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.
25 For 1.32 Testing/Documentation:
26 - Document new records in Director. SDAddress SDDeviceName, SDPassword.
27 FDPassword, FDAddress, DBAddress, DBPort, DBPassword.
28 - Document that it is 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.
34 - Enhance "update slots" to include a "scan" feature
35 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
36 - Allow a slot or range of slots on the label barcodes command.
37 - Don't print "Warning: Wrong Volume mounted ..." if mounting second volume.
42 - Get and test MySQL 4.0
43 - Do a complete audit of all pthreads_mutex, cond, ... to ensure that
44 any that are dynamically initialized are destroyed when no longer used.
45 - Write a mini-readline with history and editing.
46 - Look at how fuser works and /proc/PID/fd that is how Nic found the
47 file descriptor leak in Bacula.
48 - Implement WrapCounters in Counters.
49 - Turn on SIGHUP in dird.c and test.
50 - Use system dependent calls to get more precise info on tape errors.
51 - Add heartbeat from FD to SD if hb interval expires.
52 - Suppress read error on blank tape when doing a label.
53 - Can we dynamically change FileSets.
54 - If pool specified to label command and Label Format is specified,
55 automatically generate the Volume name.
56 - Take a careful look a the Basic recycling algorithm. When Bacula
57 chooses, the order should be:
59 - Look for Recycle or Purged
62 Instead of using lowest media Id, find the least recently used
65 When the tape is mounted and Bacula requests the status
66 - Do everything possible to use it.
68 Define a "available" status, which is the currently mounted
69 Volume and all volumes that are currently in the autochanger.
71 - Why can't SQL do the filename sort for restore?
72 - Is a pool specification really needed for a restore? Yes, and
73 you may want to exclude archive Pools.
74 - Look at libkse (man kse) for FreeBSD threading.
75 - Look into Microsoft Volume Shadowcopy Service VSS for backing
76 up system state components (Active Directory, System Volume, ...)
77 - Add ExhautiveRestoreSearch
78 - Look at the possibility of loading only the necessary
79 data into the restore tree (i.e. do it one directory at a
80 time as the user walks through the tree).
81 - Possibly use the hash code if the user selects all for a restore command.
82 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
83 - Implement some way for the File daemon to contact the Director
84 to start a job or pass its DHCP obtained IP number.
85 - Implement multiple Consoles.
86 - Add Console usr permissions.
87 - Fix "restore all" to bypass building the tree.
88 - Fix restore to list errors if Invalid block found, and if # files
89 restored does not match # expected.
90 - Prohibit backing up archive device (findlib/find_one.c:128)
91 - Implement Release Device in the Job resource to unmount a drive.
92 - Implement Acquire Device in the Job resource to mount a drive,
93 be sure this works with admin jobs so that the user can get
94 prompted to insert the correct tape. Possibly some way to say to
95 run the job but don't save the files.
96 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
97 - Implement FileOptions (see end of this document)
98 - Make things like list where a file is saved case independent for
100 - Edit the Client/Storage name into authentication failure messages.
101 - Implement job in VerifyToCatalog
103 - Implement a PostgreSQL driver.
104 - Bacula needs to propagate SD errors.
105 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
106 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
107 /dev/nst0. ERR=Input/output error.
108 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
109 Last block=5162 Current block=5164.
110 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
112 - Use autochanger to handle multiple devices.
113 - Fix packet too big problem.
114 - Add SuSE install doc to list.
115 - Check and rechedk "Invalid block number"
116 - Make bextract release the drive properly between tapes
117 so that an autochanger can be made to work.
118 - User wants to NOT backup up certain big files (email files).
119 - Maybe remove multiple simultaneous devices code in SD.
120 - On Windows with very long path names, it may be impossible to create
121 a file (and thus restore it) because the total length is too long.
122 We must cd into the directory then create the file without the
124 - lstat() is not going to work on Win32 for testing date.
125 - Something is not right in last block of fill command.
126 - Implement a Recycle command
127 - Add FileSet to command line arguments for restore.
128 - Allow multiple Storage specifications (or multiple names on
129 a single Storage specification) in the Job record. Thus a job
130 can be backed up to a number of storage devices.
131 - Add client name to cram-md5 challenge so Director can immediately
132 verify if it is the correct client.
133 - Implement ClientRunBeforeJob and ClientRunAfterJob.
