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 that it is safe to use the drive when the lights stop flashing.
30 - Document new Include/Exclude ...
31 - Document all the status codes JobLevel, JobType, JobStatus.
32 - Add test of exclusion, test multiple Include {} statements.
33 - Add counter variable test.
36 - Add GUI interface to manual
37 - Document that Volume pruning can delete last Full backup and
38 hence you will not have a valid backup.
39 - Clarify the fact that having the Bacula cygwin1.dll loaded
40 is not the same as having cygwin installed.
41 - Fix sparse file handeling so that it always reads a multiple
42 of 512. Currently, it subtracts 8 bytes (for faddr).
43 - Separate Dir heartbeat in FD from the SD heartbeat.
46 - When job rescheduled, status gives is waiting for Client Rufus
47 to connect to Storage File. Dir needs to inform SD that job
49 - Fix get_storage_from_media_type (ua_restore) to use command line
51 - Enhance "update slots" to include a "scan" feature
52 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
53 - Allow a slot or range of slots on the label barcodes command.
54 - Don't print "Warning: Wrong Volume mounted ..." if mounting second volume.
55 - Make Dmsg look at global before calling subroutine.
56 - Enable trace output at runtime for Win32
57 - Make sure that Volumes are recycled based on "Least recently used"
58 rather than lowest MediaId.
59 - Available volumes for autochangers (see patrick@baanboard.com 3 Sep 03
60 and 4 Sep) scan slots.
61 - Upgrade to cygwin 1.5
63 - Get and test MySQL 4.0
64 - Do a complete audit of all pthreads_mutex, cond, ... to ensure that
65 any that are dynamically initialized are destroyed when no longer used.
66 - Write a mini-readline with history and editing.
67 - Look at how fuser works and /proc/PID/fd that is how Nic found the
68 file descriptor leak in Bacula.
69 - Implement WrapCounters in Counters.
70 - Turn on SIGHUP in dird.c and test.
71 - Use system dependent calls to get more precise info on tape errors.
72 - Add heartbeat from FD to SD if hb interval expires.
73 - Suppress read error on blank tape when doing a label.
74 - Can we dynamically change FileSets.
75 - If pool specified to label command and Label Format is specified,
76 automatically generate the Volume name.
77 - Take a careful look a the Basic recycling algorithm. When Bacula
78 chooses, the order should be:
80 - Look for Recycle or Purged
83 Instead of using lowest media Id, find the least recently used
86 When the tape is mounted and Bacula requests the status
87 - Do everything possible to use it.
89 Define a "available" status, which is the currently mounted
90 Volume and all volumes that are currently in the autochanger.
92 - Why can't SQL do the filename sort for restore?
93 - Is a pool specification really needed for a restore? Yes, and
94 you may want to exclude archive Pools.
95 - Look at libkse (man kse) for FreeBSD threading.
96 - Look into Microsoft Volume Shadowcopy Service VSS for backing
97 up system state components (Active Directory, System Volume, ...)
98 - Add ExhautiveRestoreSearch
99 - Look at the possibility of loading only the necessary
100 data into the restore tree (i.e. do it one directory at a
101 time as the user walks through the tree).
102 - Possibly use the hash code if the user selects all for a restore command.
103 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
104 - Implement some way for the File daemon to contact the Director
105 to start a job or pass its DHCP obtained IP number.
106 - Implement multiple Consoles.
107 - Add Console usr permissions.
108 - Fix "restore all" to bypass building the tree.
109 - Fix restore to list errors if Invalid block found, and if # files
110 restored does not match # expected.
111 - Prohibit backing up archive device (findlib/find_one.c:128)
112 - Implement Release Device in the Job resource to unmount a drive.
113 - Implement Acquire Device in the Job resource to mount a drive,
114 be sure this works with admin jobs so that the user can get
115 prompted to insert the correct tape. Possibly some way to say to
116 run the job but don't save the files.
117 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
118 - Implement FileOptions (see end of this document)
119 - Make things like list where a file is saved case independent for
121 - Edit the Client/Storage name into authentication failure messages.
