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 - Document FInclude ...
17 - Document all the status codes JobLevel, JobType, JobStatus.
19 Testing to do: (painful)
20 - that ALL console command line options work and are always implemented
21 - blocksize recognition code.
22 - Test if rewind at end of tape waits for tape to rewind.
24 - Test not zeroing Autochanger slot when it is wrong.
25 - Test recycling and purging (code changed in db_find_next_volume and
27 - Figure out how to use ssh or stunnel to protect Bacula communications.
30 - Look at libkse (man kse) for FreeBSD threading.
31 - Look into Microsoft Volume Shadowcopy Service VSS for backing
32 up system state components (Active Directory, System Volume, ...)
33 - Add ExhautiveRestoreSearch and use repositioning at the beginning
35 - Add Machine type (Linux/Windows) to Status report for daemons.
36 - Document "status" in the console.
37 - Document driving console from shell script.
38 - Specify list of files to restore
39 - Implement ClientRunBeforeJob and ClientRunAfterJob.
41 - Write JobMedia records with max file size is reached on tape.
42 - Look at the possibility of loading only the necessary
43 data into the restore tree (i.e. do it one directory at a
44 time as the user walks through the tree).
45 - Possibly use the hash code if the user selects all.
46 - Handle the case of multiple JobMedia records pending (i.e. the
47 thread is slow and multiple situations requiring a JobMedia
49 - Fix maximum file size (block.c) to generate JobMedia records.
50 - Do performance analysis on the restore tree routines.
51 - Implement forward spacing between files.
52 - Make the default file size 1GB on the tape.
53 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
54 - Implement some way for the File daemon to contact the Director
55 to start a job or pass its DHCP obtained IP number.
56 - Implement multiple Consoles.
57 - Add Console usr permissions.
58 - Fix "restore all" to bypass building the tree.
59 - Fix restore to list errors if Invalid block found, and if # files
60 restored does not match # expected.
61 - Prohibit backing up archive device (findlib/find_one.c:128)
62 - Make | and < work on FD side.
63 - Implement Release Device in the Job resource to unmount a drive.
64 - Implement Acquire Device in the Job resource to mount a drive,
65 be sure this works with admin jobs so that the user can get
66 prompted to insert the correct tape. Possibly some way to say to
67 run the job but don't save the files.
68 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
69 - Implement FileOptions (see end of this document)
70 - Enhance autochanger interface to include a "scan" feature
71 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
72 when the magazine is changed.
73 - Make things like list where a file is saved case independent for
75 - Edit the Client/Storage name into authentication failure messages.
76 - Implement job in VerifyToCatalog
78 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
79 - Allow a slot or range of slots on the label barcodes command.
80 - Implement a PostgreSQL driver.
81 - Is a pool specification really needed for a restore?
82 - Bacula needs to propagate SD errors.
83 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
84 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
85 /dev/nst0. ERR=Input/output error.
86 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
87 Last block=5162 Current block=5164.
88 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
90 - Use autochanger to handle multiple devices.
91 - Fix packet too big problem.
92 - Add SuSE install doc to list.
93 - Check and rechedk "Invalid block number"
94 - Make bextract release the drive properly between tapes
95 so that an autochanger can be made to work.
96 - User wants to NOT backup up certain big files (email files).
97 - Maybe remove multiple simultaneous devices code in SD.
98 - On Windows with very long path names, it may be impossible to create
99 a file (and thus restore it) because the total length is too long.
100 We must cd into the directory then create the file without the
102 - lstat() is not going to work on Win32 for testing date.
103 - Something is not right in last block of fill command.
104 - Implement a Recycle command
105 - Add FileSet to command line arguments for restore.
106 - Do full check the command line args in update (e.g. VolStatus ...).
107 - Allow multiple Storage specifications (or multiple names on
108 a single Storage specification) in the Job record. Thus a job
109 can be backed up to a number of storage devices.
110 - Add client name to cram-md5 challenge so Director can immediately
111 verify if it is the correct client.
112 - Implement ClientRunBeforeJob and ClientRunAfterJob.
113 - Add JobLevel in FD status (but make sure it is defined).
114 - Audit all UA commands to ensure that we always prompt where possible.
