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 immediate File save configuration (walk user
27 through first save to file Volume with automatic Volume labeling?).
28 - Document OPTIMIZE TABLE in MySQL
31 - Figure out how to use ssh or stunnel to protect Bacula communications.
34 - Implement a Slot priority (loaded/not loaded).
35 - Implement Automatic Mount even in operator wait.
36 - Implement create "FileSet"?
37 - Implement Release Device in the Job resource to unmount a drive.
38 - Implement Acquire Device in the Job resource to mount a drive,
39 be sure this works with admin jobs so that the user can get
40 prompted to insert the correct tape. Possibly some way to say to
41 run the job but don't save the files.
42 - Implement all command line args on run.
43 - Implement command line "restore" args.
44 - Implement "restore current select=no"
45 - Fix watchdog pthread crash on Win32
46 - Fix "access not allowed" for backup of files on WinXP.
47 - Implement "scratch pool" where tapes are defined and can be
48 taken by any pool that needs them.
49 - Implement restore "current system", but take all files without
50 doing selection tree -- so that jobs without File records can
52 - Implement scheduling of one time "run" jobs (i.e. instead of
53 starting immediately start at some specified time).
54 - Make | and < work on FD side.
55 - Pass prefix_links to FD.
56 - Implement a M_SECURITY message class.
57 - Implement disk spooling. Two parts: 1. Spool to disk then
58 immediately to tape to speed up tape operations. 2. Spool to
59 disk only when the tape is full, then when a tape is hung move
61 - From Phil Stracchino:
62 It would probably be a per-client option, and would be called
63 something like, say, "Automatically purge obsoleted jobs". What it
64 would do is, when you successfully complete a Differential backup of a
65 client, it would automatically purge all Incremental backups for that
66 client that are rendered redundant by that Differential. Likewise,
67 when a Full backup on a client completed, it would automatically purge
68 all Differential and Incremental jobs obsoleted by that Full backup.
69 This would let people minimize the number of tapes they're keeping on
70 hand without having to master the art of retention times.
72 - Allow multiple Storage specifications (or multiple names on
73 a single Storage specification) in the Job record. Thus a job
74 can be backed up to a number of storage devices.
75 - Implement dump label to UA
76 - Add prefixlinks to where or not where absolute links to FD.
77 - Look at Python for a Bacula scripting language -- www.python.org
78 - Issue message to mount a new tape before the rewind.
79 - Simplified client job initiation for portables.
80 - If SD cannot open a drive, make it periodically retry.
81 - Implement LabelTemplate (at least first cut).
82 - Add more of the config info to the tape label.
83 - Implement a Mount Command and an Unmount Command where
84 the users could specify a system command to be performed
85 to do the mount, after which Bacula could attempt to
86 read the device. This is for Removeable media such as a CDROM.
87 - Most likely, this mount command would be invoked explicitly
88 by the user using the current Console "mount" and "unmount"
89 commands -- the Storage Daemon would do the right thing
90 depending on the exact nature of the device.
91 - As with tape drives, when Bacula wanted a new removable
92 disk mounted, it would unmount the old one, and send a message
93 to the user, who would then use "mount" as described above
94 once he had actually inserted the disk.
96 - Make some way so that if a machine is skipped because it is not up
97 that Bacula will continue retrying for a specified period of time --
99 - If tape is marked read-only, then try opening it read-only rather than
100 failing, and remember that it cannot be written.
101 - Refine SD waiting output:
102 Device is being positioned
103 > Device is being positioned for append
104 > Device is being positioned to file x
106 - Figure out some way to estimate output size and to avoid splitting
107 a backup across two Volumes -- this could be useful for writing CDROMs
108 where you really prefer not to have it split -- not serious.
109 - Add RunBeforeJob and RunAfterJob to the Client program.
110 - Have SD compute MD5 or SHA1 and compare to what FD computes.
111 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
112 actual data on the Volume and compare it.
113 - Implement FileOptions (see end of this document)
114 - Implement Bacula plugins -- design API
115 - Make bcopy read through bad tape records.
116 - Fix read_record to handle multiple sessions.
117 - Program files (i.e. execute a program to read/write files).
