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 - Don't print "Warning: Wrong Volume mounted ..." if mounting second
39 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
43 - If pool specified to label command and Label Format is specified,
44 automatically generate the Volume name.
45 - Take a careful look a the Basic recycling algorithm. When Bacula
46 chooses, the order should be:
48 - Look for Recycle or Purged
51 Instead of using lowest media Id, find the least recently used
54 When the tape is mounted and Bacula requests the status
55 - Do everything possible to use it.
57 Define a "available" status, which is the currently mounted
58 Volume and all volumes that are currently in the autochanger.
60 - Why can't SQL do the filename sort for restore?
61 - Is a pool specification really needed for a restore? Yes, and
62 you may want to exclude archive Pools.
63 - Look at libkse (man kse) for FreeBSD threading.
64 - Look into Microsoft Volume Shadowcopy Service VSS for backing
65 up system state components (Active Directory, System Volume, ...)
66 - Add ExhautiveRestoreSearch
67 - Look at the possibility of loading only the necessary
68 data into the restore tree (i.e. do it one directory at a
69 time as the user walks through the tree).
70 - Possibly use the hash code if the user selects all for a restore command.
71 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
72 - Implement some way for the File daemon to contact the Director
73 to start a job or pass its DHCP obtained IP number.
74 - Implement multiple Consoles.
75 - Add Console usr permissions.
76 - Fix "restore all" to bypass building the tree.
77 - Fix restore to list errors if Invalid block found, and if # files
78 restored does not match # expected.
79 - Prohibit backing up archive device (findlib/find_one.c:128)
80 - Implement Release Device in the Job resource to unmount a drive.
81 - Implement Acquire Device in the Job resource to mount a drive,
82 be sure this works with admin jobs so that the user can get
83 prompted to insert the correct tape. Possibly some way to say to
84 run the job but don't save the files.
85 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
86 - Implement FileOptions (see end of this document)
87 - Make things like list where a file is saved case independent for
89 - Edit the Client/Storage name into authentication failure messages.
90 - Implement job in VerifyToCatalog
92 - Implement a PostgreSQL driver.
93 - Bacula needs to propagate SD errors.
94 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
95 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
96 /dev/nst0. ERR=Input/output error.
97 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
98 Last block=5162 Current block=5164.
99 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
101 - Use autochanger to handle multiple devices.
102 - Fix packet too big problem.
103 - Add SuSE install doc to list.
104 - Check and rechedk "Invalid block number"
105 - Make bextract release the drive properly between tapes
106 so that an autochanger can be made to work.
107 - User wants to NOT backup up certain big files (email files).
108 - Maybe remove multiple simultaneous devices code in SD.
109 - On Windows with very long path names, it may be impossible to create
110 a file (and thus restore it) because the total length is too long.
111 We must cd into the directory then create the file without the
113 - lstat() is not going to work on Win32 for testing date.
114 - Something is not right in last block of fill command.
115 - Implement a Recycle command
116 - Add FileSet to command line arguments for restore.
117 - Allow multiple Storage specifications (or multiple names on
118 a single Storage specification) in the Job record. Thus a job
119 can be backed up to a number of storage devices.
120 - Add client name to cram-md5 challenge so Director can immediately
121 verify if it is the correct client.
122 - Implement ClientRunBeforeJob and ClientRunAfterJob.
123 - Add JobLevel in FD status (but make sure it is defined).
124 - Audit all UA commands to ensure that we always prompt where possible.
125 - Restrict characters permitted in a Resource name, and don't permit
127 - Check Jmsg in bnet, may not work, must dup bsock.
128 - Suppress Job Name in Jmsg for console
129 - Create Pools that are referenced in a Run statement at startup if possible.
130 - Use runbeforejob to unload, then reload a volume previously used,
131 then the next job run gets an error reading the drive.
132 - Make bootstrap filename unique.
133 - Test a second language e.g. french.
134 - Start working on Base jobs.
135 - Make "make binary-release" work from any directory.
136 - Implement UnsavedFiles DB record.
137 - Implement argc/argv for daemon command line scanning using table driven
139 - Implement table driven single argc/argv scanner to pickup all arguments.
140 Much like xxx_conf.c scan table.
141 keyword, handler(store_routine), store_address, code, flags, default.
