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.
26 - Document restore by files.
27 - Document new Include/Exclude ...
28 - Document all the status codes JobLevel, JobType, JobStatus.
29 - Add test of exclusion, test multiple Include {} statements.
30 - Enhance "update slots" to include a "scan" feature
31 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
32 - Allow a slot or range of slots on the label barcodes command.
33 when the magazine is changed.
34 - Don't print "Warning: Wrong Volume mounted ..." if mounting second
36 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
39 - Add heartbeat from FD to SD if hb interval expires.
40 - Suppress read error on blank tape when doing a label.
41 - Can we dynamically change FileSets.
42 - If pool specified to label command and Label Format is specified,
43 automatically generate the Volume name.
44 - Take a careful look a the Basic recycling algorithm. When Bacula
45 chooses, the order should be:
47 - Look for Recycle or Purged
50 Instead of using lowest media Id, find the least recently used
53 When the tape is mounted and Bacula requests the status
54 - Do everything possible to use it.
56 Define a "available" status, which is the currently mounted
57 Volume and all volumes that are currently in the autochanger.
59 - Why can't SQL do the filename sort for restore?
60 - Is a pool specification really needed for a restore? Yes, and
61 you may want to exclude archive Pools.
62 - Look at libkse (man kse) for FreeBSD threading.
63 - Look into Microsoft Volume Shadowcopy Service VSS for backing
64 up system state components (Active Directory, System Volume, ...)
65 - Add ExhautiveRestoreSearch
66 - Look at the possibility of loading only the necessary
67 data into the restore tree (i.e. do it one directory at a
68 time as the user walks through the tree).
69 - Possibly use the hash code if the user selects all for a restore command.
70 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
71 - Implement some way for the File daemon to contact the Director
72 to start a job or pass its DHCP obtained IP number.
73 - Implement multiple Consoles.
74 - Add Console usr permissions.
75 - Fix "restore all" to bypass building the tree.
76 - Fix restore to list errors if Invalid block found, and if # files
77 restored does not match # expected.
78 - Prohibit backing up archive device (findlib/find_one.c:128)
79 - Implement Release Device in the Job resource to unmount a drive.
80 - Implement Acquire Device in the Job resource to mount a drive,
81 be sure this works with admin jobs so that the user can get
82 prompted to insert the correct tape. Possibly some way to say to
83 run the job but don't save the files.
84 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
85 - Implement FileOptions (see end of this document)
86 - Make things like list where a file is saved case independent for
88 - Edit the Client/Storage name into authentication failure messages.
89 - Implement job in VerifyToCatalog
91 - Implement a PostgreSQL driver.
92 - Bacula needs to propagate SD errors.
93 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
94 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
95 /dev/nst0. ERR=Input/output error.
96 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
97 Last block=5162 Current block=5164.
98 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
100 - Use autochanger to handle multiple devices.
101 - Fix packet too big problem.
102 - Add SuSE install doc to list.
103 - Check and rechedk "Invalid block number"
104 - Make bextract release the drive properly between tapes
105 so that an autochanger can be made to work.
106 - User wants to NOT backup up certain big files (email files).
107 - Maybe remove multiple simultaneous devices code in SD.
108 - On Windows with very long path names, it may be impossible to create
109 a file (and thus restore it) because the total length is too long.
110 We must cd into the directory then create the file without the
112 - lstat() is not going to work on Win32 for testing date.
113 - Something is not right in last block of fill command.
114 - Implement a Recycle command
115 - Add FileSet to command line arguments for restore.
116 - Allow multiple Storage specifications (or multiple names on
117 a single Storage specification) in the Job record. Thus a job
118 can be backed up to a number of storage devices.
119 - Add client name to cram-md5 challenge so Director can immediately
120 verify if it is the correct client.
121 - Implement ClientRunBeforeJob and ClientRunAfterJob.
122 - Add JobLevel in FD status (but make sure it is defined).
123 - Audit all UA commands to ensure that we always prompt where possible.
124 - Restrict characters permitted in a Resource name, and don't permit
126 - Check Jmsg in bnet, may not work, must dup bsock.
127 - Suppress Job Name in Jmsg for console
128 - Create Pools that are referenced in a Run statement at startup if possible.
129 - Use runbeforejob to unload, then reload a volume previously used,
130 then the next job run gets an error reading the drive.