118 Pass read date of last backup, size of file last time.
119 - Add Signature type to File DB record.
120 - Make Restore report an error if FD or SD term codes are not OK.
121 - CD into subdirectory when open()ing files for backup to
122 speed up things. Test with testfind().
123 - Priority job to go to top of list.
124 - Find out why Full saves run slower and slower (hashing?)
125 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
126 - Implement some way for the Console to dynamically create a job.
127 - Restore to a particular time -- e.g. before date, after date.
128 - Solaris -I on tar for include list
129 - Prohibit backing up archive device (findlib/find_one.c:128)
130 - Need a verbose mode in restore, perhaps to bsr.
131 - bscan without -v is too quiet -- perhaps show jobs.
132 - Add code to reject whole blocks if not wanted on restore.
133 - Start working on Base jobs.
134 - Check if we can increase Bacula FD priorty in Win2000
135 - Make sure the MaxVolFiles is fully implemented in SD
136 - Check if both CatalogFiles and UseCatalog are set to SD.
137 - Need return status on read_cb() from read_records(). Need multiple
138 records -- one per Job, maybe a JCR or some other structure with
139 a block and a record.
140 - Figure out how to do a bare metal Windows restore
141 - Put system type returned by FD into catalog.
142 - Possibly add email to Watchdog if drive is unmounted too
143 long and a job is waiting on the drive.
144 - Use read_record.c in SD code.
145 - Why don't we get an error message from Win32 FD when bootstrap
146 file cannot be created for restore command?
147 - When Marking a file in Restore that is a hard link, also
148 mark the link so that the data will be reloaded.
149 - Restore program that errors in SD due to no tape reports
150 OK incorrectly in output.
151 - After unmount, if restore job started, ask to mount.
152 - Convert all %x substitution variables, which are hard to remember
153 and read to %(variable-name). Idea from TMDA.
154 - Add JobLevel in FD status (but make sure it is defined).
155 - Make Pool resource handle Counter resources.
156 - Remove NextId for SQLite. Optimize.
157 - Move all SQL statements into a single location.
158 - Add UA rc and history files.
159 - put termcap (used by console) in ./configure and
160 allow -with-termcap-dir.
161 - Enhance time and size scanning routines.
162 - Fix Autoprune for Volumes to respect need for full save.
163 - Fix Win32 config file definition name on /install
164 - No READLINE_SRC if found in alternate directory.
165 - Test a second language e.g. french.
166 - Compare tape to Client files (attributes, or attributes and data)
167 - Make all database Ids 64 bit.
168 - Write an applet for Linux.
169 - Add estimate to Console commands
170 - Find solution to blank filename (i.e. path only) problem.
171 - Implement new daemon communications protocol.
172 - Remove PoolId from Job table, it exists in Media.
173 - Allow console commands to detach or run in background.
174 - Fix status delay on storage daemon during rewind.
175 - Add SD message variables to control operator wait time
176 - Maximum Operator Wait
177 - Minimum Message Interval
178 - Maximum Message Interval
179 - Send Operator message when cannot read tape label.
180 - Verify level=Volume (scan only), level=Data (compare of data to file).
181 Verify level=Catalog, level=InitCatalog
183 - Add keyword search to show command in Console.
184 - Fix Win2000 error with no messages during startup.
185 - Events : tape has more than xxx bytes.
186 - Restrict characters permitted in a Resource name.
187 - Complete code in Bacula Resources -- this will permit
188 reading a new config file at any time.
189 - Handle ctl-c in Console
190 - Implement script driven addition of File daemon to config files.
191 - Think about how to make Bacula work better with File (non-tape) archives.
192 - Write Unix emulator for Windows.
193 - Implement new serialize subroutines
194 send(socket, "string", &Vol, "uint32", &i, NULL)
195 - Audit all UA commands to ensure that we always prompt where possible.
196 - If ./btape is called without /dev, assume argument is a Storage resource name.
197 - Put memory utilization in Status output of each daemon
198 if full status requested or if some level of debug on.
199 - Make database type selectable by .conf files i.e. at runtime
200 - Set flag for uname -a. Add to Volume label.
