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 - Document Maximum File Size
13 ***** Write up how to use/manage disk Volume Storage. ******
14 - Add a section to the doc on Manual cycling of Volumes.
15 - Document new MaximumConcurrentJob records (Job, Client, Storage)
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 - Fix "access not allowed" for backup of files on WinXP.
24 - Issue message to mount a new tape before the rewind.
25 - Simplified client job initiation for portables.
26 - If SD cannot open a drive, make it periodically retry.
27 - Implement LabelTemplate (at least first cut).
28 - Add more of the config info to the tape label.
29 - Implement bar code reader for autochangers
30 - Add a default File storage so that new users can do backup
31 and restores right away.
32 - Implement a Mount Command and an Unmount Command where
33 the users could specify a system command to be performed
34 to do the mount, after which Bacula could attempt to
35 read the device. This is for Removeable media such as a CDROM.
36 - Most likely, this mount command would be invoked explicitly
37 by the user using the current Console "mount" and "unmount"
38 commands -- the Storage Daemon would do the right thing
39 depending on the exact nature of the device.
40 - As with tape drives, when Bacula wanted a new removable
41 disk mounted, it would unmount the old one, and send a message
42 to the user, who would then use "mount" as described above
43 once he had actually inserted the disk.
45 - Make some way so that if a machine is skipped because it is not up
46 that Bacula will continue retrying for a specified period of time --
48 - If tape is marked read-only, then try opening it read-only rather than
49 failing, and remember that it cannot be written.
50 - Refine SD waiting output:
51 Device is being positioned
52 > Device is being positioned for append
53 > Device is being positioned to file x
55 - Figure out some way to estimate output size and to avoid splitting
56 a backup across two Volumes -- this could be useful for writing CDROMs
57 where you really prefer not to have it split -- not serious.
58 - Add RunBeforeJob and RunAfterJob to the Client program.
59 - Have SD compute MD5 or SHA1 and compare to what FD computes.
60 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
61 actual data on the Volume and compare it.
62 - Implement FileOptions (see end of this document)
63 - Implement Bacula plugins -- design API
64 - Make bcopy read through bad tape records.
65 - Fix read_record to handle multiple sessions.
66 - Program files (i.e. execute a program to read/write files).
67 Pass read date of last backup, size of file last time.
68 - Add Signature type to File DB record.
69 - Make Restore report an error if FD or SD term codes are not OK.
70 - CD into subdirectory when open()ing files for backup to
71 speed up things. Test with testfind().
72 - Add prefixlinks to where or not where absolute links to FD.
73 - Priority job to go to top of list.
74 - Find out why Full saves run slower and slower (hashing?)
75 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
76 - Implement some way for the Console to dynamically create a job.
77 - Restore to a particular time -- e.g. before date, after date.
78 - Implement disk spooling
79 - Solaris -I on tar for include list
80 - Enable avoid backing up archive device (findlib/find_one.c:128)
81 - Need a verbose mode in restore, perhaps to bsr.
82 - bscan without -v is too quiet -- perhaps show jobs.
83 - Add code to reject whole blocks if not wanted on restore.
84 - Start working on Base jobs.
85 - Make sure the MaxVolFiles is fully implemented in SD
86 - Check if both CatalogFiles and UseCatalog are set to SD.
87 - Check if we can increase Bacula FD priorty in Win2000
88 - Need return status on read_cb() from read_records(). Need multiple
89 records -- one per Job, maybe a JCR or some other structure with
91 - Figure out how to do a bare metal Windows restore
92 - Put system type returned by FD into catalog.
93 - Possibly add email to Watchdog if drive is unmounted too
94 long and a job is waiting on the drive.
95 - Use read_record.c in SD code.
96 - Why don't we get an error message from Win32 FD when bootstrap
97 file cannot be created for restore command?
98 - When Marking a file in Restore that is a hard link, also
99 mark the link so that the data will be reloaded.
100 - Restore program that errors in SD due to no tape reports
101 OK incorrectly in output.
102 - After unmount, if restore job started, ask to mount.
