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.
24 - Test connect timeouts.
26 For 1.32 Testing/Documentation:
27 - Document new records in Director. SDAddress SDDeviceName, SDPassword.
28 FDPassword, FDAddress, DBAddress, DBPort, DBPassword.
29 - Document new Include/Exclude ...
30 - Add test of exclusion, test multiple Include {} statements.
31 - Add counter variable test.
36 - When job rescheduled, status gives is waiting for Client Rufus
37 to connect to Storage File. Dir needs to inform SD that job
39 - Fix get_storage_from_media_type (ua_restore) to use command line
41 - Enhance "update slots" to include a "scan" feature
42 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
43 - Allow a slot or range of slots on the label barcodes command.
44 - Don't print "Warning: Wrong Volume mounted ..." if mounting second volume.
45 - Make Dmsg look at global before calling subroutine.
46 - Enable trace output at runtime for Win32
47 - Make sure that Volumes are recycled based on "Least recently used"
48 rather than lowest MediaId.
49 - Available volumes for autochangers (see patrick@baanboard.com 3 Sep 03
50 and 4 Sep) scan slots.
51 - Upgrade to cygwin 1.5
53 - Get and test MySQL 4.0
54 - Do a complete audit of all pthreads_mutex, cond, ... to ensure that
55 any that are dynamically initialized are destroyed when no longer used.
56 - Write a mini-readline with history and editing.
57 - Look at how fuser works and /proc/PID/fd that is how Nic found the
58 file descriptor leak in Bacula.
59 - Implement WrapCounters in Counters.
60 - Turn on SIGHUP in dird.c and test.
61 - Use system dependent calls to get more precise info on tape errors.
62 - Add heartbeat from FD to SD if hb interval expires.
63 - Suppress read error on blank tape when doing a label.
64 - Can we dynamically change FileSets.
65 - If pool specified to label command and Label Format is specified,
66 automatically generate the Volume name.
67 - Take a careful look a the Basic recycling algorithm. When Bacula
68 chooses, the order should be:
70 - Look for Recycle or Purged
73 Instead of using lowest media Id, find the least recently used
76 When the tape is mounted and Bacula requests the status
77 - Do everything possible to use it.
79 Define a "available" status, which is the currently mounted
80 Volume and all volumes that are currently in the autochanger.
82 - Why can't SQL do the filename sort for restore?
83 - Is a pool specification really needed for a restore? Yes, and
84 you may want to exclude archive Pools.
85 - Look at libkse (man kse) for FreeBSD threading.
86 - Look into Microsoft Volume Shadowcopy Service VSS for backing
87 up system state components (Active Directory, System Volume, ...)
88 - Add ExhautiveRestoreSearch
89 - Look at the possibility of loading only the necessary
90 data into the restore tree (i.e. do it one directory at a
91 time as the user walks through the tree).
92 - Possibly use the hash code if the user selects all for a restore command.
93 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
94 - Implement some way for the File daemon to contact the Director
95 to start a job or pass its DHCP obtained IP number.
96 - Implement multiple Consoles.
97 - Add Console usr permissions.
98 - Fix "restore all" to bypass building the tree.
99 - Fix restore to list errors if Invalid block found, and if # files
100 restored does not match # expected.
101 - Prohibit backing up archive device (findlib/find_one.c:128)
102 - Implement Release Device in the Job resource to unmount a drive.
103 - Implement Acquire Device in the Job resource to mount a drive,
104 be sure this works with admin jobs so that the user can get
105 prompted to insert the correct tape. Possibly some way to say to
106 run the job but don't save the files.
107 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
108 - Implement FileOptions (see end of this document)
109 - Make things like list where a file is saved case independent for
111 - Edit the Client/Storage name into authentication failure messages.
112 - Implement job in VerifyToCatalog
114 - Implement a PostgreSQL driver.
115 - Bacula needs to propagate SD errors.
116 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
117 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
118 /dev/nst0. ERR=Input/output error.
119 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
120 Last block=5162 Current block=5164.
121 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
123 - Use autochanger to handle multiple devices.
124 - Fix packet too big problem.
125 - Add SuSE install doc to list.