142 - From Phil Stracchino:
143 It would probably be a per-client option, and would be called
144 something like, say, "Automatically purge obsoleted jobs". What it
145 would do is, when you successfully complete a Differential backup of a
146 client, it would automatically purge all Incremental backups for that
147 client that are rendered redundant by that Differential. Likewise,
148 when a Full backup on a client completed, it would automatically purge
149 all Differential and Incremental jobs obsoleted by that Full backup.
150 This would let people minimize the number of tapes they're keeping on
151 hand without having to master the art of retention times.
152 - Implement new serialize subroutines
153 send(socket, "string", &Vol, "uint32", &i, NULL)
154 - Scratch Pool where the volumes can be re-assigned to any Pool.
155 - Implement a M_SECURITY message class.
156 - Implement forward spacing block/file: position_device(bsr) --
157 just before read_block_from_device();
158 - When doing a Backup send all attributes back to the Director, who
159 would then figure out what files have been deleted.
160 - Currently in mount.c:236 the SD simply creates a Volume. It should have
161 explicit permission to do so. It should also mark the tape in error
162 if there is an error.
163 - Make sure all restore counters are working correctly in the FD.
164 - SD Bytes Read is wrong.
165 - Look at ALL higher level routines that call block.c to be sure
166 they don't expect something in errmsg.
167 - Investigate doing RAW backup of Win32 partition.
168 - Add thread specific data to hold the jcr -- send error messages from
169 low level routines by accessing it and using Jmsg().
170 - Cancel waiting for Client connect in SD if FD goes away.
172 - Add Progress command that periodically reports the progress of
174 - One block was orphaned in the SD probably after cancel.
177 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
178 - Implement timeout in response() when it should come quickly.
179 - Implement console @echo command.
180 - Implement a Slot priority (loaded/not loaded).
181 - Implement "vacation" Incremental only saves.
182 - Implement single pane restore (much like the Gftp panes).
183 - Implement Automatic Mount even in operator wait.
184 - Implement create "FileSet"?
185 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
186 - Implement "scratch pool" where tapes are defined and can be
187 taken by any pool that needs them.
188 - Implement restore "current system", but take all files without
189 doing selection tree -- so that jobs without File records can
191 - Implement disk spooling. Two parts: 1. Spool to disk then
192 immediately to tape to speed up tape operations. 2. Spool to
193 disk only when the tape is full, then when a tape is hung move
195 - Implement a relocatable bacula.spec
196 - Implement dump/print label to UA
197 - Add prefixlinks to where or not where absolute links to FD.
198 - Issue message to mount a new tape before the rewind.
199 - Simplified client job initiation for portables.
200 - If SD cannot open a drive, make it periodically retry.
201 - Implement LabelTemplate (at least first cut).
202 - Add more of the config info to the tape label.
203 - Implement a Mount Command and an Unmount Command where
204 the users could specify a system command to be performed
205 to do the mount, after which Bacula could attempt to
206 read the device. This is for Removeable media such as a CDROM.
207 - Most likely, this mount command would be invoked explicitly
208 by the user using the current Console "mount" and "unmount"
209 commands -- the Storage Daemon would do the right thing
210 depending on the exact nature of the device.
211 - As with tape drives, when Bacula wanted a new removable
212 disk mounted, it would unmount the old one, and send a message
213 to the user, who would then use "mount" as described above
214 once he had actually inserted the disk.
216 - If tape is marked read-only, then try opening it read-only rather than
217 failing, and remember that it cannot be written.
218 - Refine SD waiting output:
219 Device is being positioned
220 > Device is being positioned for append
221 > Device is being positioned to file x
223 - Figure out some way to estimate output size and to avoid splitting
224 a backup across two Volumes -- this could be useful for writing CDROMs
225 where you really prefer not to have it split -- not serious.
226 - Add RunBeforeJob and RunAfterJob to the Client program.
227 - Have SD compute MD5 or SHA1 and compare to what FD computes.
228 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
229 actual data on the Volume and compare it.
230 - Implement Bacula plugins -- design API
231 - Make bcopy read through bad tape records.
232 - Program files (i.e. execute a program to read/write files).