131 - Make bootstrap filename unique.
132 - Test a second language e.g. french.
133 - Start working on Base jobs.
134 - Make "make binary-release" work from any directory.
135 - Implement UnsavedFiles DB record.
136 - Implement argc/argv for daemon command line scanning using table driven
138 - Implement table driven single argc/argv scanner to pickup all arguments.
139 Much like xxx_conf.c scan table.
140 keyword, handler(store_routine), store_address, code, flags, default.
141 - From Phil Stracchino:
142 It would probably be a per-client option, and would be called
143 something like, say, "Automatically purge obsoleted jobs". What it
144 would do is, when you successfully complete a Differential backup of a
145 client, it would automatically purge all Incremental backups for that
146 client that are rendered redundant by that Differential. Likewise,
147 when a Full backup on a client completed, it would automatically purge
148 all Differential and Incremental jobs obsoleted by that Full backup.
149 This would let people minimize the number of tapes they're keeping on
150 hand without having to master the art of retention times.
151 - Implement new serialize subroutines
152 send(socket, "string", &Vol, "uint32", &i, NULL)
153 - Scratch Pool where the volumes can be re-assigned to any Pool.
154 - Implement a M_SECURITY message class.
155 - Implement forward spacing block/file: position_device(bsr) --
156 just before read_block_from_device();
157 - When doing a Backup send all attributes back to the Director, who
158 would then figure out what files have been deleted.
159 - Currently in mount.c:236 the SD simply creates a Volume. It should have
160 explicit permission to do so. It should also mark the tape in error
161 if there is an error.
162 - Make sure all restore counters are working correctly in the FD.
163 - SD Bytes Read is wrong.
164 - Look at ALL higher level routines that call block.c to be sure
165 they don't expect something in errmsg.
166 - Investigate doing RAW backup of Win32 partition.
167 - Add thread specific data to hold the jcr -- send error messages from
168 low level routines by accessing it and using Jmsg().
169 - Cancel waiting for Client connect in SD if FD goes away.
171 - Add Progress command that periodically reports the progress of
173 - One block was orphaned in the SD probably after cancel.
176 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
177 - Implement timeout in response() when it should come quickly.
178 - Implement console @echo command.
179 - Implement a Slot priority (loaded/not loaded).
180 - Implement "vacation" Incremental only saves.
181 - Implement single pane restore (much like the Gftp panes).
182 - Implement Automatic Mount even in operator wait.
183 - Implement create "FileSet"?
184 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
185 - Implement "scratch pool" where tapes are defined and can be
186 taken by any pool that needs them.
187 - Implement restore "current system", but take all files without
188 doing selection tree -- so that jobs without File records can
190 - Implement disk spooling. Two parts: 1. Spool to disk then
191 immediately to tape to speed up tape operations. 2. Spool to
192 disk only when the tape is full, then when a tape is hung move
194 - Implement a relocatable bacula.spec
195 - Implement dump/print label to UA
196 - Add prefixlinks to where or not where absolute links to FD.
197 - Issue message to mount a new tape before the rewind.
198 - Simplified client job initiation for portables.
199 - If SD cannot open a drive, make it periodically retry.
200 - Implement LabelTemplate (at least first cut).
201 - Add more of the config info to the tape label.
202 - Implement a Mount Command and an Unmount Command where
203 the users could specify a system command to be performed
204 to do the mount, after which Bacula could attempt to
205 read the device. This is for Removeable media such as a CDROM.
206 - Most likely, this mount command would be invoked explicitly
207 by the user using the current Console "mount" and "unmount"
208 commands -- the Storage Daemon would do the right thing
209 depending on the exact nature of the device.
210 - As with tape drives, when Bacula wanted a new removable
211 disk mounted, it would unmount the old one, and send a message
212 to the user, who would then use "mount" as described above
213 once he had actually inserted the disk.
215 - If tape is marked read-only, then try opening it read-only rather than
216 failing, and remember that it cannot be written.
217 - Refine SD waiting output:
218 Device is being positioned
219 > Device is being positioned for append
220 > Device is being positioned to file x
222 - Figure out some way to estimate output size and to avoid splitting
223 a backup across two Volumes -- this could be useful for writing CDROMs
224 where you really prefer not to have it split -- not serious.
225 - Add RunBeforeJob and RunAfterJob to the Client program.