126 - Check and rechedk "Invalid block number"
127 - Make bextract release the drive properly between tapes
128 so that an autochanger can be made to work.
129 - User wants to NOT backup up certain big files (email files).
130 - Maybe remove multiple simultaneous devices code in SD.
131 - On Windows with very long path names, it may be impossible to create
132 a file (and thus restore it) because the total length is too long.
133 We must cd into the directory then create the file without the
135 - lstat() is not going to work on Win32 for testing date.
136 - Something is not right in last block of fill command.
137 - Implement a Recycle command
138 - Add FileSet to command line arguments for restore.
139 - Allow multiple Storage specifications (or multiple names on
140 a single Storage specification) in the Job record. Thus a job
141 can be backed up to a number of storage devices.
142 - Add client name to cram-md5 challenge so Director can immediately
143 verify if it is the correct client.
144 - Implement ClientRunBeforeJob and ClientRunAfterJob.
145 - Add JobLevel in FD status (but make sure it is defined).
146 - Audit all UA commands to ensure that we always prompt where possible.
147 - Restrict characters permitted in a Resource name, and don't permit
149 - Check Jmsg in bnet, may not work, must dup bsock.
150 - Suppress Job Name in Jmsg for console
151 - Create Pools that are referenced in a Run statement at startup if possible.
152 - Use runbeforejob to unload, then reload a volume previously used,
153 then the next job run gets an error reading the drive.
154 - Make bootstrap filename unique.
155 - Test a second language e.g. french.
156 - Start working on Base jobs.
157 - Make "make binary-release" work from any directory.
158 - Implement UnsavedFiles DB record.
159 - Implement argc/argv for daemon command line scanning using table driven
161 - Implement table driven single argc/argv scanner to pickup all arguments.
162 Much like xxx_conf.c scan table.
163 keyword, handler(store_routine), store_address, code, flags, default.
164 - From Phil Stracchino:
165 It would probably be a per-client option, and would be called
166 something like, say, "Automatically purge obsoleted jobs". What it
167 would do is, when you successfully complete a Differential backup of a
168 client, it would automatically purge all Incremental backups for that
169 client that are rendered redundant by that Differential. Likewise,
170 when a Full backup on a client completed, it would automatically purge
171 all Differential and Incremental jobs obsoleted by that Full backup.
172 This would let people minimize the number of tapes they're keeping on
173 hand without having to master the art of retention times.
174 - Implement new serialize subroutines
175 send(socket, "string", &Vol, "uint32", &i, NULL)
176 - Scratch Pool where the volumes can be re-assigned to any Pool.
177 - Implement a M_SECURITY message class.
178 - Implement forward spacing block/file: position_device(bsr) --
179 just before read_block_from_device();
180 - When doing a Backup send all attributes back to the Director, who
181 would then figure out what files have been deleted.
182 - Currently in mount.c:236 the SD simply creates a Volume. It should have
183 explicit permission to do so. It should also mark the tape in error
184 if there is an error.
185 - Make sure all restore counters are working correctly in the FD.
186 - SD Bytes Read is wrong.
187 - Look at ALL higher level routines that call block.c to be sure
188 they don't expect something in errmsg.
189 - Investigate doing RAW backup of Win32 partition.
190 - Add thread specific data to hold the jcr -- send error messages from
191 low level routines by accessing it and using Jmsg().
192 - Cancel waiting for Client connect in SD if FD goes away.
194 - Add Progress command that periodically reports the progress of
196 - One block was orphaned in the SD probably after cancel.
199 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
200 - Implement timeout in response() when it should come quickly.
201 - Implement console @echo command.
202 - Implement a Slot priority (loaded/not loaded).
203 - Implement "vacation" Incremental only saves.
204 - Implement single pane restore (much like the Gftp panes).
205 - Implement Automatic Mount even in operator wait.
206 - Implement create "FileSet"?
207 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
208 - Implement "scratch pool" where tapes are defined and can be
209 taken by any pool that needs them.