233 Pass read date of last backup, size of file last time.
234 - Add Signature type to File DB record.
235 - CD into subdirectory when open()ing files for backup to
236 speed up things. Test with testfind().
237 - Priority job to go to top of list.
238 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
239 - Implement some way for the Console to dynamically create a job.
240 - Restore to a particular time -- e.g. before date, after date.
241 - Solaris -I on tar for include list
242 - Need a verbose mode in restore, perhaps to bsr.
243 - bscan without -v is too quiet -- perhaps show jobs.
244 - Add code to reject whole blocks if not wanted on restore.
245 - Check if we can increase Bacula FD priorty in Win2000
246 - Make sure the MaxVolFiles is fully implemented in SD
247 - Check if both CatalogFiles and UseCatalog are set to SD.
248 - Need return status on read_cb() from read_records(). Need multiple
249 records -- one per Job, maybe a JCR or some other structure with
250 a block and a record.
251 - Figure out how to do a bare metal Windows restore
252 - Possibly add email to Watchdog if drive is unmounted too
253 long and a job is waiting on the drive.
254 - Restore program that errs in SD due to no tape, reports
255 OK incorrectly in output.
256 - After unmount, if restore job started, ask to mount.
257 - Convert all %x substitution variables, which are hard to remember
258 and read to %(variable-name). Idea from TMDA.
259 - Remove NextId for SQLite. Optimize.
260 - Move all SQL statements into a single location.
261 - Add UA rc and history files.
262 - put termcap (used by console) in ./configure and
263 allow -with-termcap-dir.
264 - Enhance time and size scanning routines.
265 - Fix Autoprune for Volumes to respect need for full save.
266 - Fix Win32 config file definition name on /install
267 - Compare tape to Client files (attributes, or attributes and data)
268 - Make all database Ids 64 bit.
269 - Write an applet for Linux.
270 - Implement new inter-daemon communications protocol.
271 - Allow console commands to detach or run in background.
272 - Fix status delay on storage daemon during rewind.
273 - Add SD message variables to control operator wait time
274 - Maximum Operator Wait
275 - Minimum Message Interval
276 - Maximum Message Interval
277 - Send Operator message when cannot read tape label.
278 - Verify level=Volume (scan only), level=Data (compare of data to file).
279 Verify level=Catalog, level=InitCatalog
281 - Add keyword search to show command in Console.
282 - Events : tape has more than xxx bytes.
283 - Complete code in Bacula Resources -- this will permit
284 reading a new config file at any time.
285 - Handle ctl-c in Console
286 - Implement script driven addition of File daemon to config files.
287 - Think about how to make Bacula work better with File (non-tape) archives.
288 - Write Unix emulator for Windows.
289 - Put memory utilization in Status output of each daemon
290 if full status requested or if some level of debug on.
291 - Make database type selectable by .conf files i.e. at runtime
292 - Set flag for uname -a. Add to Volume label.
293 - Implement throttled work queue.
294 - Restore files modified after date
295 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
296 - Implement Restore FileSet=
297 - Create a protocol.h and protocol.c where all protocol messages
299 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
300 - Timout a job or terminate if link goes down, or reopen link and query.
301 - Concept of precious tapes (cannot be reused).
302 - Make bcopy copy with a single tape drive.
303 - Permit changing ownership during restore.
306 > My suggestion: Add a feature on the systray menu-icon menu to request
307 > an immediate backup now. This would be useful for laptop users who may
308 > not be on the network when the regular scheduled backup is run.
310 > My wife's suggestion: Add a setting to the win32 client to allow it to
311 > shut down the machine after backup is complete (after, of course,
312 > displaying a "System will shut down in one minute, click here to cancel"
313 > warning dialog). This would be useful for sites that want user
314 > woorkstations to be shut down overnight to save power.
317 - From Terry Manderson <terry@apnic.net>
318 jobset { # new structure
319 name = "monthlyUnixBoxen"
322 jobs = "wakame;durian;soy;wasabi;miso" #new!
332 fileset = "genericUnixSet"
338 fileset = "genericUnixSet"
344 fileset = "UnixDevelBoxSet"
349 - Autolabel should be specified by DIR instead of SD.