134 - Add JobLevel in FD status (but make sure it is defined).
135 - Audit all UA commands to ensure that we always prompt where possible.
136 - Restrict characters permitted in a Resource name, and don't permit
138 - Check Jmsg in bnet, may not work, must dup bsock.
139 - Suppress Job Name in Jmsg for console
140 - Create Pools that are referenced in a Run statement at startup if possible.
141 - Use runbeforejob to unload, then reload a volume previously used,
142 then the next job run gets an error reading the drive.
143 - Make bootstrap filename unique.
144 - Test a second language e.g. french.
145 - Start working on Base jobs.
146 - Make "make binary-release" work from any directory.
147 - Implement UnsavedFiles DB record.
148 - Implement argc/argv for daemon command line scanning using table driven
150 - Implement table driven single argc/argv scanner to pickup all arguments.
151 Much like xxx_conf.c scan table.
152 keyword, handler(store_routine), store_address, code, flags, default.
153 - From Phil Stracchino:
154 It would probably be a per-client option, and would be called
155 something like, say, "Automatically purge obsoleted jobs". What it
156 would do is, when you successfully complete a Differential backup of a
157 client, it would automatically purge all Incremental backups for that
158 client that are rendered redundant by that Differential. Likewise,
159 when a Full backup on a client completed, it would automatically purge
160 all Differential and Incremental jobs obsoleted by that Full backup.
161 This would let people minimize the number of tapes they're keeping on
162 hand without having to master the art of retention times.
163 - Implement new serialize subroutines
164 send(socket, "string", &Vol, "uint32", &i, NULL)
165 - Scratch Pool where the volumes can be re-assigned to any Pool.
166 - Implement a M_SECURITY message class.
167 - Implement forward spacing block/file: position_device(bsr) --
168 just before read_block_from_device();
169 - When doing a Backup send all attributes back to the Director, who
170 would then figure out what files have been deleted.
171 - Currently in mount.c:236 the SD simply creates a Volume. It should have
172 explicit permission to do so. It should also mark the tape in error
173 if there is an error.
174 - Make sure all restore counters are working correctly in the FD.
175 - SD Bytes Read is wrong.
176 - Look at ALL higher level routines that call block.c to be sure
177 they don't expect something in errmsg.
178 - Investigate doing RAW backup of Win32 partition.
179 - Add thread specific data to hold the jcr -- send error messages from
180 low level routines by accessing it and using Jmsg().
181 - Cancel waiting for Client connect in SD if FD goes away.
183 - Add Progress command that periodically reports the progress of
185 - One block was orphaned in the SD probably after cancel.
188 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
189 - Implement timeout in response() when it should come quickly.
190 - Implement console @echo command.
191 - Implement a Slot priority (loaded/not loaded).
192 - Implement "vacation" Incremental only saves.
193 - Implement single pane restore (much like the Gftp panes).
194 - Implement Automatic Mount even in operator wait.
195 - Implement create "FileSet"?
196 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
197 - Implement "scratch pool" where tapes are defined and can be
198 taken by any pool that needs them.
199 - Implement restore "current system", but take all files without
200 doing selection tree -- so that jobs without File records can
202 - Implement disk spooling. Two parts: 1. Spool to disk then
203 immediately to tape to speed up tape operations. 2. Spool to
204 disk only when the tape is full, then when a tape is hung move
206 - Implement a relocatable bacula.spec
207 - Implement dump/print label to UA
208 - Add prefixlinks to where or not where absolute links to FD.
209 - Issue message to mount a new tape before the rewind.
210 - Simplified client job initiation for portables.
211 - If SD cannot open a drive, make it periodically retry.
212 - Implement LabelTemplate (at least first cut).
213 - Add more of the config info to the tape label.
214 - Implement a Mount Command and an Unmount Command where
215 the users could specify a system command to be performed
216 to do the mount, after which Bacula could attempt to
217 read the device. This is for Removeable media such as a CDROM.
218 - Most likely, this mount command would be invoked explicitly
219 by the user using the current Console "mount" and "unmount"
220 commands -- the Storage Daemon would do the right thing
221 depending on the exact nature of the device.
222 - As with tape drives, when Bacula wanted a new removable
223 disk mounted, it would unmount the old one, and send a message
224 to the user, who would then use "mount" as described above
225 once he had actually inserted the disk.
