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 - Document default file backup
17 Testing to do: (painful)
18 - that ALL console command line options work and are always implemented
19 - blocksize recognition code.
20 - multiple simultaneous Volumes
23 - Do not ignore SIGCHLD
24 - Fix Bare Metal restore problem.
25 - Fix "access not allowed" for backup of files on WinXP.
26 - Document new transparent Console commands and wait command.
27 - Document new immediate File save capability.
28 - Document OPTIMIZE TABLE in MySQL
29 - Fix watchdog pthread crash on Win32
31 - Figure out how to use ssh or stunnel to protect Bacula communications.
34 - Implement restore "current system", but take all files without
35 doing selection tree -- so that jobs without File records can
37 - Implement scheduling of one time "run" jobs (i.e. instead of
38 starting immediately start at some specified time).
39 - Make | and < work on FD side.
40 - Pass prefix_links to FD.
41 - Implement a M_SECURITY message class.
42 - Implement disk spooling. Two parts: 1. Spool to disk then
43 immediately to tape to speed up tape operations. 2. Spool to
44 disk only when the tape is full, then when a tape is hung move
46 - From Phil Stracchino:
47 It would probably be a per-client option, and would be called
48 something like, say, "Automatically purge obsoleted jobs". What it
49 would do is, when you successfully complete a Differential backup of a
50 client, it would automatically purge all Incremental backups for that
51 client that are rendered redundant by that Differential. Likewise,
52 when a Full backup on a client completed, it would automatically purge
53 all Differential and Incremental jobs obsoleted by that Full backup.
54 This would let people minimize the number of tapes they're keeping on
55 hand without having to master the art of retention times.
57 - Allow multiple Storage specifications (or multiple names on
58 a single Storage specification) in the Job record. Thus a job
59 can be backed up to a number of storage devices.
60 - Implement dump label to UA
61 - Add prefixlinks to where or not where absolute links to FD.
62 - Look at Python for a Bacula scripting language -- www.python.org
63 - Issue message to mount a new tape before the rewind.
64 - Simplified client job initiation for portables.
65 - If SD cannot open a drive, make it periodically retry.
66 - Implement LabelTemplate (at least first cut).
67 - Add more of the config info to the tape label.
68 - Implement a Mount Command and an Unmount Command where
69 the users could specify a system command to be performed
70 to do the mount, after which Bacula could attempt to
71 read the device. This is for Removeable media such as a CDROM.
72 - Most likely, this mount command would be invoked explicitly
73 by the user using the current Console "mount" and "unmount"
74 commands -- the Storage Daemon would do the right thing
75 depending on the exact nature of the device.
76 - As with tape drives, when Bacula wanted a new removable
77 disk mounted, it would unmount the old one, and send a message
78 to the user, who would then use "mount" as described above
79 once he had actually inserted the disk.
81 - Make some way so that if a machine is skipped because it is not up
82 that Bacula will continue retrying for a specified period of time --
84 - If tape is marked read-only, then try opening it read-only rather than
85 failing, and remember that it cannot be written.
86 - Refine SD waiting output:
87 Device is being positioned
88 > Device is being positioned for append
89 > Device is being positioned to file x
91 - Figure out some way to estimate output size and to avoid splitting
92 a backup across two Volumes -- this could be useful for writing CDROMs
93 where you really prefer not to have it split -- not serious.
94 - Add RunBeforeJob and RunAfterJob to the Client program.
95 - Have SD compute MD5 or SHA1 and compare to what FD computes.
96 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
97 actual data on the Volume and compare it.
98 - Implement FileOptions (see end of this document)
99 - Implement Bacula plugins -- design API
100 - Make bcopy read through bad tape records.
101 - Fix read_record to handle multiple sessions.
102 - Program files (i.e. execute a program to read/write files).
103 Pass read date of last backup, size of file last time.
104 - Add Signature type to File DB record.
105 - Make Restore report an error if FD or SD term codes are not OK.
106 - CD into subdirectory when open()ing files for backup to
107 speed up things. Test with testfind().
