table that contains:
730 JobId, BaseJobId, FileId (from Base).
731 i.e. for each base file that exists but is not saved because
732 it has not changed, the File daemon sends the JobId, BaseId,
733 and FileId back to the Director who creates the DB entry.
734 - To initiate a Base save, the Director sends the FD
735 the FileId, and full filename for each file in the Base.
736 - When the FD finds a Base file, he requests the Director to
737 send him the full File entry (stat packet plus MD5), or
738 conversely, the FD sends it to the Director and the Director
739 says yes or no. This can be quite rapid if the FileId is kept
740 by the FD for each Base Filename.
741 - It is probably better to have the comparison done by the FD
742 despite the fact that the File entry must be sent across the
744 - An alternative would be to send the FD the whole File entry
745 from the start. The disadvantage is that it requires a lot of
746 space. The advantage is that it requires less communications
748 - The Job record must be updated to indicate that one or more
750 - At end of Job, FD returns:
751 1. Count of base files/bytes not written to tape (i.e. matches)
752 2. Count of base file that were saved i.e. had changed.
753 - No tape record would be written for a Base file that matches, in the
754 same way that no tape record is written for Incremental jobs where
755 the file is not saved because it is unchanged.
756 - On a restore, all the Base file records must explicitly be
757 found from the BaseFile tape. I.e. for each Full save that is marked
758 to have one or more Base Jobs, search the BaseFile for all occurrences
760 - An optimization might be to make the BaseFile have:
766 This would avoid the need to explicitly fetch each File record for
767 the Base job. The Base Job record will be fetched to get the
768 VolSessionId and VolSessionTime.
769 =========================================================
773 =============================================================
775 Request For Comments For File Backup Options
778 Subject: File Backup Options
781 A few days ago, a Bacula user who is backing up to file volumes and
782 using compression asked if it was possible to suppress compressing
783 all .gz files since it was a waste of CPU time. Although Bacula
784 currently permits using different options (compression, ...) on
785 a directory by directory basis, it cannot do it on a file by
786 file basis, which is clearly what was desired.
788 Proposed Implementation:
789 To solve this problem, I propose the following:
791 - Add a new Director resource type called Options.
793 - The Options resource will have records for all
794 options that can currently be specified on the Include record
795 (in a FileSet). Examples below.
797 - The Options resource will permit an exclude option as well
798 as a number of additional options.
800 - The heart of the Options resource is the ability to
801 supply any number of Match records which specify POSIX
802 regular expressions. These Match regular expressions are
803 applied to the fully qualified filename (path and all). If
804 one matches, then the Options will be used.
806 - When an Match specification matches an included file, the
807 options specified in the Options resource will override
808 the default options specified on the Include record.
810 - Include records will be modified to permit referencing one or
811 more Options resources. The Options will be used
812 in the order listed on the Include record and the first
813 one that matches will be applied.
815 - Options (or specifications) currently supplied on the Include
816 record will be deprecated (i.e. removed in a later version a
817 year or so from now).
819 - The Exclude record will be deprecated as the same functionality
820 can be obtained by using an Exclude = yes in the Options.
823 The following records can appear in the Options resource. An
824 asterisk preceding the name indicates a feature not currently
828 - Compression= (GZIP, ...)
829 - Signature= (MD5, SHA1, ...)
831 - OneFs= (yes/no) - remain on one filesystem
832 - Recurse= (yes/no) - recurse into subdirectories
833 - Sparse= (yes/no) - do sparse file backup
834 - *Exclude= (yes/no) - exclude file from being saved
835 - *Reader= (filename) - external read (backup) program
836 - *Plugin= (filename) - read/write plugin module
839 - verify= (ipnougsamc5) - verify options
842 - replace= (always/ifnewer/ifolder/never) - replace options currently
844 - *Writer= (filename) - external write (restore) program
848 Currently options specifying compression, MD5 signatures, recursion,
849 ... of a FileSet are supplied on the Include record. These will now
850 all be collected into a Options resource, which will be
851 specified in the Include in place of the options. Multiple Options
852 may be specified. Since the Options may contain regular expressions
853 that are applied to the full filename, this will give the ability
854 to specify backup options on a file by file basis to whatever level
863 Include = compression=GZIP signature=MD5 {
875 Match = /*.?*/ # matches all files.
880 That's a lot more to do the same thing, but it gives the ability to
881 apply options on a file by file basis. For example, suppose you
882 want to compress all files but not any file with extensions .gz or .Z.
883 In that case, you will need to group two sets of options using
884 the Options resource as follows:
892 # Note multiple Matches are ORed
893 Match = /*.gz/ # matches .gz files */
894 Match = /*.Z/ # matches .Z files */
899 Match = /*.?*/ # matches all files
905 Now, since the no Compression option is specified in the
906 first group of Options, *.gz or *.Z file will have an MD5 signature computed,
907 but will not be compressed. For all other files, the *.gz *.Z will not
908 match, so the second group of options will be used which will include GZIP
912 - Is it necessary to provide some means of ANDing regular expressions
913 and negation? (not currently planned)
915 e.g. Match = /*.gz/ && !/big.gz/
917 - I see that Networker has a "null" module which, if specified, does not
918 backup the file, but does make an record of the file in the catalog
919 so that the catalog will reflect an exact picture of the filesystem.
