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 - Add a section to the doc on Manual cycling of Volumes.
13 Testing to do: (painful)
14 - that ALL console command line options work and are always implemented
15 - blocksize recognition code.
16 - multiple simultaneous Volumes
19 - Examine Bare Metal restore problem.
20 - Test multiple simultaneous Volumes
21 - Document FInclude ...
23 - Figure out how to use ssh or stunnel to protect Bacula communications.
26 - Fix command prompt in gnome-console by checking on Ready.
27 - Implement HEART_BEAT while SD waiting for tapes.
28 - Include RunBeforeJob and RunAfterJob output in the message
30 - Check if Job/File retentions apply to multivolume jobs.
31 - Change M_INFO to M_RESTORED for all restored files.
32 - Remove subsysdir from conf files (used only in autostart scripts).
33 - Implement console @echo command.
34 - Implement global with DB name and add to btraceback.gdb
35 - Bug: fix access problems on files restored on WinXP.
36 - Implement a Slot priority (loaded/not loaded).
37 - Implement "vacation" Incremental only saves.
38 - Implement single pane restore (much like the Gftp panes).
39 - Implement Automatic Mount even in operator wait.
40 - Implement create "FileSet"?
41 - Implement Release Device in the Job resource to unmount a drive.
42 - Implement Acquire Device in the Job resource to mount a drive,
43 be sure this works with admin jobs so that the user can get
44 prompted to insert the correct tape. Possibly some way to say to
45 run the job but don't save the files.
46 - Implement all command line args on run.
47 - Implement command line "restore" args.
48 - Implement "restore current select=no"
49 - Fix watchdog pthread crash on Win32
50 - Fix "access not allowed" for backup of files on WinXP.
51 - Implement "scratch pool" where tapes are defined and can be
52 taken by any pool that needs them.
53 - Implement restore "current system", but take all files without
54 doing selection tree -- so that jobs without File records can
56 - Make | and < work on FD side.
57 - Pass prefix_links to FD.
58 - Implement a M_SECURITY message class.
59 - Implement disk spooling. Two parts: 1. Spool to disk then
60 immediately to tape to speed up tape operations. 2. Spool to
61 disk only when the tape is full, then when a tape is hung move
63 - From Phil Stracchino:
64 It would probably be a per-client option, and would be called
65 something like, say, "Automatically purge obsoleted jobs". What it
66 would do is, when you successfully complete a Differential backup of a
67 client, it would automatically purge all Incremental backups for that
68 client that are rendered redundant by that Differential. Likewise,
69 when a Full backup on a client completed, it would automatically purge
70 all Differential and Incremental jobs obsoleted by that Full backup.
71 This would let people minimize the number of tapes they're keeping on
72 hand without having to master the art of retention times.
74 - Allow multiple Storage specifications (or multiple names on
75 a single Storage specification) in the Job record. Thus a job
76 can be backed up to a number of storage devices.
77 - Implement dump label to UA
78 - Add prefixlinks to where or not where absolute links to FD.
79 - Look at Python for a Bacula scripting language -- www.python.org
80 - Issue message to mount a new tape before the rewind.
81 - Simplified client job initiation for portables.
82 - If SD cannot open a drive, make it periodically retry.
83 - Implement LabelTemplate (at least first cut).
84 - Add more of the config info to the tape label.
85 - Implement a Mount Command and an Unmount Command where
86 the users could specify a system command to be performed
87 to do the mount, after which Bacula could attempt to
88 read the device. This is for Removeable media such as a CDROM.
89 - Most likely, this mount command would be invoked explicitly
90 by the user using the current Console "mount" and "unmount"
91 commands -- the Storage Daemon would do the right thing
92 depending on the exact nature of the device.
93 - As with tape drives, when Bacula wanted a new removable
94 disk mounted, it would unmount the old one, and send a message
95 to the user, who would then use "mount" as described above
96 once he had actually inserted the disk.
98 - Make some way so that if a machine is skipped because it is not up
99 that Bacula will continue retrying for a specified period of time --
101 - If tape is marked read-only, then try opening it read-only rather than
102 failing, and remember that it cannot be written.