108 - Priority job to go to top of list.
109 - Find out why Full saves run slower and slower (hashing?)
110 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
111 - Implement some way for the Console to dynamically create a job.
112 - Restore to a particular time -- e.g. before date, after date.
113 - Solaris -I on tar for include list
114 - Prohibit backing up archive device (findlib/find_one.c:128)
115 - Need a verbose mode in restore, perhaps to bsr.
116 - bscan without -v is too quiet -- perhaps show jobs.
117 - Add code to reject whole blocks if not wanted on restore.
118 - Start working on Base jobs.
119 - Check if we can increase Bacula FD priorty in Win2000
120 - Make sure the MaxVolFiles is fully implemented in SD
121 - Check if both CatalogFiles and UseCatalog are set to SD.
122 - Need return status on read_cb() from read_records(). Need multiple
123 records -- one per Job, maybe a JCR or some other structure with
124 a block and a record.
125 - Figure out how to do a bare metal Windows restore
126 - Put system type returned by FD into catalog.
127 - Possibly add email to Watchdog if drive is unmounted too
128 long and a job is waiting on the drive.
129 - Use read_record.c in SD code.
130 - Why don't we get an error message from Win32 FD when bootstrap
131 file cannot be created for restore command?
132 - When Marking a file in Restore that is a hard link, also
133 mark the link so that the data will be reloaded.
134 - Restore program that errors in SD due to no tape reports
135 OK incorrectly in output.
136 - After unmount, if restore job started, ask to mount.
137 - Convert all %x substitution variables, which are hard to remember
138 and read to %(variable-name). Idea from TMDA.
139 - Add JobLevel in FD status (but make sure it is defined).
140 - Make Pool resource handle Counter resources.
141 - Remove NextId for SQLite. Optimize.
142 - Move all SQL statements into a single location.
143 - Add UA rc and history files.
144 - put termcap (used by console) in ./configure and
145 allow -with-termcap-dir.
146 - Enhance time and size scanning routines.
147 - Fix Autoprune for Volumes to respect need for full save.
148 - Fix Win32 config file definition name on /install
149 - No READLINE_SRC if found in alternate directory.
150 - Test a second language e.g. french.
151 - Compare tape to Client files (attributes, or attributes and data)
152 - Make all database Ids 64 bit.
153 - Write an applet for Linux.
154 - Add estimate to Console commands
155 - Find solution to blank filename (i.e. path only) problem.
156 - Implement new daemon communications protocol.
157 - Remove PoolId from Job table, it exists in Media.
158 - Allow console commands to detach or run in background.
159 - Fix status delay on storage daemon during rewind.
160 - Add SD message variables to control operator wait time
161 - Maximum Operator Wait
162 - Minimum Message Interval
163 - Maximum Message Interval
164 - Send Operator message when cannot read tape label.
165 - Verify level=Volume (scan only), level=Data (compare of data to file).
166 Verify level=Catalog, level=InitCatalog
168 - Add keyword search to show command in Console.
169 - Fix Win2000 error with no messages during startup.
170 - Events : tape has more than xxx bytes.
171 - Restrict characters permitted in a Resource name.
172 - Complete code in Bacula Resources -- this will permit
173 reading a new config file at any time.
174 - Handle ctl-c in Console
175 - Implement script driven addition of File daemon to config files.
176 - Think about how to make Bacula work better with File (non-tape) archives.
177 - Write Unix emulator for Windows.
178 - Implement new serialize subroutines
179 send(socket, "string", &Vol, "uint32", &i, NULL)
180 - Audit all UA commands to ensure that we always prompt where possible.
181 - If ./btape is called without /dev, assume argument is a Storage resource name.
182 - Put memory utilization in Status output of each daemon
183 if full status requested or if some level of debug on.
184 - Make database type selectable by .conf files i.e. at runtime
185 - Set flag for uname -a. Add to Volume label.
186 - Implement throttled work queue.