226 - Have SD compute MD5 or SHA1 and compare to what FD computes.
227 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
228 actual data on the Volume and compare it.
229 - Implement Bacula plugins -- design API
230 - Make bcopy read through bad tape records.
231 - Program files (i.e. execute a program to read/write files).
232 Pass read date of last backup, size of file last time.
233 - Add Signature type to File DB record.
234 - CD into subdirectory when open()ing files for backup to
235 speed up things. Test with testfind().
236 - Priority job to go to top of list.
237 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
238 - Implement some way for the Console to dynamically create a job.
239 - Restore to a particular time -- e.g. before date, after date.
240 - Solaris -I on tar for include list
241 - Need a verbose mode in restore, perhaps to bsr.
242 - bscan without -v is too quiet -- perhaps show jobs.
243 - Add code to reject whole blocks if not wanted on restore.
244 - Check if we can increase Bacula FD priorty in Win2000
245 - Make sure the MaxVolFiles is fully implemented in SD
246 - Check if both CatalogFiles and UseCatalog are set to SD.
247 - Need return status on read_cb() from read_records(). Need multiple
248 records -- one per Job, maybe a JCR or some other structure with
249 a block and a record.
250 - Figure out how to do a bare metal Windows restore
251 - Possibly add email to Watchdog if drive is unmounted too
252 long and a job is waiting on the drive.
253 - Restore program that errs in SD due to no tape, reports
254 OK incorrectly in output.
255 - After unmount, if restore job started, ask to mount.
256 - Convert all %x substitution variables, which are hard to remember
257 and read to %(variable-name). Idea from TMDA.
258 - Remove NextId for SQLite. Optimize.
259 - Move all SQL statements into a single location.
260 - Add UA rc and history files.
261 - put termcap (used by console) in ./configure and
262 allow -with-termcap-dir.
263 - Enhance time and size scanning routines.
264 - Fix Autoprune for Volumes to respect need for full save.
265 - Fix Win32 config file definition name on /install
266 - Compare tape to Client files (attributes, or attributes and data)
267 - Make all database Ids 64 bit.
268 - Write an applet for Linux.
269 - Implement new inter-daemon communications protocol.
270 - Allow console commands to detach or run in background.
271 - Fix status delay on storage daemon during rewind.
272 - Add SD message variables to control operator wait time
273 - Maximum Operator Wait
274 - Minimum Message Interval
275 - Maximum Message Interval
276 - Send Operator message when cannot read tape label.
277 - Verify level=Volume (scan only), level=Data (compare of data to file).
278 Verify level=Catalog, level=InitCatalog
280 - Add keyword search to show command in Console.
281 - Events : tape has more than xxx bytes.
282 - Complete code in Bacula Resources -- this will permit
283 reading a new config file at any time.
284 - Handle ctl-c in Console
285 - Implement script driven addition of File daemon to config files.
286 - Think about how to make Bacula work better with File (non-tape) archives.
287 - Write Unix emulator for Windows.
288 - Put memory utilization in Status output of each daemon
289 if full status requested or if some level of debug on.
290 - Make database type selectable by .conf files i.e. at runtime
291 - Set flag for uname -a. Add to Volume label.
292 - Implement throttled work queue.
293 - Restore files modified after date
294 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
295 - Implement Restore FileSet=
296 - Create a protocol.h and protocol.c where all protocol messages
298 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
299 - Timout a job or terminate if link goes down, or reopen link and query.
300 - Concept of precious tapes (cannot be reused).
301 - Make bcopy copy with a single tape drive.
302 - Permit changing ownership during restore.
305 > My suggestion: Add a feature on the systray menu-icon menu to request
306 > an immediate backup now. This would be useful for laptop users who may
307 > not be on the network when the regular scheduled backup is run.
309 > My wife's suggestion: Add a setting to the win32 client to allow it to
310 > shut down the machine after backup is complete (after, of course,
311 > displaying a "System will shut down in one minute, click here to cancel"
312 > warning dialog). This would be useful for sites that want user
313 > woorkstations to be shut down overnight to save power.
316 - From Terry Manderson <terry@apnic.net>
317 jobset { # new structure
318 name = "monthlyUnixBoxen"
321 jobs = "wakame;durian;soy;wasabi;miso" #new!