227 - If tape is marked read-only, then try opening it read-only rather than
228 failing, and remember that it cannot be written.
229 - Refine SD waiting output:
230 Device is being positioned
231 > Device is being positioned for append
232 > Device is being positioned to file x
234 - Figure out some way to estimate output size and to avoid splitting
235 a backup across two Volumes -- this could be useful for writing CDROMs
236 where you really prefer not to have it split -- not serious.
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 - Figure out how to do a bare metal Windows restore
259 - Possibly add email to Watchdog if drive is unmounted too
260 long and a job is waiting on the drive.
261 - Restore program that errs in SD due to no tape, reports
262 OK incorrectly in output.
263 - After unmount, if restore job started, ask to mount.
264 - Convert all %x substitution variables, which are hard to remember
265 and read to %(variable-name). Idea from TMDA.
266 - Remove NextId for SQLite. Optimize.
267 - Move all SQL statements into a single location.
268 - Add UA rc and history files.
269 - put termcap (used by console) in ./configure and
270 allow -with-termcap-dir.
271 - Enhance time and size scanning routines.
272 - Fix Autoprune for Volumes to respect need for full save.
273 - Fix Win32 config file definition name on /install
274 - Compare tape to Client files (attributes, or attributes and data)
275 - Make all database Ids 64 bit.
276 - Write an applet for Linux.
277 - Implement new inter-daemon communications protocol.
278 - Allow console commands to detach or run in background.
279 - Fix status delay on storage daemon during rewind.
280 - Add SD message variables to control operator wait time
281 - Maximum Operator Wait
282 - Minimum Message Interval
283 - Maximum Message Interval
284 - Send Operator message when cannot read tape label.
285 - Verify level=Volume (scan only), level=Data (compare of data to file).
286 Verify level=Catalog, level=InitCatalog
288 - Add keyword search to show command in Console.
289 - Events : tape has more than xxx bytes.
290 - Complete code in Bacula Resources -- this will permit
291 reading a new config file at any time.
292 - Handle ctl-c in Console
293 - Implement script driven addition of File daemon to config files.
294 - Think about how to make Bacula work better with File (non-tape) archives.
295 - Write Unix emulator for Windows.
296 - Put memory utilization in Status output of each daemon
297 if full status requested or if some level of debug on.
298 - Make database type selectable by .conf files i.e. at runtime
299 - Set flag for uname -a. Add to Volume label.
300 - Implement throttled work queue.
301 - Restore files modified after date
302 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
303 - Implement Restore FileSet=
304 - Create a protocol.h and protocol.c where all protocol messages
306 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
307 - Timout a job or terminate if link goes down, or reopen link and query.
308 - Concept of precious tapes (cannot be reused).
309 - Make bcopy copy with a single tape drive.
310 - Permit changing ownership during restore.
313 > My suggestion: Add a feature on the systray menu-icon menu to request
314 > an immediate backup now. This would be useful for laptop users who may
315 > not be on the network when the regular scheduled backup is run.
317 > My wife's suggestion: Add a setting to the win32 client to allow it to
318 > shut down the machine after backup is complete (after, of course,
319 > displaying a "System will shut down in one minute, click here to cancel"
320 > warning dialog). This would be useful for sites that want user
321 > woorkstations to be shut down overnight to save power.
324 - From Terry Manderson <terry@apnic.net>
325 jobset { # new structure
326 name = "monthlyUnixBoxen"
329 jobs = "wakame;durian;soy;wasabi;miso" #new!
339 fileset = "genericUnixSet"
345 fileset = "genericUnixSet"
351 fileset = "UnixDevelBoxSet"
356 - Autolabel should be specified by DIR instead of SD.
359 - AutoScan (check checksum of tape)
360 - Format command = "format /dev/nst0"
364 - Seek resolution (usually corresponds to buffer size)
365 - EODErrorCode=ENOSPC or code
366 - Partial Read error code
367 - Partial write error code
368 - Nonformatted read error
369 - Nonformatted write error
370 - WriteProtected error
374 - IgnoreCloseErrors=yes
384 - FD sends unsaved file list to Director at end of job (see
386 - File daemon should build list of files skipped, and then
387 at end of save retry and report any errors.
388 - Write a Storage daemon that uses pipes and
389 standard Unix programs to write to the tape.