187 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
188 - Restore files modified after date
189 - Restore file modified before date
190 - Emergency restore info:
192 - Backup working directory
194 - Restore -- do nothing but show what would happen
195 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
196 - Implement Restore FileSet=
197 - Create a protocol.h and protocol.c where all protocol messages
199 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
200 - Timout a job or terminate if link goes down, or reopen link and query.
201 - Find general solution for sscanf size problems (as well
202 as sprintf. Do at run time?
203 - Concept of precious tapes (cannot be reused).
204 - Make bcopy copy with a single tape drive.
205 - Permit changing ownership during restore.
207 - Autolabel should be specified by DIR instead of SD.
208 - Find out how to get the system tape block limits, e.g.:
209 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
210 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
213 - AutoScan (check checksum of tape)
214 - Format command = "format /dev/nst0"
218 - Seek resolution (usually corresponds to buffer size)
219 - EODErrorCode=ENOSPC or code
220 - Partial Read error code
221 - Partial write error code
222 - Nonformatted read error
223 - Nonformatted write error
224 - WriteProtected error
228 - IgnoreCloseErrors=yes
238 - FD sends unsaved file list to Director at end of job (see
240 - File daemon should build list of files skipped, and then
241 at end of save retry and report any errors.
242 - Write a Storage daemon that uses pipes and
243 standard Unix programs to write to the tape.
245 - Need something that monitors the JCR queue and
246 times out jobs by asking the deamons where they are.
247 - Enhance Jmsg code to permit buffering and saving to disk.
248 - device driver = "xxxx" for drives.
249 - restart: paranoid: read label fsf to
250 eom read append block, and go
251 super-paranoid: read label, read all files
252 in between, read append block, and go
253 verify: backspace, read append block, and go
254 permissive: same as above but frees drive
255 if tape is not valid.
257 - Ensure that /dev/null works
258 - Need report class for messages. Perhaps
259 report resource where report=group of messages
260 - enhance scan_attrib and rename scan_jobtype, and
261 fill in code for "since" option
262 - Director needs a time after which the report status is sent
263 anyway -- or better yet, a retry time for the job.
264 Don't reschedule a job if previous incarnation is still running.
265 - Some way to automatically backup everything is needed????
266 - Need a structure for pending actions:
268 - termination status (part of buffered msgs?)
270 Read, Write, Clean, Delete
271 - Login to Bacula; Bacula users with different permissions:
272 owner, group, user, quotas
273 - Store info on each file system type (probably in the job header on tape.
274 This could be the output of df; or perhaps some sort of /etc/mtab record.
277 - Design at hierarchial storage for Bacula. Migration and Clone.
278 - Implement FSM (File System Modules).
279 - Identify unchanged or "system" files and save them to a
280 special tape thus removing them from the standard
281 backup FileSet -- BASE backup.
282 - Heartbeat between daemons.
283 - Audit M_ error codes to ensure they are correct and consistent.
284 - Add variable break characters to lex analyzer.
285 Either a bit mask or a string of chars so that
286 the caller can change the break characters.
287 - Make a single T_BREAK to replace T_COMMA, etc.
288 - Ensure that File daemon and Storage daemon can
289 continue a save if the Director goes down (this
290 is NOT currently the case). Must detect socket error,
291 buffer messages for later.
292 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
293 - Add ability to backup to two Storage devices (two SD sessions) at
294 the same time -- e.g. onsite, offsite.
295 - Add the ability to consolidate old backup sets (basically do a restore
296 to tape and appropriately update the catalog). Compress Volume sets.
297 Might need to spool via file is only one drive is available.
298 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
299 someway to do so with every volume that has less than x% valid
304 Bacula Projects Roadmap
306 last update 5 January 2003
308 Item 1: Multiple simultaneous Jobs. (done)
309 Done -- Restore part needs better implementation to work correctly
311 What: Permit multiple simultaneous jobs in Bacula.
313 Why: An enterprise level solution needs to go fast without the
314 need for the system administrator to carefully tweak
315 timing. Based on the benchmarks, during a full
316 backup, NetWorker typically hit 10 times the bandwidth to
317 the tape compared to Bacula--largely. This is probably due to
318 running parallel jobs and multi-threaded filling of buffers
319 and writing them to tape. This should also make things work
320 better when you have a mix of fast and slow machines backing
323 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
324 implementing this is a matter of some small cleanups and
328 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
329 Deferred -- not necessary yet.
331 What: If data is coming into the SD too fast, buffer it to
332 disk if the user has configured this option.