331 fileset = "genericUnixSet"
337 fileset = "genericUnixSet"
343 fileset = "UnixDevelBoxSet"
348 - Autolabel should be specified by DIR instead of SD.
351 - AutoScan (check checksum of tape)
352 - Format command = "format /dev/nst0"
356 - Seek resolution (usually corresponds to buffer size)
357 - EODErrorCode=ENOSPC or code
358 - Partial Read error code
359 - Partial write error code
360 - Nonformatted read error
361 - Nonformatted write error
362 - WriteProtected error
366 - IgnoreCloseErrors=yes
376 - FD sends unsaved file list to Director at end of job (see
378 - File daemon should build list of files skipped, and then
379 at end of save retry and report any errors.
380 - Write a Storage daemon that uses pipes and
381 standard Unix programs to write to the tape.
383 - Need something that monitors the JCR queue and
384 times out jobs by asking the deamons where they are.
385 - Enhance Jmsg code to permit buffering and saving to disk.
386 - device driver = "xxxx" for drives.
388 - Ensure that /dev/null works
389 - Need report class for messages. Perhaps
390 report resource where report=group of messages
391 - enhance scan_attrib and rename scan_jobtype, and
392 fill in code for "since" option
393 - Director needs a time after which the report status is sent
394 anyway -- or better yet, a retry time for the job.
395 - Don't reschedule a job if previous incarnation is still running.
396 - Some way to automatically backup everything is needed????
397 - Need a structure for pending actions:
399 - termination status (part of buffered msgs?)
401 Read, Write, Clean, Delete
402 - Login to Bacula; Bacula users with different permissions:
403 owner, group, user, quotas
404 - Store info on each file system type (probably in the job header on tape.
405 This could be the output of df; or perhaps some sort of /etc/mtab record.
408 - Design at hierarchial storage for Bacula. Migration and Clone.
409 - Implement FSM (File System Modules).
410 - Audit M_ error codes to ensure they are correct and consistent.
411 - Add variable break characters to lex analyzer.
412 Either a bit mask or a string of chars so that
413 the caller can change the break characters.
414 - Make a single T_BREAK to replace T_COMMA, etc.
415 - Ensure that File daemon and Storage daemon can
416 continue a save if the Director goes down (this
417 is NOT currently the case). Must detect socket error,
418 buffer messages for later.
419 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
420 - Add ability to backup to two Storage devices (two SD sessions) at
421 the same time -- e.g. onsite, offsite.
422 - Add the ability to consolidate old backup sets (basically do a restore
423 to tape and appropriately update the catalog). Compress Volume sets.
424 Might need to spool via file is only one drive is available.
425 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
426 someway to do so with every volume that has less than x% valid
430 Migration: Move a backup from one Volume to another
431 Clone: Copy a backup -- two Volumes
433 Bacula Migration is based on Jobs (apparently Networker is file by file).
435 Migration triggered by:
439 Highwater mark (keep total size)
443 Bacula Projects Roadmap
445 last update 8 May 2003
447 Item 1: Multiple simultaneous Jobs. (done)
448 Done -- Restore part needs better implementation to work correctly
449 Also, it needs considerable testing
451 What: Permit multiple simultaneous jobs in Bacula.
453 Why: An enterprise level solution needs to go fast without the
454 need for the system administrator to carefully tweak
455 timing. Based on the benchmarks, during a full
456 backup, NetWorker typically hit 10 times the bandwidth to
457 the tape compared to Bacula--largely. This is probably due to
458 running parallel jobs and multi-threaded filling of buffers
459 and writing them to tape. This should also make things work
460 better when you have a mix of fast and slow machines backing
463 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
464 implementing this is a matter of some small cleanups and
468 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
469 Deferred -- not necessary yet -- possibly implement with Migration.
471 What: If data is coming into the SD too fast, buffer it to
472 disk if the user has configured this option.
474 Why: This would be nice, especially if it more or less falls out
475 when implementing (1) above. If not, it probably should not
476 be given a high priority because fundamentally the backup time
477 is limited by the tape bandwidth. Even though you may finish a
478 client job quicker by spilling to disk, you still have to
479 eventually get it onto tape. If intermediate disk buffering
480 allows us to improve write bandwidth to tape, it may make
483 Notes: Whether or not this is implemented will depend upon performance
484 testing after item 1 is implemented.