352 - AutoScan (check checksum of tape)
353 - Format command = "format /dev/nst0"
357 - Seek resolution (usually corresponds to buffer size)
358 - EODErrorCode=ENOSPC or code
359 - Partial Read error code
360 - Partial write error code
361 - Nonformatted read error
362 - Nonformatted write error
363 - WriteProtected error
367 - IgnoreCloseErrors=yes
377 - FD sends unsaved file list to Director at end of job (see
379 - File daemon should build list of files skipped, and then
380 at end of save retry and report any errors.
381 - Write a Storage daemon that uses pipes and
382 standard Unix programs to write to the tape.
384 - Need something that monitors the JCR queue and
385 times out jobs by asking the deamons where they are.
386 - Enhance Jmsg code to permit buffering and saving to disk.
387 - device driver = "xxxx" for drives.
389 - Ensure that /dev/null works
390 - Need report class for messages. Perhaps
391 report resource where report=group of messages
392 - enhance scan_attrib and rename scan_jobtype, and
393 fill in code for "since" option
394 - Director needs a time after which the report status is sent
395 anyway -- or better yet, a retry time for the job.
396 - Don't reschedule a job if previous incarnation is still running.
397 - Some way to automatically backup everything is needed????
398 - Need a structure for pending actions:
400 - termination status (part of buffered msgs?)
402 Read, Write, Clean, Delete
403 - Login to Bacula; Bacula users with different permissions:
404 owner, group, user, quotas
405 - Store info on each file system type (probably in the job header on tape.
406 This could be the output of df; or perhaps some sort of /etc/mtab record.
409 - Design at hierarchial storage for Bacula. Migration and Clone.
410 - Implement FSM (File System Modules).
411 - Audit M_ error codes to ensure they are correct and consistent.
412 - Add variable break characters to lex analyzer.
413 Either a bit mask or a string of chars so that
414 the caller can change the break characters.
415 - Make a single T_BREAK to replace T_COMMA, etc.
416 - Ensure that File daemon and Storage daemon can
417 continue a save if the Director goes down (this
418 is NOT currently the case). Must detect socket error,
419 buffer messages for later.
420 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
421 - Add ability to backup to two Storage devices (two SD sessions) at
422 the same time -- e.g. onsite, offsite.
423 - Add the ability to consolidate old backup sets (basically do a restore
424 to tape and appropriately update the catalog). Compress Volume sets.
425 Might need to spool via file is only one drive is available.
426 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
427 someway to do so with every volume that has less than x% valid
431 Migration: Move a backup from one Volume to another
432 Clone: Copy a backup -- two Volumes
434 Bacula Migration is based on Jobs (apparently Networker is file by file).
436 Migration triggered by:
440 Highwater mark (keep total size)
444 Bacula Projects Roadmap
446 last update 8 May 2003
448 Item 1: Multiple simultaneous Jobs. (done)
449 Done -- Restore part needs better implementation to work correctly
450 Also, it needs considerable testing
452 What: Permit multiple simultaneous jobs in Bacula.
454 Why: An enterprise level solution needs to go fast without the
455 need for the system administrator to carefully tweak
456 timing. Based on the benchmarks, during a full
457 backup, NetWorker typically hit 10 times the bandwidth to
458 the tape compared to Bacula--largely. This is probably due to
459 running parallel jobs and multi-threaded filling of buffers
460 and writing them to tape. This should also make things work
461 better when you have a mix of fast and slow machines backing
464 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
465 implementing this is a matter of some small cleanups and
469 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
470 Deferred -- not necessary yet -- possibly implement with Migration.
472 What: If data is coming into the SD too fast, buffer it to
473 disk if the user has configured this option.
475 Why: This would be nice, especially if it more or less falls out
476 when implementing (1) above. If not, it probably should not
477 be given a high priority because fundamentally the backup time
478 is limited by the tape bandwidth. Even though you may finish a
479 client job quicker by spilling to disk, you still have to
480 eventually get it onto tape. If intermediate disk buffering
481 allows us to improve write bandwidth to tape, it may make
484 Notes: Whether or not this is implemented will depend upon performance
485 testing after item 1 is implemented.
488 Item 3: Write the bscan program -- also write a bcopy program.
491 What: Write a program that reads a Bacula tape and puts all the
492 appropriate data into the catalog. This allows recovery
493 from a tape that is no longer in the database, or it allows
494 re-creation of a database if lost.