122 - Implement job in VerifyToCatalog
124 - Implement a PostgreSQL driver.
125 - Bacula needs to propagate SD errors.
126 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
127 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
128 /dev/nst0. ERR=Input/output error.
129 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
130 Last block=5162 Current block=5164.
131 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
133 - Use autochanger to handle multiple devices.
134 - Fix packet too big problem.
135 - Add SuSE install doc to list.
136 - Check and rechedk "Invalid block number"
137 - Make bextract release the drive properly between tapes
138 so that an autochanger can be made to work.
139 - User wants to NOT backup up certain big files (email files).
140 - Maybe remove multiple simultaneous devices code in SD.
141 - On Windows with very long path names, it may be impossible to create
142 a file (and thus restore it) because the total length is too long.
143 We must cd into the directory then create the file without the
145 - lstat() is not going to work on Win32 for testing date.
146 - Something is not right in last block of fill command.
147 - Implement a Recycle command
148 - Add FileSet to command line arguments for restore.
149 - Allow multiple Storage specifications (or multiple names on
150 a single Storage specification) in the Job record. Thus a job
151 can be backed up to a number of storage devices.
152 - Add client name to cram-md5 challenge so Director can immediately
153 verify if it is the correct client.
154 - Implement ClientRunBeforeJob and ClientRunAfterJob.
155 - Add JobLevel in FD status (but make sure it is defined).
156 - Audit all UA commands to ensure that we always prompt where possible.
157 - Restrict characters permitted in a Resource name, and don't permit
159 - Check Jmsg in bnet, may not work, must dup bsock.
160 - Suppress Job Name in Jmsg for console
161 - Create Pools that are referenced in a Run statement at startup if possible.
162 - Use runbeforejob to unload, then reload a volume previously used,
163 then the next job run gets an error reading the drive.
164 - Make bootstrap filename unique.
165 - Test a second language e.g. french.
166 - Start working on Base jobs.
167 - Make "make binary-release" work from any directory.
168 - Implement UnsavedFiles DB record.
169 - Implement argc/argv for daemon command line scanning using table driven
171 - Implement table driven single argc/argv scanner to pickup all arguments.
172 Much like xxx_conf.c scan table.
173 keyword, handler(store_routine), store_address, code, flags, default.
174 - From Phil Stracchino:
175 It would probably be a per-client option, and would be called
176 something like, say, "Automatically purge obsoleted jobs". What it
177 would do is, when you successfully complete a Differential backup of a
178 client, it would automatically purge all Incremental backups for that
179 client that are rendered redundant by that Differential. Likewise,
180 when a Full backup on a client completed, it would automatically purge
181 all Differential and Incremental jobs obsoleted by that Full backup.
182 This would let people minimize the number of tapes they're keeping on
183 hand without having to master the art of retention times.
184 - Implement new serialize subroutines
185 send(socket, "string", &Vol, "uint32", &i, NULL)
186 - Scratch Pool where the volumes can be re-assigned to any Pool.
187 - Implement a M_SECURITY message class.
188 - Implement forward spacing block/file: position_device(bsr) --
189 just before read_block_from_device();
190 - When doing a Backup send all attributes back to the Director, who
191 would then figure out what files have been deleted.
192 - Currently in mount.c:236 the SD simply creates a Volume. It should have
193 explicit permission to do so. It should also mark the tape in error
194 if there is an error.
195 - Make sure all restore counters are working correctly in the FD.
196 - SD Bytes Read is wrong.
197 - Look at ALL higher level routines that call block.c to be sure
198 they don't expect something in errmsg.
199 - Investigate doing RAW backup of Win32 partition.
200 - Add thread specific data to hold the jcr -- send error messages from
201 low level routines by accessing it and using Jmsg().
202 - Cancel waiting for Client connect in SD if FD goes away.
204 - Add Progress command that periodically reports the progress of
206 - One block was orphaned in the SD probably after cancel.
209 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
210 - Implement timeout in response() when it should come quickly.
211 - Implement console @echo command.
212 - Implement a Slot priority (loaded/not loaded).
213 - Implement "vacation" Incremental only saves.