103 - Refine SD waiting output:
104 Device is being positioned
105 > Device is being positioned for append
106 > Device is being positioned to file x
108 - Figure out some way to estimate output size and to avoid splitting
109 a backup across two Volumes -- this could be useful for writing CDROMs
110 where you really prefer not to have it split -- not serious.
111 - Add RunBeforeJob and RunAfterJob to the Client program.
112 - Have SD compute MD5 or SHA1 and compare to what FD computes.
113 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
114 actual data on the Volume and compare it.
115 - Implement FileOptions (see end of this document)
116 - Implement Bacula plugins -- design API
117 - Make bcopy read through bad tape records.
118 - Fix read_record to handle multiple sessions.
119 - Program files (i.e. execute a program to read/write files).
120 Pass read date of last backup, size of file last time.
121 - Add Signature type to File DB record.
122 - Make Restore report an error if FD or SD term codes are not OK.
123 - CD into subdirectory when open()ing files for backup to
124 speed up things. Test with testfind().
125 - Priority job to go to top of list.
126 - Find out why Full saves run slower and slower (hashing?)
127 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
128 - Implement some way for the Console to dynamically create a job.
129 - Restore to a particular time -- e.g. before date, after date.
130 - Solaris -I on tar for include list
131 - Prohibit backing up archive device (findlib/find_one.c:128)
132 - Need a verbose mode in restore, perhaps to bsr.
133 - bscan without -v is too quiet -- perhaps show jobs.
134 - Add code to reject whole blocks if not wanted on restore.
135 - Start working on Base jobs.
136 - Check if we can increase Bacula FD priorty in Win2000
137 - Make sure the MaxVolFiles is fully implemented in SD
138 - Check if both CatalogFiles and UseCatalog are set to SD.
139 - Need return status on read_cb() from read_records(). Need multiple
140 records -- one per Job, maybe a JCR or some other structure with
141 a block and a record.
142 - Figure out how to do a bare metal Windows restore
143 - Put system type returned by FD into catalog.
144 - Possibly add email to Watchdog if drive is unmounted too
145 long and a job is waiting on the drive.
146 - Use read_record.c in SD code.
147 - Why don't we get an error message from Win32 FD when bootstrap
148 file cannot be created for restore command?
149 - When Marking a file in Restore that is a hard link, also
150 mark the link so that the data will be reloaded.
151 - Restore program that errors in SD due to no tape reports
152 OK incorrectly in output.
153 - After unmount, if restore job started, ask to mount.
154 - Convert all %x substitution variables, which are hard to remember
155 and read to %(variable-name). Idea from TMDA.
156 - Add JobLevel in FD status (but make sure it is defined).
157 - Make Pool resource handle Counter resources.
158 - Remove NextId for SQLite. Optimize.
159 - Move all SQL statements into a single location.
160 - Add UA rc and history files.
161 - put termcap (used by console) in ./configure and
162 allow -with-termcap-dir.
163 - Enhance time and size scanning routines.
164 - Fix Autoprune for Volumes to respect need for full save.
165 - Fix Win32 config file definition name on /install
166 - No READLINE_SRC if found in alternate directory.
167 - Test a second language e.g. french.
168 - Compare tape to Client files (attributes, or attributes and data)
169 - Make all database Ids 64 bit.
170 - Write an applet for Linux.
171 - Add estimate to Console commands
172 - Find solution to blank filename (i.e. path only) problem.
173 - Implement new daemon communications protocol.
174 - Remove PoolId from Job table, it exists in Media.
175 - Allow console commands to detach or run in background.
176 - Fix status delay on storage daemon during rewind.
177 - Add SD message variables to control operator wait time
178 - Maximum Operator Wait
179 - Minimum Message Interval
180 - Maximum Message Interval
181 - Send Operator message when cannot read tape label.
182 - Verify level=Volume (scan only), level=Data (compare of data to file).
183 Verify level=Catalog, level=InitCatalog
185 - Add keyword search to show command in Console.
186 - Fix Win2000 error with no messages during startup.
187 - Events : tape has more than xxx bytes.
188 - Restrict characters permitted in a Resource name.
189 - Complete code in Bacula Resources -- this will permit
190 reading a new config file at any time.
191 - Handle ctl-c in Console
192 - Implement script driven addition of File daemon to config files.
193 - Think about how to make Bacula work better with File (non-tape) archives.
194 - Write Unix emulator for Windows.