496 Why: This is a fundamental robustness and disaster recovery tool
497 which will increase the comfort level of a sysadmin
498 considering adopting Bacula.
500 Notes: A skeleton of this program already exists, but much work
501 needs to be done. Implementing this will also make apparent
502 any deficiencies in the current Bacula tape format.
505 Item 4: Implement Base jobs.
507 What: A base job is sort of like a Full save except that you
508 will want the FileSet to contain only files that are unlikely
509 to change in the future (i.e. a snapshot of most of your
510 system after installing it). After the base job has been run,
511 when you are doing a Full save, you can specify to exclude
512 all files saved by the base job that have not been modified.
514 Why: This is something none of the competition does, as far as we know
515 (except BackupPC, which is a Perl program that saves to disk
516 only). It is big win for the user, it makes Bacula stand out
517 as offering a unique optimization that immediately saves time
520 Notes: Big savings in tape usage. Will require more resources because
521 the e. DIR must send FD a list of files/attribs, and the FD must
522 search the list and compare it for each file to be saved.
525 Item 5: Implement Label templates
528 What: This is a mechanism whereby Bacula can automatically create
529 a tape label for new tapes according to a detailed specification
530 provided by the user.
532 Why: It is a major convenience item for folks who use automated label
535 Notes: Bacula already has a working form of automatic tape label
536 creation, but it is very crude. The design for the complete
537 tape labeling project is already documented in the manual.
540 Item 6: Write a regression script.
541 Done -- Continue to expand its testing.
543 What: This is an automatic script that runs and tests as many features
544 of Bacula as possible. The output is compared to previous
545 versions of Bacula and any differences are reported.
547 Why: This is an enormous help in preventing introduction of new
548 errors in parts of the program that already work correctly.
550 Notes: This probably should be ranked higher, it's something the typical
551 user doesn't see. Depending on how it's implemented, it may
552 make sense to defer it until the archival tape format and
553 user interface mature.
556 Item 7: GUI for interactive restore
557 Item 8: GUI for interactive backup
559 What: The current interactive restore is implemented with a tty
560 interface. It would be much nicer to be able to "see" the
561 list of files backed up in typical GUI tree format.
562 The same mechanism could also be used for creating
563 ad-hoc backup FileSets (item 8).
565 Why: Ease of use -- especially for the end user.
567 Notes: Rather than implementing in Gtk, we probably should go directly
568 for a Browser implementation, even if doing so meant the
569 capability wouldn't be available until much later. Not only
570 is there the question of Windows sites, most
571 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
572 without installing lots of stuff admins are very wary about.
573 Real sysadmins will always use the command line anyway, and
574 the user who's doing an interactive restore or backup of his
575 own files will in most cases be on a Windows machine running
579 Item 9: Add SSL to daemon communications.
581 What: This provides for secure communications between the daemons.
583 Why: This would allow doing backup across the Internet without
584 privacy concerns (or with much less concern).
586 Notes: The vast majority of near term potential users will be backing up
587 a single site over a LAN and, correctly or not, they probably
588 won't be concerned with security, at least not enough to go to
589 the trouble to set up keys, etc. to screw things down. We suspect
590 that many users genuinely interested in multi-site backup
591 already run some form of VPN software in their internetwork
592 connections, and are willing to delegate security to that layer.
595 Item 10: Define definitive tape format.
598 What: Define that definitive tape format that will not change
599 for the next millennium.
601 Why: Stability, security.
603 Notes: See notes for item 11 below.
606 Item 11: New daemon communication protocol.
608 What: The current daemon to daemon protocol is basically an ASCII
609 printf() and sending the buffer. On the receiving end, the
610 buffer is sscanf()ed to unpack it. The new scheme would
611 be a binary format that allows quick packing and unpacking
612 of any data type with named fields.
614 Why: Using binary packing would be faster. Named fields will permit
615 error checking to ensure that what is sent is what the
616 receiver really wants.