103 - Convert all %x substitution variables, which are hard to remember
104 and read to %(variable-name). Idea from TMDA.
105 - Add JobLevel in FD status (but make sure it is defined).
106 - Make Pool resource handle Counter resources.
107 - Remove NextId for SQLite. Optimize.
108 - Strip trailing / from Include
109 - Move all SQL statements into a single location.
110 - Add UA rc and history files.
111 - put termcap (used by console) in ./configure and
112 allow -with-termcap-dir.
113 - Enhance time and size scanning routines.
114 - Fix Autoprune for Volumes to respect need for full save.
115 - Fix Win32 config file definition name on /install
116 - No READLINE_SRC if found in alternate directory.
117 - Add Client FS/OS id (Linux, Win95/98, ...).
118 - Test a second language e.g. french.
119 - Compare tape to Client files (attributes, or attributes and data)
120 - Make all database Ids 64 bit.
121 - Write an applet for Linux.
122 - Add estimate to Console commands
123 - Find solution to blank filename (i.e. path only) problem.
124 - Implement new daemon communications protocol.
125 - Remove PoolId from Job table, it exists in Media.
126 - Allow console commands to detach or run in background.
127 - Fix status delay on storage daemon during rewind.
128 - Add SD message variables to control operator wait time
129 - Maximum Operator Wait
130 - Minimum Message Interval
131 - Maximum Message Interval
132 - Send Operator message when cannot read tape label.
133 - Verify level=Volume (scan only), level=Data (compare of data to file).
134 Verify level=Catalog, level=InitCatalog
136 - Add keyword search to show command in Console.
137 - Fix Win2000 error with no messages during startup.
138 - Events : tape has more than xxx bytes.
139 - Restrict characters permitted in a Resource name.
140 - Complete code in Bacula Resources -- this will permit
141 reading a new config file at any time.
142 - Handle ctl-c in Console
143 - Implement script driven addition of File daemon to config files.
144 - Think about how to make Bacula work better with File (non-tape) archives.
145 - Write Unix emulator for Windows.
147 - Implement new serialize subroutines
148 send(socket, "string", &Vol, "uint32", &i, NULL)
149 - Audit all UA commands to ensure that we always prompt where possible.
150 - If ./btape is called without /dev, assume argument is a Storage resource name.
151 - Put memory utilization in Status output of each daemon
152 if full status requested or if some level of debug on.
153 - Make database type selectable by .conf files i.e. at runtime
154 - Set flag for uname -a. Add to Volume label.
155 - Implement throttled work queue.
156 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
157 - Allow multiple Storage specifications (or multiple names on
158 a single Storage specification) in the Job record. Thus a job
159 can be backed up to a number of storage devices.
160 - Implement dump label to UA
161 - Concept of VolumeSet during restore which is a list
162 of Volume names needed.
163 - Restore files modified after date
164 - Restore file modified before date
165 - Emergency restore info:
167 - Backup working directory
169 - Restore -- do nothing but show what would happen
170 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
171 - Implement Restore FileSet=
172 - Create a protocol.h and protocol.c where all protocol messages
174 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
175 - Timout a job or terminate if link goes down, or reopen link and query.
176 - Find general solution for sscanf size problems (as well
177 as sprintf. Do at run time?
178 - Concept of precious tapes (cannot be reused).
179 - Make bcopy copy with a single tape drive.
180 - Permit changing ownership during restore.
182 - Autolabel should be specified by DIR instead of SD.
183 - Find out how to get the system tape block limits, e.g.:
184 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
185 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
188 - AutoScan (check checksum of tape)
189 - Format command = "format /dev/nst0"
193 - Seek resolution (usually corresponds to buffer size)
194 - EODErrorCode=ENOSPC or code
195 - Partial Read error code
196 - Partial write error code
197 - Nonformatted read error
198 - Nonformatted write error
199 - WriteProtected error
203 - IgnoreCloseErrors=yes
213 - FD sends unsaved file list to Director at end of job (see
215 - File daemon should build list of files skipped, and then
216 at end of save retry and report any errors.
