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 - Change M_INFO to M_RESTORED for all restored files.
27 - Remove subsysdir from conf files (used only in autostart scripts).
28 - Implement console @echo command.
29 - Implement global with DB name and add to btraceback.gdb
30 - Bug: fix access problems on files restored on WinXP.
31 - Implement a Slot priority (loaded/not loaded).
32 - Implement "vacation" Incremental only saves.
33 - Implement single pane restore (much like the Gftp panes).
34 - Implement Automatic Mount even in operator wait.
35 - Implement create "FileSet"?
36 - Implement Release Device in the Job resource to unmount a drive.
37 - Implement Acquire Device in the Job resource to mount a drive,
38 be sure this works with admin jobs so that the user can get
39 prompted to insert the correct tape. Possibly some way to say to
40 run the job but don't save the files.
41 - Implement all command line args on run.
42 - Implement command line "restore" args.
43 - Implement "restore current select=no"
44 - Fix watchdog pthread crash on Win32
45 - Fix "access not allowed" for backup of files on WinXP.
46 - Implement "scratch pool" where tapes are defined and can be
47 taken by any pool that needs them.
48 - Implement restore "current system", but take all files without
49 doing selection tree -- so that jobs without File records can
51 - Make | and < work on FD side.
52 - Pass prefix_links to FD.
53 - Implement a M_SECURITY message class.
54 - Implement disk spooling. Two parts: 1. Spool to disk then
55 immediately to tape to speed up tape operations. 2. Spool to
56 disk only when the tape is full, then when a tape is hung move
58 - From Phil Stracchino:
59 It would probably be a per-client option, and would be called
60 something like, say, "Automatically purge obsoleted jobs". What it
61 would do is, when you successfully complete a Differential backup of a
62 client, it would automatically purge all Incremental backups for that
63 client that are rendered redundant by that Differential. Likewise,
64 when a Full backup on a client completed, it would automatically purge
65 all Differential and Incremental jobs obsoleted by that Full backup.
66 This would let people minimize the number of tapes they're keeping on
67 hand without having to master the art of retention times.
69 - Allow multiple Storage specifications (or multiple names on
70 a single Storage specification) in the Job record. Thus a job
71 can be backed up to a number of storage devices.
72 - Implement dump label to UA
73 - Add prefixlinks to where or not where absolute links to FD.
74 - Look at Python for a Bacula scripting language -- www.python.org
75 - Issue message to mount a new tape before the rewind.
76 - Simplified client job initiation for portables.
77 - If SD cannot open a drive, make it periodically retry.
78 - Implement LabelTemplate (at least first cut).
79 - Add more of the config info to the tape label.
80 - Implement a Mount Command and an Unmount Command where
81 the users could specify a system command to be performed
82 to do the mount, after which Bacula could attempt to
83 read the device. This is for Removeable media such as a CDROM.
84 - Most likely, this mount command would be invoked explicitly
85 by the user using the current Console "mount" and "unmount"
86 commands -- the Storage Daemon would do the right thing
87 depending on the exact nature of the device.
88 - As with tape drives, when Bacula wanted a new removable
89 disk mounted, it would unmount the old one, and send a message
90 to the user, who would then use "mount" as described above
91 once he had actually inserted the disk.
93 - Make some way so that if a machine is skipped because it is not up
94 that Bacula will continue retrying for a specified period of time --
96 - If tape is marked read-only, then try opening it read-only rather than
97 failing, and remember that it cannot be written.
98 - Refine SD waiting output:
99 Device is being positioned
100 > Device is being positioned for append
101 > Device is being positioned to file x
103 - Figure out some way to estimate output size and to avoid splitting
104 a backup across two Volumes -- this could be useful for writing CDROMs
105 where you really prefer not to have it split -- not serious.
106 - Add RunBeforeJob and RunAfterJob to the Client program.
107 - Have SD compute MD5 or SHA1 and compare to what FD computes.
108 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
109 actual data on the Volume and compare it.
110 - Implement FileOptions (see end of this document)
111 - Implement Bacula plugins -- design API
112 - Make bcopy read through bad tape records.
113 - Fix read_record to handle multiple sessions.
114 - Program files (i.e. execute a program to read/write files).