210 - Implement restore "current system", but take all files without
211 doing selection tree -- so that jobs without File records can
213 - Implement disk spooling. Two parts: 1. Spool to disk then
214 immediately to tape to speed up tape operations. 2. Spool to
215 disk only when the tape is full, then when a tape is hung move
217 - Implement a relocatable bacula.spec
218 - Implement dump/print label to UA
219 - Add prefixlinks to where or not where absolute links to FD.
220 - Issue message to mount a new tape before the rewind.
221 - Simplified client job initiation for portables.
222 - If SD cannot open a drive, make it periodically retry.
223 - Implement LabelTemplate (at least first cut).
224 - Add more of the config info to the tape label.
225 - Implement a Mount Command and an Unmount Command where
226 the users could specify a system command to be performed
227 to do the mount, after which Bacula could attempt to
228 read the device. This is for Removeable media such as a CDROM.
229 - Most likely, this mount command would be invoked explicitly
230 by the user using the current Console "mount" and "unmount"
231 commands -- the Storage Daemon would do the right thing
232 depending on the exact nature of the device.
233 - As with tape drives, when Bacula wanted a new removable
234 disk mounted, it would unmount the old one, and send a message
235 to the user, who would then use "mount" as described above
236 once he had actually inserted the disk.
238 - If tape is marked read-only, then try opening it read-only rather than
239 failing, and remember that it cannot be written.
240 - Refine SD waiting output:
241 Device is being positioned
242 > Device is being positioned for append
243 > Device is being positioned to file x
245 - Figure out some way to estimate output size and to avoid splitting
246 a backup across two Volumes -- this could be useful for writing CDROMs
247 where you really prefer not to have it split -- not serious.
248 - Have SD compute MD5 or SHA1 and compare to what FD computes.
249 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
250 actual data on the Volume and compare it.
251 - Implement Bacula plugins -- design API
252 - Make bcopy read through bad tape records.
253 - Program files (i.e. execute a program to read/write files).
254 Pass read date of last backup, size of file last time.
255 - Add Signature type to File DB record.
256 - CD into subdirectory when open()ing files for backup to
257 speed up things. Test with testfind().
258 - Priority job to go to top of list.
259 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
260 - Implement some way for the Console to dynamically create a job.
261 - Restore to a particular time -- e.g. before date, after date.
262 - Solaris -I on tar for include list
263 - Need a verbose mode in restore, perhaps to bsr.
264 - bscan without -v is too quiet -- perhaps show jobs.
265 - Add code to reject whole blocks if not wanted on restore.
266 - Check if we can increase Bacula FD priorty in Win2000
267 - Make sure the MaxVolFiles is fully implemented in SD
268 - Check if both CatalogFiles and UseCatalog are set to SD.
269 - Figure out how to do a bare metal Windows restore
270 - Possibly add email to Watchdog if drive is unmounted too
271 long and a job is waiting on the drive.
272 - Restore program that errs in SD due to no tape, reports
273 OK incorrectly in output.
274 - After unmount, if restore job started, ask to mount.
275 - Convert all %x substitution variables, which are hard to remember
276 and read to %(variable-name). Idea from TMDA.
277 - Remove NextId for SQLite. Optimize.
278 - Move all SQL statements into a single location.
279 - Add UA rc and history files.
280 - put termcap (used by console) in ./configure and
281 allow -with-termcap-dir.
282 - Enhance time and size scanning routines.
283 - Fix Autoprune for Volumes to respect need for full save.
284 - Fix Win32 config file definition name on /install
285 - Compare tape to Client files (attributes, or attributes and data)
286 - Make all database Ids 64 bit.
287 - Write an applet for Linux.
288 - Implement new inter-daemon communications protocol.
289 - Allow console commands to detach or run in background.
290 - Fix status delay on storage daemon during rewind.
291 - Add SD message variables to control operator wait time
292 - Maximum Operator Wait
293 - Minimum Message Interval
294 - Maximum Message Interval
295 - Send Operator message when cannot read tape label.
296 - Verify level=Volume (scan only), level=Data (compare of data to file).
297 Verify level=Catalog, level=InitCatalog
299 - Add keyword search to show command in Console.
300 - Events : tape has more than xxx bytes.
301 - Complete code in Bacula Resources -- this will permit
302 reading a new config file at any time.
