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.
18 - Document SDConnectTimeout (in FD).
20 Testing to do: (painful)
21 - that ALL console command line options work and are always implemented
22 - blocksize recognition code.
23 - Test if rewind at end of tape waits for tape to rewind.
25 - Test not zeroing Autochanger slot when it is wrong.
26 - Test recycling and purging (code changed in db_find_next_volume and
28 - Figure out how to use ssh or stunnel to protect Bacula communications.
31 - Add restore by filename test.
32 - Add test of exclusion, test multiple Include {} statements.
33 - Enhance "update slots" to include a "scan" feature
34 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
35 - Allow a slot or range of slots on the label barcodes command.
36 when the magazine is changed.
37 - Figure out what is interrupting sql command in console.
38 - Don't print "Warning: Wrong Volume mounted ..." if mounting second
40 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
41 - Take a careful look a the Basic recycling algorithm. When Bacula
42 chooses, the order should be:
44 - Look for Recycle or Purged
47 Instead of using lowest media Id, find the least recently used
50 When the tape is mounted and Bacula requests the status
51 - Do everything possible to use it.
53 Define a "available" status, which is the currently mounted
54 Volume and all volumes that are currently in the autochanger.
58 - Why can't SQL do the filename sort for restore?
59 - Is a pool specification really needed for a restore? Yes, and
60 you may want to exclude archive Pools.
61 - Look at libkse (man kse) for FreeBSD threading.
62 - Look into Microsoft Volume Shadowcopy Service VSS for backing
63 up system state components (Active Directory, System Volume, ...)
64 - Add ExhautiveRestoreSearch
65 - Look at the possibility of loading only the necessary
66 data into the restore tree (i.e. do it one directory at a
67 time as the user walks through the tree).
68 - Possibly use the hash code if the user selects all for a restore command.
69 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
70 - Implement some way for the File daemon to contact the Director
71 to start a job or pass its DHCP obtained IP number.
72 - Implement multiple Consoles.
73 - Add Console usr permissions.
74 - Fix "restore all" to bypass building the tree.
75 - Fix restore to list errors if Invalid block found, and if # files
76 restored does not match # expected.
77 - Prohibit backing up archive device (findlib/find_one.c:128)
78 - Implement Release Device in the Job resource to unmount a drive.
79 - Implement Acquire Device in the Job resource to mount a drive,
80 be sure this works with admin jobs so that the user can get
81 prompted to insert the correct tape. Possibly some way to say to
82 run the job but don't save the files.
83 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
84 - Implement FileOptions (see end of this document)
85 - Make things like list where a file is saved case independent for
87 - Edit the Client/Storage name into authentication failure messages.
88 - Implement job in VerifyToCatalog
90 - Implement a PostgreSQL driver.
91 - Bacula needs to propagate SD errors.
92 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
93 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
94 /dev/nst0. ERR=Input/output error.
95 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
96 Last block=5162 Current block=5164.
97 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
99 - Use autochanger to handle multiple devices.
100 - Fix packet too big problem.
101 - Add SuSE install doc to list.
102 - Check and rechedk "Invalid block number"
103 - Make bextract release the drive properly between tapes
104 so that an autochanger can be made to work.
105 - User wants to NOT backup up certain big files (email files).
106 - Maybe remove multiple simultaneous devices code in SD.
107 - On Windows with very long path names, it may be impossible to create
108 a file (and thus restore it) because the total length is too long.
109 We must cd into the directory then create the file without the
111 - lstat() is not going to work on Win32 for testing date.
112 - Something is not right in last block of fill command.
113 - Implement a Recycle command
114 - Add FileSet to command line arguments for restore.
115 - Allow multiple Storage specifications (or multiple names on
116 a single Storage specification) in the Job record. Thus a job
117 can be backed up to a number of storage devices.
118 - Add client name to cram-md5 challenge so Director can immediately
119 verify if it is the correct client.
120 - Implement ClientRunBeforeJob and ClientRunAfterJob.
121 - Add JobLevel in FD status (but make sure it is defined).
122 - Audit all UA commands to ensure that we always prompt where possible.
123 - Restrict characters permitted in a Resource name, and don't permit
125 - Check Jmsg in bnet, may not work, must dup bsock.
126 - Suppress Job Name in Jmsg for console
127 - Create Pools that are referenced in a Run statement at startup if possible.