195 - Implement new serialize subroutines
196 send(socket, "string", &Vol, "uint32", &i, NULL)
197 - Audit all UA commands to ensure that we always prompt where possible.
198 - If ./btape is called without /dev, assume argument is a Storage resource name.
199 - Put memory utilization in Status output of each daemon
200 if full status requested or if some level of debug on.
201 - Make database type selectable by .conf files i.e. at runtime
202 - Set flag for uname -a. Add to Volume label.
203 - Implement throttled work queue.
204 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
205 - Restore files modified after date
206 - Restore file modified before date
207 - Emergency restore info:
209 - Backup working directory
211 - Restore -- do nothing but show what would happen
212 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
213 - Implement Restore FileSet=
214 - Create a protocol.h and protocol.c where all protocol messages
216 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
217 - Timout a job or terminate if link goes down, or reopen link and query.
218 - Find general solution for sscanf size problems (as well
219 as sprintf. Do at run time?
220 - Concept of precious tapes (cannot be reused).
221 - Make bcopy copy with a single tape drive.
222 - Permit changing ownership during restore.
224 - Autolabel should be specified by DIR instead of SD.
225 - Find out how to get the system tape block limits, e.g.:
226 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
227 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
230 - AutoScan (check checksum of tape)
231 - Format command = "format /dev/nst0"
235 - Seek resolution (usually corresponds to buffer size)
236 - EODErrorCode=ENOSPC or code
237 - Partial Read error code
238 - Partial write error code
239 - Nonformatted read error
240 - Nonformatted write error
241 - WriteProtected error
245 - IgnoreCloseErrors=yes
255 - FD sends unsaved file list to Director at end of job (see
257 - File daemon should build list of files skipped, and then
258 at end of save retry and report any errors.
259 - Write a Storage daemon that uses pipes and
260 standard Unix programs to write to the tape.
262 - Need something that monitors the JCR queue and
263 times out jobs by asking the deamons where they are.
264 - Enhance Jmsg code to permit buffering and saving to disk.
265 - device driver = "xxxx" for drives.
266 - restart: paranoid: read label fsf to
267 eom read append block, and go
268 super-paranoid: read label, read all files
269 in between, read append block, and go
270 verify: backspace, read append block, and go
271 permissive: same as above but frees drive
272 if tape is not valid.
274 - Ensure that /dev/null works
275 - Need report class for messages. Perhaps
276 report resource where report=group of messages
277 - enhance scan_attrib and rename scan_jobtype, and
278 fill in code for "since" option
279 - Director needs a time after which the report status is sent
280 anyway -- or better yet, a retry time for the job.
281 Don't reschedule a job if previous incarnation is still running.
282 - Some way to automatically backup everything is needed????
283 - Need a structure for pending actions:
285 - termination status (part of buffered msgs?)
287 Read, Write, Clean, Delete
288 - Login to Bacula; Bacula users with different permissions:
289 owner, group, user, quotas
290 - Store info on each file system type (probably in the job header on tape.
291 This could be the output of df; or perhaps some sort of /etc/mtab record.
294 - Design at hierarchial storage for Bacula. Migration and Clone.
295 - Implement FSM (File System Modules).
296 - Identify unchanged or "system" files and save them to a
297 special tape thus removing them from the standard
298 backup FileSet -- BASE backup.
299 - Heartbeat between daemons.
300 - Audit M_ error codes to ensure they are correct and consistent.
301 - Add variable break characters to lex analyzer.
302 Either a bit mask or a string of chars so that
303 the caller can change the break characters.
304 - Make a single T_BREAK to replace T_COMMA, etc.
305 - Ensure that File daemon and Storage daemon can
306 continue a save if the Director goes down (this
307 is NOT currently the case). Must detect socket error,
308 buffer messages for later.
309 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
310 - Add ability to backup to two Storage devices (two SD sessions) at
311 the same time -- e.g. onsite, offsite.
312 - Add the ability to consolidate old backup sets (basically do a restore
313 to tape and appropriately update the catalog). Compress Volume sets.