115 Pass read date of last backup, size of file last time.
116 - Add Signature type to File DB record.
117 - Make Restore report an error if FD or SD term codes are not OK.
118 - CD into subdirectory when open()ing files for backup to
119 speed up things. Test with testfind().
120 - Priority job to go to top of list.
121 - Find out why Full saves run slower and slower (hashing?)
122 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
123 - Implement some way for the Console to dynamically create a job.
124 - Restore to a particular time -- e.g. before date, after date.
125 - Solaris -I on tar for include list
126 - Prohibit backing up archive device (findlib/find_one.c:128)
127 - Need a verbose mode in restore, perhaps to bsr.
128 - bscan without -v is too quiet -- perhaps show jobs.
129 - Add code to reject whole blocks if not wanted on restore.
130 - Start working on Base jobs.
131 - Check if we can increase Bacula FD priorty in Win2000
132 - Make sure the MaxVolFiles is fully implemented in SD
133 - Check if both CatalogFiles and UseCatalog are set to SD.
134 - Need return status on read_cb() from read_records(). Need multiple
135 records -- one per Job, maybe a JCR or some other structure with
136 a block and a record.
137 - Figure out how to do a bare metal Windows restore
138 - Put system type returned by FD into catalog.
139 - Possibly add email to Watchdog if drive is unmounted too
140 long and a job is waiting on the drive.
141 - Use read_record.c in SD code.
142 - Why don't we get an error message from Win32 FD when bootstrap
143 file cannot be created for restore command?
144 - When Marking a file in Restore that is a hard link, also
145 mark the link so that the data will be reloaded.
146 - Restore program that errors in SD due to no tape reports
147 OK incorrectly in output.
148 - After unmount, if restore job started, ask to mount.
149 - Convert all %x substitution variables, which are hard to remember
150 and read to %(variable-name). Idea from TMDA.
151 - Add JobLevel in FD status (but make sure it is defined).
152 - Make Pool resource handle Counter resources.
153 - Remove NextId for SQLite. Optimize.
154 - Move all SQL statements into a single location.
155 - Add UA rc and history files.
156 - put termcap (used by console) in ./configure and
157 allow -with-termcap-dir.
158 - Enhance time and size scanning routines.
159 - Fix Autoprune for Volumes to respect need for full save.
160 - Fix Win32 config file definition name on /install
161 - No READLINE_SRC if found in alternate directory.
162 - Test a second language e.g. french.
163 - Compare tape to Client files (attributes, or attributes and data)
164 - Make all database Ids 64 bit.
165 - Write an applet for Linux.
166 - Add estimate to Console commands
167 - Find solution to blank filename (i.e. path only) problem.
168 - Implement new daemon communications protocol.
169 - Remove PoolId from Job table, it exists in Media.
170 - Allow console commands to detach or run in background.
171 - Fix status delay on storage daemon during rewind.
172 - Add SD message variables to control operator wait time
173 - Maximum Operator Wait
174 - Minimum Message Interval
175 - Maximum Message Interval
176 - Send Operator message when cannot read tape label.
177 - Verify level=Volume (scan only), level=Data (compare of data to file).
178 Verify level=Catalog, level=InitCatalog
180 - Add keyword search to show command in Console.
181 - Fix Win2000 error with no messages during startup.
182 - Events : tape has more than xxx bytes.
183 - Restrict characters permitted in a Resource name.
184 - Complete code in Bacula Resources -- this will permit
185 reading a new config file at any time.
186 - Handle ctl-c in Console
187 - Implement script driven addition of File daemon to config files.
188 - Think about how to make Bacula work better with File (non-tape) archives.
189 - Write Unix emulator for Windows.
190 - Implement new serialize subroutines
191 send(socket, "string", &Vol, "uint32", &i, NULL)
192 - Audit all UA commands to ensure that we always prompt where possible.
193 - If ./btape is called without /dev, assume argument is a Storage resource name.
194 - Put memory utilization in Status output of each daemon
195 if full status requested or if some level of debug on.
196 - Make database type selectable by .conf files i.e. at runtime
197 - Set flag for uname -a. Add to Volume label.
198 - Implement throttled work queue.