128 - Use runbeforejob to unload, then reload a volume previously used,
129 then the next job run gets an error reading the drive.
130 - Make bootstrap filename unique.
131 - Test a second language e.g. french.
132 - Start working on Base jobs.
133 - Make "make binary-release" work from any directory.
134 - Implement UnsavedFiles DB record.
135 - Implement argc/argv for daemon command line scanning using table driven
137 - Implement table driven single argc/argv scanner to pickup all arguments.
138 Much like xxx_conf.c scan table.
139 keyword, handler(store_routine), store_address, code, flags, default.
140 - From Phil Stracchino:
141 It would probably be a per-client option, and would be called
142 something like, say, "Automatically purge obsoleted jobs". What it
143 would do is, when you successfully complete a Differential backup of a
144 client, it would automatically purge all Incremental backups for that
145 client that are rendered redundant by that Differential. Likewise,
146 when a Full backup on a client completed, it would automatically purge
147 all Differential and Incremental jobs obsoleted by that Full backup.
148 This would let people minimize the number of tapes they're keeping on
149 hand without having to master the art of retention times.
150 - Implement new serialize subroutines
151 send(socket, "string", &Vol, "uint32", &i, NULL)
152 - Scratch Pool where the volumes can be re-assigned to any Pool.
153 - Implement a M_SECURITY message class.
154 - Implement forward spacing block/file: position_device(bsr) --
155 just before read_block_from_device();
156 - When doing a Backup send all attributes back to the Director, who
157 would then figure out what files have been deleted.
158 - Currently in mount.c:236 the SD simply creates a Volume. It should have
159 explicit permission to do so. It should also mark the tape in error
160 if there is an error.
161 - Make sure all restore counters are working correctly in the FD.
162 - SD Bytes Read is wrong.
163 - Look at ALL higher level routines that call block.c to be sure
164 they don't expect something in errmsg.
165 - Investigate doing RAW backup of Win32 partition.
166 - Add thread specific data to hold the jcr -- send error messages from
167 low level routines by accessing it and using Jmsg().
168 - Cancel waiting for Client connect in SD if FD goes away.
170 - Add Progress command that periodically reports the progress of
172 - One block was orphaned in the SD probably after cancel.
175 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
176 - Implement timeout in response() when it should come quickly.
177 - Implement console @echo command.
178 - Implement a Slot priority (loaded/not loaded).
179 - Implement "vacation" Incremental only saves.
180 - Implement single pane restore (much like the Gftp panes).
181 - Implement Automatic Mount even in operator wait.
182 - Implement create "FileSet"?
183 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
184 - Implement "scratch pool" where tapes are defined and can be
185 taken by any pool that needs them.
186 - Implement restore "current system", but take all files without
187 doing selection tree -- so that jobs without File records can
189 - Implement disk spooling. Two parts: 1. Spool to disk then
190 immediately to tape to speed up tape operations. 2. Spool to
191 disk only when the tape is full, then when a tape is hung move
193 - Implement a relocatable bacula.spec
194 - Implement dump/print label to UA
195 - Add prefixlinks to where or not where absolute links to FD.
196 - Issue message to mount a new tape before the rewind.
197 - Simplified client job initiation for portables.
198 - If SD cannot open a drive, make it periodically retry.
199 - Implement LabelTemplate (at least first cut).
200 - Add more of the config info to the tape label.
201 - Implement a Mount Command and an Unmount Command where
202 the users could specify a system command to be performed
203 to do the mount, after which Bacula could attempt to
204 read the device. This is for Removeable media such as a CDROM.
205 - Most likely, this mount command would be invoked explicitly
206 by the user using the current Console "mount" and "unmount"
207 commands -- the Storage Daemon would do the right thing
208 depending on the exact nature of the device.
209 - As with tape drives, when Bacula wanted a new removable
210 disk mounted, it would unmount the old one, and send a message
211 to the user, who would then use "mount" as described above
212 once he had actually inserted the disk.
214 - If tape is marked read-only, then try opening it read-only rather than
215 failing, and remember that it cannot be written.
216 - Refine SD waiting output:
217 Device is being positioned
218 > Device is being positioned for append
219 > Device is being positioned to file x
221 - Figure out some way to estimate output size and to avoid splitting
222 a backup across two Volumes -- this could be useful for writing CDROMs
223 where you really prefer not to have it split -- not serious.