487 Item 3: Write the bscan program -- also write a bcopy program.
490 What: Write a program that reads a Bacula tape and puts all the
491 appropriate data into the catalog. This allows recovery
492 from a tape that is no longer in the database, or it allows
493 re-creation of a database if lost.
495 Why: This is a fundamental robustness and disaster recovery tool
496 which will increase the comfort level of a sysadmin
497 considering adopting Bacula.
499 Notes: A skeleton of this program already exists, but much work
500 needs to be done. Implementing this will also make apparent
501 any deficiencies in the current Bacula tape format.
504 Item 4: Implement Base jobs.
506 What: A base job is sort of like a Full save except that you
507 will want the FileSet to contain only files that are unlikely
508 to change in the future (i.e. a snapshot of most of your
509 system after installing it). After the base job has been run,
510 when you are doing a Full save, you can specify to exclude
511 all files saved by the base job that have not been modified.
513 Why: This is something none of the competition does, as far as we know
514 (except BackupPC, which is a Perl program that saves to disk
515 only). It is big win for the user, it makes Bacula stand out
516 as offering a unique optimization that immediately saves time
519 Notes: Big savings in tape usage. Will require more resources because
520 the e. DIR must send FD a list of files/attribs, and the FD must
521 search the list and compare it for each file to be saved.
524 Item 5: Implement Label templates
527 What: This is a mechanism whereby Bacula can automatically create
528 a tape label for new tapes according to a detailed specification
529 provided by the user.
531 Why: It is a major convenience item for folks who use automated label
534 Notes: Bacula already has a working form of automatic tape label
535 creation, but it is very crude. The design for the complete
536 tape labeling project is already documented in the manual.
539 Item 6: Write a regression script.
540 Done -- Continue to expand its testing.
542 What: This is an automatic script that runs and tests as many features
543 of Bacula as possible. The output is compared to previous
544 versions of Bacula and any differences are reported.
546 Why: This is an enormous help in preventing introduction of new
547 errors in parts of the program that already work correctly.
549 Notes: This probably should be ranked higher, it's something the typical
550 user doesn't see. Depending on how it's implemented, it may
551 make sense to defer it until the archival tape format and
552 user interface mature.
555 Item 7: GUI for interactive restore
556 Item 8: GUI for interactive backup
558 What: The current interactive restore is implemented with a tty
559 interface. It would be much nicer to be able to "see" the
560 list of files backed up in typical GUI tree format.
561 The same mechanism could also be used for creating
562 ad-hoc backup FileSets (item 8).
564 Why: Ease of use -- especially for the end user.
566 Notes: Rather than implementing in Gtk, we probably should go directly
567 for a Browser implementation, even if doing so meant the
568 capability wouldn't be available until much later. Not only
569 is there the question of Windows sites, most
570 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
571 without installing lots of stuff admins are very wary about.
572 Real sysadmins will always use the command line anyway, and
573 the user who's doing an interactive restore or backup of his
574 own files will in most cases be on a Windows machine running
578 Item 9: Add SSL to daemon communications.
580 What: This provides for secure communications between the daemons.
582 Why: This would allow doing backup across the Internet without
583 privacy concerns (or with much less concern).
585 Notes: The vast majority of near term potential users will be backing up
586 a single site over a LAN and, correctly or not, they probably
587 won't be concerned with security, at least not enough to go to
588 the trouble to set up keys, etc. to screw things down. We suspect
589 that many users genuinely interested in multi-site backup
590 already run some form of VPN software in their internetwork
591 connections, and are willing to delegate security to that layer.
594 Item 10: Define definitive tape format.
597 What: Define that definitive tape format that will not change
598 for the next millennium.
600 Why: Stability, security.
602 Notes: See notes for item 11 below.
605 Item 11: New daemon communication protocol.
607 What: The current daemon to daemon protocol is basically an ASCII
608 printf() and sending the buffer. On the receiving end, the
609 buffer is sscanf()ed to unpack it. The new scheme would
610 be a binary format that allows quick packing and unpacking
611 of any data type with named fields.
613 Why: Using binary packing would be faster. Named fields will permit
614 error checking to ensure that what is sent is what the
615 receiver really wants.