115 - Restrict characters permitted in a Resource name, and don't permit
117 - Check Jmsg in bnet, may not work, must dup bsock.
118 - Suppress Job Name in Jmsg for console
119 - Create Pools that are referenced in a Run statement at startup if possible.
120 - Use runbeforejob to unload, then reload a volume previously used,
121 then the next job run gets an error reading the drive.
122 - Make bootstrap filename unique.
123 - Test a second language e.g. french.
124 - Start working on Base jobs.
125 - Make "make binary-release" work from any directory.
126 - Implement UnsavedFiles DB record.
127 - Implement argc/argv for daemon command line scanning using table driven
129 - Implement table driven single argc/argv scanner to pickup all arguments.
130 Much like xxx_conf.c scan table.
131 keyword, handler(store_routine), store_address, code, flags, default.
132 - From Phil Stracchino:
133 It would probably be a per-client option, and would be called
134 something like, say, "Automatically purge obsoleted jobs". What it
135 would do is, when you successfully complete a Differential backup of a
136 client, it would automatically purge all Incremental backups for that
137 client that are rendered redundant by that Differential. Likewise,
138 when a Full backup on a client completed, it would automatically purge
139 all Differential and Incremental jobs obsoleted by that Full backup.
140 This would let people minimize the number of tapes they're keeping on
141 hand without having to master the art of retention times.
142 - Implement new serialize subroutines
143 send(socket, "string", &Vol, "uint32", &i, NULL)
144 - Scratch Pool where the volumes can be re-assigned to any Pool.
145 - Implement a M_SECURITY message class.
146 - Implement forward spacing block/file: position_device(bsr) --
147 just before read_block_from_device();
148 - When doing a Backup send all attributes back to the Director, who
149 would then figure out what files have been deleted.
150 - Currently in mount.c:236 the SD simply creates a Volume. It should have
151 explicit permission to do so. It should also mark the tape in error
152 if there is an error.
153 - Make sure all restore counters are working correctly in the FD.
154 - SD Bytes Read is wrong.
155 - Look at ALL higher level routines that call block.c to be sure
156 they don't expect something in errmsg.
157 - Investigate doing RAW backup of Win32 partition.
158 - Add thread specific data to hold the jcr -- send error messages from
159 low level routines by accessing it and using Jmsg().
160 - Cancel waiting for Client connect in SD if FD goes away.
162 - Add Progress command that periodically reports the progress of
164 - One block was orphaned in the SD probably after cancel.
167 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
168 - Implement timeout in response() when it should come quickly.
169 - Implement console @echo command.
170 - Implement a Slot priority (loaded/not loaded).
171 - Implement "vacation" Incremental only saves.
172 - Implement single pane restore (much like the Gftp panes).
173 - Implement Automatic Mount even in operator wait.
174 - Implement create "FileSet"?
175 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
176 - Implement "scratch pool" where tapes are defined and can be
177 taken by any pool that needs them.
178 - Implement restore "current system", but take all files without
179 doing selection tree -- so that jobs without File records can
181 - Implement disk spooling. Two parts: 1. Spool to disk then
182 immediately to tape to speed up tape operations. 2. Spool to
183 disk only when the tape is full, then when a tape is hung move
185 - Implement a relocatable bacula.spec
186 - Implement dump/print label to UA
187 - Add prefixlinks to where or not where absolute links to FD.
188 - Issue message to mount a new tape before the rewind.
189 - Simplified client job initiation for portables.
190 - If SD cannot open a drive, make it periodically retry.
191 - Implement LabelTemplate (at least first cut).
192 - Add more of the config info to the tape label.
193 - Implement a Mount Command and an Unmount Command where
194 the users could specify a system command to be performed
195 to do the mount, after which Bacula could attempt to
196 read the device. This is for Removeable media such as a CDROM.
197 - Most likely, this mount command would be invoked explicitly
198 by the user using the current Console "mount" and "unmount"
199 commands -- the Storage Daemon would do the right thing
200 depending on the exact nature of the device.
201 - As with tape drives, when Bacula wanted a new removable
202 disk mounted, it would unmount the old one, and send a message
203 to the user, who would then use "mount" as described above
204 once he had actually inserted the disk.
