4 Documentation to do: (a little bit at a time)
5 - Document running a test version.
6 - Document query file format.
7 - Document static linking
8 - Document how to automatically backup all local partitions
9 - Document problems with Verify and pruning.
10 - Document how to use multiple databases.
11 - Add a section to the doc on Manual cycling of Volumes.
14 Testing to do: (painful)
15 - that ALL console command line options work and are always implemented
16 - blocksize recognition code.
17 - multiple simultaneous Volumes
20 - Bug: up arrow prints garbage in command line on gnome-console!
21 - Bug: fix access problems on files restored on WinXP (possibly do
22 this in a later update.
23 - Fix Bare Metal restore problem.
24 - Test multiple simultaneous Volumes
26 - Document new transparent Console commands and wait command.
27 - Document new immediate File save configuration (walk user
28 through first save to file Volume with automatic Volume labeling?).
29 - Document OPTIMIZE TABLE in MySQL
30 - Document labeling a whole magazine using "cat"
31 - Document Lutz Kittler's trick of using "Run Before Job" to
32 abort a job on a particular day.
33 - Document Ludovic Strappazon's Win32 raw device save/restore.
34 - Document not to restore .journal .autofsck
36 - Document . and @ commands
39 - Figure out how to use ssh or stunnel to protect Bacula communications.
42 - Implement a Slot priority (loaded/not loaded).
43 - Implement Automatic Mount even in operator wait.
44 - Implement create "FileSet"?
45 - Implement Release Device in the Job resource to unmount a drive.
46 - Implement Acquire Device in the Job resource to mount a drive,
47 be sure this works with admin jobs so that the user can get
48 prompted to insert the correct tape. Possibly some way to say to
49 run the job but don't save the files.
50 - Implement all command line args on run.
51 - Implement command line "restore" args.
52 - Implement "restore current select=no"
53 - Fix watchdog pthread crash on Win32
54 - Fix "access not allowed" for backup of files on WinXP.
55 - Implement "scratch pool" where tapes are defined and can be
56 taken by any pool that needs them.
57 - Implement restore "current system", but take all files without
58 doing selection tree -- so that jobs without File records can
60 - Implement scheduling of one time "run" jobs (i.e. instead of
61 starting immediately start at some specified time).
62 - Make | and < work on FD side.
63 - Pass prefix_links to FD.
64 - Implement a M_SECURITY message class.
65 - Implement disk spooling. Two parts: 1. Spool to disk then
66 immediately to tape to speed up tape operations. 2. Spool to
67 disk only when the tape is full, then when a tape is hung move
69 - From Phil Stracchino:
70 It would probably be a per-client option, and would be called
71 something like, say, "Automatically purge obsoleted jobs". What it
72 would do is, when you successfully complete a Differential backup of a
73 client, it would automatically purge all Incremental backups for that
74 client that are rendered redundant by that Differential. Likewise,
75 when a Full backup on a client completed, it would automatically purge
76 all Differential and Incremental jobs obsoleted by that Full backup.
77 This would let people minimize the number of tapes they're keeping on
78 hand without having to master the art of retention times.
80 - Allow multiple Storage specifications (or multiple names on
81 a single Storage specification) in the Job record. Thus a job
82 can be backed up to a number of storage devices.
83 - Implement dump label to UA
84 - Add prefixlinks to where or not where absolute links to FD.
85 - Look at Python for a Bacula scripting language -- www.python.org
86 - Issue message to mount a new tape before the rewind.
87 - Simplified client job initiation for portables.
88 - If SD cannot open a drive, make it periodically retry.
89 - Implement LabelTemplate (at least first cut).
90 - Add more of the config info to the tape label.
91 - Implement a Mount Command and an Unmount Command where
92 the users could specify a system command to be performed
93 to do the mount, after which Bacula could attempt to
94 read the device. This is for Removeable media such as a CDROM.
95 - Most likely, this mount command would be invoked explicitly
96 by the user using the current Console "mount" and "unmount"
97 commands -- the Storage Daemon would do the right thing
98 depending on the exact nature of the device.
99 - As with tape drives, when Bacula wanted a new removable
100 disk mounted, it would unmount the old one, and send a message
101 to the user, who would then use "mount" as described above
102 once he had actually inserted the disk.