224 - Add RunBeforeJob and RunAfterJob to the Client program.
225 - Have SD compute MD5 or SHA1 and compare to what FD computes.
226 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
227 actual data on the Volume and compare it.
228 - Implement Bacula plugins -- design API
229 - Make bcopy read through bad tape records.
230 - Program files (i.e. execute a program to read/write files).
231 Pass read date of last backup, size of file last time.
232 - Add Signature type to File DB record.
233 - CD into subdirectory when open()ing files for backup to
234 speed up things. Test with testfind().
235 - Priority job to go to top of list.
236 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
237 - Implement some way for the Console to dynamically create a job.
238 - Restore to a particular time -- e.g. before date, after date.
239 - Solaris -I on tar for include list
240 - Need a verbose mode in restore, perhaps to bsr.
241 - bscan without -v is too quiet -- perhaps show jobs.
242 - Add code to reject whole blocks if not wanted on restore.
243 - Check if we can increase Bacula FD priorty in Win2000
244 - Make sure the MaxVolFiles is fully implemented in SD
245 - Check if both CatalogFiles and UseCatalog are set to SD.
246 - Need return status on read_cb() from read_records(). Need multiple
247 records -- one per Job, maybe a JCR or some other structure with
248 a block and a record.
249 - Figure out how to do a bare metal Windows restore
250 - Possibly add email to Watchdog if drive is unmounted too
251 long and a job is waiting on the drive.
252 - Restore program that errs in SD due to no tape, reports
253 OK incorrectly in output.
254 - After unmount, if restore job started, ask to mount.
255 - Convert all %x substitution variables, which are hard to remember
256 and read to %(variable-name). Idea from TMDA.
257 - Remove NextId for SQLite. Optimize.
258 - Move all SQL statements into a single location.
259 - Add UA rc and history files.
260 - put termcap (used by console) in ./configure and
261 allow -with-termcap-dir.
262 - Enhance time and size scanning routines.
263 - Fix Autoprune for Volumes to respect need for full save.
264 - Fix Win32 config file definition name on /install
265 - Compare tape to Client files (attributes, or attributes and data)
266 - Make all database Ids 64 bit.
267 - Write an applet for Linux.
268 - Implement new inter-daemon communications protocol.
269 - Allow console commands to detach or run in background.
270 - Fix status delay on storage daemon during rewind.
271 - Add SD message variables to control operator wait time
272 - Maximum Operator Wait
273 - Minimum Message Interval
274 - Maximum Message Interval
275 - Send Operator message when cannot read tape label.
276 - Verify level=Volume (scan only), level=Data (compare of data to file).
277 Verify level=Catalog, level=InitCatalog
279 - Add keyword search to show command in Console.
280 - Events : tape has more than xxx bytes.
281 - Complete code in Bacula Resources -- this will permit
282 reading a new config file at any time.
283 - Handle ctl-c in Console
284 - Implement script driven addition of File daemon to config files.
285 - Think about how to make Bacula work better with File (non-tape) archives.
286 - Write Unix emulator for Windows.
287 - Put memory utilization in Status output of each daemon
288 if full status requested or if some level of debug on.
289 - Make database type selectable by .conf files i.e. at runtime
290 - Set flag for uname -a. Add to Volume label.
291 - Implement throttled work queue.
292 - Restore files modified after date
293 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
294 - Implement Restore FileSet=
295 - Create a protocol.h and protocol.c where all protocol messages
297 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
298 - Timout a job or terminate if link goes down, or reopen link and query.
299 - Concept of precious tapes (cannot be reused).
300 - Make bcopy copy with a single tape drive.
301 - Permit changing ownership during restore.
304 > My suggestion: Add a feature on the systray menu-icon menu to request
305 > an immediate backup now. This would be useful for laptop users who may
306 > not be on the network when the regular scheduled backup is run.
308 > My wife's suggestion: Add a setting to the win32 client to allow it to
309 > shut down the machine after backup is complete (after, of course,
310 > displaying a "System will shut down in one minute, click here to cancel"
311 > warning dialog). This would be useful for sites that want user
312 > woorkstations to be shut down overnight to save power.