206 - If tape is marked read-only, then try opening it read-only rather than
207 failing, and remember that it cannot be written.
208 - Refine SD waiting output:
209 Device is being positioned
210 > Device is being positioned for append
211 > Device is being positioned to file x
213 - Figure out some way to estimate output size and to avoid splitting
214 a backup across two Volumes -- this could be useful for writing CDROMs
215 where you really prefer not to have it split -- not serious.
216 - Add RunBeforeJob and RunAfterJob to the Client program.
217 - Have SD compute MD5 or SHA1 and compare to what FD computes.
218 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
219 actual data on the Volume and compare it.
220 - Implement Bacula plugins -- design API
221 - Make bcopy read through bad tape records.
222 - Program files (i.e. execute a program to read/write files).
223 Pass read date of last backup, size of file last time.
224 - Add Signature type to File DB record.
225 - CD into subdirectory when open()ing files for backup to
226 speed up things. Test with testfind().
227 - Priority job to go to top of list.
228 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
229 - Implement some way for the Console to dynamically create a job.
230 - Restore to a particular time -- e.g. before date, after date.
231 - Solaris -I on tar for include list
232 - Need a verbose mode in restore, perhaps to bsr.
233 - bscan without -v is too quiet -- perhaps show jobs.
234 - Add code to reject whole blocks if not wanted on restore.
235 - Check if we can increase Bacula FD priorty in Win2000
236 - Make sure the MaxVolFiles is fully implemented in SD
237 - Check if both CatalogFiles and UseCatalog are set to SD.
238 - Need return status on read_cb() from read_records(). Need multiple
239 records -- one per Job, maybe a JCR or some other structure with
240 a block and a record.
241 - Figure out how to do a bare metal Windows restore
242 - Possibly add email to Watchdog if drive is unmounted too
243 long and a job is waiting on the drive.
244 - Restore program that errs in SD due to no tape, reports
245 OK incorrectly in output.
246 - After unmount, if restore job started, ask to mount.
247 - Convert all %x substitution variables, which are hard to remember
248 and read to %(variable-name). Idea from TMDA.
249 - Remove NextId for SQLite. Optimize.
250 - Move all SQL statements into a single location.
251 - Add UA rc and history files.
252 - put termcap (used by console) in ./configure and
253 allow -with-termcap-dir.
254 - Enhance time and size scanning routines.
255 - Fix Autoprune for Volumes to respect need for full save.
256 - Fix Win32 config file definition name on /install
257 - Compare tape to Client files (attributes, or attributes and data)
258 - Make all database Ids 64 bit.
259 - Write an applet for Linux.
260 - Implement new inter-daemon communications protocol.
261 - Allow console commands to detach or run in background.
262 - Fix status delay on storage daemon during rewind.
263 - Add SD message variables to control operator wait time
264 - Maximum Operator Wait
265 - Minimum Message Interval
266 - Maximum Message Interval
267 - Send Operator message when cannot read tape label.
268 - Verify level=Volume (scan only), level=Data (compare of data to file).
269 Verify level=Catalog, level=InitCatalog
271 - Add keyword search to show command in Console.
272 - Events : tape has more than xxx bytes.
273 - Complete code in Bacula Resources -- this will permit
274 reading a new config file at any time.
275 - Handle ctl-c in Console
276 - Implement script driven addition of File daemon to config files.
277 - Think about how to make Bacula work better with File (non-tape) archives.
278 - Write Unix emulator for Windows.
279 - Put memory utilization in Status output of each daemon
280 if full status requested or if some level of debug on.
281 - Make database type selectable by .conf files i.e. at runtime
282 - Set flag for uname -a. Add to Volume label.
283 - Implement throttled work queue.
284 - Restore files modified after date
285 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
286 - Implement Restore FileSet=
287 - Create a protocol.h and protocol.c where all protocol messages
289 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
290 - Timout a job or terminate if link goes down, or reopen link and query.
291 - Concept of precious tapes (cannot be reused).
292 - Make bcopy copy with a single tape drive.
293 - Permit changing ownership during restore.
296 > My suggestion: Add a feature on the systray menu-icon menu to request
297 > an immediate backup now. This would be useful for laptop users who may
298 > not be on the network when the regular scheduled backup is run.