303 - Handle ctl-c in Console
304 - Implement script driven addition of File daemon to config files.
305 - Think about how to make Bacula work better with File (non-tape) archives.
306 - Write Unix emulator for Windows.
307 - Put memory utilization in Status output of each daemon
308 if full status requested or if some level of debug on.
309 - Make database type selectable by .conf files i.e. at runtime
310 - Set flag for uname -a. Add to Volume label.
311 - Implement throttled work queue.
312 - Restore files modified after date
313 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
314 - Implement Restore FileSet=
315 - Create a protocol.h and protocol.c where all protocol messages
317 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
318 - Timout a job or terminate if link goes down, or reopen link and query.
319 - Concept of precious tapes (cannot be reused).
320 - Make bcopy copy with a single tape drive.
321 - Permit changing ownership during restore.
324 > My suggestion: Add a feature on the systray menu-icon menu to request
325 > an immediate backup now. This would be useful for laptop users who may
326 > not be on the network when the regular scheduled backup is run.
328 > My wife's suggestion: Add a setting to the win32 client to allow it to
329 > shut down the machine after backup is complete (after, of course,
330 > displaying a "System will shut down in one minute, click here to cancel"
331 > warning dialog). This would be useful for sites that want user
332 > woorkstations to be shut down overnight to save power.
335 - From Terry Manderson <terry@apnic.net>
336 jobset { # new structure
337 name = "monthlyUnixBoxen"
340 jobs = "wakame;durian;soy;wasabi;miso" #new!
350 fileset = "genericUnixSet"
356 fileset = "genericUnixSet"
362 fileset = "UnixDevelBoxSet"
367 - Autolabel should be specified by DIR instead of SD.
370 - AutoScan (check checksum of tape)
371 - Format command = "format /dev/nst0"
375 - Seek resolution (usually corresponds to buffer size)
376 - EODErrorCode=ENOSPC or code
377 - Partial Read error code
378 - Partial write error code
379 - Nonformatted read error
380 - Nonformatted write error
381 - WriteProtected error
385 - IgnoreCloseErrors=yes
395 - FD sends unsaved file list to Director at end of job (see
397 - File daemon should build list of files skipped, and then
398 at end of save retry and report any errors.
399 - Write a Storage daemon that uses pipes and
400 standard Unix programs to write to the tape.
402 - Need something that monitors the JCR queue and
403 times out jobs by asking the deamons where they are.
404 - Enhance Jmsg code to permit buffering and saving to disk.
405 - device driver = "xxxx" for drives.
407 - Ensure that /dev/null works
408 - Need report class for messages. Perhaps
409 report resource where report=group of messages
410 - enhance scan_attrib and rename scan_jobtype, and
411 fill in code for "since" option
412 - Director needs a time after which the report status is sent
413 anyway -- or better yet, a retry time for the job.
414 - Don't reschedule a job if previous incarnation is still running.
415 - Some way to automatically backup everything is needed????
416 - Need a structure for pending actions:
418 - termination status (part of buffered msgs?)
420 Read, Write, Clean, Delete
421 - Login to Bacula; Bacula users with different permissions:
422 owner, group, user, quotas
423 - Store info on each file system type (probably in the job header on tape.
424 This could be the output of df; or perhaps some sort of /etc/mtab record.
427 - Design at hierarchial storage for Bacula. Migration and Clone.
428 - Implement FSM (File System Modules).
429 - Audit M_ error codes to ensure they are correct and consistent.
430 - Add variable break characters to lex analyzer.
431 Either a bit mask or a string of chars so that
432 the caller can change the break characters.
433 - Make a single T_BREAK to replace T_COMMA, etc.
434 - Ensure that File daemon and Storage daemon can
435 continue a save if the Director goes down (this
436 is NOT currently the case). Must detect socket error,
437 buffer messages for later.
438 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
439 - Add ability to backup to two Storage devices (two SD sessions) at
440 the same time -- e.g. onsite, offsite.
441 - Add the ability to consolidate old backup sets (basically do a restore
442 to tape and appropriately update the catalog). Compress Volume sets.