315 - From Terry Manderson <terry@apnic.net>
316 jobset { # new structure
317 name = "monthlyUnixBoxen"
320 jobs = "wakame;durian;soy;wasabi;miso" #new!
330 fileset = "genericUnixSet"
336 fileset = "genericUnixSet"
342 fileset = "UnixDevelBoxSet"
347 - Autolabel should be specified by DIR instead of SD.
350 - AutoScan (check checksum of tape)
351 - Format command = "format /dev/nst0"
355 - Seek resolution (usually corresponds to buffer size)
356 - EODErrorCode=ENOSPC or code
357 - Partial Read error code
358 - Partial write error code
359 - Nonformatted read error
360 - Nonformatted write error
361 - WriteProtected error
365 - IgnoreCloseErrors=yes
375 - FD sends unsaved file list to Director at end of job (see
377 - File daemon should build list of files skipped, and then
378 at end of save retry and report any errors.
379 - Write a Storage daemon that uses pipes and
380 standard Unix programs to write to the tape.
382 - Need something that monitors the JCR queue and
383 times out jobs by asking the deamons where they are.
384 - Enhance Jmsg code to permit buffering and saving to disk.
385 - device driver = "xxxx" for drives.
387 - Ensure that /dev/null works
388 - Need report class for messages. Perhaps
389 report resource where report=group of messages
390 - enhance scan_attrib and rename scan_jobtype, and
391 fill in code for "since" option
392 - Director needs a time after which the report status is sent
393 anyway -- or better yet, a retry time for the job.
394 - Don't reschedule a job if previous incarnation is still running.
395 - Some way to automatically backup everything is needed????
396 - Need a structure for pending actions:
398 - termination status (part of buffered msgs?)
400 Read, Write, Clean, Delete
401 - Login to Bacula; Bacula users with different permissions:
402 owner, group, user, quotas
403 - Store info on each file system type (probably in the job header on tape.
404 This could be the output of df; or perhaps some sort of /etc/mtab record.
407 - Design at hierarchial storage for Bacula. Migration and Clone.
408 - Implement FSM (File System Modules).
409 - Audit M_ error codes to ensure they are correct and consistent.
410 - Add variable break characters to lex analyzer.
411 Either a bit mask or a string of chars so that
412 the caller can change the break characters.
413 - Make a single T_BREAK to replace T_COMMA, etc.
414 - Ensure that File daemon and Storage daemon can
415 continue a save if the Director goes down (this
416 is NOT currently the case). Must detect socket error,
417 buffer messages for later.
418 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
419 - Add ability to backup to two Storage devices (two SD sessions) at
420 the same time -- e.g. onsite, offsite.
421 - Add the ability to consolidate old backup sets (basically do a restore
422 to tape and appropriately update the catalog). Compress Volume sets.
423 Might need to spool via file is only one drive is available.
424 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
425 someway to do so with every volume that has less than x% valid
429 Migration: Move a backup from one Volume to another
430 Clone: Copy a backup -- two Volumes
432 Bacula Migration is based on Jobs (apparently Networker is file by file).
434 Migration triggered by:
438 Highwater mark (keep total size)
442 Bacula Projects Roadmap
444 last update 8 May 2003
446 Item 1: Multiple simultaneous Jobs. (done)
447 Done -- Restore part needs better implementation to work correctly
448 Also, it needs considerable testing
450 What: Permit multiple simultaneous jobs in Bacula.
452 Why: An enterprise level solution needs to go fast without the
453 need for the system administrator to carefully tweak
454 timing. Based on the benchmarks, during a full
455 backup, NetWorker typically hit 10 times the bandwidth to
456 the tape compared to Bacula--largely. This is probably due to
457 running parallel jobs and multi-threaded filling of buffers
458 and writing them to tape. This should also make things work
459 better when you have a mix of fast and slow machines backing
462 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
463 implementing this is a matter of some small cleanups and
467 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
468 Deferred -- not necessary yet -- possibly implement with Migration.
470 What: If data is coming into the SD too fast, buffer it to
471 disk if the user has configured this option.
473 Why: This would be nice, especially if it more or less falls out
474 when implementing (1) above. If not, it probably should not
475 be given a high priority because fundamentally the backup time
476 is limited by the tape bandwidth. Even though you may finish a
477 client job quicker by spilling to disk, you still have to
478 eventually get it onto tape. If intermediate disk buffering
479 allows us to improve write bandwidth to tape, it may make
482 Notes: Whether or not this is implemented will depend upon performance
483 testing after item 1 is implemented.