300 > My wife's suggestion: Add a setting to the win32 client to allow it to
301 > shut down the machine after backup is complete (after, of course,
302 > displaying a "System will shut down in one minute, click here to cancel"
303 > warning dialog). This would be useful for sites that want user
304 > woorkstations to be shut down overnight to save power.
307 - From Terry Manderson <terry@apnic.net>
308 jobset { # new structure
309 name = "monthlyUnixBoxen"
312 jobs = "wakame;durian;soy;wasabi;miso" #new!
322 fileset = "genericUnixSet"
328 fileset = "genericUnixSet"
334 fileset = "UnixDevelBoxSet"
339 - Autolabel should be specified by DIR instead of SD.
342 - AutoScan (check checksum of tape)
343 - Format command = "format /dev/nst0"
347 - Seek resolution (usually corresponds to buffer size)
348 - EODErrorCode=ENOSPC or code
349 - Partial Read error code
350 - Partial write error code
351 - Nonformatted read error
352 - Nonformatted write error
353 - WriteProtected error
357 - IgnoreCloseErrors=yes
367 - FD sends unsaved file list to Director at end of job (see
369 - File daemon should build list of files skipped, and then
370 at end of save retry and report any errors.
371 - Write a Storage daemon that uses pipes and
372 standard Unix programs to write to the tape.
374 - Need something that monitors the JCR queue and
375 times out jobs by asking the deamons where they are.
376 - Enhance Jmsg code to permit buffering and saving to disk.
377 - device driver = "xxxx" for drives.
379 - Ensure that /dev/null works
380 - Need report class for messages. Perhaps
381 report resource where report=group of messages
382 - enhance scan_attrib and rename scan_jobtype, and
383 fill in code for "since" option
384 - Director needs a time after which the report status is sent
385 anyway -- or better yet, a retry time for the job.
386 - Don't reschedule a job if previous incarnation is still running.
387 - Some way to automatically backup everything is needed????
388 - Need a structure for pending actions:
390 - termination status (part of buffered msgs?)
392 Read, Write, Clean, Delete
393 - Login to Bacula; Bacula users with different permissions:
394 owner, group, user, quotas
395 - Store info on each file system type (probably in the job header on tape.
396 This could be the output of df; or perhaps some sort of /etc/mtab record.
399 - Design at hierarchial storage for Bacula. Migration and Clone.
400 - Implement FSM (File System Modules).
401 - Audit M_ error codes to ensure they are correct and consistent.
402 - Add variable break characters to lex analyzer.
403 Either a bit mask or a string of chars so that
404 the caller can change the break characters.
405 - Make a single T_BREAK to replace T_COMMA, etc.
406 - Ensure that File daemon and Storage daemon can
407 continue a save if the Director goes down (this
408 is NOT currently the case). Must detect socket error,
409 buffer messages for later.
410 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
411 - Add ability to backup to two Storage devices (two SD sessions) at
412 the same time -- e.g. onsite, offsite.
413 - Add the ability to consolidate old backup sets (basically do a restore
414 to tape and appropriately update the catalog). Compress Volume sets.
415 Might need to spool via file is only one drive is available.
416 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
417 someway to do so with every volume that has less than x% valid
421 Migration: Move a backup from one Volume to another
422 Clone: Copy a backup -- two Volumes
424 Bacula Migration is based on Jobs (apparently Networker is file by file).
426 Migration triggered by:
430 Highwater mark (keep total size)
434 Bacula Projects Roadmap
436 last update 8 May 2003
438 Item 1: Multiple simultaneous Jobs. (done)
439 Done -- Restore part needs better implementation to work correctly
440 Also, it needs considerable testing
442 What: Permit multiple simultaneous jobs in Bacula.
444 Why: An enterprise level solution needs to go fast without the
445 need for the system administrator to carefully tweak
446 timing. Based on the benchmarks, during a full
447 backup, NetWorker typically hit 10 times the bandwidth to
448 the tape compared to Bacula--largely. This is probably due to
449 running parallel jobs and multi-threaded filling of buffers
450 and writing them to tape. This should also make things work
451 better when you have a mix of fast and slow machines backing
454 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
455 implementing this is a matter of some small cleanups and
459 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
460 Deferred -- not necessary yet -- possibly implement with Migration.
462 What: If data is coming into the SD too fast, buffer it to
463 disk if the user has configured this option.