443 Might need to spool via file is only one drive is available.
444 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
445 someway to do so with every volume that has less than x% valid
449 Migration: Move a backup from one Volume to another
450 Clone: Copy a backup -- two Volumes
452 Bacula Migration is based on Jobs (apparently Networker is file by file).
454 Migration triggered by:
458 Highwater mark (keep total size)
462 Bacula Projects Roadmap
464 last update 8 May 2003
466 Item 1: Multiple simultaneous Jobs. (done)
467 Done -- Restore part needs better implementation to work correctly
468 Also, it needs considerable testing
470 What: Permit multiple simultaneous jobs in Bacula.
472 Why: An enterprise level solution needs to go fast without the
473 need for the system administrator to carefully tweak
474 timing. Based on the benchmarks, during a full
475 backup, NetWorker typically hit 10 times the bandwidth to
476 the tape compared to Bacula--largely. This is probably due to
477 running parallel jobs and multi-threaded filling of buffers
478 and writing them to tape. This should also make things work
479 better when you have a mix of fast and slow machines backing
482 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
483 implementing this is a matter of some small cleanups and
487 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
488 Deferred -- not necessary yet -- possibly implement with Migration.
490 What: If data is coming into the SD too fast, buffer it to
491 disk if the user has configured this option.
493 Why: This would be nice, especially if it more or less falls out
494 when implementing (1) above. If not, it probably should not
495 be given a high priority because fundamentally the backup time
496 is limited by the tape bandwidth. Even though you may finish a
497 client job quicker by spilling to disk, you still have to
498 eventually get it onto tape. If intermediate disk buffering
499 allows us to improve write bandwidth to tape, it may make
502 Notes: Whether or not this is implemented will depend upon performance
503 testing after item 1 is implemented.
506 Item 3: Write the bscan program -- also write a bcopy program.
509 What: Write a program that reads a Bacula tape and puts all the
510 appropriate data into the catalog. This allows recovery
511 from a tape that is no longer in the database, or it allows
512 re-creation of a database if lost.
514 Why: This is a fundamental robustness and disaster recovery tool
515 which will increase the comfort level of a sysadmin
516 considering adopting Bacula.
518 Notes: A skeleton of this program already exists, but much work
519 needs to be done. Implementing this will also make apparent
520 any deficiencies in the current Bacula tape format.
523 Item 4: Implement Base jobs.
525 What: A base job is sort of like a Full save except that you
526 will want the FileSet to contain only files that are unlikely
527 to change in the future (i.e. a snapshot of most of your
528 system after installing it). After the base job has been run,
529 when you are doing a Full save, you can specify to exclude
530 all files saved by the base job that have not been modified.
532 Why: This is something none of the competition does, as far as we know
533 (except BackupPC, which is a Perl program that saves to disk
534 only). It is big win for the user, it makes Bacula stand out
535 as offering a unique optimization that immediately saves time
538 Notes: Big savings in tape usage. Will require more resources because
539 the e. DIR must send FD a list of files/attribs, and the FD must
540 search the list and compare it for each file to be saved.
543 Item 5: Implement Label templates
546 What: This is a mechanism whereby Bacula can automatically create
547 a tape label for new tapes according to a detailed specification
548 provided by the user.
550 Why: It is a major convenience item for folks who use automated label
553 Notes: Bacula already has a working form of automatic tape label
554 creation, but it is very crude. The design for the complete
555 tape labeling project is already documented in the manual.
558 Item 6: Write a regression script.
559 Done -- Continue to expand its testing.
561 What: This is an automatic script that runs and tests as many features
562 of Bacula as possible. The output is compared to previous
563 versions of Bacula and any differences are reported.
565 Why: This is an enormous help in preventing introduction of new
566 errors in parts of the program that already work correctly.
568 Notes: This probably should be ranked higher, it's something the typical
569 user doesn't see. Depending on how it's implemented, it may
570 make sense to defer it until the archival tape format and
571 user interface mature.
574 Item 7: GUI for interactive restore
575 Item 8: GUI for interactive backup
577 What: The current interactive restore is implemented with a tty
578 interface. It would be much nicer to be able to "see" the
579 list of files backed up in typical GUI tree format.