201 - Implement throttled work queue.
202 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
203 - Restore files modified after date
204 - Restore file modified before date
205 - Emergency restore info:
207 - Backup working directory
209 - Restore -- do nothing but show what would happen
210 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
211 - Implement Restore FileSet=
212 - Create a protocol.h and protocol.c where all protocol messages
214 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
215 - Timout a job or terminate if link goes down, or reopen link and query.
216 - Find general solution for sscanf size problems (as well
217 as sprintf. Do at run time?
218 - Concept of precious tapes (cannot be reused).
219 - Make bcopy copy with a single tape drive.
220 - Permit changing ownership during restore.
222 - Autolabel should be specified by DIR instead of SD.
223 - Find out how to get the system tape block limits, e.g.:
224 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
225 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
228 - AutoScan (check checksum of tape)
229 - Format command = "format /dev/nst0"
233 - Seek resolution (usually corresponds to buffer size)
234 - EODErrorCode=ENOSPC or code
235 - Partial Read error code
236 - Partial write error code
237 - Nonformatted read error
238 - Nonformatted write error
239 - WriteProtected error
243 - IgnoreCloseErrors=yes
253 - FD sends unsaved file list to Director at end of job (see
255 - File daemon should build list of files skipped, and then
256 at end of save retry and report any errors.
257 - Write a Storage daemon that uses pipes and
258 standard Unix programs to write to the tape.
260 - Need something that monitors the JCR queue and
261 times out jobs by asking the deamons where they are.
262 - Enhance Jmsg code to permit buffering and saving to disk.
263 - device driver = "xxxx" for drives.
264 - restart: paranoid: read label fsf to
265 eom read append block, and go
266 super-paranoid: read label, read all files
267 in between, read append block, and go
268 verify: backspace, read append block, and go
269 permissive: same as above but frees drive
270 if tape is not valid.
272 - Ensure that /dev/null works
273 - Need report class for messages. Perhaps
274 report resource where report=group of messages
275 - enhance scan_attrib and rename scan_jobtype, and
276 fill in code for "since" option
277 - Director needs a time after which the report status is sent
278 anyway -- or better yet, a retry time for the job.
279 Don't reschedule a job if previous incarnation is still running.
280 - Some way to automatically backup everything is needed????
281 - Need a structure for pending actions:
283 - termination status (part of buffered msgs?)
285 Read, Write, Clean, Delete
286 - Login to Bacula; Bacula users with different permissions:
287 owner, group, user, quotas
288 - Store info on each file system type (probably in the job header on tape.
289 This could be the output of df; or perhaps some sort of /etc/mtab record.
292 - Design at hierarchial storage for Bacula. Migration and Clone.
293 - Implement FSM (File System Modules).
294 - Identify unchanged or "system" files and save them to a
295 special tape thus removing them from the standard
296 backup FileSet -- BASE backup.
297 - Heartbeat between daemons.
298 - Audit M_ error codes to ensure they are correct and consistent.
299 - Add variable break characters to lex analyzer.
300 Either a bit mask or a string of chars so that
301 the caller can change the break characters.
302 - Make a single T_BREAK to replace T_COMMA, etc.
303 - Ensure that File daemon and Storage daemon can
304 continue a save if the Director goes down (this
305 is NOT currently the case). Must detect socket error,
306 buffer messages for later.
307 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
308 - Add ability to backup to two Storage devices (two SD sessions) at
309 the same time -- e.g. onsite, offsite.
310 - Add the ability to consolidate old backup sets (basically do a restore
311 to tape and appropriately update the catalog). Compress Volume sets.
312 Might need to spool via file is only one drive is available.
313 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
314 someway to do so with every volume that has less than x% valid
319 Bacula Projects Roadmap
321 last update 5 January 2003
323 Item 1: Multiple simultaneous Jobs. (done)
324 Done -- Restore part needs better implementation to work correctly
326 What: Permit multiple simultaneous jobs in Bacula.
328 Why: An enterprise level solution needs to go fast without the
329 need for the system administrator to carefully tweak
330 timing. Based on the benchmarks, during a full
331 backup, NetWorker typically hit 10 times the bandwidth to
332 the tape compared to Bacula--largely. This is probably due to
333 running parallel jobs and multi-threaded filling of buffers
334 and writing them to tape. This should also make things work
335 better when you have a mix of fast and slow machines backing
338 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
339 implementing this is a matter of some small cleanups and
343 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
344 Deferred -- not necessary yet.
346 What: If data is coming into the SD too fast, buffer it to
347 disk if the user has configured this option.