465 Why: This would be nice, especially if it more or less falls out
466 when implementing (1) above. If not, it probably should not
467 be given a high priority because fundamentally the backup time
468 is limited by the tape bandwidth. Even though you may finish a
469 client job quicker by spilling to disk, you still have to
470 eventually get it onto tape. If intermediate disk buffering
471 allows us to improve write bandwidth to tape, it may make
474 Notes: Whether or not this is implemented will depend upon performance
475 testing after item 1 is implemented.
478 Item 3: Write the bscan program -- also write a bcopy program.
481 What: Write a program that reads a Bacula tape and puts all the
482 appropriate data into the catalog. This allows recovery
483 from a tape that is no longer in the database, or it allows
484 re-creation of a database if lost.
486 Why: This is a fundamental robustness and disaster recovery tool
487 which will increase the comfort level of a sysadmin
488 considering adopting Bacula.
490 Notes: A skeleton of this program already exists, but much work
491 needs to be done. Implementing this will also make apparent
492 any deficiencies in the current Bacula tape format.
495 Item 4: Implement Base jobs.
497 What: A base job is sort of like a Full save except that you
498 will want the FileSet to contain only files that are unlikely
499 to change in the future (i.e. a snapshot of most of your
500 system after installing it). After the base job has been run,
501 when you are doing a Full save, you can specify to exclude
502 all files saved by the base job that have not been modified.
504 Why: This is something none of the competition does, as far as we know
505 (except BackupPC, which is a Perl program that saves to disk
506 only). It is big win for the user, it makes Bacula stand out
507 as offering a unique optimization that immediately saves time
510 Notes: Big savings in tape usage. Will require more resources because
511 the e. DIR must send FD a list of files/attribs, and the FD must
512 search the list and compare it for each file to be saved.
515 Item 5: Implement Label templates
518 What: This is a mechanism whereby Bacula can automatically create
519 a tape label for new tapes according to a detailed specification
520 provided by the user.
522 Why: It is a major convenience item for folks who use automated label
525 Notes: Bacula already has a working form of automatic tape label
526 creation, but it is very crude. The design for the complete
527 tape labeling project is already documented in the manual.
530 Item 6: Write a regression script.
531 Done -- Continue to expand its testing.
533 What: This is an automatic script that runs and tests as many features
534 of Bacula as possible. The output is compared to previous
535 versions of Bacula and any differences are reported.
537 Why: This is an enormous help in preventing introduction of new
538 errors in parts of the program that already work correctly.
540 Notes: This probably should be ranked higher, it's something the typical
541 user doesn't see. Depending on how it's implemented, it may
542 make sense to defer it until the archival tape format and
543 user interface mature.
546 Item 7: GUI for interactive restore
547 Item 8: GUI for interactive backup
549 What: The current interactive restore is implemented with a tty
550 interface. It would be much nicer to be able to "see" the
551 list of files backed up in typical GUI tree format.
552 The same mechanism could also be used for creating
553 ad-hoc backup FileSets (item 8).
555 Why: Ease of use -- especially for the end user.
557 Notes: Rather than implementing in Gtk, we probably should go directly
558 for a Browser implementation, even if doing so meant the
559 capability wouldn't be available until much later. Not only
560 is there the question of Windows sites, most
561 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
562 without installing lots of stuff admins are very wary about.
563 Real sysadmins will always use the command line anyway, and
564 the user who's doing an interactive restore or backup of his
565 own files will in most cases be on a Windows machine running
569 Item 9: Add SSL to daemon communications.
571 What: This provides for secure communications between the daemons.
573 Why: This would allow doing backup across the Internet without
574 privacy concerns (or with much less concern).
576 Notes: The vast majority of near term potential users will be backing up
577 a single site over a LAN and, correctly or not, they probably
578 won't be concerned with security, at least not enough to go to
579 the trouble to set up keys, etc. to screw things down. We suspect
580 that many users genuinely interested in multi-site backup
581 already run some form of VPN software in their internetwork
582 connections, and are willing to delegate security to that layer.