334 Why: This would be nice, especially if it more or less falls out
335 when implementing (1) above. If not, it probably should not
336 be given a high priority because fundamentally the backup time
337 is limited by the tape bandwidth. Even though you may finish a
338 client job quicker by spilling to disk, you still have to
339 eventually get it onto tape. If intermediate disk buffering
340 allows us to improve write bandwidth to tape, it may make
343 Notes: Whether or not this is implemented will depend upon performance
344 testing after item 1 is implemented.
347 Item 3: Write the bscan program -- also write a bcopy program.
350 What: Write a program that reads a Bacula tape and puts all the
351 appropriate data into the catalog. This allows recovery
352 from a tape that is no longer in the database, or it allows
353 re-creation of a database if lost.
355 Why: This is a fundamental robustness and disaster recovery tool
356 which will increase the comfort level of a sysadmin
357 considering adopting Bacula.
359 Notes: A skeleton of this program already exists, but much work
360 needs to be done. Implementing this will also make apparent
361 any deficiencies in the current Bacula tape format.
364 Item 4: Implement Base jobs.
366 What: A base job is sort of like a Full save except that you
367 will want the FileSet to contain only files that are unlikely
368 to change in the future (i.e. a snapshot of most of your
369 system after installing it). After the base job has been run,
370 when you are doing a Full save, you can specify to exclude
371 all files saved by the base job that have not been modified.
373 Why: This is something none of the competition does, as far as we know
374 (except BackupPC, which is a Perl program that saves to disk
375 only). It is big win for the user, it makes Bacula stand out
376 as offering a unique optimization that immediately saves time
379 Notes: Big savings in tape usage. Will require more resources because
380 the e. DIR must send FD a list of files/attribs, and the FD must
381 search the list and compare it for each file to be saved.
384 Item 5: Implement Label templates
386 What: This is a mechanism whereby Bacula can automatically create
387 a tape label for new tapes according to a detailed specification
388 provided by the user.
390 Why: It is a major convenience item for folks who use automated label
393 Notes: Bacula already has a working form of automatic tape label
394 creation, but it is very crude. The design for the complete
395 tape labeling project is already documented in the manual.
398 Item 6: Write a regression script.
399 Done -- Continue to expand its testing.
401 What: This is an automatic script that runs and tests as many features
402 of Bacula as possible. The output is compared to previous
403 versions of Bacula and any differences are reported.
405 Why: This is an enormous help in preventing introduction of new
406 errors in parts of the program that already work correctly.
408 Notes: This probably should be ranked higher, it's something the typical
409 user doesn't see. Depending on how it's implemented, it may
410 make sense to defer it until the archival tape format and
411 user interface mature.
414 Item 7: GUI for interactive restore
415 Item 8: GUI for interactive backup
417 What: The current interactive restore is implemented with a tty
418 interface. It would be much nicer to be able to "see" the
419 list of files backed up in typical GUI tree format.
420 The same mechanism could also be used for creating
421 ad-hoc backup FileSets (item 8).
423 Why: Ease of use -- especially for the end user.
425 Notes: Rather than implementing in Gtk, we probably should go directly
426 for a Browser implementation, even if doing so meant the
427 capability wouldn't be available until much later. Not only
428 is there the question of Windows sites, most
429 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
430 without installing lots of stuff admins are very wary about.
431 Real sysadmins will always use the command line anyway, and
432 the user who's doing an interactive restore or backup of his
433 own files will in most cases be on a Windows machine running
437 Item 9: Add SSL to daemon communications.
439 What: This provides for secure communications between the daemons.
441 Why: This would allow doing backup across the Internet without
442 privacy concerns (or with much less concern).