314 Might need to spool via file is only one drive is available.
315 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
316 someway to do so with every volume that has less than x% valid
321 Bacula Projects Roadmap
323 last update 5 January 2003
325 Item 1: Multiple simultaneous Jobs. (done)
326 Done -- Restore part needs better implementation to work correctly
328 What: Permit multiple simultaneous jobs in Bacula.
330 Why: An enterprise level solution needs to go fast without the
331 need for the system administrator to carefully tweak
332 timing. Based on the benchmarks, during a full
333 backup, NetWorker typically hit 10 times the bandwidth to
334 the tape compared to Bacula--largely. This is probably due to
335 running parallel jobs and multi-threaded filling of buffers
336 and writing them to tape. This should also make things work
337 better when you have a mix of fast and slow machines backing
340 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
341 implementing this is a matter of some small cleanups and
345 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
346 Deferred -- not necessary yet.
348 What: If data is coming into the SD too fast, buffer it to
349 disk if the user has configured this option.
351 Why: This would be nice, especially if it more or less falls out
352 when implementing (1) above. If not, it probably should not
353 be given a high priority because fundamentally the backup time
354 is limited by the tape bandwidth. Even though you may finish a
355 client job quicker by spilling to disk, you still have to
356 eventually get it onto tape. If intermediate disk buffering
357 allows us to improve write bandwidth to tape, it may make
360 Notes: Whether or not this is implemented will depend upon performance
361 testing after item 1 is implemented.
364 Item 3: Write the bscan program -- also write a bcopy program.
367 What: Write a program that reads a Bacula tape and puts all the
368 appropriate data into the catalog. This allows recovery
369 from a tape that is no longer in the database, or it allows
370 re-creation of a database if lost.
372 Why: This is a fundamental robustness and disaster recovery tool
373 which will increase the comfort level of a sysadmin
374 considering adopting Bacula.
376 Notes: A skeleton of this program already exists, but much work
377 needs to be done. Implementing this will also make apparent
378 any deficiencies in the current Bacula tape format.
381 Item 4: Implement Base jobs.
383 What: A base job is sort of like a Full save except that you
384 will want the FileSet to contain only files that are unlikely
385 to change in the future (i.e. a snapshot of most of your
386 system after installing it). After the base job has been run,
387 when you are doing a Full save, you can specify to exclude
388 all files saved by the base job that have not been modified.
390 Why: This is something none of the competition does, as far as we know
391 (except BackupPC, which is a Perl program that saves to disk
392 only). It is big win for the user, it makes Bacula stand out
393 as offering a unique optimization that immediately saves time
396 Notes: Big savings in tape usage. Will require more resources because
397 the e. DIR must send FD a list of files/attribs, and the FD must
398 search the list and compare it for each file to be saved.
401 Item 5: Implement Label templates
403 What: This is a mechanism whereby Bacula can automatically create
404 a tape label for new tapes according to a detailed specification
405 provided by the user.
407 Why: It is a major convenience item for folks who use automated label
410 Notes: Bacula already has a working form of automatic tape label
411 creation, but it is very crude. The design for the complete
412 tape labeling project is already documented in the manual.
415 Item 6: Write a regression script.
416 Done -- Continue to expand its testing.
418 What: This is an automatic script that runs and tests as many features
419 of Bacula as possible. The output is compared to previous
420 versions of Bacula and any differences are reported.
422 Why: This is an enormous help in preventing introduction of new
423 errors in parts of the program that already work correctly.
425 Notes: This probably should be ranked higher, it's something the typical
426 user doesn't see. Depending on how it's implemented, it may
427 make sense to defer it until the archival tape format and
428 user interface mature.
431 Item 7: GUI for interactive restore
432 Item 8: GUI for interactive backup
434 What: The current interactive restore is implemented with a tty
435 interface. It would be much nicer to be able to "see" the
436 list of files backed up in typical GUI tree format.
437 The same mechanism could also be used for creating
438 ad-hoc backup FileSets (item 8).
440 Why: Ease of use -- especially for the end user.
442 Notes: Rather than implementing in Gtk, we probably should go directly
443 for a Browser implementation, even if doing so meant the
444 capability wouldn't be available until much later. Not only
445 is there the question of Windows sites, most
446 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
447 without installing lots of stuff admins are very wary about.
448 Real sysadmins will always use the command line anyway, and
449 the user who's doing an interactive restore or backup of his
450 own files will in most cases be on a Windows machine running
454 Item 9: Add SSL to daemon communications.
456 What: This provides for secure communications between the daemons.
458 Why: This would allow doing backup across the Internet without
459 privacy concerns (or with much less concern).