486 Item 3: Write the bscan program -- also write a bcopy program.
489 What: Write a program that reads a Bacula tape and puts all the
490 appropriate data into the catalog. This allows recovery
491 from a tape that is no longer in the database, or it allows
492 re-creation of a database if lost.
494 Why: This is a fundamental robustness and disaster recovery tool
495 which will increase the comfort level of a sysadmin
496 considering adopting Bacula.
498 Notes: A skeleton of this program already exists, but much work
499 needs to be done. Implementing this will also make apparent
500 any deficiencies in the current Bacula tape format.
503 Item 4: Implement Base jobs.
505 What: A base job is sort of like a Full save except that you
506 will want the FileSet to contain only files that are unlikely
507 to change in the future (i.e. a snapshot of most of your
508 system after installing it). After the base job has been run,
509 when you are doing a Full save, you can specify to exclude
510 all files saved by the base job that have not been modified.
512 Why: This is something none of the competition does, as far as we know
513 (except BackupPC, which is a Perl program that saves to disk
514 only). It is big win for the user, it makes Bacula stand out
515 as offering a unique optimization that immediately saves time
518 Notes: Big savings in tape usage. Will require more resources because
519 the e. DIR must send FD a list of files/attribs, and the FD must
520 search the list and compare it for each file to be saved.
523 Item 5: Implement Label templates
526 What: This is a mechanism whereby Bacula can automatically create
527 a tape label for new tapes according to a detailed specification
528 provided by the user.
530 Why: It is a major convenience item for folks who use automated label
533 Notes: Bacula already has a working form of automatic tape label
534 creation, but it is very crude. The design for the complete
535 tape labeling project is already documented in the manual.
538 Item 6: Write a regression script.
539 Done -- Continue to expand its testing.
541 What: This is an automatic script that runs and tests as many features
542 of Bacula as possible. The output is compared to previous
543 versions of Bacula and any differences are reported.
545 Why: This is an enormous help in preventing introduction of new
546 errors in parts of the program that already work correctly.
548 Notes: This probably should be ranked higher, it's something the typical
549 user doesn't see. Depending on how it's implemented, it may
550 make sense to defer it until the archival tape format and
551 user interface mature.
554 Item 7: GUI for interactive restore
555 Item 8: GUI for interactive backup
557 What: The current interactive restore is implemented with a tty
558 interface. It would be much nicer to be able to "see" the
559 list of files backed up in typical GUI tree format.
560 The same mechanism could also be used for creating
561 ad-hoc backup FileSets (item 8).
563 Why: Ease of use -- especially for the end user.
565 Notes: Rather than implementing in Gtk, we probably should go directly
566 for a Browser implementation, even if doing so meant the
567 capability wouldn't be available until much later. Not only
568 is there the question of Windows sites, most
569 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
570 without installing lots of stuff admins are very wary about.
571 Real sysadmins will always use the command line anyway, and
572 the user who's doing an interactive restore or backup of his
573 own files will in most cases be on a Windows machine running
577 Item 9: Add SSL to daemon communications.
579 What: This provides for secure communications between the daemons.
581 Why: This would allow doing backup across the Internet without
582 privacy concerns (or with much less concern).
584 Notes: The vast majority of near term potential users will be backing up
585 a single site over a LAN and, correctly or not, they probably
586 won't be concerned with security, at least not enough to go to
587 the trouble to set up keys, etc. to screw things down. We suspect
588 that many users genuinely interested in multi-site backup
589 already run some form of VPN software in their internetwork
590 connections, and are willing to delegate security to that layer.
593 Item 10: Define definitive tape format.
596 What: Define that definitive tape format that will not change
597 for the next millennium.
599 Why: Stability, security.
601 Notes: See notes for item 11 below.
604 Item 11: New daemon communication protocol.
606 What: The current daemon to daemon protocol is basically an ASCII
607 printf() and sending the buffer. On the receiving end, the
608 buffer is sscanf()ed to unpack it. The new scheme would
609 be a binary format that allows quick packing and unpacking
610 of any data type with named fields.