444 Notes: The vast majority of near term potential users will be backing up
445 a single site over a LAN and, correctly or not, they probably
446 won't be concerned with security, at least not enough to go to
447 the trouble to set up keys, etc. to screw things down. We suspect
448 that many users genuinely interested in multi-site backup
449 already run some form of VPN software in their internetwork
450 connections, and are willing to delegate security to that layer.
453 Item 10: Define definitive tape format.
456 What: Define that definitive tape format that will not change
457 for the next millennium.
459 Why: Stability, security.
461 Notes: See notes for item 11 below.
464 Item 11: New daemon communication protocol.
466 What: The current daemon to daemon protocol is basically an ASCII
467 printf() and sending the buffer. On the receiving end, the
468 buffer is sscanf()ed to unpack it. The new scheme would
469 be a binary format that allows quick packing and unpacking
470 of any data type with named fields.
472 Why: Using binary packing would be faster. Named fields will permit
473 error checking to ensure that what is sent is what the
474 receiver really wants.
476 Notes: These are internal improvements in the interest of the
477 long-term stability and evolution of the program. On the one
478 hand, the sooner they're done, the less code we have to rip
479 up when the time comes to install them. On the other hand, they
480 don't bring an immediately perceptible benefit to potential
481 users. Item 10 and possibly item 11 should be deferred until Bacula
482 is well established with a growing user community more or
483 less happy with the feature set. At that time, it will make a
484 good "next generation" upgrade in the interest of data
490 ======================================================
492 It is somewhat like a Full save becomes an incremental since
493 the Base job (or jobs) plus other non-base files.
495 - New BaseFile table that contains:
496 JobId, BaseJobId, FileId (from Base).
497 i.e. for each base file that exists but is not saved because
498 it has not changed, the File daemon sends the JobId, BaseId,
499 and FileId back to the Director who creates the DB entry.
500 - To initiate a Base save, the Director sends the FD
501 the FileId, and full filename for each file in the Base.
502 - When the FD finds a Base file, he requests the Director to
503 send him the full File entry (stat packet plus MD5), or
504 conversely, the FD sends it to the Director and the Director
505 says yes or no. This can be quite rapid if the FileId is kept
506 by the FD for each Base Filename.
507 - It is probably better to have the comparison done by the FD
508 despite the fact that the File entry must be sent across the
510 - An alternative would be to send the FD the whole File entry
511 from the start. The disadvantage is that it requires a lot of
512 space. The advantage is that it requires less communications
514 - The Job record must be updated to indicate that one or more
516 - At end of Job, FD returns:
517 1. Count of base files/bytes not written to tape (i.e. matches)
518 2. Count of base file that were saved i.e. had changed.
519 - No tape record would be written for a Base file that matches, in the
520 same way that no tape record is written for Incremental jobs where
521 the file is not saved because it is unchanged.
522 - On a restore, all the Base file records must explicitly be
523 found from the BaseFile tape. I.e. for each Full save that is marked
524 to have one or more Base Jobs, search the BaseFile for all occurrences
526 - An optimization might be to make the BaseFile have:
532 This would avoid the need to explicitly fetch each File record for
533 the Base job. The Base Job record will be fetched to get the
534 VolSessionId and VolSessionTime.
535 =========================================================
537 ==========================================================
539 For each Incremental job that is run, there may be files that
540 were found but not saved because they were locked (this applies
541 only to Windows). Such a system could send back to the Director
542 a list of Unsaved files.
544 - New UnSavedFiles table that contains:
548 - Then in the next Incremental job, the list of Unsaved Files will be
549 feed to the FD, who will ensure that they are explicitly chosen even
550 if standard date/time check would not have selected them.