349 Why: This would be nice, especially if it more or less falls out
350 when implementing (1) above. If not, it probably should not
351 be given a high priority because fundamentally the backup time
352 is limited by the tape bandwidth. Even though you may finish a
353 client job quicker by spilling to disk, you still have to
354 eventually get it onto tape. If intermediate disk buffering
355 allows us to improve write bandwidth to tape, it may make
358 Notes: Whether or not this is implemented will depend upon performance
359 testing after item 1 is implemented.
362 Item 3: Write the bscan program -- also write a bcopy program.
365 What: Write a program that reads a Bacula tape and puts all the
366 appropriate data into the catalog. This allows recovery
367 from a tape that is no longer in the database, or it allows
368 re-creation of a database if lost.
370 Why: This is a fundamental robustness and disaster recovery tool
371 which will increase the comfort level of a sysadmin
372 considering adopting Bacula.
374 Notes: A skeleton of this program already exists, but much work
375 needs to be done. Implementing this will also make apparent
376 any deficiencies in the current Bacula tape format.
379 Item 4: Implement Base jobs.
381 What: A base job is sort of like a Full save except that you
382 will want the FileSet to contain only files that are unlikely
383 to change in the future (i.e. a snapshot of most of your
384 system after installing it). After the base job has been run,
385 when you are doing a Full save, you can specify to exclude
386 all files saved by the base job that have not been modified.
388 Why: This is something none of the competition does, as far as we know
389 (except BackupPC, which is a Perl program that saves to disk
390 only). It is big win for the user, it makes Bacula stand out
391 as offering a unique optimization that immediately saves time
394 Notes: Big savings in tape usage. Will require more resources because
395 the e. DIR must send FD a list of files/attribs, and the FD must
396 search the list and compare it for each file to be saved.
399 Item 5: Implement Label templates
401 What: This is a mechanism whereby Bacula can automatically create
402 a tape label for new tapes according to a detailed specification
403 provided by the user.
405 Why: It is a major convenience item for folks who use automated label
408 Notes: Bacula already has a working form of automatic tape label
409 creation, but it is very crude. The design for the complete
410 tape labeling project is already documented in the manual.
413 Item 6: Write a regression script.
414 Done -- Continue to expand its testing.
416 What: This is an automatic script that runs and tests as many features
417 of Bacula as possible. The output is compared to previous
418 versions of Bacula and any differences are reported.
420 Why: This is an enormous help in preventing introduction of new
421 errors in parts of the program that already work correctly.
423 Notes: This probably should be ranked higher, it's something the typical
424 user doesn't see. Depending on how it's implemented, it may
425 make sense to defer it until the archival tape format and
426 user interface mature.
429 Item 7: GUI for interactive restore
430 Item 8: GUI for interactive backup
432 What: The current interactive restore is implemented with a tty
433 interface. It would be much nicer to be able to "see" the
434 list of files backed up in typical GUI tree format.
435 The same mechanism could also be used for creating
436 ad-hoc backup FileSets (item 8).
438 Why: Ease of use -- especially for the end user.
440 Notes: Rather than implementing in Gtk, we probably should go directly
441 for a Browser implementation, even if doing so meant the
442 capability wouldn't be available until much later. Not only
443 is there the question of Windows sites, most
444 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
445 without installing lots of stuff admins are very wary about.
446 Real sysadmins will always use the command line anyway, and
447 the user who's doing an interactive restore or backup of his
448 own files will in most cases be on a Windows machine running
452 Item 9: Add SSL to daemon communications.
454 What: This provides for secure communications between the daemons.
456 Why: This would allow doing backup across the Internet without
457 privacy concerns (or with much less concern).