617 Notes: These are internal improvements in the interest of the
618 long-term stability and evolution of the program. On the one
619 hand, the sooner they're done, the less code we have to rip
620 up when the time comes to install them. On the other hand, they
621 don't bring an immediately perceptible benefit to potential
622 users. Item 10 and possibly item 11 should be deferred until Bacula
623 is well established with a growing user community more or
624 less happy with the feature set. At that time, it will make a
625 good "next generation" upgrade in the interest of data
631 ======================================================
633 It is somewhat like a Full save becomes an incremental since
634 the Base job (or jobs) plus other non-base files.
636 - New BaseFile table that contains:
637 JobId, BaseJobId, FileId (from Base).
638 i.e. for each base file that exists but is not saved because
639 it has not changed, the File daemon sends the JobId, BaseId,
640 and FileId back to the Director who creates the DB entry.
641 - To initiate a Base save, the Director sends the FD
642 the FileId, and full filename for each file in the Base.
643 - When the FD finds a Base file, he requests the Director to
644 send him the full File entry (stat packet plus MD5), or
645 conversely, the FD sends it to the Director and the Director
646 says yes or no. This can be quite rapid if the FileId is kept
647 by the FD for each Base Filename.
648 - It is probably better to have the comparison done by the FD
649 despite the fact that the File entry must be sent across the
651 - An alternative would be to send the FD the whole File entry
652 from the start. The disadvantage is that it requires a lot of
653 space. The advantage is that it requires less communications
655 - The Job record must be updated to indicate that one or more
657 - At end of Job, FD returns:
658 1. Count of base files/bytes not written to tape (i.e. matches)
659 2. Count of base file that were saved i.e. had changed.
660 - No tape record would be written for a Base file that matches, in the
661 same way that no tape record is written for Incremental jobs where
662 the file is not saved because it is unchanged.
663 - On a restore, all the Base file records must explicitly be
664 found from the BaseFile tape. I.e. for each Full save that is marked
665 to have one or more Base Jobs, search the BaseFile for all occurrences
667 - An optimization might be to make the BaseFile have:
673 This would avoid the need to explicitly fetch each File record for
674 the Base job. The Base Job record will be fetched to get the
675 VolSessionId and VolSessionTime.
676 =========================================================
680 =============================================================
682 Request For Comments For File Backup Options
685 Subject: File Backup Options
688 A few days ago, a Bacula user who is backing up to file volumes and
689 using compression asked if it was possible to suppress compressing
690 all .gz files since it was a waste of CPU time. Although Bacula
691 currently permits using different options (compression, ...) on
692 a directory by directory basis, it cannot do it on a file by
693 file basis, which is clearly what was desired.
695 Proposed Implementation:
696 To solve this problem, I propose the following:
698 - Add a new Director resource type called FileOptions.
700 - The FileOptions resource will have records for all
701 options that can currently be specified on the Include record
702 (in a FileSet). Examples below.
704 - The FileOptions resource will permit an exclude option as well
705 as a number of additional options.
707 - The heart of the FileOptions resource is the ability to
708 supply any number of ApplyTo records which specify POSIX
709 regular expressions. These ApplyTo regular expressions are
710 applied to the fully qualified filename (path and all). If
711 one matches, then the FileOptions will be used.
713 - When an ApplyTo specification matches an included file, the
714 options specified in the FileOptions resource will override
715 the default options specified on the Include record.
717 - Include records will be modified to permit referencing one or
718 more FileOptions resources. The FileOptions will be used
719 in the order listed on the Include record and the first
720 one that matches will be applied.
722 - Options (or specifications) currently supplied on the Include
723 record will be deprecated (i.e. removed in a later version a
724 year or so from now).
726 - The Exclude record will be deprecated as the same functionality
727 can be obtained by using an Exclude = yes in the FileOptions.
730 The following records can appear in the FileOptions resource. An
731 asterisk preceding the name indicates a feature not currently
735 - Compression= (GZIP, ...)
736 - Signature= (MD5, SHA1, ...)