920 The result is that the file can be "seen" when "browsing" the save
921 sets, but it cannot be restored.
923 Is this really useful? Should it be implemented in Bacula?
926 After implementing the above, the user will be able to specify
927 on a file by file basis (using regular expressions) what options are
928 applied for the backup.
931 =============================================
933 ==========================================================
935 For each Incremental job that is run, there may be files that
936 were found but not saved because they were locked (this applies
937 only to Windows). Such a system could send back to the Director
938 a list of Unsaved files.
940 - New UnSavedFiles table that contains:
944 - Then in the next Incremental job, the list of Unsaved Files will be
945 feed to the FD, who will ensure that they are explicitly chosen even
946 if standard date/time check would not have selected them.
947 =============================================================
950 Done: (see kernsdone for more)
951 - Implement new alist in FileSet scanning.
952 - bls should continue reading even if it finds Win32 data on the tape.
953 The error should be Warning rather the Error.
954 - Add user configurable timeout for connecting to SD.
955 - Unsaved Flag in Job record (use JobMissingFiles).
956 - Base Flag in Job record.
957 - Configure mtx-changer to have correct path to mtx.
958 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
959 - Make some way so that if a machine is skipped because it is not up
960 that Bacula will continue retrying for a specified period of time --
962 - Implement all command line args on run.
963 - Implement command line "restore" args.
964 - Implement "restore current select=no"
965 - Restore file modified before date
966 - Restore -- do nothing but show what would happen
967 - Add estimate to Console commands
968 - Use read_record.c in SD code.
969 - Fix read_record to handle multiple sessions.
970 - Tip from Steve Allam
971 mt -f /dev/nst0 defblksize 0
972 - Document "status" in the console.
973 - Document driving console from shell script.
974 - Write JobMedia records with max file size is reached on tape.
975 - Handle the case of multiple JobMedia records pending (i.e. the
976 thread is slow and multiple situations requiring a JobMedia
978 - Do performance analysis on the restore tree routines.
979 - Fix maximum file size (block.c) to generate JobMedia records.
980 - Make the default file size 1GB on the tape.
981 - Implement forward spacing between files.
982 - Add Machine type (Linux/Windows) to Status report for daemons.
984 - Use repositioning at the beginning of the tape.
985 - Do full check the command line args in update (e.g. VolStatus ...).
986 - Specify list of files to restore
987 - Implement ClientRunBeforeJob and ClientRunAfterJob.
988 - Make | and < work on FD side.
989 - Check to see if "blocked" is set during restore.
990 - Figure out what is interrupting sql command in console.
991 - Make new job print warning User Unmounted Tape.
992 - Test recycling and purging (code changed in db_find_next_volume and
994 - Document SDConnectTimeout (in FD).
995 - Add restore by filename test.
996 - Document restore by files.
997 - Make variable expansion work correctly.
998 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
999 - Copy static programs into install directory.
1000 - Think about changing Storage resource Device record to be
1002 - Add RunBeforeJob and RunAfterJob to the Client program.
1003 - Need return status on read_cb() from read_records(). Need multiple
1004 records -- one per Job, maybe a JCR or some other structure with
1005 a block and a record.
1006 - LabelFormat on tape volume apparently creates the db record but
1007 never actually labels the volume.
1008 - Recycling a volume when two jobs are using it is going to break. Fixed.
1009 - Document list nextvol and new format status dir.
1010 - Client files in Win32 with Unix eol conventions doesn't work.
1011 - Either fix or document that fill command in btape can be
1012 compressed enormously by the hardware - a 36GB tape wrote 750GB!
1013 - Add multiple character duration qualifiers.
1014 - Require some modifer.
1015 - Restrict characters permitted in a Resource name, and don't permit
1017 - Figure out some way to ignore or get past checksum errors in
1019 - The SD spooling file gets created even if it is not used.
1020 - Look at Cleaning tape in ua_label.c for media create/update
1021 - Add regression testing to the manual
1022 - End time: in job output of rescheduled job is time of first run.
1023 - Document list nextvol and status output.
1024 - Separate Dir heartbeat in FD from the SD heartbeat.
1025 - Fix sparse file handeling so that it always reads a multiple
1026 of 512. Currently, it subtracts 8 bytes (for faddr).
1027 Kludged with #ifdef for FreeBSD.
1028 - Document that Volume pruning can delete last Full backup and
1029 hence you will not have a valid backup.
1030 - Clarify the fact that having the Bacula cygwin1.dll loaded
1031 is not the same as having cygwin installed.
1032 - Document that it is safe to use the drive when the lights stop flashing.
1033 - Document all the status codes JobLevel, JobType, JobStatus.
1034 - Add GUI interface to manual
1035 - Combine the 3 places that search run records for the next
1036 job. Use find_job_pool() modified in ua_output.c
1037 - Test connect timeouts.
1038 - Fix FreeBSD build with tcp_wrapper -- should not have -lnsl