199 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
200 - Restore files modified after date
201 - Restore file modified before date
202 - Emergency restore info:
204 - Backup working directory
206 - Restore -- do nothing but show what would happen
207 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
208 - Implement Restore FileSet=
209 - Create a protocol.h and protocol.c where all protocol messages
211 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
212 - Timout a job or terminate if link goes down, or reopen link and query.
213 - Find general solution for sscanf size problems (as well
214 as sprintf. Do at run time?
215 - Concept of precious tapes (cannot be reused).
216 - Make bcopy copy with a single tape drive.
217 - Permit changing ownership during restore.
219 - Autolabel should be specified by DIR instead of SD.
220 - Find out how to get the system tape block limits, e.g.:
221 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
222 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
225 - AutoScan (check checksum of tape)
226 - Format command = "format /dev/nst0"
230 - Seek resolution (usually corresponds to buffer size)
231 - EODErrorCode=ENOSPC or code
232 - Partial Read error code
233 - Partial write error code
234 - Nonformatted read error
235 - Nonformatted write error
236 - WriteProtected error
240 - IgnoreCloseErrors=yes
250 - FD sends unsaved file list to Director at end of job (see
252 - File daemon should build list of files skipped, and then
253 at end of save retry and report any errors.
254 - Write a Storage daemon that uses pipes and
255 standard Unix programs to write to the tape.
257 - Need something that monitors the JCR queue and
258 times out jobs by asking the deamons where they are.
259 - Enhance Jmsg code to permit buffering and saving to disk.
260 - device driver = "xxxx" for drives.
261 - restart: paranoid: read label fsf to
262 eom read append block, and go
263 super-paranoid: read label, read all files
264 in between, read append block, and go
265 verify: backspace, read append block, and go
266 permissive: same as above but frees drive
267 if tape is not valid.
269 - Ensure that /dev/null works
270 - Need report class for messages. Perhaps
271 report resource where report=group of messages
272 - enhance scan_attrib and rename scan_jobtype, and
273 fill in code for "since" option
274 - Director needs a time after which the report status is sent
275 anyway -- or better yet, a retry time for the job.
276 Don't reschedule a job if previous incarnation is still running.
277 - Some way to automatically backup everything is needed????
278 - Need a structure for pending actions:
280 - termination status (part of buffered msgs?)
282 Read, Write, Clean, Delete
283 - Login to Bacula; Bacula users with different permissions:
284 owner, group, user, quotas
285 - Store info on each file system type (probably in the job header on tape.
286 This could be the output of df; or perhaps some sort of /etc/mtab record.
289 - Design at hierarchial storage for Bacula. Migration and Clone.
290 - Implement FSM (File System Modules).
291 - Identify unchanged or "system" files and save them to a
292 special tape thus removing them from the standard
293 backup FileSet -- BASE backup.
294 - Heartbeat between daemons.
295 - Audit M_ error codes to ensure they are correct and consistent.
296 - Add variable break characters to lex analyzer.
297 Either a bit mask or a string of chars so that
298 the caller can change the break characters.
299 - Make a single T_BREAK to replace T_COMMA, etc.
300 - Ensure that File daemon and Storage daemon can
301 continue a save if the Director goes down (this
302 is NOT currently the case). Must detect socket error,
303 buffer messages for later.
304 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
305 - Add ability to backup to two Storage devices (two SD sessions) at
306 the same time -- e.g. onsite, offsite.
307 - Add the ability to consolidate old backup sets (basically do a restore
308 to tape and appropriately update the catalog). Compress Volume sets.
309 Might need to spool via file is only one drive is available.
310 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
311 someway to do so with every volume that has less than x% valid
316 Bacula Projects Roadmap
318 last update 5 January 2003
320 Item 1: Multiple simultaneous Jobs. (done)
321 Done -- Restore part needs better implementation to work correctly
323 What: Permit multiple simultaneous jobs in Bacula.
325 Why: An enterprise level solution needs to go fast without the
326 need for the system administrator to carefully tweak
327 timing. Based on the benchmarks, during a full
328 backup, NetWorker typically hit 10 times the bandwidth to
329 the tape compared to Bacula--largely. This is probably due to
330 running parallel jobs and multi-threaded filling of buffers
331 and writing them to tape. This should also make things work
332 better when you have a mix of fast and slow machines backing
335 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
336 implementing this is a matter of some small cleanups and
340 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
341 Deferred -- not necessary yet.
343 What: If data is coming into the SD too fast, buffer it to
344 disk if the user has configured this option.