459 Notes: The vast majority of near term potential users will be backing up
460 a single site over a LAN and, correctly or not, they probably
461 won't be concerned with security, at least not enough to go to
462 the trouble to set up keys, etc. to screw things down. We suspect
463 that many users genuinely interested in multi-site backup
464 already run some form of VPN software in their internetwork
465 connections, and are willing to delegate security to that layer.
468 Item 10: Define definitive tape format.
471 What: Define that definitive tape format that will not change
472 for the next millennium.
474 Why: Stability, security.
476 Notes: See notes for item 11 below.
479 Item 11: New daemon communication protocol.
481 What: The current daemon to daemon protocol is basically an ASCII
482 printf() and sending the buffer. On the receiving end, the
483 buffer is sscanf()ed to unpack it. The new scheme would
484 be a binary format that allows quick packing and unpacking
485 of any data type with named fields.
487 Why: Using binary packing would be faster. Named fields will permit
488 error checking to ensure that what is sent is what the
489 receiver really wants.
491 Notes: These are internal improvements in the interest of the
492 long-term stability and evolution of the program. On the one
493 hand, the sooner they're done, the less code we have to rip
494 up when the time comes to install them. On the other hand, they
495 don't bring an immediately perceptible benefit to potential
496 users. Item 10 and possibly item 11 should be deferred until Bacula
497 is well established with a growing user community more or
498 less happy with the feature set. At that time, it will make a
499 good "next generation" upgrade in the interest of data
505 ======================================================
507 It is somewhat like a Full save becomes an incremental since
508 the Base job (or jobs) plus other non-base files.
510 - New BaseFile table that contains:
511 JobId, BaseJobId, FileId (from Base).
512 i.e. for each base file that exists but is not saved because
513 it has not changed, the File daemon sends the JobId, BaseId,
514 and FileId back to the Director who creates the DB entry.
515 - To initiate a Base save, the Director sends the FD
516 the FileId, and full filename for each file in the Base.
517 - When the FD finds a Base file, he requests the Director to
518 send him the full File entry (stat packet plus MD5), or
519 conversely, the FD sends it to the Director and the Director
520 says yes or no. This can be quite rapid if the FileId is kept
521 by the FD for each Base Filename.
522 - It is probably better to have the comparison done by the FD
523 despite the fact that the File entry must be sent across the
525 - An alternative would be to send the FD the whole File entry
526 from the start. The disadvantage is that it requires a lot of
527 space. The advantage is that it requires less communications
529 - The Job record must be updated to indicate that one or more
531 - At end of Job, FD returns:
532 1. Count of base files/bytes not written to tape (i.e. matches)
533 2. Count of base file that were saved i.e. had changed.
534 - No tape record would be written for a Base file that matches, in the
535 same way that no tape record is written for Incremental jobs where
536 the file is not saved because it is unchanged.
537 - On a restore, all the Base file records must explicitly be
538 found from the BaseFile tape. I.e. for each Full save that is marked
539 to have one or more Base Jobs, search the BaseFile for all occurrences
541 - An optimization might be to make the BaseFile have:
547 This would avoid the need to explicitly fetch each File record for
548 the Base job. The Base Job record will be fetched to get the
549 VolSessionId and VolSessionTime.
550 =========================================================
552 ==========================================================
554 For each Incremental job that is run, there may be files that
555 were found but not saved because they were locked (this applies
556 only to Windows). Such a system could send back to the Director
557 a list of Unsaved files.
559 - New UnSavedFiles table that contains:
563 - Then in the next Incremental job, the list of Unsaved Files will be
564 feed to the FD, who will ensure that they are explicitly chosen even
565 if standard date/time check would not have selected them.
566 =============================================================
570 =============================================================
572 Request For Comments For File Backup Options
575 Subject: File Backup Options
578 A few days ago, a Bacula user who is backing up to file volumes and
579 using compression asked if it was possible to suppress compressing
580 all .gz files since it was a waste of CPU time. Although Bacula
581 currently permits using different options (compression, ...) on
582 a directory by directory basis, it cannot do it on a file by
583 file basis, which is clearly what was desired.