551 =============================================================
555 =============================================================
557 Request For Comments For File Backup Options
560 Subject: File Backup Options
563 A few days ago, a Bacula user who is backing up to file volumes and
564 using compression asked if it was possible to suppress compressing
565 all .gz files since it was a waste of CPU time. Although Bacula
566 currently permits using different options (compression, ...) on
567 a directory by directory basis, it cannot do it on a file by
568 file basis, which is clearly what was desired.
570 Proposed Implementation:
571 To solve this problem, I propose the following:
573 - Add a new Director resource type called FileOptions.
575 - The FileOptions resource will have records for all
576 options that can currently be specified on the Include record
577 (in a FileSet). Examples below.
579 - The FileOptions resource will permit an exclude option as well
580 as a number of additional options.
582 - The heart of the FileOptions resource is the ability to
583 supply any number of ApplyTo records which specify POSIX
584 regular expressions. These ApplyTo regular expressions are
585 applied to the fully qualified filename (path and all). If
586 one matches, then the FileOptions will be used.
588 - When an ApplyTo specification matches an included file, the
589 options specified in the FileOptions resource will override
590 the default options specified on the Include record.
592 - Include records will be modified to permit referencing one or
593 more FileOptions resources. The FileOptions will be used
594 in the order listed on the Include record and the first
595 one that matches will be applied.
597 - Options (or specifications) currently supplied on the Include
598 record will be deprecated (i.e. removed in a later version a
599 year or so from now).
601 - The Exclude record will be deprecated as the same functionality
602 can be obtained by using an Exclude = yes in the FileOptions.
605 The following records can appear in the FileOptions resource. An
606 asterisk preceding the name indicates a feature not currently
610 - Compression= (GZIP, ...)
611 - Signature= (MD5, SHA1, ...)
613 - OneFs= (yes/no) - remain on one filesystem
614 - Recurse= (yes/no) - recurse into subdirectories
615 - Sparse= (yes/no) - do sparse file backup
616 - *Exclude= (yes/no) - exclude file from being saved
617 - *Reader= (filename) - external read (backup) program
618 - *Plugin= (filename) - read/write plugin module
621 - verify= (ipnougsamc5) - verify options
624 - replace= (always/ifnewer/ifolder/never) - replace options currently
626 - *Writer= (filename) - external write (restore) program
630 Currently options specifying compression, MD5 signatures, recursion,
631 ... of a FileSet are supplied on the Include record. These will now
632 all be collected into a FileOptions resource, which will be
633 specified on the Include in place of the options. Multiple FileOptions
634 may be specified. Since the FileOptions contain regular expressions
635 that are applied to the full filename, this will give the ability
636 to specify backup options on a file by file basis to whatever level
645 Include = compression=GZIP signature=MD5 {
654 Include = FileOptions=Opts {
665 That's a lot more to do the same thing, but it gives the ability to
666 apply options on a file by file basis. For example, suppose you
667 want to compress all files but not any file with extensions .gz or .Z.
668 You could do so as follows:
672 Include = FileOptions=NoCompress FileOptions=Opts {
680 ApplyTo = /*.?*/ # matches all files
685 # Note multiple ApplyTos are ORed
686 ApplyTo = /*.gz/ # matches .gz files */
687 ApplyTo = /*.Z/ # matches .Z files */
690 Now, since the NoCompress FileOptions is specified first on the
691 Include line, any *.gz or *.Z file will have an MD5 signature computed,
692 but will not be compressed. For all other files, the NoCompress will not
693 match, so the Opts options will be used which will include GZIP
697 - Is it necessary to provide some means of ANDing regular expressions
698 and negation? (not currently planned)
700 e.g. ApplyTo = /*.gz/ && !/big.gz/
702 - I see that Networker has a "null" module which, if specified, does not
703 backup the file, but does make an record of the file in the catalog
704 so that the catalog will reflect an exact picture of the filesystem.
705 The result is that the file can be "seen" when "browsing" the save
706 sets, but it cannot be restored.
708 Is this really useful? Should it be implemented in Bacula?
711 After implementing the above, the user will be able to specify
712 on a file by file basis (using regular expressions) what options are
713 applied for the backup.