214 - Implement single pane restore (much like the Gftp panes).
215 - Implement Automatic Mount even in operator wait.
216 - Implement create "FileSet"?
217 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
218 - Implement "scratch pool" where tapes are defined and can be
219 taken by any pool that needs them.
220 - Implement restore "current system", but take all files without
221 doing selection tree -- so that jobs without File records can
223 - Implement disk spooling. Two parts: 1. Spool to disk then
224 immediately to tape to speed up tape operations. 2. Spool to
225 disk only when the tape is full, then when a tape is hung move
227 - Implement a relocatable bacula.spec
228 - Implement dump/print label to UA
229 - Add prefixlinks to where or not where absolute links to FD.
230 - Issue message to mount a new tape before the rewind.
231 - Simplified client job initiation for portables.
232 - If SD cannot open a drive, make it periodically retry.
233 - Implement LabelTemplate (at least first cut).
234 - Add more of the config info to the tape label.
235 - Implement a Mount Command and an Unmount Command where
236 the users could specify a system command to be performed
237 to do the mount, after which Bacula could attempt to
238 read the device. This is for Removeable media such as a CDROM.
239 - Most likely, this mount command would be invoked explicitly
240 by the user using the current Console "mount" and "unmount"
241 commands -- the Storage Daemon would do the right thing
242 depending on the exact nature of the device.
243 - As with tape drives, when Bacula wanted a new removable
244 disk mounted, it would unmount the old one, and send a message
245 to the user, who would then use "mount" as described above
246 once he had actually inserted the disk.
248 - If tape is marked read-only, then try opening it read-only rather than
249 failing, and remember that it cannot be written.
250 - Refine SD waiting output:
251 Device is being positioned
252 > Device is being positioned for append
253 > Device is being positioned to file x
255 - Figure out some way to estimate output size and to avoid splitting
256 a backup across two Volumes -- this could be useful for writing CDROMs
257 where you really prefer not to have it split -- not serious.
258 - Have SD compute MD5 or SHA1 and compare to what FD computes.
259 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
260 actual data on the Volume and compare it.
261 - Implement Bacula plugins -- design API
262 - Make bcopy read through bad tape records.
263 - Program files (i.e. execute a program to read/write files).
264 Pass read date of last backup, size of file last time.
265 - Add Signature type to File DB record.
266 - CD into subdirectory when open()ing files for backup to
267 speed up things. Test with testfind().
268 - Priority job to go to top of list.
269 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
270 - Implement some way for the Console to dynamically create a job.
271 - Restore to a particular time -- e.g. before date, after date.
272 - Solaris -I on tar for include list
273 - Need a verbose mode in restore, perhaps to bsr.
274 - bscan without -v is too quiet -- perhaps show jobs.
275 - Add code to reject whole blocks if not wanted on restore.
276 - Check if we can increase Bacula FD priorty in Win2000
277 - Make sure the MaxVolFiles is fully implemented in SD
278 - Check if both CatalogFiles and UseCatalog are set to SD.
279 - Figure out how to do a bare metal Windows restore
280 - Possibly add email to Watchdog if drive is unmounted too
281 long and a job is waiting on the drive.
282 - Restore program that errs in SD due to no tape, reports
283 OK incorrectly in output.
284 - After unmount, if restore job started, ask to mount.
285 - Convert all %x substitution variables, which are hard to remember
286 and read to %(variable-name). Idea from TMDA.
287 - Remove NextId for SQLite. Optimize.
288 - Move all SQL statements into a single location.
289 - Add UA rc and history files.
290 - put termcap (used by console) in ./configure and
291 allow -with-termcap-dir.
292 - Enhance time and size scanning routines.
293 - Fix Autoprune for Volumes to respect need for full save.
294 - Fix Win32 config file definition name on /install
295 - Compare tape to Client files (attributes, or attributes and data)
296 - Make all database Ids 64 bit.
297 - Write an applet for Linux.
298 - Implement new inter-daemon communications protocol.
299 - Allow console commands to detach or run in background.
300 - Fix status delay on storage daemon during rewind.