217 - Write a Storage daemon that uses pipes and
218 standard Unix programs to write to the tape.
220 - Need something that monitors the JCR queue and
221 times out jobs by asking the deamons where they are.
222 - Enhance Jmsg code to permit buffering and saving to disk.
223 - device driver = "xxxx" for drives.
224 - restart: paranoid: read label fsf to
225 eom read append block, and go
226 super-paranoid: read label, read all files
227 in between, read append block, and go
228 verify: backspace, read append block, and go
229 permissive: same as above but frees drive
230 if tape is not valid.
232 - Ensure that /dev/null works
233 - Need report class for messages. Perhaps
234 report resource where report=group of messages
235 - enhance scan_attrib and rename scan_jobtype, and
236 fill in code for "since" option
237 - Director needs a time after which the report status is sent
238 anyway -- or better yet, a retry time for the job.
239 Don't reschedule a job if previous incarnation is still running.
240 - Some way to automatically backup everything is needed????
241 - Need a structure for pending actions:
243 - termination status (part of buffered msgs?)
244 - Concept of grouping Storage devices and job can use
245 any of a number of devices
247 Read, Write, Clean, Delete
248 - Login to Bacula; Bacula users with different permissions:
249 owner, group, user, quotas
250 - Store info on each file system type (probably in the job header on tape.
251 This could be the output of df; or perhaps some sort of /etc/mtab record.
254 - Design at hierarchial storage for Bacula. Migration and Clone.
255 - Implement FSM (File System Modules).
256 - Identify unchanged or "system" files and save them to a
257 special tape thus removing them from the standard
258 backup FileSet -- BASE backup.
259 - Turn virutally all sprintfs into snprintfs.
260 - Heartbeat between daemons.
261 - Audit M_ error codes to ensure they are correct and consistent.
262 - Add variable break characters to lex analyzer.
263 Either a bit mask or a string of chars so that
264 the caller can change the break characters.
265 - Make a single T_BREAK to replace T_COMMA, etc.
266 - Ensure that File daemon and Storage daemon can
267 continue a save if the Director goes down (this
268 is NOT currently the case). Must detect socket error,
269 buffer messages for later.
270 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
271 - Add ability to backup to two Storage devices (two SD sessions) at
272 the same time -- e.g. onsite, offsite.
273 - Add the ability to consolidate old backup sets (basically do a restore
274 to tape and appropriately update the catalog). Compress Volume sets.
275 Might need to spool via file is only one drive is available.
276 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
277 someway to do so with every volume that has less than x% valid
282 Bacula Projects Roadmap
284 last update 5 January 2003
286 Item 1: Multiple simultaneous Jobs. (done)
287 Done -- Restore part needs better implementation to work correctly
289 What: Permit multiple simultaneous jobs in Bacula.
291 Why: An enterprise level solution needs to go fast without the
292 need for the system administrator to carefully tweak
293 timing. Based on the benchmarks, during a full
294 backup, NetWorker typically hit 10 times the bandwidth to
295 the tape compared to Bacula--largely. This is probably due to
296 running parallel jobs and multi-threaded filling of buffers
297 and writing them to tape. This should also make things work
298 better when you have a mix of fast and slow machines backing
301 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
302 implementing this is a matter of some small cleanups and
306 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
307 Deferred -- not necessary yet.
309 What: If data is coming into the SD too fast, buffer it to
310 disk if the user has configured this option.
312 Why: This would be nice, especially if it more or less falls out
313 when implementing (1) above. If not, it probably should not
314 be given a high priority because fundamentally the backup time
315 is limited by the tape bandwidth. Even though you may finish a
316 client job quicker by spilling to disk, you still have to
317 eventually get it onto tape. If intermediate disk buffering
318 allows us to improve write bandwidth to tape, it may make
321 Notes: Whether or not this is implemented will depend upon performance
322 testing after item 1 is implemented.