714 ====================================
716 =========================================
717 Proposal by Bill Sellers
719 Return-Path: <w.a.sellers@larc.nasa.gov>
720 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
721 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)
722 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
723 X-Sender: w.a.sellers@pop.larc.nasa.gov
724 X-Mailer: QUALCOMM Windows Eudora Version 5.1
725 Date: Tue, 14 Jan 2003 16:30:18 -0500
726 To: Kern Sibbald <kern@sibbald.com>
727 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
728 Subject: Re: [Bacula-users] Bacula remote storage?
729 In-Reply-To: <1042565382.1845.177.camel@rufus>
730 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>
732 Content-Type: text/plain; charset="us-ascii"; format=flowed
733 X-Annoyance-Filter-Junk-Probability: 0
734 X-Annoyance-Filter-Classification: Mail
735 At 06:29 PM 1/14/2003 +0100, you wrote:
738 >Well, if you cannot put a Bacula client on the machine,
739 >then it is a big problem. If you know of some software
740 >that can do what you want, let me know, because I
741 >really just don't know how to do it -- at least not
747 We have been able to get Amanda to use the HSM as a storage
748 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
749 handle Windows systems very well (or at all without Samba). So I am
750 looking for a backup system that has a Windows client. I really like the
751 Windows integration of Bacula.
753 From the command line, its rather trivial to move the data around. We use
756 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
758 or if you use GNU tar:
760 tar czf hsm:path/file.tgz ./files
762 One idea for you to consider; Sendmail offers pipes in the aliases file;
763 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
764 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
765 pipe available, as a storage device? Then we could use any command that
766 handles stdin as a storage destination.
773 #Password is not used in rsh, but might be used in ftp.
774 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
781 Password = "foobar&-"
782 Device = "| ncftpput -c hsm /path/file.bacula"
786 >If you have some local storage available, you could
787 >use Bacula to backup to disk volumes, then use some
788 >other software (ftp, scp) to move them to the HSM
789 >machine. However, this is a bit kludgy.
792 It is, but maybe worth a try. Is there some function in Bacula to put
793 variables in filenames? i.e. backup.2003-01-15.root
800 w.a.sellers@larc.nasa.gov
802 ==============================================
803 The Project for the above
805 I finally realized that this is not at all
806 the same as reader/writer programs or plugins,
807 which are alternate ways of accessing the
808 files to be backed up. Rather, it is an alternate
809 form of storage device, and I have always planned
810 that Bacula should be able to handle all sorts
813 So, I propose the following phases:
815 1. OK from you to invest some time in testing
816 this as I implement it (requires that you
817 know how to download from the SourceForge
818 cvs -- which I imagine is a piece of cake
821 2. Dumb implementation by allowing a device to
822 be a fifo for write only.
823 Reason: easy to implement, proof of concept.
825 3. Try reading from fifo but with fixed block
827 Reason: proof of concept, easy to implement.
829 4. Extend reading from fifo (restores) to handle
831 Reason: requires some delicate low level coding
832 which could destabilize all of Bacula.
834 5. Implementation of above but to a program. E.g.
835 Device = "|program" (not full pipeline).
836 Reason: routines already exist, and program can
837 be a shell script which contains anything.
839 6. Add full pipeline as a possibility. E.g.
840 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
841 Reason: needs additional coding to implement full
842 pipeline (must fire off either a shell or all
843 programs and connect their pipes).
845 There are a good number of details in each step
846 that I have left out, but I will specify them at
847 every stage, and there may be a few changes as things
848 evolve. I expect that to get to stage 5 will take a
849 few weeks, and at that point, you will have
850 everything you need (just inside a script).
851 Stage 6 will probably take longer, but if this
852 project pleases you, what we do for 5 should
853 be adequate for some time.
857 =============================================
859 Done: (see kernsdone for more)
860 - Look into Pruning/purging problems or why there seem to
861 be so many files listed each night.
862 - Fix cancel in find_one -- need jcr.
863 - Cancel does not work for restore in FD.