346 Why: This would be nice, especially if it more or less falls out
347 when implementing (1) above. If not, it probably should not
348 be given a high priority because fundamentally the backup time
349 is limited by the tape bandwidth. Even though you may finish a
350 client job quicker by spilling to disk, you still have to
351 eventually get it onto tape. If intermediate disk buffering
352 allows us to improve write bandwidth to tape, it may make
355 Notes: Whether or not this is implemented will depend upon performance
356 testing after item 1 is implemented.
359 Item 3: Write the bscan program -- also write a bcopy program.
362 What: Write a program that reads a Bacula tape and puts all the
363 appropriate data into the catalog. This allows recovery
364 from a tape that is no longer in the database, or it allows
365 re-creation of a database if lost.
367 Why: This is a fundamental robustness and disaster recovery tool
368 which will increase the comfort level of a sysadmin
369 considering adopting Bacula.
371 Notes: A skeleton of this program already exists, but much work
372 needs to be done. Implementing this will also make apparent
373 any deficiencies in the current Bacula tape format.
376 Item 4: Implement Base jobs.
378 What: A base job is sort of like a Full save except that you
379 will want the FileSet to contain only files that are unlikely
380 to change in the future (i.e. a snapshot of most of your
381 system after installing it). After the base job has been run,
382 when you are doing a Full save, you can specify to exclude
383 all files saved by the base job that have not been modified.
385 Why: This is something none of the competition does, as far as we know
386 (except BackupPC, which is a Perl program that saves to disk
387 only). It is big win for the user, it makes Bacula stand out
388 as offering a unique optimization that immediately saves time
391 Notes: Big savings in tape usage. Will require more resources because
392 the e. DIR must send FD a list of files/attribs, and the FD must
393 search the list and compare it for each file to be saved.
396 Item 5: Implement Label templates
398 What: This is a mechanism whereby Bacula can automatically create
399 a tape label for new tapes according to a detailed specification
400 provided by the user.
402 Why: It is a major convenience item for folks who use automated label
405 Notes: Bacula already has a working form of automatic tape label
406 creation, but it is very crude. The design for the complete
407 tape labeling project is already documented in the manual.
410 Item 6: Write a regression script.
411 Done -- Continue to expand its testing.
413 What: This is an automatic script that runs and tests as many features
414 of Bacula as possible. The output is compared to previous
415 versions of Bacula and any differences are reported.
417 Why: This is an enormous help in preventing introduction of new
418 errors in parts of the program that already work correctly.
420 Notes: This probably should be ranked higher, it's something the typical
421 user doesn't see. Depending on how it's implemented, it may
422 make sense to defer it until the archival tape format and
423 user interface mature.
426 Item 7: GUI for interactive restore
427 Item 8: GUI for interactive backup
429 What: The current interactive restore is implemented with a tty
430 interface. It would be much nicer to be able to "see" the
431 list of files backed up in typical GUI tree format.
432 The same mechanism could also be used for creating
433 ad-hoc backup FileSets (item 8).
435 Why: Ease of use -- especially for the end user.
437 Notes: Rather than implementing in Gtk, we probably should go directly
438 for a Browser implementation, even if doing so meant the
439 capability wouldn't be available until much later. Not only
440 is there the question of Windows sites, most
441 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
442 without installing lots of stuff admins are very wary about.
443 Real sysadmins will always use the command line anyway, and
444 the user who's doing an interactive restore or backup of his
445 own files will in most cases be on a Windows machine running
449 Item 9: Add SSL to daemon communications.
451 What: This provides for secure communications between the daemons.
453 Why: This would allow doing backup across the Internet without
454 privacy concerns (or with much less concern).