612 Why: Using binary packing would be faster. Named fields will permit
613 error checking to ensure that what is sent is what the
614 receiver really wants.
616 Notes: These are internal improvements in the interest of the
617 long-term stability and evolution of the program. On the one
618 hand, the sooner they're done, the less code we have to rip
619 up when the time comes to install them. On the other hand, they
620 don't bring an immediately perceptible benefit to potential
621 users. Item 10 and possibly item 11 should be deferred until Bacula
622 is well established with a growing user community more or
623 less happy with the feature set. At that time, it will make a
624 good "next generation" upgrade in the interest of data
630 ======================================================
632 It is somewhat like a Full save becomes an incremental since
633 the Base job (or jobs) plus other non-base files.
635 - New BaseFile table that contains:
636 JobId, BaseJobId, FileId (from Base).
637 i.e. for each base file that exists but is not saved because
638 it has not changed, the File daemon sends the JobId, BaseId,
639 and FileId back to the Director who creates the DB entry.
640 - To initiate a Base save, the Director sends the FD
641 the FileId, and full filename for each file in the Base.
642 - When the FD finds a Base file, he requests the Director to
643 send him the full File entry (stat packet plus MD5), or
644 conversely, the FD sends it to the Director and the Director
645 says yes or no. This can be quite rapid if the FileId is kept
646 by the FD for each Base Filename.
647 - It is probably better to have the comparison done by the FD
648 despite the fact that the File entry must be sent across the
650 - An alternative would be to send the FD the whole File entry
651 from the start. The disadvantage is that it requires a lot of
652 space. The advantage is that it requires less communications
654 - The Job record must be updated to indicate that one or more
656 - At end of Job, FD returns:
657 1. Count of base files/bytes not written to tape (i.e. matches)
658 2. Count of base file that were saved i.e. had changed.
659 - No tape record would be written for a Base file that matches, in the
660 same way that no tape record is written for Incremental jobs where
661 the file is not saved because it is unchanged.
662 - On a restore, all the Base file records must explicitly be
663 found from the BaseFile tape. I.e. for each Full save that is marked
664 to have one or more Base Jobs, search the BaseFile for all occurrences
666 - An optimization might be to make the BaseFile have:
672 This would avoid the need to explicitly fetch each File record for
673 the Base job. The Base Job record will be fetched to get the
674 VolSessionId and VolSessionTime.
675 =========================================================
679 =============================================================
681 Request For Comments For File Backup Options
684 Subject: File Backup Options
687 A few days ago, a Bacula user who is backing up to file volumes and
688 using compression asked if it was possible to suppress compressing
689 all .gz files since it was a waste of CPU time. Although Bacula
690 currently permits using different options (compression, ...) on
691 a directory by directory basis, it cannot do it on a file by
692 file basis, which is clearly what was desired.
694 Proposed Implementation:
695 To solve this problem, I propose the following:
697 - Add a new Director resource type called FileOptions.
699 - The FileOptions resource will have records for all
700 options that can currently be specified on the Include record
701 (in a FileSet). Examples below.
703 - The FileOptions resource will permit an exclude option as well
704 as a number of additional options.
706 - The heart of the FileOptions resource is the ability to
707 supply any number of ApplyTo records which specify POSIX
708 regular expressions. These ApplyTo regular expressions are
709 applied to the fully qualified filename (path and all). If
710 one matches, then the FileOptions will be used.
712 - When an ApplyTo specification matches an included file, the
713 options specified in the FileOptions resource will override
714 the default options specified on the Include record.
716 - Include records will be modified to permit referencing one or
717 more FileOptions resources. The FileOptions will be used
718 in the order listed on the Include record and the first
719 one that matches will be applied.
721 - Options (or specifications) currently supplied on the Include
722 record will be deprecated (i.e. removed in a later version a
723 year or so from now).
725 - The Exclude record will be deprecated as the same functionality
726 can be obtained by using an Exclude = yes in the FileOptions.
729 The following records can appear in the FileOptions resource. An
730 asterisk preceding the name indicates a feature not currently
734 - Compression= (GZIP, ...)
735 - Signature= (MD5, SHA1, ...)