461 Notes: The vast majority of near term potential users will be backing up
462 a single site over a LAN and, correctly or not, they probably
463 won't be concerned with security, at least not enough to go to
464 the trouble to set up keys, etc. to screw things down. We suspect
465 that many users genuinely interested in multi-site backup
466 already run some form of VPN software in their internetwork
467 connections, and are willing to delegate security to that layer.
470 Item 10: Define definitive tape format.
473 What: Define that definitive tape format that will not change
474 for the next millennium.
476 Why: Stability, security.
478 Notes: See notes for item 11 below.
481 Item 11: New daemon communication protocol.
483 What: The current daemon to daemon protocol is basically an ASCII
484 printf() and sending the buffer. On the receiving end, the
485 buffer is sscanf()ed to unpack it. The new scheme would
486 be a binary format that allows quick packing and unpacking
487 of any data type with named fields.
489 Why: Using binary packing would be faster. Named fields will permit
490 error checking to ensure that what is sent is what the
491 receiver really wants.
493 Notes: These are internal improvements in the interest of the
494 long-term stability and evolution of the program. On the one
495 hand, the sooner they're done, the less code we have to rip
496 up when the time comes to install them. On the other hand, they
497 don't bring an immediately perceptible benefit to potential
498 users. Item 10 and possibly item 11 should be deferred until Bacula
499 is well established with a growing user community more or
500 less happy with the feature set. At that time, it will make a
501 good "next generation" upgrade in the interest of data
507 ======================================================
509 It is somewhat like a Full save becomes an incremental since
510 the Base job (or jobs) plus other non-base files.
512 - New BaseFile table that contains:
513 JobId, BaseJobId, FileId (from Base).
514 i.e. for each base file that exists but is not saved because
515 it has not changed, the File daemon sends the JobId, BaseId,
516 and FileId back to the Director who creates the DB entry.
517 - To initiate a Base save, the Director sends the FD
518 the FileId, and full filename for each file in the Base.
519 - When the FD finds a Base file, he requests the Director to
520 send him the full File entry (stat packet plus MD5), or
521 conversely, the FD sends it to the Director and the Director
522 says yes or no. This can be quite rapid if the FileId is kept
523 by the FD for each Base Filename.
524 - It is probably better to have the comparison done by the FD
525 despite the fact that the File entry must be sent across the
527 - An alternative would be to send the FD the whole File entry
528 from the start. The disadvantage is that it requires a lot of
529 space. The advantage is that it requires less communications
531 - The Job record must be updated to indicate that one or more
533 - At end of Job, FD returns:
534 1. Count of base files/bytes not written to tape (i.e. matches)
535 2. Count of base file that were saved i.e. had changed.
536 - No tape record would be written for a Base file that matches, in the
537 same way that no tape record is written for Incremental jobs where
538 the file is not saved because it is unchanged.
539 - On a restore, all the Base file records must explicitly be
540 found from the BaseFile tape. I.e. for each Full save that is marked
541 to have one or more Base Jobs, search the BaseFile for all occurrences
543 - An optimization might be to make the BaseFile have:
549 This would avoid the need to explicitly fetch each File record for
550 the Base job. The Base Job record will be fetched to get the
551 VolSessionId and VolSessionTime.
552 =========================================================
554 ==========================================================
556 For each Incremental job that is run, there may be files that
557 were found but not saved because they were locked (this applies
558 only to Windows). Such a system could send back to the Director
559 a list of Unsaved files.
561 - New UnSavedFiles table that contains:
565 - Then in the next Incremental job, the list of Unsaved Files will be
566 feed to the FD, who will ensure that they are explicitly chosen even
567 if standard date/time check would not have selected them.
568 =============================================================
572 =============================================================
574 Request For Comments For File Backup Options
577 Subject: File Backup Options
580 A few days ago, a Bacula user who is backing up to file volumes and
581 using compression asked if it was possible to suppress compressing
582 all .gz files since it was a waste of CPU time. Although Bacula
583 currently permits using different options (compression, ...) on
584 a directory by directory basis, it cannot do it on a file by
585 file basis, which is clearly what was desired.