585 Proposed Implementation:
586 To solve this problem, I propose the following:
588 - Add a new Director resource type called FileOptions.
590 - The FileOptions resource will have records for all
591 options that can currently be specified on the Include record
592 (in a FileSet). Examples below.
594 - The FileOptions resource will permit an exclude option as well
595 as a number of additional options.
597 - The heart of the FileOptions resource is the ability to
598 supply any number of ApplyTo records which specify POSIX
599 regular expressions. These ApplyTo regular expressions are
600 applied to the fully qualified filename (path and all). If
601 one matches, then the FileOptions will be used.
603 - When an ApplyTo specification matches an included file, the
604 options specified in the FileOptions resource will override
605 the default options specified on the Include record.
607 - Include records will be modified to permit referencing one or
608 more FileOptions resources. The FileOptions will be used
609 in the order listed on the Include record and the first
610 one that matches will be applied.
612 - Options (or specifications) currently supplied on the Include
613 record will be deprecated (i.e. removed in a later version a
614 year or so from now).
616 - The Exclude record will be deprecated as the same functionality
617 can be obtained by using an Exclude = yes in the FileOptions.
620 The following records can appear in the FileOptions resource. An
621 asterisk preceding the name indicates a feature not currently
625 - Compression= (GZIP, ...)
626 - Signature= (MD5, SHA1, ...)
628 - OneFs= (yes/no) - remain on one filesystem
629 - Recurse= (yes/no) - recurse into subdirectories
630 - Sparse= (yes/no) - do sparse file backup
631 - *Exclude= (yes/no) - exclude file from being saved
632 - *Reader= (filename) - external read (backup) program
633 - *Plugin= (filename) - read/write plugin module
636 - verify= (ipnougsamc5) - verify options
639 - replace= (always/ifnewer/ifolder/never) - replace options currently
641 - *Writer= (filename) - external write (restore) program
645 Currently options specifying compression, MD5 signatures, recursion,
646 ... of a FileSet are supplied on the Include record. These will now
647 all be collected into a FileOptions resource, which will be
648 specified on the Include in place of the options. Multiple FileOptions
649 may be specified. Since the FileOptions contain regular expressions
650 that are applied to the full filename, this will give the ability
651 to specify backup options on a file by file basis to whatever level
660 Include = compression=GZIP signature=MD5 {
669 Include = FileOptions=Opts {
680 That's a lot more to do the same thing, but it gives the ability to
681 apply options on a file by file basis. For example, suppose you
682 want to compress all files but not any file with extensions .gz or .Z.
683 You could do so as follows:
687 Include = FileOptions=NoCompress FileOptions=Opts {
695 ApplyTo = /*.?*/ # matches all files
700 # Note multiple ApplyTos are ORed
701 ApplyTo = /*.gz/ # matches .gz files */
702 ApplyTo = /*.Z/ # matches .Z files */
705 Now, since the NoCompress FileOptions is specified first on the
706 Include line, any *.gz or *.Z file will have an MD5 signature computed,
707 but will not be compressed. For all other files, the NoCompress will not
708 match, so the Opts options will be used which will include GZIP
712 - Is it necessary to provide some means of ANDing regular expressions
713 and negation? (not currently planned)
715 e.g. ApplyTo = /*.gz/ && !/big.gz/
717 - I see that Networker has a "null" module which, if specified, does not
718 backup the file, but does make an record of the file in the catalog
719 so that the catalog will reflect an exact picture of the filesystem.
720 The result is that the file can be "seen" when "browsing" the save
721 sets, but it cannot be restored.
723 Is this really useful? Should it be implemented in Bacula?
726 After implementing the above, the user will be able to specify
727 on a file by file basis (using regular expressions) what options are
728 applied for the backup.