456 Notes: The vast majority of near term potential users will be backing up
457 a single site over a LAN and, correctly or not, they probably
458 won't be concerned with security, at least not enough to go to
459 the trouble to set up keys, etc. to screw things down. We suspect
460 that many users genuinely interested in multi-site backup
461 already run some form of VPN software in their internetwork
462 connections, and are willing to delegate security to that layer.
465 Item 10: Define definitive tape format.
468 What: Define that definitive tape format that will not change
469 for the next millennium.
471 Why: Stability, security.
473 Notes: See notes for item 11 below.
476 Item 11: New daemon communication protocol.
478 What: The current daemon to daemon protocol is basically an ASCII
479 printf() and sending the buffer. On the receiving end, the
480 buffer is sscanf()ed to unpack it. The new scheme would
481 be a binary format that allows quick packing and unpacking
482 of any data type with named fields.
484 Why: Using binary packing would be faster. Named fields will permit
485 error checking to ensure that what is sent is what the
486 receiver really wants.
488 Notes: These are internal improvements in the interest of the
489 long-term stability and evolution of the program. On the one
490 hand, the sooner they're done, the less code we have to rip
491 up when the time comes to install them. On the other hand, they
492 don't bring an immediately perceptible benefit to potential
493 users. Item 10 and possibly item 11 should be deferred until Bacula
494 is well established with a growing user community more or
495 less happy with the feature set. At that time, it will make a
496 good "next generation" upgrade in the interest of data
502 ======================================================
504 It is somewhat like a Full save becomes an incremental since
505 the Base job (or jobs) plus other non-base files.
507 - New BaseFile table that contains:
508 JobId, BaseJobId, FileId (from Base).
509 i.e. for each base file that exists but is not saved because
510 it has not changed, the File daemon sends the JobId, BaseId,
511 and FileId back to the Director who creates the DB entry.
512 - To initiate a Base save, the Director sends the FD
513 the FileId, and full filename for each file in the Base.
514 - When the FD finds a Base file, he requests the Director to
515 send him the full File entry (stat packet plus MD5), or
516 conversely, the FD sends it to the Director and the Director
517 says yes or no. This can be quite rapid if the FileId is kept
518 by the FD for each Base Filename.
519 - It is probably better to have the comparison done by the FD
520 despite the fact that the File entry must be sent across the
522 - An alternative would be to send the FD the whole File entry
523 from the start. The disadvantage is that it requires a lot of
524 space. The advantage is that it requires less communications
526 - The Job record must be updated to indicate that one or more
528 - At end of Job, FD returns:
529 1. Count of base files/bytes not written to tape (i.e. matches)
530 2. Count of base file that were saved i.e. had changed.
531 - No tape record would be written for a Base file that matches, in the
532 same way that no tape record is written for Incremental jobs where
533 the file is not saved because it is unchanged.
534 - On a restore, all the Base file records must explicitly be
535 found from the BaseFile tape. I.e. for each Full save that is marked
536 to have one or more Base Jobs, search the BaseFile for all occurrences
538 - An optimization might be to make the BaseFile have:
544 This would avoid the need to explicitly fetch each File record for
545 the Base job. The Base Job record will be fetched to get the
546 VolSessionId and VolSessionTime.
547 =========================================================
549 ==========================================================
551 For each Incremental job that is run, there may be files that
552 were found but not saved because they were locked (this applies
553 only to Windows). Such a system could send back to the Director
554 a list of Unsaved files.
556 - New UnSavedFiles table that contains:
560 - Then in the next Incremental job, the list of Unsaved Files will be
561 feed to the FD, who will ensure that they are explicitly chosen even
562 if standard date/time check would not have selected them.
563 =============================================================
567 =============================================================
569 Request For Comments For File Backup Options
572 Subject: File Backup Options
575 A few days ago, a Bacula user who is backing up to file volumes and
576 using compression asked if it was possible to suppress compressing
577 all .gz files since it was a waste of CPU time. Although Bacula
578 currently permits using different options (compression, ...) on
579 a directory by directory basis, it cannot do it on a file by
580 file basis, which is clearly what was desired.
582 Proposed Implementation:
583 To solve this problem, I propose the following:
585 - Add a new Director resource type called FileOptions.
587 - The FileOptions resource will have records for all
588 options that can currently be specified on the Include record
589 (in a FileSet). Examples below.