618 Notes: These are internal improvements in the interest of the
619 long-term stability and evolution of the program. On the one
620 hand, the sooner they're done, the less code we have to rip
621 up when the time comes to install them. On the other hand, they
622 don't bring an immediately perceptible benefit to potential
623 users. Item 10 and possibly item 11 should be deferred until Bacula
624 is well established with a growing user community more or
625 less happy with the feature set. At that time, it will make a
626 good "next generation" upgrade in the interest of data
632 ======================================================
634 It is somewhat like a Full save becomes an incremental since
635 the Base job (or jobs) plus other non-base files.
637 - New BaseFile table that contains:
638 JobId, BaseJobId, FileId (from Base).
639 i.e. for each base file that exists but is not saved because
640 it has not changed, the File daemon sends the JobId, BaseId,
641 and FileId back to the Director who creates the DB entry.
642 - To initiate a Base save, the Director sends the FD
643 the FileId, and full filename for each file in the Base.
644 - When the FD finds a Base file, he requests the Director to
645 send him the full File entry (stat packet plus MD5), or
646 conversely, the FD sends it to the Director and the Director
647 says yes or no. This can be quite rapid if the FileId is kept
648 by the FD for each Base Filename.
649 - It is probably better to have the comparison done by the FD
650 despite the fact that the File entry must be sent across the
652 - An alternative would be to send the FD the whole File entry
653 from the start. The disadvantage is that it requires a lot of
654 space. The advantage is that it requires less communications
656 - The Job record must be updated to indicate that one or more
658 - At end of Job, FD returns:
659 1. Count of base files/bytes not written to tape (i.e. matches)
660 2. Count of base file that were saved i.e. had changed.
661 - No tape record would be written for a Base file that matches, in the
662 same way that no tape record is written for Incremental jobs where
663 the file is not saved because it is unchanged.
664 - On a restore, all the Base file records must explicitly be
665 found from the BaseFile tape. I.e. for each Full save that is marked
666 to have one or more Base Jobs, search the BaseFile for all occurrences
668 - An optimization might be to make the BaseFile have:
674 This would avoid the need to explicitly fetch each File record for
675 the Base job. The Base Job record will be fetched to get the
676 VolSessionId and VolSessionTime.
677 =========================================================
681 =============================================================
683 Request For Comments For File Backup Options
686 Subject: File Backup Options
689 A few days ago, a Bacula user who is backing up to file volumes and
690 using compression asked if it was possible to suppress compressing
691 all .gz files since it was a waste of CPU time. Although Bacula
692 currently permits using different options (compression, ...) on
693 a directory by directory basis, it cannot do it on a file by
694 file basis, which is clearly what was desired.
696 Proposed Implementation:
697 To solve this problem, I propose the following:
699 - Add a new Director resource type called FileOptions.
701 - The FileOptions resource will have records for all
702 options that can currently be specified on the Include record
703 (in a FileSet). Examples below.
705 - The FileOptions resource will permit an exclude option as well
706 as a number of additional options.
708 - The heart of the FileOptions resource is the ability to
709 supply any number of ApplyTo records which specify POSIX
710 regular expressions. These ApplyTo regular expressions are
711 applied to the fully qualified filename (path and all). If
712 one matches, then the FileOptions will be used.
714 - When an ApplyTo specification matches an included file, the
715 options specified in the FileOptions resource will override
716 the default options specified on the Include record.
718 - Include records will be modified to permit referencing one or
719 more FileOptions resources. The FileOptions will be used
720 in the order listed on the Include record and the first
721 one that matches will be applied.
723 - Options (or specifications) currently supplied on the Include
724 record will be deprecated (i.e. removed in a later version a
725 year or so from now).
727 - The Exclude record will be deprecated as the same functionality
728 can be obtained by using an Exclude = yes in the FileOptions.
731 The following records can appear in the FileOptions resource. An
732 asterisk preceding the name indicates a feature not currently
736 - Compression= (GZIP, ...)
737 - Signature= (MD5, SHA1, ...)