104 - Make some way so that if a machine is skipped because it is not up
105 that Bacula will continue retrying for a specified period of time --
107 - If tape is marked read-only, then try opening it read-only rather than
108 failing, and remember that it cannot be written.
109 - Refine SD waiting output:
110 Device is being positioned
111 > Device is being positioned for append
112 > Device is being positioned to file x
114 - Figure out some way to estimate output size and to avoid splitting
115 a backup across two Volumes -- this could be useful for writing CDROMs
116 where you really prefer not to have it split -- not serious.
117 - Add RunBeforeJob and RunAfterJob to the Client program.
118 - Have SD compute MD5 or SHA1 and compare to what FD computes.
119 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
120 actual data on the Volume and compare it.
121 - Implement FileOptions (see end of this document)
122 - Implement Bacula plugins -- design API
123 - Make bcopy read through bad tape records.
124 - Fix read_record to handle multiple sessions.
125 - Program files (i.e. execute a program to read/write files).
126 Pass read date of last backup, size of file last time.
127 - Add Signature type to File DB record.
128 - Make Restore report an error if FD or SD term codes are not OK.
129 - CD into subdirectory when open()ing files for backup to
130 speed up things. Test with testfind().
131 - Priority job to go to top of list.
132 - Find out why Full saves run slower and slower (hashing?)
133 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
134 - Implement some way for the Console to dynamically create a job.
135 - Restore to a particular time -- e.g. before date, after date.
136 - Solaris -I on tar for include list
137 - Prohibit backing up archive device (findlib/find_one.c:128)
138 - Need a verbose mode in restore, perhaps to bsr.
139 - bscan without -v is too quiet -- perhaps show jobs.
140 - Add code to reject whole blocks if not wanted on restore.
141 - Start working on Base jobs.
142 - Check if we can increase Bacula FD priorty in Win2000
143 - Make sure the MaxVolFiles is fully implemented in SD
144 - Check if both CatalogFiles and UseCatalog are set to SD.
145 - Need return status on read_cb() from read_records(). Need multiple
146 records -- one per Job, maybe a JCR or some other structure with
147 a block and a record.
148 - Figure out how to do a bare metal Windows restore
149 - Put system type returned by FD into catalog.
150 - Possibly add email to Watchdog if drive is unmounted too
151 long and a job is waiting on the drive.
152 - Use read_record.c in SD code.
153 - Why don't we get an error message from Win32 FD when bootstrap
154 file cannot be created for restore command?
155 - When Marking a file in Restore that is a hard link, also
156 mark the link so that the data will be reloaded.
157 - Restore program that errors in SD due to no tape reports
158 OK incorrectly in output.
159 - After unmount, if restore job started, ask to mount.
160 - Convert all %x substitution variables, which are hard to remember
161 and read to %(variable-name). Idea from TMDA.
162 - Add JobLevel in FD status (but make sure it is defined).
163 - Make Pool resource handle Counter resources.
164 - Remove NextId for SQLite. Optimize.
165 - Move all SQL statements into a single location.
166 - Add UA rc and history files.
167 - put termcap (used by console) in ./configure and
168 allow -with-termcap-dir.
169 - Enhance time and size scanning routines.
170 - Fix Autoprune for Volumes to respect need for full save.
171 - Fix Win32 config file definition name on /install
172 - No READLINE_SRC if found in alternate directory.
173 - Test a second language e.g. french.
174 - Compare tape to Client files (attributes, or attributes and data)
175 - Make all database Ids 64 bit.
176 - Write an applet for Linux.
177 - Add estimate to Console commands
178 - Find solution to blank filename (i.e. path only) problem.
179 - Implement new daemon communications protocol.
180 - Remove PoolId from Job table, it exists in Media.
181 - Allow console commands to detach or run in background.
182 - Fix status delay on storage daemon during rewind.
183 - Add SD message variables to control operator wait time
184 - Maximum Operator Wait
185 - Minimum Message Interval
186 - Maximum Message Interval
187 - Send Operator message when cannot read tape label.
188 - Verify level=Volume (scan only), level=Data (compare of data to file).
189 Verify level=Catalog, level=InitCatalog
191 - Add keyword search to show command in Console.
192 - Fix Win2000 error with no messages during startup.
193 - Events : tape has more than xxx bytes.
194 - Restrict characters permitted in a Resource name.
195 - Complete code in Bacula Resources -- this will permit
196 reading a new config file at any time.