737 - OneFs= (yes/no) - remain on one filesystem
738 - Recurse= (yes/no) - recurse into subdirectories
739 - Sparse= (yes/no) - do sparse file backup
740 - *Exclude= (yes/no) - exclude file from being saved
741 - *Reader= (filename) - external read (backup) program
742 - *Plugin= (filename) - read/write plugin module
745 - verify= (ipnougsamc5) - verify options
748 - replace= (always/ifnewer/ifolder/never) - replace options currently
750 - *Writer= (filename) - external write (restore) program
754 Currently options specifying compression, MD5 signatures, recursion,
755 ... of a FileSet are supplied on the Include record. These will now
756 all be collected into a FileOptions resource, which will be
757 specified on the Include in place of the options. Multiple FileOptions
758 may be specified. Since the FileOptions contain regular expressions
759 that are applied to the full filename, this will give the ability
760 to specify backup options on a file by file basis to whatever level
769 Include = compression=GZIP signature=MD5 {
781 Match = /*.?*/ # matches all files.
786 That's a lot more to do the same thing, but it gives the ability to
787 apply options on a file by file basis. For example, suppose you
788 want to compress all files but not any file with extensions .gz or .Z.
789 You could do so as follows:
796 # Note multiple Matches are ORed
797 Match = /*.gz/ # matches .gz files */
798 Match = /*.Z/ # matches .Z files */
803 Match = /*.?*/ # matches all files
809 Now, since the NoCompress FileOptions is specified first on the
810 Include line, any *.gz or *.Z file will have an MD5 signature computed,
811 but will not be compressed. For all other files, the NoCompress will not
812 match, so the Opts options will be used which will include GZIP
816 - Is it necessary to provide some means of ANDing regular expressions
817 and negation? (not currently planned)
819 e.g. Match = /*.gz/ && !/big.gz/
821 - I see that Networker has a "null" module which, if specified, does not
822 backup the file, but does make an record of the file in the catalog
823 so that the catalog will reflect an exact picture of the filesystem.
824 The result is that the file can be "seen" when "browsing" the save
825 sets, but it cannot be restored.
827 Is this really useful? Should it be implemented in Bacula?
830 After implementing the above, the user will be able to specify
831 on a file by file basis (using regular expressions) what options are
832 applied for the backup.
835 =============================================
837 ==========================================================
839 For each Incremental job that is run, there may be files that
840 were found but not saved because they were locked (this applies
841 only to Windows). Such a system could send back to the Director
842 a list of Unsaved files.
844 - New UnSavedFiles table that contains:
848 - Then in the next Incremental job, the list of Unsaved Files will be
849 feed to the FD, who will ensure that they are explicitly chosen even
850 if standard date/time check would not have selected them.
851 =============================================================
854 Done: (see kernsdone for more)
855 - Implement new alist in FileSet scanning.
856 - bls should continue reading even if it finds Win32 data on the tape.
857 The error should be Warning rather the Error.
858 - Add user configurable timeout for connecting to SD.
859 - Unsaved Flag in Job record (use JobMissingFiles).
860 - Base Flag in Job record.
861 - Configure mtx-changer to have correct path to mtx.
862 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
863 - Make some way so that if a machine is skipped because it is not up
864 that Bacula will continue retrying for a specified period of time --
866 - Implement all command line args on run.
867 - Implement command line "restore" args.
868 - Implement "restore current select=no"
869 - Restore file modified before date
870 - Restore -- do nothing but show what would happen
871 - Add estimate to Console commands
872 - Use read_record.c in SD code.
873 - Fix read_record to handle multiple sessions.
874 - Tip from Steve Allam
875 mt -f /dev/nst0 defblksize 0
876 - Document "status" in the console.
877 - Document driving console from shell script.
878 - Write JobMedia records with max file size is reached on tape.
879 - Handle the case of multiple JobMedia records pending (i.e. the
880 thread is slow and multiple situations requiring a JobMedia
882 - Do performance analysis on the restore tree routines.
883 - Fix maximum file size (block.c) to generate JobMedia records.
884 - Make the default file size 1GB on the tape.
885 - Implement forward spacing between files.
886 - Add Machine type (Linux/Windows) to Status report for daemons.
888 - Use repositioning at the beginning of the tape.
889 - Do full check the command line args in update (e.g. VolStatus ...).
890 - Specify list of files to restore
891 - Implement ClientRunBeforeJob and ClientRunAfterJob.
892 - Make | and < work on FD side.