580 The same mechanism could also be used for creating
581 ad-hoc backup FileSets (item 8).
583 Why: Ease of use -- especially for the end user.
585 Notes: Rather than implementing in Gtk, we probably should go directly
586 for a Browser implementation, even if doing so meant the
587 capability wouldn't be available until much later. Not only
588 is there the question of Windows sites, most
589 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
590 without installing lots of stuff admins are very wary about.
591 Real sysadmins will always use the command line anyway, and
592 the user who's doing an interactive restore or backup of his
593 own files will in most cases be on a Windows machine running
597 Item 9: Add SSL to daemon communications.
599 What: This provides for secure communications between the daemons.
601 Why: This would allow doing backup across the Internet without
602 privacy concerns (or with much less concern).
604 Notes: The vast majority of near term potential users will be backing up
605 a single site over a LAN and, correctly or not, they probably
606 won't be concerned with security, at least not enough to go to
607 the trouble to set up keys, etc. to screw things down. We suspect
608 that many users genuinely interested in multi-site backup
609 already run some form of VPN software in their internetwork
610 connections, and are willing to delegate security to that layer.
613 Item 10: Define definitive tape format.
616 What: Define that definitive tape format that will not change
617 for the next millennium.
619 Why: Stability, security.
621 Notes: See notes for item 11 below.
624 Item 11: New daemon communication protocol.
626 What: The current daemon to daemon protocol is basically an ASCII
627 printf() and sending the buffer. On the receiving end, the
628 buffer is sscanf()ed to unpack it. The new scheme would
629 be a binary format that allows quick packing and unpacking
630 of any data type with named fields.
632 Why: Using binary packing would be faster. Named fields will permit
633 error checking to ensure that what is sent is what the
634 receiver really wants.
636 Notes: These are internal improvements in the interest of the
637 long-term stability and evolution of the program. On the one
638 hand, the sooner they're done, the less code we have to rip
639 up when the time comes to install them. On the other hand, they
640 don't bring an immediately perceptible benefit to potential
641 users. Item 10 and possibly item 11 should be deferred until Bacula
642 is well established with a growing user community more or
643 less happy with the feature set. At that time, it will make a
644 good "next generation" upgrade in the interest of data
650 ======================================================
652 It is somewhat like a Full save becomes an incremental since
653 the Base job (or jobs) plus other non-base files.
655 - New BaseFile table that contains:
656 JobId, BaseJobId, FileId (from Base).
657 i.e. for each base file that exists but is not saved because
658 it has not changed, the File daemon sends the JobId, BaseId,
659 and FileId back to the Director who creates the DB entry.
660 - To initiate a Base save, the Director sends the FD
661 the FileId, and full filename for each file in the Base.
662 - When the FD finds a Base file, he requests the Director to
663 send him the full File entry (stat packet plus MD5), or
664 conversely, the FD sends it to the Director and the Director
665 says yes or no. This can be quite rapid if the FileId is kept
666 by the FD for each Base Filename.
667 - It is probably better to have the comparison done by the FD
668 despite the fact that the File entry must be sent across the
670 - An alternative would be to send the FD the whole File entry
671 from the start. The disadvantage is that it requires a lot of
672 space. The advantage is that it requires less communications
674 - The Job record must be updated to indicate that one or more
676 - At end of Job, FD returns:
677 1. Count of base files/bytes not written to tape (i.e. matches)
678 2. Count of base file that were saved i.e. had changed.
679 - No tape record would be written for a Base file that matches, in the
680 same way that no tape record is written for Incremental jobs where
681 the file is not saved because it is unchanged.
682 - On a restore, all the Base file records must explicitly be
683 found from the BaseFile tape. I.e. for each Full save that is marked
684 to have one or more Base Jobs, search the BaseFile for all occurrences
686 - An optimization might be to make the BaseFile have:
692 This would avoid the need to explicitly fetch each File record for
693 the Base job. The Base Job record will be fetched to get the
694 VolSessionId and VolSessionTime.
695 =========================================================
699 =============================================================
701 Request For Comments For File Backup Options
704 Subject: File Backup Options
707 A few days ago, a Bacula user who is backing up to file volumes and
708 using compression asked if it was possible to suppress compressing
709 all .gz files since it was a waste of CPU time. Although Bacula
710 currently permits using different options (compression, ...) on
711 a directory by directory basis, it cannot do it on a file by
712 file basis, which is clearly what was desired.