391 - Need something that monitors the JCR queue and
392 times out jobs by asking the deamons where they are.
393 - Enhance Jmsg code to permit buffering and saving to disk.
394 - device driver = "xxxx" for drives.
396 - Ensure that /dev/null works
397 - Need report class for messages. Perhaps
398 report resource where report=group of messages
399 - enhance scan_attrib and rename scan_jobtype, and
400 fill in code for "since" option
401 - Director needs a time after which the report status is sent
402 anyway -- or better yet, a retry time for the job.
403 - Don't reschedule a job if previous incarnation is still running.
404 - Some way to automatically backup everything is needed????
405 - Need a structure for pending actions:
407 - termination status (part of buffered msgs?)
409 Read, Write, Clean, Delete
410 - Login to Bacula; Bacula users with different permissions:
411 owner, group, user, quotas
412 - Store info on each file system type (probably in the job header on tape.
413 This could be the output of df; or perhaps some sort of /etc/mtab record.
416 - Design at hierarchial storage for Bacula. Migration and Clone.
417 - Implement FSM (File System Modules).
418 - Audit M_ error codes to ensure they are correct and consistent.
419 - Add variable break characters to lex analyzer.
420 Either a bit mask or a string of chars so that
421 the caller can change the break characters.
422 - Make a single T_BREAK to replace T_COMMA, etc.
423 - Ensure that File daemon and Storage daemon can
424 continue a save if the Director goes down (this
425 is NOT currently the case). Must detect socket error,
426 buffer messages for later.
427 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
428 - Add ability to backup to two Storage devices (two SD sessions) at
429 the same time -- e.g. onsite, offsite.
430 - Add the ability to consolidate old backup sets (basically do a restore
431 to tape and appropriately update the catalog). Compress Volume sets.
432 Might need to spool via file is only one drive is available.
433 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
434 someway to do so with every volume that has less than x% valid
438 Migration: Move a backup from one Volume to another
439 Clone: Copy a backup -- two Volumes
441 Bacula Migration is based on Jobs (apparently Networker is file by file).
443 Migration triggered by:
447 Highwater mark (keep total size)
451 Bacula Projects Roadmap
453 last update 8 May 2003
455 Item 1: Multiple simultaneous Jobs. (done)
456 Done -- Restore part needs better implementation to work correctly
457 Also, it needs considerable testing
459 What: Permit multiple simultaneous jobs in Bacula.
461 Why: An enterprise level solution needs to go fast without the
462 need for the system administrator to carefully tweak
463 timing. Based on the benchmarks, during a full
464 backup, NetWorker typically hit 10 times the bandwidth to
465 the tape compared to Bacula--largely. This is probably due to
466 running parallel jobs and multi-threaded filling of buffers
467 and writing them to tape. This should also make things work
468 better when you have a mix of fast and slow machines backing
471 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
472 implementing this is a matter of some small cleanups and
476 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
477 Deferred -- not necessary yet -- possibly implement with Migration.
479 What: If data is coming into the SD too fast, buffer it to
480 disk if the user has configured this option.
482 Why: This would be nice, especially if it more or less falls out
483 when implementing (1) above. If not, it probably should not
484 be given a high priority because fundamentally the backup time
485 is limited by the tape bandwidth. Even though you may finish a
486 client job quicker by spilling to disk, you still have to
487 eventually get it onto tape. If intermediate disk buffering
488 allows us to improve write bandwidth to tape, it may make
491 Notes: Whether or not this is implemented will depend upon performance
492 testing after item 1 is implemented.
495 Item 3: Write the bscan program -- also write a bcopy program.
498 What: Write a program that reads a Bacula tape and puts all the
499 appropriate data into the catalog. This allows recovery
500 from a tape that is no longer in the database, or it allows
501 re-creation of a database if lost.
503 Why: This is a fundamental robustness and disaster recovery tool
504 which will increase the comfort level of a sysadmin
505 considering adopting Bacula.
507 Notes: A skeleton of this program already exists, but much work
508 needs to be done. Implementing this will also make apparent
509 any deficiencies in the current Bacula tape format.
512 Item 4: Implement Base jobs.
514 What: A base job is sort of like a Full save except that you
515 will want the FileSet to contain only files that are unlikely
516 to change in the future (i.e. a snapshot of most of your
517 system after installing it). After the base job has been run,
518 when you are doing a Full save, you can specify to exclude
519 all files saved by the base job that have not been modified.