591 - The FileOptions resource will permit an exclude option as well
592 as a number of additional options.
594 - The heart of the FileOptions resource is the ability to
595 supply any number of ApplyTo records which specify POSIX
596 regular expressions. These ApplyTo regular expressions are
597 applied to the fully qualified filename (path and all). If
598 one matches, then the FileOptions will be used.
600 - When an ApplyTo specification matches an included file, the
601 options specified in the FileOptions resource will override
602 the default options specified on the Include record.
604 - Include records will be modified to permit referencing one or
605 more FileOptions resources. The FileOptions will be used
606 in the order listed on the Include record and the first
607 one that matches will be applied.
609 - Options (or specifications) currently supplied on the Include
610 record will be deprecated (i.e. removed in a later version a
611 year or so from now).
613 - The Exclude record will be deprecated as the same functionality
614 can be obtained by using an Exclude = yes in the FileOptions.
617 The following records can appear in the FileOptions resource. An
618 asterisk preceding the name indicates a feature not currently
622 - Compression= (GZIP, ...)
623 - Signature= (MD5, SHA1, ...)
625 - OneFs= (yes/no) - remain on one filesystem
626 - Recurse= (yes/no) - recurse into subdirectories
627 - Sparse= (yes/no) - do sparse file backup
628 - *Exclude= (yes/no) - exclude file from being saved
629 - *Reader= (filename) - external read (backup) program
630 - *Plugin= (filename) - read/write plugin module
633 - verify= (ipnougsamc5) - verify options
636 - replace= (always/ifnewer/ifolder/never) - replace options currently
638 - *Writer= (filename) - external write (restore) program
642 Currently options specifying compression, MD5 signatures, recursion,
643 ... of a FileSet are supplied on the Include record. These will now
644 all be collected into a FileOptions resource, which will be
645 specified on the Include in place of the options. Multiple FileOptions
646 may be specified. Since the FileOptions contain regular expressions
647 that are applied to the full filename, this will give the ability
648 to specify backup options on a file by file basis to whatever level
657 Include = compression=GZIP signature=MD5 {
666 Include = FileOptions=Opts {
677 That's a lot more to do the same thing, but it gives the ability to
678 apply options on a file by file basis. For example, suppose you
679 want to compress all files but not any file with extensions .gz or .Z.
680 You could do so as follows:
684 Include = FileOptions=NoCompress FileOptions=Opts {
692 ApplyTo = /*.?*/ # matches all files
697 # Note multiple ApplyTos are ORed
698 ApplyTo = /*.gz/ # matches .gz files */
699 ApplyTo = /*.Z/ # matches .Z files */
702 Now, since the NoCompress FileOptions is specified first on the
703 Include line, any *.gz or *.Z file will have an MD5 signature computed,
704 but will not be compressed. For all other files, the NoCompress will not
705 match, so the Opts options will be used which will include GZIP
709 - Is it necessary to provide some means of ANDing regular expressions
710 and negation? (not currently planned)
712 e.g. ApplyTo = /*.gz/ && !/big.gz/
714 - I see that Networker has a "null" module which, if specified, does not
715 backup the file, but does make an record of the file in the catalog
716 so that the catalog will reflect an exact picture of the filesystem.
717 The result is that the file can be "seen" when "browsing" the save
718 sets, but it cannot be restored.
720 Is this really useful? Should it be implemented in Bacula?
723 After implementing the above, the user will be able to specify
724 on a file by file basis (using regular expressions) what options are
725 applied for the backup.
726 ====================================
728 =========================================
729 Proposal by Bill Sellers
731 Return-Path: <w.a.sellers@larc.nasa.gov>
732 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
733 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)