585 Item 10: Define definitive tape format.
588 What: Define that definitive tape format that will not change
589 for the next millennium.
591 Why: Stability, security.
593 Notes: See notes for item 11 below.
596 Item 11: New daemon communication protocol.
598 What: The current daemon to daemon protocol is basically an ASCII
599 printf() and sending the buffer. On the receiving end, the
600 buffer is sscanf()ed to unpack it. The new scheme would
601 be a binary format that allows quick packing and unpacking
602 of any data type with named fields.
604 Why: Using binary packing would be faster. Named fields will permit
605 error checking to ensure that what is sent is what the
606 receiver really wants.
608 Notes: These are internal improvements in the interest of the
609 long-term stability and evolution of the program. On the one
610 hand, the sooner they're done, the less code we have to rip
611 up when the time comes to install them. On the other hand, they
612 don't bring an immediately perceptible benefit to potential
613 users. Item 10 and possibly item 11 should be deferred until Bacula
614 is well established with a growing user community more or
615 less happy with the feature set. At that time, it will make a
616 good "next generation" upgrade in the interest of data
622 ======================================================
624 It is somewhat like a Full save becomes an incremental since
625 the Base job (or jobs) plus other non-base files.
627 - New BaseFile table that contains:
628 JobId, BaseJobId, FileId (from Base).
629 i.e. for each base file that exists but is not saved because
630 it has not changed, the File daemon sends the JobId, BaseId,
631 and FileId back to the Director who creates the DB entry.
632 - To initiate a Base save, the Director sends the FD
633 the FileId, and full filename for each file in the Base.
634 - When the FD finds a Base file, he requests the Director to
635 send him the full File entry (stat packet plus MD5), or
636 conversely, the FD sends it to the Director and the Director
637 says yes or no. This can be quite rapid if the FileId is kept
638 by the FD for each Base Filename.
639 - It is probably better to have the comparison done by the FD
640 despite the fact that the File entry must be sent across the
642 - An alternative would be to send the FD the whole File entry
643 from the start. The disadvantage is that it requires a lot of
644 space. The advantage is that it requires less communications
646 - The Job record must be updated to indicate that one or more
648 - At end of Job, FD returns:
649 1. Count of base files/bytes not written to tape (i.e. matches)
650 2. Count of base file that were saved i.e. had changed.
651 - No tape record would be written for a Base file that matches, in the
652 same way that no tape record is written for Incremental jobs where
653 the file is not saved because it is unchanged.
654 - On a restore, all the Base file records must explicitly be
655 found from the BaseFile tape. I.e. for each Full save that is marked
656 to have one or more Base Jobs, search the BaseFile for all occurrences
658 - An optimization might be to make the BaseFile have:
664 This would avoid the need to explicitly fetch each File record for
665 the Base job. The Base Job record will be fetched to get the
666 VolSessionId and VolSessionTime.
667 =========================================================
671 =============================================================
673 Request For Comments For File Backup Options
676 Subject: File Backup Options
679 A few days ago, a Bacula user who is backing up to file volumes and
680 using compression asked if it was possible to suppress compressing
681 all .gz files since it was a waste of CPU time. Although Bacula
682 currently permits using different options (compression, ...) on
683 a directory by directory basis, it cannot do it on a file by
684 file basis, which is clearly what was desired.
686 Proposed Implementation:
687 To solve this problem, I propose the following:
689 - Add a new Director resource type called FileOptions.
691 - The FileOptions resource will have records for all
692 options that can currently be specified on the Include record
693 (in a FileSet). Examples below.
695 - The FileOptions resource will permit an exclude option as well
696 as a number of additional options.
698 - The heart of the FileOptions resource is the ability to
699 supply any number of ApplyTo records which specify POSIX
700 regular expressions. These ApplyTo regular expressions are
701 applied to the fully qualified filename (path and all). If
702 one matches, then the FileOptions will be used.
704 - When an ApplyTo specification matches an included file, the
705 options specified in the FileOptions resource will override
706 the default options specified on the Include record.
708 - Include records will be modified to permit referencing one or
709 more FileOptions resources. The FileOptions will be used
710 in the order listed on the Include record and the first
711 one that matches will be applied.