521 Why: This is something none of the competition does, as far as we know
522 (except BackupPC, which is a Perl program that saves to disk
523 only). It is big win for the user, it makes Bacula stand out
524 as offering a unique optimization that immediately saves time
527 Notes: Big savings in tape usage. Will require more resources because
528 the e. DIR must send FD a list of files/attribs, and the FD must
529 search the list and compare it for each file to be saved.
532 Item 5: Implement Label templates
535 What: This is a mechanism whereby Bacula can automatically create
536 a tape label for new tapes according to a detailed specification
537 provided by the user.
539 Why: It is a major convenience item for folks who use automated label
542 Notes: Bacula already has a working form of automatic tape label
543 creation, but it is very crude. The design for the complete
544 tape labeling project is already documented in the manual.
547 Item 6: Write a regression script.
548 Done -- Continue to expand its testing.
550 What: This is an automatic script that runs and tests as many features
551 of Bacula as possible. The output is compared to previous
552 versions of Bacula and any differences are reported.
554 Why: This is an enormous help in preventing introduction of new
555 errors in parts of the program that already work correctly.
557 Notes: This probably should be ranked higher, it's something the typical
558 user doesn't see. Depending on how it's implemented, it may
559 make sense to defer it until the archival tape format and
560 user interface mature.
563 Item 7: GUI for interactive restore
564 Item 8: GUI for interactive backup
566 What: The current interactive restore is implemented with a tty
567 interface. It would be much nicer to be able to "see" the
568 list of files backed up in typical GUI tree format.
569 The same mechanism could also be used for creating
570 ad-hoc backup FileSets (item 8).
572 Why: Ease of use -- especially for the end user.
574 Notes: Rather than implementing in Gtk, we probably should go directly
575 for a Browser implementation, even if doing so meant the
576 capability wouldn't be available until much later. Not only
577 is there the question of Windows sites, most
578 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
579 without installing lots of stuff admins are very wary about.
580 Real sysadmins will always use the command line anyway, and
581 the user who's doing an interactive restore or backup of his
582 own files will in most cases be on a Windows machine running
586 Item 9: Add SSL to daemon communications.
588 What: This provides for secure communications between the daemons.
590 Why: This would allow doing backup across the Internet without
591 privacy concerns (or with much less concern).
593 Notes: The vast majority of near term potential users will be backing up
594 a single site over a LAN and, correctly or not, they probably
595 won't be concerned with security, at least not enough to go to
596 the trouble to set up keys, etc. to screw things down. We suspect
597 that many users genuinely interested in multi-site backup
598 already run some form of VPN software in their internetwork
599 connections, and are willing to delegate security to that layer.
602 Item 10: Define definitive tape format.
605 What: Define that definitive tape format that will not change
606 for the next millennium.
608 Why: Stability, security.
610 Notes: See notes for item 11 below.
613 Item 11: New daemon communication protocol.
615 What: The current daemon to daemon protocol is basically an ASCII
616 printf() and sending the buffer. On the receiving end, the
617 buffer is sscanf()ed to unpack it. The new scheme would
618 be a binary format that allows quick packing and unpacking
619 of any data type with named fields.
621 Why: Using binary packing would be faster. Named fields will permit
622 error checking to ensure that what is sent is what the
623 receiver really wants.
625 Notes: These are internal improvements in the interest of the
626 long-term stability and evolution of the program. On the one
627 hand, the sooner they're done, the less code we have to rip
628 up when the time comes to install them. On the other hand, they
629 don't bring an immediately perceptible benefit to potential
630 users. Item 10 and possibly item 11 should be deferred until Bacula
631 is well established with a growing user community more or
632 less happy with the feature set. At that time, it will make a
633 good "next generation" upgrade in the interest of data
639 ======================================================
641 It is somewhat like a Full save becomes an incremental since
642 the Base job (or jobs) plus other non-base files.
644 - New BaseFile table that contains:
645 JobId, BaseJobId, FileId (from Base).
646 i.e. for each base file that exists but is not saved because
647 it has not changed, the File daemon sends the JobId, BaseId,
648 and FileId back to the Director who creates the DB entry.
649 - To initiate a Base save, the Director sends the FD
650 the FileId, and full filename for each file in the Base.