738 - OneFs= (yes/no) - remain on one filesystem
739 - Recurse= (yes/no) - recurse into subdirectories
740 - Sparse= (yes/no) - do sparse file backup
741 - *Exclude= (yes/no) - exclude file from being saved
742 - *Reader= (filename) - external read (backup) program
743 - *Plugin= (filename) - read/write plugin module
746 - verify= (ipnougsamc5) - verify options
749 - replace= (always/ifnewer/ifolder/never) - replace options currently
751 - *Writer= (filename) - external write (restore) program
755 Currently options specifying compression, MD5 signatures, recursion,
756 ... of a FileSet are supplied on the Include record. These will now
757 all be collected into a FileOptions resource, which will be
758 specified on the Include in place of the options. Multiple FileOptions
759 may be specified. Since the FileOptions contain regular expressions
760 that are applied to the full filename, this will give the ability
761 to specify backup options on a file by file basis to whatever level
770 Include = compression=GZIP signature=MD5 {
782 Match = /*.?*/ # matches all files.
787 That's a lot more to do the same thing, but it gives the ability to
788 apply options on a file by file basis. For example, suppose you
789 want to compress all files but not any file with extensions .gz or .Z.
790 You could do so as follows:
797 # Note multiple Matches are ORed
798 Match = /*.gz/ # matches .gz files */
799 Match = /*.Z/ # matches .Z files */
804 Match = /*.?*/ # matches all files
810 Now, since the NoCompress FileOptions is specified first on the
811 Include line, any *.gz or *.Z file will have an MD5 signature computed,
812 but will not be compressed. For all other files, the NoCompress will not
813 match, so the Opts options will be used which will include GZIP
817 - Is it necessary to provide some means of ANDing regular expressions
818 and negation? (not currently planned)
820 e.g. Match = /*.gz/ && !/big.gz/
822 - I see that Networker has a "null" module which, if specified, does not
823 backup the file, but does make an record of the file in the catalog
824 so that the catalog will reflect an exact picture of the filesystem.
825 The result is that the file can be "seen" when "browsing" the save
826 sets, but it cannot be restored.
828 Is this really useful? Should it be implemented in Bacula?
831 After implementing the above, the user will be able to specify
832 on a file by file basis (using regular expressions) what options are
833 applied for the backup.
836 =============================================
838 ==========================================================
840 For each Incremental job that is run, there may be files that
841 were found but not saved because they were locked (this applies
842 only to Windows). Such a system could send back to the Director
843 a list of Unsaved files.
845 - New UnSavedFiles table that contains:
849 - Then in the next Incremental job, the list of Unsaved Files will be
850 feed to the FD, who will ensure that they are explicitly chosen even
851 if standard date/time check would not have selected them.
852 =============================================================
855 Done: (see kernsdone for more)
856 - Implement new alist in FileSet scanning.
857 - bls should continue reading even if it finds Win32 data on the tape.
858 The error should be Warning rather the Error.
859 - Add user configurable timeout for connecting to SD.
860 - Unsaved Flag in Job record (use JobMissingFiles).
861 - Base Flag in Job record.
862 - Configure mtx-changer to have correct path to mtx.
863 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
864 - Make some way so that if a machine is skipped because it is not up
865 that Bacula will continue retrying for a specified period of time --
867 - Implement all command line args on run.
868 - Implement command line "restore" args.
869 - Implement "restore current select=no"
870 - Restore file modified before date
871 - Restore -- do nothing but show what would happen
872 - Add estimate to Console commands
873 - Use read_record.c in SD code.
874 - Fix read_record to handle multiple sessions.
875 - Tip from Steve Allam
876 mt -f /dev/nst0 defblksize 0
877 - Document "status" in the console.
878 - Document driving console from shell script.
879 - Write JobMedia records with max file size is reached on tape.
880 - Handle the case of multiple JobMedia records pending (i.e. the
881 thread is slow and multiple situations requiring a JobMedia
883 - Do performance analysis on the restore tree routines.
884 - Fix maximum file size (block.c) to generate JobMedia records.
885 - Make the default file size 1GB on the tape.
886 - Implement forward spacing between files.
887 - Add Machine type (Linux/Windows) to Status report for daemons.
889 - Use repositioning at the beginning of the tape.
890 - Do full check the command line args in update (e.g. VolStatus ...).
891 - Specify list of files to restore
892 - Implement ClientRunBeforeJob and ClientRunAfterJob.
893 - Make | and < work on FD side.
894 - Check to see if "blocked" is set during restore.
895 - Figure out what is interrupting sql command in console.
896 - Make new job print warning User Unmounted Tape.
897 - Test recycling and purging (code changed in db_find_next_volume and
899 - Document SDConnectTimeout (in FD).
900 - Add restore by filename test.