714 Proposed Implementation:
715 To solve this problem, I propose the following:
717 - Add a new Director resource type called Options.
719 - The Options resource will have records for all
720 options that can currently be specified on the Include record
721 (in a FileSet). Examples below.
723 - The Options resource will permit an exclude option as well
724 as a number of additional options.
726 - The heart of the Options resource is the ability to
727 supply any number of Match records which specify POSIX
728 regular expressions. These Match regular expressions are
729 applied to the fully qualified filename (path and all). If
730 one matches, then the Options will be used.
732 - When an Match specification matches an included file, the
733 options specified in the Options resource will override
734 the default options specified on the Include record.
736 - Include records will be modified to permit referencing one or
737 more Options resources. The Options will be used
738 in the order listed on the Include record and the first
739 one that matches will be applied.
741 - Options (or specifications) currently supplied on the Include
742 record will be deprecated (i.e. removed in a later version a
743 year or so from now).
745 - The Exclude record will be deprecated as the same functionality
746 can be obtained by using an Exclude = yes in the Options.
749 The following records can appear in the Options resource. An
750 asterisk preceding the name indicates a feature not currently
754 - Compression= (GZIP, ...)
755 - Signature= (MD5, SHA1, ...)
757 - OneFs= (yes/no) - remain on one filesystem
758 - Recurse= (yes/no) - recurse into subdirectories
759 - Sparse= (yes/no) - do sparse file backup
760 - *Exclude= (yes/no) - exclude file from being saved
761 - *Reader= (filename) - external read (backup) program
762 - *Plugin= (filename) - read/write plugin module
765 - verify= (ipnougsamc5) - verify options
768 - replace= (always/ifnewer/ifolder/never) - replace options currently
770 - *Writer= (filename) - external write (restore) program
774 Currently options specifying compression, MD5 signatures, recursion,
775 ... of a FileSet are supplied on the Include record. These will now
776 all be collected into a Options resource, which will be
777 specified in the Include in place of the options. Multiple Options
778 may be specified. Since the Options may contain regular expressions
779 that are applied to the full filename, this will give the ability
780 to specify backup options on a file by file basis to whatever level
789 Include = compression=GZIP signature=MD5 {
801 Match = /*.?*/ # matches all files.
806 That's a lot more to do the same thing, but it gives the ability to
807 apply options on a file by file basis. For example, suppose you
808 want to compress all files but not any file with extensions .gz or .Z.
809 In that case, you will need to group two sets of options using
810 the Options resource as follows:
818 # Note multiple Matches are ORed
819 Match = /*.gz/ # matches .gz files */
820 Match = /*.Z/ # matches .Z files */
825 Match = /*.?*/ # matches all files
831 Now, since the no Compression option is specified in the
832 first group of Options, *.gz or *.Z file will have an MD5 signature computed,
833 but will not be compressed. For all other files, the *.gz *.Z will not
834 match, so the second group of options will be used which will include GZIP
838 - Is it necessary to provide some means of ANDing regular expressions
839 and negation? (not currently planned)
841 e.g. Match = /*.gz/ && !/big.gz/
843 - I see that Networker has a "null" module which, if specified, does not
844 backup the file, but does make an record of the file in the catalog
845 so that the catalog will reflect an exact picture of the filesystem.
846 The result is that the file can be "seen" when "browsing" the save
847 sets, but it cannot be restored.
849 Is this really useful? Should it be implemented in Bacula?
852 After implementing the above, the user will be able to specify
853 on a file by file basis (using regular expressions) what options are
854 applied for the backup.
857 =============================================
859 ==========================================================
861 For each Incremental job that is run, there may be files that
862 were found but not saved because they were locked (this applies
863 only to Windows). Such a system could send back to the Director
864 a list of Unsaved files.