651 - When the FD finds a Base file, he requests the Director to
652 send him the full File entry (stat packet plus MD5), or
653 conversely, the FD sends it to the Director and the Director
654 says yes or no. This can be quite rapid if the FileId is kept
655 by the FD for each Base Filename.
656 - It is probably better to have the comparison done by the FD
657 despite the fact that the File entry must be sent across the
659 - An alternative would be to send the FD the whole File entry
660 from the start. The disadvantage is that it requires a lot of
661 space. The advantage is that it requires less communications
663 - The Job record must be updated to indicate that one or more
665 - At end of Job, FD returns:
666 1. Count of base files/bytes not written to tape (i.e. matches)
667 2. Count of base file that were saved i.e. had changed.
668 - No tape record would be written for a Base file that matches, in the
669 same way that no tape record is written for Incremental jobs where
670 the file is not saved because it is unchanged.
671 - On a restore, all the Base file records must explicitly be
672 found from the BaseFile tape. I.e. for each Full save that is marked
673 to have one or more Base Jobs, search the BaseFile for all occurrences
675 - An optimization might be to make the BaseFile have:
681 This would avoid the need to explicitly fetch each File record for
682 the Base job. The Base Job record will be fetched to get the
683 VolSessionId and VolSessionTime.
684 =========================================================
688 =============================================================
690 Request For Comments For File Backup Options
693 Subject: File Backup Options
696 A few days ago, a Bacula user who is backing up to file volumes and
697 using compression asked if it was possible to suppress compressing
698 all .gz files since it was a waste of CPU time. Although Bacula
699 currently permits using different options (compression, ...) on
700 a directory by directory basis, it cannot do it on a file by
701 file basis, which is clearly what was desired.
703 Proposed Implementation:
704 To solve this problem, I propose the following:
706 - Add a new Director resource type called Options.
708 - The Options resource will have records for all
709 options that can currently be specified on the Include record
710 (in a FileSet). Examples below.
712 - The Options resource will permit an exclude option as well
713 as a number of additional options.
715 - The heart of the Options resource is the ability to
716 supply any number of Match records which specify POSIX
717 regular expressions. These Match regular expressions are
718 applied to the fully qualified filename (path and all). If
719 one matches, then the Options will be used.
721 - When an Match specification matches an included file, the
722 options specified in the Options resource will override
723 the default options specified on the Include record.
725 - Include records will be modified to permit referencing one or
726 more Options resources. The Options will be used
727 in the order listed on the Include record and the first
728 one that matches will be applied.
730 - Options (or specifications) currently supplied on the Include
731 record will be deprecated (i.e. removed in a later version a
732 year or so from now).
734 - The Exclude record will be deprecated as the same functionality
735 can be obtained by using an Exclude = yes in the Options.
738 The following records can appear in the Options resource. An
739 asterisk preceding the name indicates a feature not currently
743 - Compression= (GZIP, ...)
744 - Signature= (MD5, SHA1, ...)
746 - OneFs= (yes/no) - remain on one filesystem
747 - Recurse= (yes/no) - recurse into subdirectories
748 - Sparse= (yes/no) - do sparse file backup
749 - *Exclude= (yes/no) - exclude file from being saved
750 - *Reader= (filename) - external read (backup) program
751 - *Plugin= (filename) - read/write plugin module
754 - verify= (ipnougsamc5) - verify options
757 - replace= (always/ifnewer/ifolder/never) - replace options currently
759 - *Writer= (filename) - external write (restore) program
763 Currently options specifying compression, MD5 signatures, recursion,
764 ... of a FileSet are supplied on the Include record. These will now
765 all be collected into a Options resource, which will be
766 specified in the Include in place of the options. Multiple Options
767 may be specified. Since the Options may contain regular expressions
768 that are applied to the full filename, this will give the ability
769 to specify backup options on a file by file basis to whatever level
778 Include = compression=GZIP signature=MD5 {
790 Match = /*.?*/ # matches all files.
795 That's a lot more to do the same thing, but it gives the ability to
796 apply options on a file by file basis. For example, suppose you
797 want to compress all files but not any file with extensions .gz or .Z.