587 Proposed Implementation:
588 To solve this problem, I propose the following:
590 - Add a new Director resource type called FileOptions.
592 - The FileOptions resource will have records for all
593 options that can currently be specified on the Include record
594 (in a FileSet). Examples below.
596 - The FileOptions resource will permit an exclude option as well
597 as a number of additional options.
599 - The heart of the FileOptions resource is the ability to
600 supply any number of ApplyTo records which specify POSIX
601 regular expressions. These ApplyTo regular expressions are
602 applied to the fully qualified filename (path and all). If
603 one matches, then the FileOptions will be used.
605 - When an ApplyTo specification matches an included file, the
606 options specified in the FileOptions resource will override
607 the default options specified on the Include record.
609 - Include records will be modified to permit referencing one or
610 more FileOptions resources. The FileOptions will be used
611 in the order listed on the Include record and the first
612 one that matches will be applied.
614 - Options (or specifications) currently supplied on the Include
615 record will be deprecated (i.e. removed in a later version a
616 year or so from now).
618 - The Exclude record will be deprecated as the same functionality
619 can be obtained by using an Exclude = yes in the FileOptions.
622 The following records can appear in the FileOptions resource. An
623 asterisk preceding the name indicates a feature not currently
627 - Compression= (GZIP, ...)
628 - Signature= (MD5, SHA1, ...)
630 - OneFs= (yes/no) - remain on one filesystem
631 - Recurse= (yes/no) - recurse into subdirectories
632 - Sparse= (yes/no) - do sparse file backup
633 - *Exclude= (yes/no) - exclude file from being saved
634 - *Reader= (filename) - external read (backup) program
635 - *Plugin= (filename) - read/write plugin module
638 - verify= (ipnougsamc5) - verify options
641 - replace= (always/ifnewer/ifolder/never) - replace options currently
643 - *Writer= (filename) - external write (restore) program
647 Currently options specifying compression, MD5 signatures, recursion,
648 ... of a FileSet are supplied on the Include record. These will now
649 all be collected into a FileOptions resource, which will be
650 specified on the Include in place of the options. Multiple FileOptions
651 may be specified. Since the FileOptions contain regular expressions
652 that are applied to the full filename, this will give the ability
653 to specify backup options on a file by file basis to whatever level
662 Include = compression=GZIP signature=MD5 {
671 Include = FileOptions=Opts {
682 That's a lot more to do the same thing, but it gives the ability to
683 apply options on a file by file basis. For example, suppose you
684 want to compress all files but not any file with extensions .gz or .Z.
685 You could do so as follows:
689 Include = FileOptions=NoCompress FileOptions=Opts {
697 ApplyTo = /*.?*/ # matches all files
702 # Note multiple ApplyTos are ORed
703 ApplyTo = /*.gz/ # matches .gz files */
704 ApplyTo = /*.Z/ # matches .Z files */
707 Now, since the NoCompress FileOptions is specified first on the
708 Include line, any *.gz or *.Z file will have an MD5 signature computed,
709 but will not be compressed. For all other files, the NoCompress will not
710 match, so the Opts options will be used which will include GZIP
714 - Is it necessary to provide some means of ANDing regular expressions
715 and negation? (not currently planned)
717 e.g. ApplyTo = /*.gz/ && !/big.gz/
719 - I see that Networker has a "null" module which, if specified, does not
720 backup the file, but does make an record of the file in the catalog
721 so that the catalog will reflect an exact picture of the filesystem.
722 The result is that the file can be "seen" when "browsing" the save
723 sets, but it cannot be restored.
725 Is this really useful? Should it be implemented in Bacula?
728 After implementing the above, the user will be able to specify
729 on a file by file basis (using regular expressions) what options are
730 applied for the backup.