866 - New UnSavedFiles table that contains:
870 - Then in the next Incremental job, the list of Unsaved Files will be
871 feed to the FD, who will ensure that they are explicitly chosen even
872 if standard date/time check would not have selected them.
873 =============================================================
876 Done: (see kernsdone for more)
877 - Implement new alist in FileSet scanning.
878 - bls should continue reading even if it finds Win32 data on the tape.
879 The error should be Warning rather the Error.
880 - Add user configurable timeout for connecting to SD.
881 - Unsaved Flag in Job record (use JobMissingFiles).
882 - Base Flag in Job record.
883 - Configure mtx-changer to have correct path to mtx.
884 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
885 - Make some way so that if a machine is skipped because it is not up
886 that Bacula will continue retrying for a specified period of time --
888 - Implement all command line args on run.
889 - Implement command line "restore" args.
890 - Implement "restore current select=no"
891 - Restore file modified before date
892 - Restore -- do nothing but show what would happen
893 - Add estimate to Console commands
894 - Use read_record.c in SD code.
895 - Fix read_record to handle multiple sessions.
896 - Tip from Steve Allam
897 mt -f /dev/nst0 defblksize 0
898 - Document "status" in the console.
899 - Document driving console from shell script.
900 - Write JobMedia records with max file size is reached on tape.
901 - Handle the case of multiple JobMedia records pending (i.e. the
902 thread is slow and multiple situations requiring a JobMedia
904 - Do performance analysis on the restore tree routines.
905 - Fix maximum file size (block.c) to generate JobMedia records.
906 - Make the default file size 1GB on the tape.
907 - Implement forward spacing between files.
908 - Add Machine type (Linux/Windows) to Status report for daemons.
910 - Use repositioning at the beginning of the tape.
911 - Do full check the command line args in update (e.g. VolStatus ...).
912 - Specify list of files to restore
913 - Implement ClientRunBeforeJob and ClientRunAfterJob.
914 - Make | and < work on FD side.
915 - Check to see if "blocked" is set during restore.
916 - Figure out what is interrupting sql command in console.
917 - Make new job print warning User Unmounted Tape.
918 - Test recycling and purging (code changed in db_find_next_volume and
920 - Document SDConnectTimeout (in FD).
921 - Add restore by filename test.
922 - Document restore by files.
923 - Make variable expansion work correctly.
924 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
925 - Copy static programs into install directory.
926 - Think about changing Storage resource Device record to be
928 - Add RunBeforeJob and RunAfterJob to the Client program.
929 - Need return status on read_cb() from read_records(). Need multiple
930 records -- one per Job, maybe a JCR or some other structure with
931 a block and a record.
932 - LabelFormat on tape volume apparently creates the db record but
933 never actually labels the volume.
934 - Recycling a volume when two jobs are using it is going to break. Fixed.
935 - Document list nextvol and new format status dir.
936 - Client files in Win32 with Unix eol conventions doesn't work.
937 - Either fix or document that fill command in btape can be
938 compressed enormously by the hardware - a 36GB tape wrote 750GB!
939 - Add multiple character duration qualifiers.
940 - Require some modifer.
941 - Restrict characters permitted in a Resource name, and don't permit
943 - Figure out some way to ignore or get past checksum errors in
945 - The SD spooling file gets created even if it is not used.
946 - Look at Cleaning tape in ua_label.c for media create/update
947 - Add regression testing to the manual
948 - End time: in job output of rescheduled job is time of first run.
949 - Document list nextvol and status output.
950 - Separate Dir heartbeat in FD from the SD heartbeat.
951 - Fix sparse file handeling so that it always reads a multiple
952 of 512. Currently, it subtracts 8 bytes (for faddr).
953 Kludged with #ifdef for FreeBSD.
954 - Document that Volume pruning can delete last Full backup and
955 hence you will not have a valid backup.
956 - Clarify the fact that having the Bacula cygwin1.dll loaded
957 is not the same as having cygwin installed.
958 - Document that it is safe to use the drive when the lights stop flashing.
959 - Document all the status codes JobLevel, JobType, JobStatus.
960 - Add GUI interface to manual