197 - Handle ctl-c in Console
198 - Implement script driven addition of File daemon to config files.
199 - Think about how to make Bacula work better with File (non-tape) archives.
200 - Write Unix emulator for Windows.
201 - Implement new serialize subroutines
202 send(socket, "string", &Vol, "uint32", &i, NULL)
203 - Audit all UA commands to ensure that we always prompt where possible.
204 - If ./btape is called without /dev, assume argument is a Storage resource name.
205 - Put memory utilization in Status output of each daemon
206 if full status requested or if some level of debug on.
207 - Make database type selectable by .conf files i.e. at runtime
208 - Set flag for uname -a. Add to Volume label.
209 - Implement throttled work queue.
210 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
211 - Restore files modified after date
212 - Restore file modified before date
213 - Emergency restore info:
215 - Backup working directory
217 - Restore -- do nothing but show what would happen
218 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
219 - Implement Restore FileSet=
220 - Create a protocol.h and protocol.c where all protocol messages
222 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
223 - Timout a job or terminate if link goes down, or reopen link and query.
224 - Find general solution for sscanf size problems (as well
225 as sprintf. Do at run time?
226 - Concept of precious tapes (cannot be reused).
227 - Make bcopy copy with a single tape drive.
228 - Permit changing ownership during restore.
230 - Autolabel should be specified by DIR instead of SD.
231 - Find out how to get the system tape block limits, e.g.:
232 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
233 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
236 - AutoScan (check checksum of tape)
237 - Format command = "format /dev/nst0"
241 - Seek resolution (usually corresponds to buffer size)
242 - EODErrorCode=ENOSPC or code
243 - Partial Read error code
244 - Partial write error code
245 - Nonformatted read error
246 - Nonformatted write error
247 - WriteProtected error
251 - IgnoreCloseErrors=yes
261 - FD sends unsaved file list to Director at end of job (see
263 - File daemon should build list of files skipped, and then
264 at end of save retry and report any errors.
265 - Write a Storage daemon that uses pipes and
266 standard Unix programs to write to the tape.
268 - Need something that monitors the JCR queue and
269 times out jobs by asking the deamons where they are.
270 - Enhance Jmsg code to permit buffering and saving to disk.
271 - device driver = "xxxx" for drives.
272 - restart: paranoid: read label fsf to
273 eom read append block, and go
274 super-paranoid: read label, read all files
275 in between, read append block, and go
276 verify: backspace, read append block, and go
277 permissive: same as above but frees drive
278 if tape is not valid.
280 - Ensure that /dev/null works
281 - Need report class for messages. Perhaps
282 report resource where report=group of messages
283 - enhance scan_attrib and rename scan_jobtype, and
284 fill in code for "since" option
285 - Director needs a time after which the report status is sent
286 anyway -- or better yet, a retry time for the job.
287 Don't reschedule a job if previous incarnation is still running.
288 - Some way to automatically backup everything is needed????
289 - Need a structure for pending actions:
291 - termination status (part of buffered msgs?)
293 Read, Write, Clean, Delete
294 - Login to Bacula; Bacula users with different permissions:
295 owner, group, user, quotas
296 - Store info on each file system type (probably in the job header on tape.
297 This could be the output of df; or perhaps some sort of /etc/mtab record.
300 - Design at hierarchial storage for Bacula. Migration and Clone.
301 - Implement FSM (File System Modules).
302 - Identify unchanged or "system" files and save them to a
303 special tape thus removing them from the standard
304 backup FileSet -- BASE backup.
305 - Heartbeat between daemons.
306 - Audit M_ error codes to ensure they are correct and consistent.
307 - Add variable break characters to lex analyzer.
308 Either a bit mask or a string of chars so that
309 the caller can change the break characters.
310 - Make a single T_BREAK to replace T_COMMA, etc.
311 - Ensure that File daemon and Storage daemon can
312 continue a save if the Director goes down (this
313 is NOT currently the case). Must detect socket error,
314 buffer messages for later.
315 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
316 - Add ability to backup to two Storage devices (two SD sessions) at
317 the same time -- e.g. onsite, offsite.
318 - Add the ability to consolidate old backup sets (basically do a restore
319 to tape and appropriately update the catalog). Compress Volume sets.