739 - OneFs= (yes/no) - remain on one filesystem
740 - Recurse= (yes/no) - recurse into subdirectories
741 - Sparse= (yes/no) - do sparse file backup
742 - *Exclude= (yes/no) - exclude file from being saved
743 - *Reader= (filename) - external read (backup) program
744 - *Plugin= (filename) - read/write plugin module
747 - verify= (ipnougsamc5) - verify options
750 - replace= (always/ifnewer/ifolder/never) - replace options currently
752 - *Writer= (filename) - external write (restore) program
756 Currently options specifying compression, MD5 signatures, recursion,
757 ... of a FileSet are supplied on the Include record. These will now
758 all be collected into a FileOptions resource, which will be
759 specified on the Include in place of the options. Multiple FileOptions
760 may be specified. Since the FileOptions contain regular expressions
761 that are applied to the full filename, this will give the ability
762 to specify backup options on a file by file basis to whatever level
771 Include = compression=GZIP signature=MD5 {
783 Match = /*.?*/ # matches all files.
788 That's a lot more to do the same thing, but it gives the ability to
789 apply options on a file by file basis. For example, suppose you
790 want to compress all files but not any file with extensions .gz or .Z.
791 You could do so as follows:
798 # Note multiple Matches are ORed
799 Match = /*.gz/ # matches .gz files */
800 Match = /*.Z/ # matches .Z files */
805 Match = /*.?*/ # matches all files
811 Now, since the NoCompress FileOptions is specified first on the
812 Include line, any *.gz or *.Z file will have an MD5 signature computed,
813 but will not be compressed. For all other files, the NoCompress will not
814 match, so the Opts options will be used which will include GZIP
818 - Is it necessary to provide some means of ANDing regular expressions
819 and negation? (not currently planned)
821 e.g. Match = /*.gz/ && !/big.gz/
823 - I see that Networker has a "null" module which, if specified, does not
824 backup the file, but does make an record of the file in the catalog
825 so that the catalog will reflect an exact picture of the filesystem.
826 The result is that the file can be "seen" when "browsing" the save
827 sets, but it cannot be restored.
829 Is this really useful? Should it be implemented in Bacula?
832 After implementing the above, the user will be able to specify
833 on a file by file basis (using regular expressions) what options are
834 applied for the backup.
837 =============================================
839 ==========================================================
841 For each Incremental job that is run, there may be files that
842 were found but not saved because they were locked (this applies
843 only to Windows). Such a system could send back to the Director
844 a list of Unsaved files.
846 - New UnSavedFiles table that contains:
850 - Then in the next Incremental job, the list of Unsaved Files will be
851 feed to the FD, who will ensure that they are explicitly chosen even
852 if standard date/time check would not have selected them.
853 =============================================================
856 Done: (see kernsdone for more)
857 - Implement new alist in FileSet scanning.
858 - bls should continue reading even if it finds Win32 data on the tape.
859 The error should be Warning rather the Error.
860 - Add user configurable timeout for connecting to SD.
861 - Unsaved Flag in Job record (use JobMissingFiles).
862 - Base Flag in Job record.
863 - Configure mtx-changer to have correct path to mtx.
864 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
865 - Make some way so that if a machine is skipped because it is not up
866 that Bacula will continue retrying for a specified period of time --
868 - Implement all command line args on run.
869 - Implement command line "restore" args.
870 - Implement "restore current select=no"
871 - Restore file modified before date
872 - Restore -- do nothing but show what would happen
873 - Add estimate to Console commands
874 - Use read_record.c in SD code.
875 - Fix read_record to handle multiple sessions.
876 - Tip from Steve Allam
877 mt -f /dev/nst0 defblksize 0
878 - Document "status" in the console.
879 - Document driving console from shell script.
880 - Write JobMedia records with max file size is reached on tape.
881 - Handle the case of multiple JobMedia records pending (i.e. the
882 thread is slow and multiple situations requiring a JobMedia
884 - Do performance analysis on the restore tree routines.
885 - Fix maximum file size (block.c) to generate JobMedia records.
886 - Make the default file size 1GB on the tape.
887 - Implement forward spacing between files.
888 - Add Machine type (Linux/Windows) to Status report for daemons.
890 - Use repositioning at the beginning of the tape.
891 - Do full check the command line args in update (e.g. VolStatus ...).
892 - Specify list of files to restore
893 - Implement ClientRunBeforeJob and ClientRunAfterJob.
894 - Make | and < work on FD side.
895 - Check to see if "blocked" is set during restore.
896 - Figure out what is interrupting sql command in console.
897 - Make new job print warning User Unmounted Tape.