734 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
735 X-Sender: w.a.sellers@pop.larc.nasa.gov
736 X-Mailer: QUALCOMM Windows Eudora Version 5.1
737 Date: Tue, 14 Jan 2003 16:30:18 -0500
738 To: Kern Sibbald <kern@sibbald.com>
739 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
740 Subject: Re: [Bacula-users] Bacula remote storage?
741 In-Reply-To: <1042565382.1845.177.camel@rufus>
742 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>
744 Content-Type: text/plain; charset="us-ascii"; format=flowed
745 X-Annoyance-Filter-Junk-Probability: 0
746 X-Annoyance-Filter-Classification: Mail
747 At 06:29 PM 1/14/2003 +0100, you wrote:
750 >Well, if you cannot put a Bacula client on the machine,
751 >then it is a big problem. If you know of some software
752 >that can do what you want, let me know, because I
753 >really just don't know how to do it -- at least not
759 We have been able to get Amanda to use the HSM as a storage
760 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
761 handle Windows systems very well (or at all without Samba). So I am
762 looking for a backup system that has a Windows client. I really like the
763 Windows integration of Bacula.
765 From the command line, its rather trivial to move the data around. We use
768 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
770 or if you use GNU tar:
772 tar czf hsm:path/file.tgz ./files
774 One idea for you to consider; Sendmail offers pipes in the aliases file;
775 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
776 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
777 pipe available, as a storage device? Then we could use any command that
778 handles stdin as a storage destination.
785 #Password is not used in rsh, but might be used in ftp.
786 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
793 Password = "foobar&-"
794 Device = "| ncftpput -c hsm /path/file.bacula"
798 >If you have some local storage available, you could
799 >use Bacula to backup to disk volumes, then use some
800 >other software (ftp, scp) to move them to the HSM
801 >machine. However, this is a bit kludgy.
804 It is, but maybe worth a try. Is there some function in Bacula to put
805 variables in filenames? i.e. backup.2003-01-15.root
812 w.a.sellers@larc.nasa.gov
814 ==============================================
815 The Project for the above
817 I finally realized that this is not at all
818 the same as reader/writer programs or plugins,
819 which are alternate ways of accessing the
820 files to be backed up. Rather, it is an alternate
821 form of storage device, and I have always planned
822 that Bacula should be able to handle all sorts
825 So, I propose the following phases:
827 1. OK from you to invest some time in testing
828 this as I implement it (requires that you
829 know how to download from the SourceForge
830 cvs -- which I imagine is a piece of cake
833 2. Dumb implementation by allowing a device to
834 be a fifo for write only.
835 Reason: easy to implement, proof of concept.
837 3. Try reading from fifo but with fixed block
839 Reason: proof of concept, easy to implement.
841 4. Extend reading from fifo (restores) to handle
843 Reason: requires some delicate low level coding
844 which could destabilize all of Bacula.
846 5. Implementation of above but to a program. E.g.
847 Device = "|program" (not full pipeline).
848 Reason: routines already exist, and program can
849 be a shell script which contains anything.
851 6. Add full pipeline as a possibility. E.g.
852 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
853 Reason: needs additional coding to implement full
854 pipeline (must fire off either a shell or all
855 programs and connect their pipes).
857 There are a good number of details in each step
858 that I have left out, but I will specify them at
859 every stage, and there may be a few changes as things
860 evolve. I expect that to get to stage 5 will take a
861 few weeks, and at that point, you will have
862 everything you need (just inside a script).
863 Stage 6 will probably take longer, but if this
864 project pleases you, what we do for 5 should
865 be adequate for some time.
869 =============================================
871 Done: (see kernsdone for more)
872 - Look into Pruning/purging problems or why there seem to
873 be so many files listed each night.
874 - Fix cancel in find_one -- need jcr.
875 - Cancel does not work for restore in FD.
876 - Write SetJobStatus() function so cancel status not lost.
877 - Add include list to end of chain in findlib
878 - Zap sd_auth_key after use
879 - Add Bar code reading capabilities (new mtx-changer)
880 - Figure out some way to automatically backup all local partitions
881 - Make hash table for linked files in findlib/find_one.c:161
883 - Rewrite find_one.c to use only pool_memory instead of
884 alloca and malloc (probably not necessary).
885 - Make sure btraceback goes into /sbin not sysconf directory.
886 - InitVerify is getting pruned and it shouldn't (document it)
887 - Make 1.28c release ??? NO do 1.29 directly
888 - Set timeout on opening fifo for save or restore (findlib)
889 - Document FIFO storage device.
890 - Document fifo and | and <
893 - Get correct error status from run_program or open_bpipe().
894 - Restrict permissions on File Volumes (now 0640).