320 Might need to spool via file is only one drive is available.
321 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
322 someway to do so with every volume that has less than x% valid
327 Bacula Projects Roadmap
329 last update 5 January 2003
331 Item 1: Multiple simultaneous Jobs. (done)
332 Done -- Restore part needs better implementation to work correctly
334 What: Permit multiple simultaneous jobs in Bacula.
336 Why: An enterprise level solution needs to go fast without the
337 need for the system administrator to carefully tweak
338 timing. Based on the benchmarks, during a full
339 backup, NetWorker typically hit 10 times the bandwidth to
340 the tape compared to Bacula--largely. This is probably due to
341 running parallel jobs and multi-threaded filling of buffers
342 and writing them to tape. This should also make things work
343 better when you have a mix of fast and slow machines backing
346 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
347 implementing this is a matter of some small cleanups and
351 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
352 Deferred -- not necessary yet.
354 What: If data is coming into the SD too fast, buffer it to
355 disk if the user has configured this option.
357 Why: This would be nice, especially if it more or less falls out
358 when implementing (1) above. If not, it probably should not
359 be given a high priority because fundamentally the backup time
360 is limited by the tape bandwidth. Even though you may finish a
361 client job quicker by spilling to disk, you still have to
362 eventually get it onto tape. If intermediate disk buffering
363 allows us to improve write bandwidth to tape, it may make
366 Notes: Whether or not this is implemented will depend upon performance
367 testing after item 1 is implemented.
370 Item 3: Write the bscan program -- also write a bcopy program.
373 What: Write a program that reads a Bacula tape and puts all the
374 appropriate data into the catalog. This allows recovery
375 from a tape that is no longer in the database, or it allows
376 re-creation of a database if lost.
378 Why: This is a fundamental robustness and disaster recovery tool
379 which will increase the comfort level of a sysadmin
380 considering adopting Bacula.
382 Notes: A skeleton of this program already exists, but much work
383 needs to be done. Implementing this will also make apparent
384 any deficiencies in the current Bacula tape format.
387 Item 4: Implement Base jobs.
389 What: A base job is sort of like a Full save except that you
390 will want the FileSet to contain only files that are unlikely
391 to change in the future (i.e. a snapshot of most of your
392 system after installing it). After the base job has been run,
393 when you are doing a Full save, you can specify to exclude
394 all files saved by the base job that have not been modified.
396 Why: This is something none of the competition does, as far as we know
397 (except BackupPC, which is a Perl program that saves to disk
398 only). It is big win for the user, it makes Bacula stand out
399 as offering a unique optimization that immediately saves time
402 Notes: Big savings in tape usage. Will require more resources because
403 the e. DIR must send FD a list of files/attribs, and the FD must
404 search the list and compare it for each file to be saved.
407 Item 5: Implement Label templates
409 What: This is a mechanism whereby Bacula can automatically create
410 a tape label for new tapes according to a detailed specification
411 provided by the user.
413 Why: It is a major convenience item for folks who use automated label
416 Notes: Bacula already has a working form of automatic tape label
417 creation, but it is very crude. The design for the complete
418 tape labeling project is already documented in the manual.
421 Item 6: Write a regression script.
422 Done -- Continue to expand its testing.
424 What: This is an automatic script that runs and tests as many features
425 of Bacula as possible. The output is compared to previous
426 versions of Bacula and any differences are reported.
428 Why: This is an enormous help in preventing introduction of new
429 errors in parts of the program that already work correctly.
431 Notes: This probably should be ranked higher, it's something the typical
432 user doesn't see. Depending on how it's implemented, it may
433 make sense to defer it until the archival tape format and
434 user interface mature.
437 Item 7: GUI for interactive restore
438 Item 8: GUI for interactive backup
440 What: The current interactive restore is implemented with a tty
441 interface. It would be much nicer to be able to "see" the
442 list of files backed up in typical GUI tree format.
443 The same mechanism could also be used for creating
444 ad-hoc backup FileSets (item 8).
446 Why: Ease of use -- especially for the end user.
448 Notes: Rather than implementing in Gtk, we probably should go directly
449 for a Browser implementation, even if doing so meant the
450 capability wouldn't be available until much later. Not only
451 is there the question of Windows sites, most
452 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
453 without installing lots of stuff admins are very wary about.