325 Item 3: Write the bscan program -- also write a bcopy program.
328 What: Write a program that reads a Bacula tape and puts all the
329 appropriate data into the catalog. This allows recovery
330 from a tape that is no longer in the database, or it allows
331 re-creation of a database if lost.
333 Why: This is a fundamental robustness and disaster recovery tool
334 which will increase the comfort level of a sysadmin
335 considering adopting Bacula.
337 Notes: A skeleton of this program already exists, but much work
338 needs to be done. Implementing this will also make apparent
339 any deficiencies in the current Bacula tape format.
342 Item 4: Implement Base jobs.
344 What: A base job is sort of like a Full save except that you
345 will want the FileSet to contain only files that are unlikely
346 to change in the future (i.e. a snapshot of most of your
347 system after installing it). After the base job has been run,
348 when you are doing a Full save, you can specify to exclude
349 all files saved by the base job that have not been modified.
351 Why: This is something none of the competition does, as far as we know
352 (except BackupPC, which is a Perl program that saves to disk
353 only). It is big win for the user, it makes Bacula stand out
354 as offering a unique optimization that immediately saves time
357 Notes: Big savings in tape usage. Will require more resources because
358 the e. DIR must send FD a list of files/attribs, and the FD must
359 search the list and compare it for each file to be saved.
362 Item 5: Implement Label templates
364 What: This is a mechanism whereby Bacula can automatically create
365 a tape label for new tapes according to a detailed specification
366 provided by the user.
368 Why: It is a major convenience item for folks who use automated label
371 Notes: Bacula already has a working form of automatic tape label
372 creation, but it is very crude. The design for the complete
373 tape labeling project is already documented in the manual.
376 Item 6: Write a regression script.
379 What: This is an automatic script that runs and tests as many features
380 of Bacula as possible. The output is compared to previous
381 versions of Bacula and any differences are reported.
383 Why: This is an enormous help in preventing introduction of new
384 errors in parts of the program that already work correctly.
386 Notes: This probably should be ranked higher, it's something the typical
387 user doesn't see. Depending on how it's implemented, it may
388 make sense to defer it until the archival tape format and
389 user interface mature.
392 Item 7: GUI for interactive restore
393 Item 8: GUI for interactive backup
395 What: The current interactive restore is implemented with a tty
396 interface. It would be much nicer to be able to "see" the
397 list of files backed up in typical GUI tree format.
398 The same mechanism could also be used for creating
399 ad-hoc backup FileSets (item 8).
401 Why: Ease of use -- especially for the end user.
403 Notes: Rather than implementing in Gtk, we probably should go directly
404 for a Browser implementation, even if doing so meant the
405 capability wouldn't be available until much later. Not only
406 is there the question of Windows sites, most
407 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
408 without installing lots of stuff admins are very wary about.
409 Real sysadmins will always use the command line anyway, and
410 the user who's doing an interactive restore or backup of his
411 own files will in most cases be on a Windows machine running
415 Item 9: Add SSL to daemon communications.
417 What: This provides for secure communications between the daemons.
419 Why: This would allow doing backup across the Internet without
420 privacy concerns (or with much less concern).
422 Notes: The vast majority of near term potential users will be backing up
423 a single site over a LAN and, correctly or not, they probably
424 won't be concerned with security, at least not enough to go to
425 the trouble to set up keys, etc. to screw things down. We suspect
426 that many users genuinely interested in multi-site backup
427 already run some form of VPN software in their internetwork
428 connections, and are willing to delegate security to that layer.
431 Item 10: Define definitive tape format.
434 What: Define that definitive tape format that will not change
435 for the next millennium.
437 Why: Stability, security.
439 Notes: See notes for item 11 below.
442 Item 11: New daemon communication protocol.
444 What: The current daemon to daemon protocol is basically an ASCII
445 printf() and sending the buffer. On the receiving end, the
446 buffer is sscanf()ed to unpack it. The new scheme would
447 be a binary format that allows quick packing and unpacking
448 of any data type with named fields.
450 Why: Using binary packing would be faster. Named fields will permit
451 error checking to ensure that what is sent is what the
452 receiver really wants.