301 - Add SD message variables to control operator wait time
302 - Maximum Operator Wait
303 - Minimum Message Interval
304 - Maximum Message Interval
305 - Send Operator message when cannot read tape label.
306 - Verify level=Volume (scan only), level=Data (compare of data to file).
307 Verify level=Catalog, level=InitCatalog
309 - Add keyword search to show command in Console.
310 - Events : tape has more than xxx bytes.
311 - Complete code in Bacula Resources -- this will permit
312 reading a new config file at any time.
313 - Handle ctl-c in Console
314 - Implement script driven addition of File daemon to config files.
315 - Think about how to make Bacula work better with File (non-tape) archives.
316 - Write Unix emulator for Windows.
317 - Put memory utilization in Status output of each daemon
318 if full status requested or if some level of debug on.
319 - Make database type selectable by .conf files i.e. at runtime
320 - Set flag for uname -a. Add to Volume label.
321 - Implement throttled work queue.
322 - Restore files modified after date
323 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
324 - Implement Restore FileSet=
325 - Create a protocol.h and protocol.c where all protocol messages
327 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
328 - Timout a job or terminate if link goes down, or reopen link and query.
329 - Concept of precious tapes (cannot be reused).
330 - Make bcopy copy with a single tape drive.
331 - Permit changing ownership during restore.
334 > My suggestion: Add a feature on the systray menu-icon menu to request
335 > an immediate backup now. This would be useful for laptop users who may
336 > not be on the network when the regular scheduled backup is run.
338 > My wife's suggestion: Add a setting to the win32 client to allow it to
339 > shut down the machine after backup is complete (after, of course,
340 > displaying a "System will shut down in one minute, click here to cancel"
341 > warning dialog). This would be useful for sites that want user
342 > woorkstations to be shut down overnight to save power.
345 - From Terry Manderson <terry@apnic.net>
346 jobset { # new structure
347 name = "monthlyUnixBoxen"
350 jobs = "wakame;durian;soy;wasabi;miso" #new!
360 fileset = "genericUnixSet"
366 fileset = "genericUnixSet"
372 fileset = "UnixDevelBoxSet"
377 - Autolabel should be specified by DIR instead of SD.
380 - AutoScan (check checksum of tape)
381 - Format command = "format /dev/nst0"
385 - Seek resolution (usually corresponds to buffer size)
386 - EODErrorCode=ENOSPC or code
387 - Partial Read error code
388 - Partial write error code
389 - Nonformatted read error
390 - Nonformatted write error
391 - WriteProtected error
395 - IgnoreCloseErrors=yes
405 - FD sends unsaved file list to Director at end of job (see
407 - File daemon should build list of files skipped, and then
408 at end of save retry and report any errors.
409 - Write a Storage daemon that uses pipes and
410 standard Unix programs to write to the tape.
412 - Need something that monitors the JCR queue and
413 times out jobs by asking the deamons where they are.
414 - Enhance Jmsg code to permit buffering and saving to disk.
415 - device driver = "xxxx" for drives.
417 - Ensure that /dev/null works
418 - Need report class for messages. Perhaps
419 report resource where report=group of messages
420 - enhance scan_attrib and rename scan_jobtype, and
421 fill in code for "since" option
422 - Director needs a time after which the report status is sent
423 anyway -- or better yet, a retry time for the job.
424 - Don't reschedule a job if previous incarnation is still running.
425 - Some way to automatically backup everything is needed????
426 - Need a structure for pending actions:
428 - termination status (part of buffered msgs?)
430 Read, Write, Clean, Delete
431 - Login to Bacula; Bacula users with different permissions:
432 owner, group, user, quotas
433 - Store info on each file system type (probably in the job header on tape.
434 This could be the output of df; or perhaps some sort of /etc/mtab record.
437 - Design at hierarchial storage for Bacula. Migration and Clone.
438 - Implement FSM (File System Modules).
439 - Audit M_ error codes to ensure they are correct and consistent.
440 - Add variable break characters to lex analyzer.
441 Either a bit mask or a string of chars so that
442 the caller can change the break characters.