454 Real sysadmins will always use the command line anyway, and
455 the user who's doing an interactive restore or backup of his
456 own files will in most cases be on a Windows machine running
460 Item 9: Add SSL to daemon communications.
462 What: This provides for secure communications between the daemons.
464 Why: This would allow doing backup across the Internet without
465 privacy concerns (or with much less concern).
467 Notes: The vast majority of near term potential users will be backing up
468 a single site over a LAN and, correctly or not, they probably
469 won't be concerned with security, at least not enough to go to
470 the trouble to set up keys, etc. to screw things down. We suspect
471 that many users genuinely interested in multi-site backup
472 already run some form of VPN software in their internetwork
473 connections, and are willing to delegate security to that layer.
476 Item 10: Define definitive tape format.
479 What: Define that definitive tape format that will not change
480 for the next millennium.
482 Why: Stability, security.
484 Notes: See notes for item 11 below.
487 Item 11: New daemon communication protocol.
489 What: The current daemon to daemon protocol is basically an ASCII
490 printf() and sending the buffer. On the receiving end, the
491 buffer is sscanf()ed to unpack it. The new scheme would
492 be a binary format that allows quick packing and unpacking
493 of any data type with named fields.
495 Why: Using binary packing would be faster. Named fields will permit
496 error checking to ensure that what is sent is what the
497 receiver really wants.
499 Notes: These are internal improvements in the interest of the
500 long-term stability and evolution of the program. On the one
501 hand, the sooner they're done, the less code we have to rip
502 up when the time comes to install them. On the other hand, they
503 don't bring an immediately perceptible benefit to potential
504 users. Item 10 and possibly item 11 should be deferred until Bacula
505 is well established with a growing user community more or
506 less happy with the feature set. At that time, it will make a
507 good "next generation" upgrade in the interest of data
513 ======================================================
515 It is somewhat like a Full save becomes an incremental since
516 the Base job (or jobs) plus other non-base files.
518 - New BaseFile table that contains:
519 JobId, BaseJobId, FileId (from Base).
520 i.e. for each base file that exists but is not saved because
521 it has not changed, the File daemon sends the JobId, BaseId,
522 and FileId back to the Director who creates the DB entry.
523 - To initiate a Base save, the Director sends the FD
524 the FileId, and full filename for each file in the Base.
525 - When the FD finds a Base file, he requests the Director to
526 send him the full File entry (stat packet plus MD5), or
527 conversely, the FD sends it to the Director and the Director
528 says yes or no. This can be quite rapid if the FileId is kept
529 by the FD for each Base Filename.
530 - It is probably better to have the comparison done by the FD
531 despite the fact that the File entry must be sent across the
533 - An alternative would be to send the FD the whole File entry
534 from the start. The disadvantage is that it requires a lot of
535 space. The advantage is that it requires less communications
537 - The Job record must be updated to indicate that one or more
539 - At end of Job, FD returns:
540 1. Count of base files/bytes not written to tape (i.e. matches)
541 2. Count of base file that were saved i.e. had changed.
542 - No tape record would be written for a Base file that matches, in the
543 same way that no tape record is written for Incremental jobs where
544 the file is not saved because it is unchanged.
545 - On a restore, all the Base file records must explicitly be
546 found from the BaseFile tape. I.e. for each Full save that is marked
547 to have one or more Base Jobs, search the BaseFile for all occurrences
549 - An optimization might be to make the BaseFile have:
555 This would avoid the need to explicitly fetch each File record for
556 the Base job. The Base Job record will be fetched to get the
557 VolSessionId and VolSessionTime.
558 =========================================================
560 ==========================================================
562 For each Incremental job that is run, there may be files that
563 were found but not saved because they were locked (this applies
564 only to Windows). Such a system could send back to the Director
565 a list of Unsaved files.
567 - New UnSavedFiles table that contains:
571 - Then in the next Incremental job, the list of Unsaved Files will be
572 feed to the FD, who will ensure that they are explicitly chosen even
573 if standard date/time check would not have selected them.
574 =============================================================
578 =============================================================
580 Request For Comments For File Backup Options
583 Subject: File Backup Options
586 A few days ago, a Bacula user who is backing up to file volumes and
587 using compression asked if it was possible to suppress compressing
588 all .gz files since it was a waste of CPU time. Although Bacula
589 currently permits using different options (compression, ...) on
590 a directory by directory basis, it cannot do it on a file by
591 file basis, which is clearly what was desired.