798 In that case, you will need to group two sets of options using
799 the Options resource as follows:
807 # Note multiple Matches are ORed
808 Match = /*.gz/ # matches .gz files */
809 Match = /*.Z/ # matches .Z files */
814 Match = /*.?*/ # matches all files
820 Now, since the no Compression option is specified in the
821 first group of Options, *.gz or *.Z file will have an MD5 signature computed,
822 but will not be compressed. For all other files, the *.gz *.Z will not
823 match, so the second group of options will be used which will include GZIP
827 - Is it necessary to provide some means of ANDing regular expressions
828 and negation? (not currently planned)
830 e.g. Match = /*.gz/ && !/big.gz/
832 - I see that Networker has a "null" module which, if specified, does not
833 backup the file, but does make an record of the file in the catalog
834 so that the catalog will reflect an exact picture of the filesystem.
835 The result is that the file can be "seen" when "browsing" the save
836 sets, but it cannot be restored.
838 Is this really useful? Should it be implemented in Bacula?
841 After implementing the above, the user will be able to specify
842 on a file by file basis (using regular expressions) what options are
843 applied for the backup.
846 =============================================
848 ==========================================================
850 For each Incremental job that is run, there may be files that
851 were found but not saved because they were locked (this applies
852 only to Windows). Such a system could send back to the Director
853 a list of Unsaved files.
855 - New UnSavedFiles table that contains:
859 - Then in the next Incremental job, the list of Unsaved Files will be
860 feed to the FD, who will ensure that they are explicitly chosen even
861 if standard date/time check would not have selected them.
862 =============================================================
865 Done: (see kernsdone for more)
866 - Implement new alist in FileSet scanning.
867 - bls should continue reading even if it finds Win32 data on the tape.
868 The error should be Warning rather the Error.
869 - Add user configurable timeout for connecting to SD.
870 - Unsaved Flag in Job record (use JobMissingFiles).
871 - Base Flag in Job record.
872 - Configure mtx-changer to have correct path to mtx.
873 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
874 - Make some way so that if a machine is skipped because it is not up
875 that Bacula will continue retrying for a specified period of time --
877 - Implement all command line args on run.
878 - Implement command line "restore" args.
879 - Implement "restore current select=no"
880 - Restore file modified before date
881 - Restore -- do nothing but show what would happen
882 - Add estimate to Console commands
883 - Use read_record.c in SD code.
884 - Fix read_record to handle multiple sessions.
885 - Tip from Steve Allam
886 mt -f /dev/nst0 defblksize 0
887 - Document "status" in the console.
888 - Document driving console from shell script.
889 - Write JobMedia records with max file size is reached on tape.
890 - Handle the case of multiple JobMedia records pending (i.e. the
891 thread is slow and multiple situations requiring a JobMedia
893 - Do performance analysis on the restore tree routines.
894 - Fix maximum file size (block.c) to generate JobMedia records.
895 - Make the default file size 1GB on the tape.
896 - Implement forward spacing between files.
897 - Add Machine type (Linux/Windows) to Status report for daemons.
899 - Use repositioning at the beginning of the tape.
900 - Do full check the command line args in update (e.g. VolStatus ...).
901 - Specify list of files to restore
902 - Implement ClientRunBeforeJob and ClientRunAfterJob.
903 - Make | and < work on FD side.
904 - Check to see if "blocked" is set during restore.
905 - Figure out what is interrupting sql command in console.
906 - Make new job print warning User Unmounted Tape.
907 - Test recycling and purging (code changed in db_find_next_volume and
909 - Document SDConnectTimeout (in FD).
910 - Add restore by filename test.
911 - Document restore by files.
912 - Make variable expansion work correctly.
913 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
914 - Copy static programs into install directory.
915 - Think about changing Storage resource Device record to be
917 - Add RunBeforeJob and RunAfterJob to the Client program.
918 - Need return status on read_cb() from read_records(). Need multiple
919 records -- one per Job, maybe a JCR or some other structure with
920 a block and a record.
921 - LabelFormat on tape volume apparently creates the db record but
922 never actually labels the volume.
923 - Recycling a volume when two jobs are using it is going to break. Fixed.
924 - Document list nextvol and new format status dir.
925 - Client files in Win32 with Unix eol conventions doesn't work.
926 - Either fix or document that fill command in btape can be
927 compressed enormously by the hardware - a 36GB tape wrote 750GB!
928 - Add multiple character duration qualifiers.
929 - Require some modifer.
930 - Restrict characters permitted in a Resource name, and don't permit