443 - Make a single T_BREAK to replace T_COMMA, etc.
444 - Ensure that File daemon and Storage daemon can
445 continue a save if the Director goes down (this
446 is NOT currently the case). Must detect socket error,
447 buffer messages for later.
448 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
449 - Add ability to backup to two Storage devices (two SD sessions) at
450 the same time -- e.g. onsite, offsite.
451 - Add the ability to consolidate old backup sets (basically do a restore
452 to tape and appropriately update the catalog). Compress Volume sets.
453 Might need to spool via file is only one drive is available.
454 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
455 someway to do so with every volume that has less than x% valid
459 Migration: Move a backup from one Volume to another
460 Clone: Copy a backup -- two Volumes
462 Bacula Migration is based on Jobs (apparently Networker is file by file).
464 Migration triggered by:
468 Highwater mark (keep total size)
472 Bacula Projects Roadmap
474 last update 8 May 2003
476 Item 1: Multiple simultaneous Jobs. (done)
477 Done -- Restore part needs better implementation to work correctly
478 Also, it needs considerable testing
480 What: Permit multiple simultaneous jobs in Bacula.
482 Why: An enterprise level solution needs to go fast without the
483 need for the system administrator to carefully tweak
484 timing. Based on the benchmarks, during a full
485 backup, NetWorker typically hit 10 times the bandwidth to
486 the tape compared to Bacula--largely. This is probably due to
487 running parallel jobs and multi-threaded filling of buffers
488 and writing them to tape. This should also make things work
489 better when you have a mix of fast and slow machines backing
492 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
493 implementing this is a matter of some small cleanups and
497 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
498 Deferred -- not necessary yet -- possibly implement with Migration.
500 What: If data is coming into the SD too fast, buffer it to
501 disk if the user has configured this option.
503 Why: This would be nice, especially if it more or less falls out
504 when implementing (1) above. If not, it probably should not
505 be given a high priority because fundamentally the backup time
506 is limited by the tape bandwidth. Even though you may finish a
507 client job quicker by spilling to disk, you still have to
508 eventually get it onto tape. If intermediate disk buffering
509 allows us to improve write bandwidth to tape, it may make
512 Notes: Whether or not this is implemented will depend upon performance
513 testing after item 1 is implemented.
516 Item 3: Write the bscan program -- also write a bcopy program.
519 What: Write a program that reads a Bacula tape and puts all the
520 appropriate data into the catalog. This allows recovery
521 from a tape that is no longer in the database, or it allows
522 re-creation of a database if lost.
524 Why: This is a fundamental robustness and disaster recovery tool
525 which will increase the comfort level of a sysadmin
526 considering adopting Bacula.
528 Notes: A skeleton of this program already exists, but much work
529 needs to be done. Implementing this will also make apparent
530 any deficiencies in the current Bacula tape format.
533 Item 4: Implement Base jobs.
535 What: A base job is sort of like a Full save except that you
536 will want the FileSet to contain only files that are unlikely
537 to change in the future (i.e. a snapshot of most of your
538 system after installing it). After the base job has been run,
539 when you are doing a Full save, you can specify to exclude
540 all files saved by the base job that have not been modified.
542 Why: This is something none of the competition does, as far as we know
543 (except BackupPC, which is a Perl program that saves to disk
544 only). It is big win for the user, it makes Bacula stand out
545 as offering a unique optimization that immediately saves time
548 Notes: Big savings in tape usage. Will require more resources because
549 the e. DIR must send FD a list of files/attribs, and the FD must
550 search the list and compare it for each file to be saved.
553 Item 5: Implement Label templates
556 What: This is a mechanism whereby Bacula can automatically create
557 a tape label for new tapes according to a detailed specification
558 provided by the user.
560 Why: It is a major convenience item for folks who use automated label
563 Notes: Bacula already has a working form of automatic tape label
564 creation, but it is very crude. The design for the complete
565 tape labeling project is already documented in the manual.
568 Item 6: Write a regression script.
569 Done -- Continue to expand its testing.
571 What: This is an automatic script that runs and tests as many features
572 of Bacula as possible. The output is compared to previous
573 versions of Bacula and any differences are reported.