895 - Umasked 022 daemons
896 - Fix restore of hard linked file.
897 - Figure out how to allow multiple simultaneous file Volumes on a single device.
898 - Implement multiple simultaneous file Volumes on a single device.
899 - Cleanup db_update_media and db_update_pool
900 - Flush all the daemon messages at the end of every job.
901 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
902 run_program -- bpipe.c
903 - Apparently cancel does not work for jobs waiting to be
905 - Implement TCP/IP connection for MySQL
906 - Pull a canceled job from the Scheduling queue.
907 - Implement max_file_size in block.c (already done, just tweaked).
908 - Look at purge jobs volume (at least document it, and see if it is
910 - Add list volumes does all pools. list volumes pool=xxx now works.
911 - Add pool= to "list media" in ua_output.c
912 - Strip trailing slashes from Include directory names in the FD.
913 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
914 ERR=Operation not permited loop.
915 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
916 - Produce better error messages in when error/eof writing block.
917 - Cancelling of a queued job does NOT work!!!!!!
919 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
920 rufus-sd: Recycled volume File0003 on device /home/kern/bacula/working, all previous data lost.
921 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
922 - Ability to backup to a file then later transfer to a tape -- Migration.
923 Migration based on MaxJobs(MinJobs),MaxVols(MinVols),AgeJobs,MaxBytes(MinBytes)
924 (i.e. HighwaterMark, LowwaterMark).
925 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
926 set of volumes recycling the oldest volume when it is needed.
927 - gethostbyname failure in bnet_connect() continues
928 generating errors -- should stop.
929 - Add chflags() code for FreeBSD file flags
930 - Bevan Anderson suggests having a run queue for each device
931 so that multiple simultaneous jobs can run but each writing
932 to a different Volume.
933 - Look at handling <> in smtp doesn't work with exim.
934 - Need to specify MaximumConcurrentJobs in the Job resource.
935 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
936 - Implement finer multiprocessing options.
937 - Implement | and < in Exclude statements.
938 - Figure out some way to specify a retention period for files
939 that no longer exist on the machine -- so that we maintain
940 say backups for 30 days, but if the file is deleted, we maintain
941 the last copy for 1 year. -- answer Volume retention.
942 - Make non-zero status from RunJobBefore/After error the job.
943 - Need define of int_least16_t in sha1.h for SuSE.
944 - Implement bar code reader for autochangers
945 - Document new MaximumConcurrentJob records (Job, Client, Storage)
946 - Write up how to use/manage disk Volume Storage. ******
947 - Remove kern and kelvin from mysql_grant...
948 - Install grant_mysql...
949 - Strip trailing / from Include
950 - add #define ENABLE_NLS for Gnome compile on SuSE.
951 - Add Client FS/OS id (Linux, Win95/98, ...).
952 - Concept of VolumeSet during restore which is a list
953 of Volume names needed.
954 - Turn virutally all sprintfs into snprintfs.
955 - Update volume=Test01 requests pool, then lists volumes.
956 **** Test select_pool_and_media ...
958 - Add IP address to authentication failures.
959 - Add a default File storage so that new users can do backup
960 and restores right away.
961 - Forbid sbindir and with-subsys-dir from being the same (otherwise
962 the binary gets deleted when the daemon is stopped in the
963 rc.d/inid.d directory.
964 - Do not ignore SIGCHLD
965 - Add Cleaning to list of volume statuses
966 - Implement run at "xxx"
967 - Document new transparent Console commands and wait command.
968 - Document . and @ commands
970 - Document Lutz Kittler's trick of using "Run Before Job" to
971 abort a job on a particular day.
972 - Document Ludovic Strappazon's Win32 raw device save/restore.
973 - Document not to restore .journal .autofsck
974 - Document labeling a whole magazine using "cat"
975 - Document how to automatically backup all local partitions
977 - Document OPTIMIZE TABLE in MySQL
978 - Document new immediate File save configuration (walk user
979 through first save to file Volume with automatic Volume labeling?).
980 - Implement scheduling of one time "run" jobs (i.e. instead of
981 starting immediately start at some specified time).
982 - Bug: up arrow prints garbage in command line on gnome-console!