593 Proposed Implementation:
594 To solve this problem, I propose the following:
596 - Add a new Director resource type called FileOptions.
598 - The FileOptions resource will have records for all
599 options that can currently be specified on the Include record
600 (in a FileSet). Examples below.
602 - The FileOptions resource will permit an exclude option as well
603 as a number of additional options.
605 - The heart of the FileOptions resource is the ability to
606 supply any number of ApplyTo records which specify POSIX
607 regular expressions. These ApplyTo regular expressions are
608 applied to the fully qualified filename (path and all). If
609 one matches, then the FileOptions will be used.
611 - When an ApplyTo specification matches an included file, the
612 options specified in the FileOptions resource will override
613 the default options specified on the Include record.
615 - Include records will be modified to permit referencing one or
616 more FileOptions resources. The FileOptions will be used
617 in the order listed on the Include record and the first
618 one that matches will be applied.
620 - Options (or specifications) currently supplied on the Include
621 record will be deprecated (i.e. removed in a later version a
622 year or so from now).
624 - The Exclude record will be deprecated as the same functionality
625 can be obtained by using an Exclude = yes in the FileOptions.
628 The following records can appear in the FileOptions resource. An
629 asterisk preceding the name indicates a feature not currently
633 - Compression= (GZIP, ...)
634 - Signature= (MD5, SHA1, ...)
636 - OneFs= (yes/no) - remain on one filesystem
637 - Recurse= (yes/no) - recurse into subdirectories
638 - Sparse= (yes/no) - do sparse file backup
639 - *Exclude= (yes/no) - exclude file from being saved
640 - *Reader= (filename) - external read (backup) program
641 - *Plugin= (filename) - read/write plugin module
644 - verify= (ipnougsamc5) - verify options
647 - replace= (always/ifnewer/ifolder/never) - replace options currently
649 - *Writer= (filename) - external write (restore) program
653 Currently options specifying compression, MD5 signatures, recursion,
654 ... of a FileSet are supplied on the Include record. These will now
655 all be collected into a FileOptions resource, which will be
656 specified on the Include in place of the options. Multiple FileOptions
657 may be specified. Since the FileOptions contain regular expressions
658 that are applied to the full filename, this will give the ability
659 to specify backup options on a file by file basis to whatever level
668 Include = compression=GZIP signature=MD5 {
677 Include = FileOptions=Opts {
688 That's a lot more to do the same thing, but it gives the ability to
689 apply options on a file by file basis. For example, suppose you
690 want to compress all files but not any file with extensions .gz or .Z.
691 You could do so as follows:
695 Include = FileOptions=NoCompress FileOptions=Opts {
703 ApplyTo = /*.?*/ # matches all files
708 # Note multiple ApplyTos are ORed
709 ApplyTo = /*.gz/ # matches .gz files */
710 ApplyTo = /*.Z/ # matches .Z files */
713 Now, since the NoCompress FileOptions is specified first on the
714 Include line, any *.gz or *.Z file will have an MD5 signature computed,
715 but will not be compressed. For all other files, the NoCompress will not
716 match, so the Opts options will be used which will include GZIP
720 - Is it necessary to provide some means of ANDing regular expressions
721 and negation? (not currently planned)
723 e.g. ApplyTo = /*.gz/ && !/big.gz/
725 - I see that Networker has a "null" module which, if specified, does not
726 backup the file, but does make an record of the file in the catalog
727 so that the catalog will reflect an exact picture of the filesystem.
728 The result is that the file can be "seen" when "browsing" the save
729 sets, but it cannot be restored.
731 Is this really useful? Should it be implemented in Bacula?
734 After implementing the above, the user will be able to specify
735 on a file by file basis (using regular expressions) what options are
736 applied for the backup.