575 Why: This is an enormous help in preventing introduction of new
576 errors in parts of the program that already work correctly.
578 Notes: This probably should be ranked higher, it's something the typical
579 user doesn't see. Depending on how it's implemented, it may
580 make sense to defer it until the archival tape format and
581 user interface mature.
584 Item 7: GUI for interactive restore
585 Item 8: GUI for interactive backup
587 What: The current interactive restore is implemented with a tty
588 interface. It would be much nicer to be able to "see" the
589 list of files backed up in typical GUI tree format.
590 The same mechanism could also be used for creating
591 ad-hoc backup FileSets (item 8).
593 Why: Ease of use -- especially for the end user.
595 Notes: Rather than implementing in Gtk, we probably should go directly
596 for a Browser implementation, even if doing so meant the
597 capability wouldn't be available until much later. Not only
598 is there the question of Windows sites, most
599 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
600 without installing lots of stuff admins are very wary about.
601 Real sysadmins will always use the command line anyway, and
602 the user who's doing an interactive restore or backup of his
603 own files will in most cases be on a Windows machine running
607 Item 9: Add SSL to daemon communications.
609 What: This provides for secure communications between the daemons.
611 Why: This would allow doing backup across the Internet without
612 privacy concerns (or with much less concern).
614 Notes: The vast majority of near term potential users will be backing up
615 a single site over a LAN and, correctly or not, they probably
616 won't be concerned with security, at least not enough to go to
617 the trouble to set up keys, etc. to screw things down. We suspect
618 that many users genuinely interested in multi-site backup
619 already run some form of VPN software in their internetwork
620 connections, and are willing to delegate security to that layer.
623 Item 10: Define definitive tape format.
626 What: Define that definitive tape format that will not change
627 for the next millennium.
629 Why: Stability, security.
631 Notes: See notes for item 11 below.
634 Item 11: New daemon communication protocol.
636 What: The current daemon to daemon protocol is basically an ASCII
637 printf() and sending the buffer. On the receiving end, the
638 buffer is sscanf()ed to unpack it. The new scheme would
639 be a binary format that allows quick packing and unpacking
640 of any data type with named fields.
642 Why: Using binary packing would be faster. Named fields will permit
643 error checking to ensure that what is sent is what the
644 receiver really wants.
646 Notes: These are internal improvements in the interest of the
647 long-term stability and evolution of the program. On the one
648 hand, the sooner they're done, the less code we have to rip
649 up when the time comes to install them. On the other hand, they
650 don't bring an immediately perceptible benefit to potential
651 users. Item 10 and possibly item 11 should be deferred until Bacula
652 is well established with a growing user community more or
653 less happy with the feature set. At that time, it will make a
654 good "next generation" upgrade in the interest of data
660 ======================================================
662 It is somewhat like a Full save becomes an incremental since
663 the Base job (or jobs) plus other non-base files.
665 - New BaseFile table that contains:
666 JobId, BaseJobId, FileId (from Base).
667 i.e. for each base file that exists but is not saved because
668 it has not changed, the File daemon sends the JobId, BaseId,
669 and FileId back to the Director who creates the DB entry.
670 - To initiate a Base save, the Director sends the FD
671 the FileId, and full filename for each file in the Base.
672 - When the FD finds a Base file, he requests the Director to
673 send him the full File entry (stat packet plus MD5), or
674 conversely, the FD sends it to the Director and the Director
675 says yes or no. This can be quite rapid if the FileId is kept
676 by the FD for each Base Filename.
677 - It is probably better to have the comparison done by the FD
678 despite the fact that the File entry must be sent across the
680 - An alternative would be to send the FD the whole File entry
681 from the start. The disadvantage is that it requires a lot of
682 space. The advantage is that it requires less communications
684 - The Job record must be updated to indicate that one or more
686 - At end of Job, FD returns:
687 1. Count of base files/bytes not written to tape (i.e. matches)
688 2. Count of base file that were saved i.e. had changed.