713 - Options (or specifications) currently supplied on the Include
714 record will be deprecated (i.e. removed in a later version a
715 year or so from now).
717 - The Exclude record will be deprecated as the same functionality
718 can be obtained by using an Exclude = yes in the FileOptions.
721 The following records can appear in the FileOptions resource. An
722 asterisk preceding the name indicates a feature not currently
726 - Compression= (GZIP, ...)
727 - Signature= (MD5, SHA1, ...)
729 - OneFs= (yes/no) - remain on one filesystem
730 - Recurse= (yes/no) - recurse into subdirectories
731 - Sparse= (yes/no) - do sparse file backup
732 - *Exclude= (yes/no) - exclude file from being saved
733 - *Reader= (filename) - external read (backup) program
734 - *Plugin= (filename) - read/write plugin module
737 - verify= (ipnougsamc5) - verify options
740 - replace= (always/ifnewer/ifolder/never) - replace options currently
742 - *Writer= (filename) - external write (restore) program
746 Currently options specifying compression, MD5 signatures, recursion,
747 ... of a FileSet are supplied on the Include record. These will now
748 all be collected into a FileOptions resource, which will be
749 specified on the Include in place of the options. Multiple FileOptions
750 may be specified. Since the FileOptions contain regular expressions
751 that are applied to the full filename, this will give the ability
752 to specify backup options on a file by file basis to whatever level
761 Include = compression=GZIP signature=MD5 {
773 Match = /*.?*/ # matches all files.
778 That's a lot more to do the same thing, but it gives the ability to
779 apply options on a file by file basis. For example, suppose you
780 want to compress all files but not any file with extensions .gz or .Z.
781 You could do so as follows:
788 # Note multiple Matches are ORed
789 Match = /*.gz/ # matches .gz files */
790 Match = /*.Z/ # matches .Z files */
795 Match = /*.?*/ # matches all files
801 Now, since the NoCompress FileOptions is specified first on the
802 Include line, any *.gz or *.Z file will have an MD5 signature computed,
803 but will not be compressed. For all other files, the NoCompress will not
804 match, so the Opts options will be used which will include GZIP
808 - Is it necessary to provide some means of ANDing regular expressions
809 and negation? (not currently planned)
811 e.g. Match = /*.gz/ && !/big.gz/
813 - I see that Networker has a "null" module which, if specified, does not
814 backup the file, but does make an record of the file in the catalog
815 so that the catalog will reflect an exact picture of the filesystem.
816 The result is that the file can be "seen" when "browsing" the save
817 sets, but it cannot be restored.
819 Is this really useful? Should it be implemented in Bacula?
822 After implementing the above, the user will be able to specify
823 on a file by file basis (using regular expressions) what options are
824 applied for the backup.
827 =============================================
829 ==========================================================
831 For each Incremental job that is run, there may be files that
832 were found but not saved because they were locked (this applies
833 only to Windows). Such a system could send back to the Director
834 a list of Unsaved files.
836 - New UnSavedFiles table that contains:
840 - Then in the next Incremental job, the list of Unsaved Files will be
841 feed to the FD, who will ensure that they are explicitly chosen even
842 if standard date/time check would not have selected them.
843 =============================================================
846 Done: (see kernsdone for more)
847 - Implement new alist in FileSet scanning.
848 - bls should continue reading even if it finds Win32 data on the tape.
849 The error should be Warning rather the Error.
850 - Add user configurable timeout for connecting to SD.
851 - Unsaved Flag in Job record (use JobMissingFiles).
852 - Base Flag in Job record.
853 - Configure mtx-changer to have correct path to mtx.
854 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
855 - Make some way so that if a machine is skipped because it is not up
856 that Bacula will continue retrying for a specified period of time --
858 - Implement all command line args on run.
859 - Implement command line "restore" args.
860 - Implement "restore current select=no"
861 - Restore file modified before date
862 - Restore -- do nothing but show what would happen
863 - Add estimate to Console commands
864 - Use read_record.c in SD code.
865 - Fix read_record to handle multiple sessions.
866 - Tip from Steve Allam
867 mt -f /dev/nst0 defblksize 0