737 ====================================
739 =========================================
740 Proposal by Bill Sellers
742 Return-Path: <w.a.sellers@larc.nasa.gov>
743 Received: from post.larc.nasa.gov (post.larc.nasa.gov [128.155.4.45]) by matou.sibbald.com (8.11.6/8.11.6) with ESMTP id h0ELUIm07622 for <kern@sibbald.com>; Tue, 14 Jan 2003 22:30:18 +0100
744 Received: from Baron.larc.nasa.gov (baron.larc.nasa.gov [128.155.40.132]) by post.larc.nasa.gov (pohub4.6) with ESMTP id QAA09768 for <kern@sibbald.com>; Tue, 14 Jan 2003 16:30:14 -0500 (EST)
745 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
746 X-Sender: w.a.sellers@pop.larc.nasa.gov
747 X-Mailer: QUALCOMM Windows Eudora Version 5.1
748 Date: Tue, 14 Jan 2003 16:30:18 -0500
749 To: Kern Sibbald <kern@sibbald.com>
750 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
751 Subject: Re: [Bacula-users] Bacula remote storage?
752 In-Reply-To: <1042565382.1845.177.camel@rufus>
753 References: <5.1.0.14.2.20030114113004.0293a210@pop.larc.nasa.gov> <5.1.0.14.2.20030113170650.028dad88@pop.larc.nasa.gov> <5.1.0.14.2.20030113170650.028dad88@pop.larc.nasa.gov> <5.1.0.14.2.20030114113004.0293a210@pop.larc.nasa.gov>
755 Content-Type: text/plain; charset="us-ascii"; format=flowed
756 X-Annoyance-Filter-Junk-Probability: 0
757 X-Annoyance-Filter-Classification: Mail
758 At 06:29 PM 1/14/2003 +0100, you wrote:
761 >Well, if you cannot put a Bacula client on the machine,
762 >then it is a big problem. If you know of some software
763 >that can do what you want, let me know, because I
764 >really just don't know how to do it -- at least not
770 We have been able to get Amanda to use the HSM as a storage
771 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
772 handle Windows systems very well (or at all without Samba). So I am
773 looking for a backup system that has a Windows client. I really like the
774 Windows integration of Bacula.
776 From the command line, its rather trivial to move the data around. We use
779 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
781 or if you use GNU tar:
783 tar czf hsm:path/file.tgz ./files
785 One idea for you to consider; Sendmail offers pipes in the aliases file;
786 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
787 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
788 pipe available, as a storage device? Then we could use any command that
789 handles stdin as a storage destination.
796 #Password is not used in rsh, but might be used in ftp.
797 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
804 Password = "foobar&-"
805 Device = "| ncftpput -c hsm /path/file.bacula"
809 >If you have some local storage available, you could
810 >use Bacula to backup to disk volumes, then use some
811 >other software (ftp, scp) to move them to the HSM
812 >machine. However, this is a bit kludgy.
815 It is, but maybe worth a try. Is there some function in Bacula to put
816 variables in filenames? i.e. backup.2003-01-15.root
823 w.a.sellers@larc.nasa.gov
825 ==============================================
826 The Project for the above
828 I finally realized that this is not at all
829 the same as reader/writer programs or plugins,
830 which are alternate ways of accessing the
831 files to be backed up. Rather, it is an alternate
832 form of storage device, and I have always planned
833 that Bacula should be able to handle all sorts
836 So, I propose the following phases:
838 1. OK from you to invest some time in testing
839 this as I implement it (requires that you
840 know how to download from the SourceForge
841 cvs -- which I imagine is a piece of cake
844 2. Dumb implementation by allowing a device to
845 be a fifo for write only.
846 Reason: easy to implement, proof of concept.
848 3. Try reading from fifo but with fixed block
850 Reason: proof of concept, easy to implement.
852 4. Extend reading from fifo (restores) to handle
854 Reason: requires some delicate low level coding
855 which could destabilize all of Bacula.
857 5. Implementation of above but to a program. E.g.
858 Device = "|program" (not full pipeline).
859 Reason: routines already exist, and program can
860 be a shell script which contains anything.
862 6. Add full pipeline as a possibility. E.g.
863 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
864 Reason: needs additional coding to implement full
865 pipeline (must fire off either a shell or all
866 programs and connect their pipes).
868 There are a good number of details in each step
869 that I have left out, but I will specify them at
870 every stage, and there may be a few changes as things
871 evolve. I expect that to get to stage 5 will take a
872 few weeks, and at that point, you will have
873 everything you need (just inside a script).
874 Stage 6 will probably take longer, but if this
875 project pleases you, what we do for 5 should
876 be adequate for some time.
880 =============================================
882 Done: (see kernsdone for more)
883 - Look into Pruning/purging problems or why there seem to
884 be so many files listed each night.
885 - Fix cancel in find_one -- need jcr.