731 ====================================
733 =========================================
734 Proposal by Bill Sellers
736 Return-Path: <w.a.sellers@larc.nasa.gov>
737 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
738 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)
739 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
740 X-Sender: w.a.sellers@pop.larc.nasa.gov
741 X-Mailer: QUALCOMM Windows Eudora Version 5.1
742 Date: Tue, 14 Jan 2003 16:30:18 -0500
743 To: Kern Sibbald <kern@sibbald.com>
744 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
745 Subject: Re: [Bacula-users] Bacula remote storage?
746 In-Reply-To: <1042565382.1845.177.camel@rufus>
747 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>
749 Content-Type: text/plain; charset="us-ascii"; format=flowed
750 X-Annoyance-Filter-Junk-Probability: 0
751 X-Annoyance-Filter-Classification: Mail
752 At 06:29 PM 1/14/2003 +0100, you wrote:
755 >Well, if you cannot put a Bacula client on the machine,
756 >then it is a big problem. If you know of some software
757 >that can do what you want, let me know, because I
758 >really just don't know how to do it -- at least not
764 We have been able to get Amanda to use the HSM as a storage
765 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
766 handle Windows systems very well (or at all without Samba). So I am
767 looking for a backup system that has a Windows client. I really like the
768 Windows integration of Bacula.
770 From the command line, its rather trivial to move the data around. We use
773 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
775 or if you use GNU tar:
777 tar czf hsm:path/file.tgz ./files
779 One idea for you to consider; Sendmail offers pipes in the aliases file;
780 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
781 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
782 pipe available, as a storage device? Then we could use any command that
783 handles stdin as a storage destination.
790 #Password is not used in rsh, but might be used in ftp.
791 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
798 Password = "foobar&-"
799 Device = "| ncftpput -c hsm /path/file.bacula"
803 >If you have some local storage available, you could
804 >use Bacula to backup to disk volumes, then use some
805 >other software (ftp, scp) to move them to the HSM
806 >machine. However, this is a bit kludgy.
809 It is, but maybe worth a try. Is there some function in Bacula to put
810 variables in filenames? i.e. backup.2003-01-15.root
817 w.a.sellers@larc.nasa.gov
819 ==============================================
820 The Project for the above
822 I finally realized that this is not at all
823 the same as reader/writer programs or plugins,
824 which are alternate ways of accessing the
825 files to be backed up. Rather, it is an alternate
826 form of storage device, and I have always planned
827 that Bacula should be able to handle all sorts
830 So, I propose the following phases:
832 1. OK from you to invest some time in testing
833 this as I implement it (requires that you
834 know how to download from the SourceForge
835 cvs -- which I imagine is a piece of cake
838 2. Dumb implementation by allowing a device to
839 be a fifo for write only.
840 Reason: easy to implement, proof of concept.
842 3. Try reading from fifo but with fixed block
844 Reason: proof of concept, easy to implement.
846 4. Extend reading from fifo (restores) to handle
848 Reason: requires some delicate low level coding
849 which could destabilize all of Bacula.
851 5. Implementation of above but to a program. E.g.
852 Device = "|program" (not full pipeline).
853 Reason: routines already exist, and program can
854 be a shell script which contains anything.
856 6. Add full pipeline as a possibility. E.g.
857 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
858 Reason: needs additional coding to implement full
859 pipeline (must fire off either a shell or all
860 programs and connect their pipes).
862 There are a good number of details in each step
863 that I have left out, but I will specify them at
864 every stage, and there may be a few changes as things
865 evolve. I expect that to get to stage 5 will take a
866 few weeks, and at that point, you will have
867 everything you need (just inside a script).
868 Stage 6 will probably take longer, but if this
869 project pleases you, what we do for 5 should
870 be adequate for some time.
874 =============================================
876 Done: (see kernsdone for more)
877 - Look into Pruning/purging problems or why there seem to
878 be so many files listed each night.
879 - Fix cancel in find_one -- need jcr.
880 - Cancel does not work for restore in FD.
881 - Write SetJobStatus() function so cancel status not lost.
882 - Add include list to end of chain in findlib
883 - Zap sd_auth_key after use
884 - Add Bar code reading capabilities (new mtx-changer)
885 - Figure out some way to automatically backup all local partitions
886 - Make hash table for linked files in findlib/find_one.c:161
888 - Rewrite find_one.c to use only pool_memory instead of
889 alloca and malloc (probably not necessary).
890 - Make sure btraceback goes into /sbin not sysconf directory.
891 - InitVerify is getting pruned and it shouldn't (document it)
892 - Make 1.28c release ??? NO do 1.29 directly
893 - Set timeout on opening fifo for save or restore (findlib)
894 - Document FIFO storage device.
895 - Document fifo and | and <
898 - Get correct error status from run_program or open_bpipe().
