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