864 - Write SetJobStatus() function so cancel status not lost.
865 - Add include list to end of chain in findlib
866 - Zap sd_auth_key after use
867 - Add Bar code reading capabilities (new mtx-changer)
868 - Figure out some way to automatically backup all local partitions
869 - Make hash table for linked files in findlib/find_one.c:161
871 - Rewrite find_one.c to use only pool_memory instead of
872 alloca and malloc (probably not necessary).
873 - Make sure btraceback goes into /sbin not sysconf directory.
874 - InitVerify is getting pruned and it shouldn't (document it)
875 - Make 1.28c release ??? NO do 1.29 directly
876 - Set timeout on opening fifo for save or restore (findlib)
877 - Document FIFO storage device.
878 - Document fifo and | and <
881 - Get correct error status from run_program or open_bpipe().
882 - Restrict permissions on File Volumes (now 0640).
883 - Umasked 022 daemons
884 - Fix restore of hard linked file.
885 - Figure out how to allow multiple simultaneous file Volumes on a single device.
886 - Implement multiple simultaneous file Volumes on a single device.
887 - Cleanup db_update_media and db_update_pool
888 - Flush all the daemon messages at the end of every job.
889 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
890 run_program -- bpipe.c
891 - Apparently cancel does not work for jobs waiting to be
893 - Implement TCP/IP connection for MySQL
894 - Pull a canceled job from the Scheduling queue.
895 - Implement max_file_size in block.c (already done, just tweaked).
896 - Look at purge jobs volume (at least document it, and see if it is
898 - Add list volumes does all pools. list volumes pool=xxx now works.
899 - Add pool= to "list media" in ua_output.c
900 - Strip trailing slashes from Include directory names in the FD.
901 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
902 ERR=Operation not permited loop.
903 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
904 - Produce better error messages in when error/eof writing block.
905 - Cancelling of a queued job does NOT work!!!!!!
907 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
908 rufus-sd: Recycled volume File0003 on device /home/kern/bacula/working, all previous data lost.
909 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
910 - Ability to backup to a file then later transfer to a tape -- Migration.
911 Migration based on MaxJobs(MinJobs),MaxVols(MinVols),AgeJobs,MaxBytes(MinBytes)
912 (i.e. HighwaterMark, LowwaterMark).
913 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
914 set of volumes recycling the oldest volume when it is needed.
915 - gethostbyname failure in bnet_connect() continues
916 generating errors -- should stop.
917 - Add chflags() code for FreeBSD file flags
918 - Bevan Anderson suggests having a run queue for each device
919 so that multiple simultaneous jobs can run but each writing
920 to a different Volume.
921 - Look at handling <> in smtp doesn't work with exim.
922 - Need to specify MaximumConcurrentJobs in the Job resource.
923 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
924 - Implement finer multiprocessing options.
925 - Implement | and < in Exclude statements.
926 - Figure out some way to specify a retention period for files
927 that no longer exist on the machine -- so that we maintain
928 say backups for 30 days, but if the file is deleted, we maintain
929 the last copy for 1 year. -- answer Volume retention.
930 - Make non-zero status from RunJobBefore/After error the job.
931 - Need define of int_least16_t in sha1.h for SuSE.
932 - Implement bar code reader for autochangers
933 - Document new MaximumConcurrentJob records (Job, Client, Storage)
934 - Write up how to use/manage disk Volume Storage. ******
935 - Remove kern and kelvin from mysql_grant...
936 - Install grant_mysql...
937 - Strip trailing / from Include
938 - add #define ENABLE_NLS for Gnome compile on SuSE.
939 - Add Client FS/OS id (Linux, Win95/98, ...).
940 - Concept of VolumeSet during restore which is a list
941 of Volume names needed.
942 - Turn virutally all sprintfs into snprintfs.
943 - Update volume=Test01 requests pool, then lists volumes.
944 **** Test select_pool_and_media ...
946 - Add IP address to authentication failures.
947 - Add a default File storage so that new users can do backup
948 and restores right away.