689 - No tape record would be written for a Base file that matches, in the
690 same way that no tape record is written for Incremental jobs where
691 the file is not saved because it is unchanged.
692 - On a restore, all the Base file records must explicitly be
693 found from the BaseFile tape. I.e. for each Full save that is marked
694 to have one or more Base Jobs, search the BaseFile for all occurrences
696 - An optimization might be to make the BaseFile have:
702 This would avoid the need to explicitly fetch each File record for
703 the Base job. The Base Job record will be fetched to get the
704 VolSessionId and VolSessionTime.
705 =========================================================
709 =============================================================
711 Request For Comments For File Backup Options
714 Subject: File Backup Options
717 A few days ago, a Bacula user who is backing up to file volumes and
718 using compression asked if it was possible to suppress compressing
719 all .gz files since it was a waste of CPU time. Although Bacula
720 currently permits using different options (compression, ...) on
721 a directory by directory basis, it cannot do it on a file by
722 file basis, which is clearly what was desired.
724 Proposed Implementation:
725 To solve this problem, I propose the following:
727 - Add a new Director resource type called Options.
729 - The Options resource will have records for all
730 options that can currently be specified on the Include record
731 (in a FileSet). Examples below.
733 - The Options resource will permit an exclude option as well
734 as a number of additional options.
736 - The heart of the Options resource is the ability to
737 supply any number of Match records which specify POSIX
738 regular expressions. These Match regular expressions are
739 applied to the fully qualified filename (path and all). If
740 one matches, then the Options will be used.
742 - When an Match specification matches an included file, the
743 options specified in the Options resource will override
744 the default options specified on the Include record.
746 - Include records will be modified to permit referencing one or
747 more Options resources. The Options will be used
748 in the order listed on the Include record and the first
749 one that matches will be applied.
751 - Options (or specifications) currently supplied on the Include
752 record will be deprecated (i.e. removed in a later version a
753 year or so from now).
755 - The Exclude record will be deprecated as the same functionality
756 can be obtained by using an Exclude = yes in the Options.
759 The following records can appear in the Options resource. An
760 asterisk preceding the name indicates a feature not currently
764 - Compression= (GZIP, ...)
765 - Signature= (MD5, SHA1, ...)
767 - OneFs= (yes/no) - remain on one filesystem
768 - Recurse= (yes/no) - recurse into subdirectories
769 - Sparse= (yes/no) - do sparse file backup
770 - *Exclude= (yes/no) - exclude file from being saved
771 - *Reader= (filename) - external read (backup) program
772 - *Plugin= (filename) - read/write plugin module
775 - verify= (ipnougsamc5) - verify options
778 - replace= (always/ifnewer/ifolder/never) - replace options currently
780 - *Writer= (filename) - external write (restore) program
784 Currently options specifying compression, MD5 signatures, recursion,
785 ... of a FileSet are supplied on the Include record. These will now
786 all be collected into a Options resource, which will be
787 specified in the Include in place of the options. Multiple Options
788 may be specified. Since the Options may contain regular expressions
789 that are applied to the full filename, this will give the ability
790 to specify backup options on a file by file basis to whatever level
799 Include = compression=GZIP signature=MD5 {
811 Match = /*.?*/ # matches all files.
816 That's a lot more to do the same thing, but it gives the ability to
817 apply options on a file by file basis. For example, suppose you
818 want to compress all files but not any file with extensions .gz or .Z.
819 In that case, you will need to group two sets of options using
820 the Options resource as follows:
828 # Note multiple Matches are ORed
829 Match = /*.gz/ # matches .gz files */
830 Match = /*.Z/ # matches .Z files */
835 Match = /*.?*/ # matches all files
841 Now, since the no Compression option is specified in the
842 first group of Options, *.gz or *.Z file will have an MD5 signature computed,
843 but will not be compressed. For all other files, the *.gz *.Z will not
844 match, so the second group of options will be used which will include GZIP
848 - Is it necessary to provide some means of ANDing regular expressions
849 and negation? (not currently planned)
851 e.g. Match = /*.gz/ && !/big.gz/
853 - I see that Networker has a "null" module which, if specified, does not
854 backup the file, but does make an record of the file in the catalog
855 so that the catalog will reflect an exact picture of the filesystem.