729 ====================================
731 =========================================
732 Proposal by Bill Sellers
734 Return-Path: <w.a.sellers@larc.nasa.gov>
735 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
736 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)
737 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
738 X-Sender: w.a.sellers@pop.larc.nasa.gov
739 X-Mailer: QUALCOMM Windows Eudora Version 5.1
740 Date: Tue, 14 Jan 2003 16:30:18 -0500
741 To: Kern Sibbald <kern@sibbald.com>
742 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
743 Subject: Re: [Bacula-users] Bacula remote storage?
744 In-Reply-To: <1042565382.1845.177.camel@rufus>
745 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>
747 Content-Type: text/plain; charset="us-ascii"; format=flowed
748 X-Annoyance-Filter-Junk-Probability: 0
749 X-Annoyance-Filter-Classification: Mail
750 At 06:29 PM 1/14/2003 +0100, you wrote:
753 >Well, if you cannot put a Bacula client on the machine,
754 >then it is a big problem. If you know of some software
755 >that can do what you want, let me know, because I
756 >really just don't know how to do it -- at least not
762 We have been able to get Amanda to use the HSM as a storage
763 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
764 handle Windows systems very well (or at all without Samba). So I am
765 looking for a backup system that has a Windows client. I really like the
766 Windows integration of Bacula.
768 From the command line, its rather trivial to move the data around. We use
771 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
773 or if you use GNU tar:
775 tar czf hsm:path/file.tgz ./files
777 One idea for you to consider; Sendmail offers pipes in the aliases file;
778 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
779 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
780 pipe available, as a storage device? Then we could use any command that
781 handles stdin as a storage destination.
788 #Password is not used in rsh, but might be used in ftp.
789 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
796 Password = "foobar&-"
797 Device = "| ncftpput -c hsm /path/file.bacula"
801 >If you have some local storage available, you could
802 >use Bacula to backup to disk volumes, then use some
803 >other software (ftp, scp) to move them to the HSM
804 >machine. However, this is a bit kludgy.
807 It is, but maybe worth a try. Is there some function in Bacula to put
808 variables in filenames? i.e. backup.2003-01-15.root
815 w.a.sellers@larc.nasa.gov
817 ==============================================
818 The Project for the above
820 I finally realized that this is not at all
821 the same as reader/writer programs or plugins,
822 which are alternate ways of accessing the
823 files to be backed up. Rather, it is an alternate
824 form of storage device, and I have always planned
825 that Bacula should be able to handle all sorts
828 So, I propose the following phases:
830 1. OK from you to invest some time in testing
831 this as I implement it (requires that you
832 know how to download from the SourceForge
833 cvs -- which I imagine is a piece of cake
836 2. Dumb implementation by allowing a device to
837 be a fifo for write only.
838 Reason: easy to implement, proof of concept.
840 3. Try reading from fifo but with fixed block
842 Reason: proof of concept, easy to implement.
844 4. Extend reading from fifo (restores) to handle
846 Reason: requires some delicate low level coding
847 which could destabilize all of Bacula.
849 5. Implementation of above but to a program. E.g.
850 Device = "|program" (not full pipeline).
851 Reason: routines already exist, and program can
852 be a shell script which contains anything.
854 6. Add full pipeline as a possibility. E.g.
855 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
856 Reason: needs additional coding to implement full
857 pipeline (must fire off either a shell or all
858 programs and connect their pipes).
860 There are a good number of details in each step
861 that I have left out, but I will specify them at
862 every stage, and there may be a few changes as things
863 evolve. I expect that to get to stage 5 will take a
864 few weeks, and at that point, you will have
865 everything you need (just inside a script).
866 Stage 6 will probably take longer, but if this
867 project pleases you, what we do for 5 should
868 be adequate for some time.
872 =============================================
874 Done: (see kernsdone for more)
875 - Look into Pruning/purging problems or why there seem to
876 be so many files listed each night.
877 - Fix cancel in find_one -- need jcr.
878 - Cancel does not work for restore in FD.
879 - Write SetJobStatus() function so cancel status not lost.
880 - Add include list to end of chain in findlib
881 - Zap sd_auth_key after use
882 - Add Bar code reading capabilities (new mtx-changer)
883 - Figure out some way to automatically backup all local partitions
884 - Make hash table for linked files in findlib/find_one.c:161
886 - Rewrite find_one.c to use only pool_memory instead of
887 alloca and malloc (probably not necessary).
888 - Make sure btraceback goes into /sbin not sysconf directory.