899 - Restrict permissions on File Volumes (now 0640).
900 - Umasked 022 daemons
901 - Fix restore of hard linked file.
902 - Figure out how to allow multiple simultaneous file Volumes on a single device.
903 - Implement multiple simultaneous file Volumes on a single device.
904 - Cleanup db_update_media and db_update_pool
905 - Flush all the daemon messages at the end of every job.
906 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
907 run_program -- bpipe.c
908 - Apparently cancel does not work for jobs waiting to be
910 - Implement TCP/IP connection for MySQL
911 - Pull a canceled job from the Scheduling queue.
912 - Implement max_file_size in block.c (already done, just tweaked).
913 - Look at purge jobs volume (at least document it, and see if it is
915 - Add list volumes does all pools. list volumes pool=xxx now works.
916 - Add pool= to "list media" in ua_output.c
917 - Strip trailing slashes from Include directory names in the FD.
918 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
919 ERR=Operation not permited loop.
920 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
921 - Produce better error messages in when error/eof writing block.
922 - Cancelling of a queued job does NOT work!!!!!!
924 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
925 rufus-sd: Recycled volume File0003 on device /home/kern/bacula/working, all previous data lost.
926 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
927 - Ability to backup to a file then later transfer to a tape -- Migration.
928 Migration based on MaxJobs(MinJobs),MaxVols(MinVols),AgeJobs,MaxBytes(MinBytes)
929 (i.e. HighwaterMark, LowwaterMark).
930 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
931 set of volumes recycling the oldest volume when it is needed.
932 - gethostbyname failure in bnet_connect() continues
933 generating errors -- should stop.
934 - Add chflags() code for FreeBSD file flags
935 - Bevan Anderson suggests having a run queue for each device
936 so that multiple simultaneous jobs can run but each writing
937 to a different Volume.
938 - Look at handling <> in smtp doesn't work with exim.
939 - Need to specify MaximumConcurrentJobs in the Job resource.
940 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
941 - Implement finer multiprocessing options.
942 - Implement | and < in Exclude statements.
943 - Figure out some way to specify a retention period for files
944 that no longer exist on the machine -- so that we maintain
945 say backups for 30 days, but if the file is deleted, we maintain
946 the last copy for 1 year. -- answer Volume retention.
947 - Make non-zero status from RunJobBefore/After error the job.
948 - Need define of int_least16_t in sha1.h for SuSE.
949 - Implement bar code reader for autochangers
950 - Document new MaximumConcurrentJob records (Job, Client, Storage)
951 - Write up how to use/manage disk Volume Storage. ******
952 - Remove kern and kelvin from mysql_grant...
953 - Install grant_mysql...
954 - Strip trailing / from Include
955 - add #define ENABLE_NLS for Gnome compile on SuSE.
956 - Add Client FS/OS id (Linux, Win95/98, ...).
957 - Concept of VolumeSet during restore which is a list
958 of Volume names needed.
959 - Turn virutally all sprintfs into snprintfs.
960 - Update volume=Test01 requests pool, then lists volumes.
961 **** Test select_pool_and_media ...
963 - Add IP address to authentication failures.
964 - Add a default File storage so that new users can do backup
965 and restores right away.
966 - Forbid sbindir and with-subsys-dir from being the same (otherwise
967 the binary gets deleted when the daemon is stopped in the
968 rc.d/inid.d directory.
969 - Do not ignore SIGCHLD
970 - Add Cleaning to list of volume statuses
971 - Implement run at "xxx"
972 - Document new transparent Console commands and wait command.
973 - Document . and @ commands
975 - Document Lutz Kittler's trick of using "Run Before Job" to
976 abort a job on a particular day.
977 - Document Ludovic Strappazon's Win32 raw device save/restore.
978 - Document not to restore .journal .autofsck
979 - Document labeling a whole magazine using "cat"
980 - Document how to automatically backup all local partitions
982 - Document OPTIMIZE TABLE in MySQL
983 - Document new immediate File save configuration (walk user
984 through first save to file Volume with automatic Volume labeling?).
985 - Implement scheduling of one time "run" jobs (i.e. instead of
986 starting immediately start at some specified time).
987 - Bug: up arrow prints garbage in command line on gnome-console!