856 The result is that the file can be "seen" when "browsing" the save
857 sets, but it cannot be restored.
859 Is this really useful? Should it be implemented in Bacula?
862 After implementing the above, the user will be able to specify
863 on a file by file basis (using regular expressions) what options are
864 applied for the backup.
867 =============================================
869 ==========================================================
871 For each Incremental job that is run, there may be files that
872 were found but not saved because they were locked (this applies
873 only to Windows). Such a system could send back to the Director
874 a list of Unsaved files.
876 - New UnSavedFiles table that contains:
880 - Then in the next Incremental job, the list of Unsaved Files will be
881 feed to the FD, who will ensure that they are explicitly chosen even
882 if standard date/time check would not have selected them.
883 =============================================================
886 Done: (see kernsdone for more)
887 - Implement new alist in FileSet scanning.
888 - bls should continue reading even if it finds Win32 data on the tape.
889 The error should be Warning rather the Error.
890 - Add user configurable timeout for connecting to SD.
891 - Unsaved Flag in Job record (use JobMissingFiles).
892 - Base Flag in Job record.
893 - Configure mtx-changer to have correct path to mtx.
894 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
895 - Make some way so that if a machine is skipped because it is not up
896 that Bacula will continue retrying for a specified period of time --
898 - Implement all command line args on run.
899 - Implement command line "restore" args.
900 - Implement "restore current select=no"
901 - Restore file modified before date
902 - Restore -- do nothing but show what would happen
903 - Add estimate to Console commands
904 - Use read_record.c in SD code.
905 - Fix read_record to handle multiple sessions.
906 - Tip from Steve Allam
907 mt -f /dev/nst0 defblksize 0
908 - Document "status" in the console.
909 - Document driving console from shell script.
910 - Write JobMedia records with max file size is reached on tape.
911 - Handle the case of multiple JobMedia records pending (i.e. the
912 thread is slow and multiple situations requiring a JobMedia
914 - Do performance analysis on the restore tree routines.
915 - Fix maximum file size (block.c) to generate JobMedia records.
916 - Make the default file size 1GB on the tape.
917 - Implement forward spacing between files.
918 - Add Machine type (Linux/Windows) to Status report for daemons.
920 - Use repositioning at the beginning of the tape.
921 - Do full check the command line args in update (e.g. VolStatus ...).
922 - Specify list of files to restore
923 - Implement ClientRunBeforeJob and ClientRunAfterJob.
924 - Make | and < work on FD side.
925 - Check to see if "blocked" is set during restore.
926 - Figure out what is interrupting sql command in console.
927 - Make new job print warning User Unmounted Tape.
928 - Test recycling and purging (code changed in db_find_next_volume and
930 - Document SDConnectTimeout (in FD).
931 - Add restore by filename test.
932 - Document restore by files.
933 - Make variable expansion work correctly.
934 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
935 - Copy static programs into install directory.
936 - Think about changing Storage resource Device record to be
938 - Add RunBeforeJob and RunAfterJob to the Client program.
939 - Need return status on read_cb() from read_records(). Need multiple
940 records -- one per Job, maybe a JCR or some other structure with
941 a block and a record.
942 - LabelFormat on tape volume apparently creates the db record but
943 never actually labels the volume.
944 - Recycling a volume when two jobs are using it is going to break. Fixed.
945 - Document list nextvol and new format status dir.
946 - Client files in Win32 with Unix eol conventions doesn't work.
947 - Either fix or document that fill command in btape can be
948 compressed enormously by the hardware - a 36GB tape wrote 750GB!
949 - Add multiple character duration qualifiers.
950 - Require some modifer.
951 - Restrict characters permitted in a Resource name, and don't permit
953 - Figure out some way to ignore or get past checksum errors in
955 - The SD spooling file gets created even if it is not used.
956 - Look at Cleaning tape in ua_label.c for media create/update
957 - Add regression testing to the manual
958 - End time: in job output of rescheduled job is time of first run.
959 - Document list nextvol and status output.