889 - InitVerify is getting pruned and it shouldn't (document it)
890 - Make 1.28c release ??? NO do 1.29 directly
891 - Set timeout on opening fifo for save or restore (findlib)
892 - Document FIFO storage device.
893 - Document fifo and | and <
896 - Get correct error status from run_program or open_bpipe().
897 - Restrict permissions on File Volumes (now 0640).
898 - Umasked 022 daemons
899 - Fix restore of hard linked file.
900 - Figure out how to allow multiple simultaneous file Volumes on a single device.
901 - Implement multiple simultaneous file Volumes on a single device.
902 - Cleanup db_update_media and db_update_pool
903 - Flush all the daemon messages at the end of every job.
904 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
905 run_program -- bpipe.c
906 - Apparently cancel does not work for jobs waiting to be
908 - Implement TCP/IP connection for MySQL
909 - Pull a canceled job from the Scheduling queue.
910 - Implement max_file_size in block.c (already done, just tweaked).
911 - Look at purge jobs volume (at least document it, and see if it is
913 - Add list volumes does all pools. list volumes pool=xxx now works.
914 - Add pool= to "list media" in ua_output.c
915 - Strip trailing slashes from Include directory names in the FD.
916 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
917 ERR=Operation not permited loop.
918 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
919 - Produce better error messages in when error/eof writing block.
920 - Cancelling of a queued job does NOT work!!!!!!
922 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
923 rufus-sd: Recycled volume File0003 on device /home/kern/bacula/working, all previous data lost.
924 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
925 - Ability to backup to a file then later transfer to a tape -- Migration.
926 Migration based on MaxJobs(MinJobs),MaxVols(MinVols),AgeJobs,MaxBytes(MinBytes)
927 (i.e. HighwaterMark, LowwaterMark).
928 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
929 set of volumes recycling the oldest volume when it is needed.
930 - gethostbyname failure in bnet_connect() continues
931 generating errors -- should stop.
932 - Add chflags() code for FreeBSD file flags
933 - Bevan Anderson suggests having a run queue for each device
934 so that multiple simultaneous jobs can run but each writing
935 to a different Volume.
936 - Look at handling <> in smtp doesn't work with exim.
937 - Need to specify MaximumConcurrentJobs in the Job resource.
938 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
939 - Implement finer multiprocessing options.
940 - Implement | and < in Exclude statements.
941 - Figure out some way to specify a retention period for files
942 that no longer exist on the machine -- so that we maintain
943 say backups for 30 days, but if the file is deleted, we maintain
944 the last copy for 1 year. -- answer Volume retention.
945 - Make non-zero status from RunJobBefore/After error the job.
946 - Need define of int_least16_t in sha1.h for SuSE.
947 - Implement bar code reader for autochangers
948 - Document new MaximumConcurrentJob records (Job, Client, Storage)
949 - Write up how to use/manage disk Volume Storage. ******
950 - Remove kern and kelvin from mysql_grant...
951 - Install grant_mysql...
952 - Strip trailing / from Include
953 - add #define ENABLE_NLS for Gnome compile on SuSE.
954 - Add Client FS/OS id (Linux, Win95/98, ...).
955 - Concept of VolumeSet during restore which is a list
956 of Volume names needed.
957 - Turn virutally all sprintfs into snprintfs.
958 - Update volume=Test01 requests pool, then lists volumes.
959 **** Test select_pool_and_media ...
961 - Add IP address to authentication failures.
962 - Add a default File storage so that new users can do backup
963 and restores right away.
964 - Forbid sbindir and with-subsys-dir from being the same (otherwise
965 the binary gets deleted when the daemon is stopped in the
966 rc.d/inid.d directory.
967 - Do not ignore SIGCHLD
968 - Add Cleaning to list of volume statuses
969 - Implement run at "xxx"
970 - Document new transparent Console commands and wait command.
971 - Document . and @ commands
973 - Document Lutz Kittler's trick of using "Run Before Job" to
974 abort a job on a particular day.
975 - Document Ludovic Strappazon's Win32 raw device save/restore.
976 - Document not to restore .journal .autofsck
977 - Document labeling a whole magazine using "cat"