886 - Cancel does not work for restore in FD.
887 - Write SetJobStatus() function so cancel status not lost.
888 - Add include list to end of chain in findlib
889 - Zap sd_auth_key after use
890 - Add Bar code reading capabilities (new mtx-changer)
891 - Figure out some way to automatically backup all local partitions
892 - Make hash table for linked files in findlib/find_one.c:161
894 - Rewrite find_one.c to use only pool_memory instead of
895 alloca and malloc (probably not necessary).
896 - Make sure btraceback goes into /sbin not sysconf directory.
897 - InitVerify is getting pruned and it shouldn't (document it)
898 - Make 1.28c release ??? NO do 1.29 directly
899 - Set timeout on opening fifo for save or restore (findlib)
900 - Document FIFO storage device.
901 - Document fifo and | and <
904 - Get correct error status from run_program or open_bpipe().
905 - Restrict permissions on File Volumes (now 0640).
906 - Umasked 022 daemons
907 - Fix restore of hard linked file.
908 - Figure out how to allow multiple simultaneous file Volumes on a single device.
909 - Implement multiple simultaneous file Volumes on a single device.
910 - Cleanup db_update_media and db_update_pool
911 - Flush all the daemon messages at the end of every job.
912 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
913 run_program -- bpipe.c
914 - Apparently cancel does not work for jobs waiting to be
916 - Implement TCP/IP connection for MySQL
917 - Pull a canceled job from the Scheduling queue.
918 - Implement max_file_size in block.c (already done, just tweaked).
919 - Look at purge jobs volume (at least document it, and see if it is
921 - Add list volumes does all pools. list volumes pool=xxx now works.
922 - Add pool= to "list media" in ua_output.c
923 - Strip trailing slashes from Include directory names in the FD.
924 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
925 ERR=Operation not permited loop.
926 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
927 - Produce better error messages in when error/eof writing block.
928 - Cancelling of a queued job does NOT work!!!!!!
930 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
931 rufus-sd: Recycled volume File0003 on device /home/kern/bacula/working, all previous data lost.
932 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
933 - Ability to backup to a file then later transfer to a tape -- Migration.
934 Migration based on MaxJobs(MinJobs),MaxVols(MinVols),AgeJobs,MaxBytes(MinBytes)
935 (i.e. HighwaterMark, LowwaterMark).
936 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
937 set of volumes recycling the oldest volume when it is needed.
938 - gethostbyname failure in bnet_connect() continues
939 generating errors -- should stop.
940 - Add chflags() code for FreeBSD file flags
941 - Bevan Anderson suggests having a run queue for each device
942 so that multiple simultaneous jobs can run but each writing
943 to a different Volume.
944 - Look at handling <> in smtp doesn't work with exim.
945 - Need to specify MaximumConcurrentJobs in the Job resource.
946 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
947 - Implement finer multiprocessing options.
948 - Implement | and < in Exclude statements.
949 - Figure out some way to specify a retention period for files
950 that no longer exist on the machine -- so that we maintain
951 say backups for 30 days, but if the file is deleted, we maintain
952 the last copy for 1 year. -- answer Volume retention.
953 - Make non-zero status from RunJobBefore/After error the job.
954 - Need define of int_least16_t in sha1.h for SuSE.
955 - Implement bar code reader for autochangers
956 - Document new MaximumConcurrentJob records (Job, Client, Storage)
957 - Write up how to use/manage disk Volume Storage. ******
958 - Remove kern and kelvin from mysql_grant...
959 - Install grant_mysql...
960 - Strip trailing / from Include
961 - add #define ENABLE_NLS for Gnome compile on SuSE.
962 - Add Client FS/OS id (Linux, Win95/98, ...).
963 - Concept of VolumeSet during restore which is a list
964 of Volume names needed.
965 - Turn virutally all sprintfs into snprintfs.
966 - Update volume=Test01 requests pool, then lists volumes.
967 **** Test select_pool_and_media ...
969 - Add IP address to authentication failures.
970 - Add a default File storage so that new users can do backup
971 and restores right away.
972 - Forbid sbindir and with-subsys-dir from being the same (otherwise
973 the binary gets deleted when the daemon is stopped in the
974 rc.d/inid.d directory.
975 - Do not ignore SIGCHLD
976 - Add Cleaning to list of volume statuses
977 - Implement run at "xxx"