454 Notes: These are internal improvements in the interest of the
455 long-term stability and evolution of the program. On the one
456 hand, the sooner they're done, the less code we have to rip
457 up when the time comes to install them. On the other hand, they
458 don't bring an immediately perceptible benefit to potential
459 users. Item 10 and possibly item 11 should be deferred until Bacula
460 is well established with a growing user community more or
461 less happy with the feature set. At that time, it will make a
462 good "next generation" upgrade in the interest of data
468 ======================================================
470 It is somewhat like a Full save becomes an incremental since
471 the Base job (or jobs) plus other non-base files.
473 - New BaseFile table that contains:
474 JobId, BaseJobId, FileId (from Base).
475 i.e. for each base file that exists but is not saved because
476 it has not changed, the File daemon sends the JobId, BaseId,
477 and FileId back to the Director who creates the DB entry.
478 - To initiate a Base save, the Director sends the FD
479 the FileId, and full filename for each file in the Base.
480 - When the FD finds a Base file, he requests the Director to
481 send him the full File entry (stat packet plus MD5), or
482 conversely, the FD sends it to the Director and the Director
483 says yes or no. This can be quite rapid if the FileId is kept
484 by the FD for each Base Filename.
485 - It is probably better to have the comparison done by the FD
486 despite the fact that the File entry must be sent across the
488 - An alternative would be to send the FD the whole File entry
489 from the start. The disadvantage is that it requires a lot of
490 space. The advantage is that it requires less communications
492 - The Job record must be updated to indicate that one or more
494 - At end of Job, FD returns:
495 1. Count of base files/bytes not written to tape (i.e. matches)
496 2. Count of base file that were saved i.e. had changed.
497 - No tape record would be written for a Base file that matches, in the
498 same way that no tape record is written for Incremental jobs where
499 the file is not saved because it is unchanged.
500 - On a restore, all the Base file records must explicitly be
501 found from the BaseFile tape. I.e. for each Full save that is marked
502 to have one or more Base Jobs, search the BaseFile for all occurrences
504 - An optimization might be to make the BaseFile have:
510 This would avoid the need to explicitly fetch each File record for
511 the Base job. The Base Job record will be fetched to get the
512 VolSessionId and VolSessionTime.
513 =========================================================
515 ==========================================================
517 For each Incremental job that is run, there may be files that
518 were found but not saved because they were locked (this applies
519 only to Windows). Such a system could send back to the Director
520 a list of Unsaved files.
522 - New UnSavedFiles table that contains:
526 - Then in the next Incremental job, the list of Unsaved Files will be
527 feed to the FD, who will ensure that they are explicitly chosen even
528 if standard date/time check would not have selected them.
529 =============================================================
533 =============================================================
535 Request For Comments For File Backup Options
538 Subject: File Backup Options
541 A few days ago, a Bacula user who is backing up to file volumes and
542 using compression asked if it was possible to suppress compressing
543 all .gz files since it was a waste of CPU time. Although Bacula
544 currently permits using different options (compression, ...) on
545 a directory by directory basis, it cannot do it on a file by
546 file basis, which is clearly what was desired.
548 Proposed Implementation:
549 To solve this problem, I propose the following:
551 - Add a new Director resource type called FileOptions.
553 - The FileOptions resource will have records for all
554 options that can currently be specified on the Include record
555 (in a FileSet). Examples below.
557 - The FileOptions resource will permit an exclude option as well
558 as a number of additional options.
560 - The heart of the FileOptions resource is the ability to
561 supply any number of ApplyTo records which specify POSIX
562 regular expressions. These ApplyTo regular expressions are
563 applied to the fully qualified filename (path and all). If
564 one matches, then the FileOptions will be used.
566 - When an ApplyTo specification matches an included file, the
567 options specified in the FileOptions resource will override
568 the default options specified on the Include record.
570 - Include records will be modified to permit referencing one or
571 more FileOptions resources. The FileOptions will be used
572 in the order listed on the Include record and the first
573 one that matches will be applied.
575 - Options (or specifications) currently supplied on the Include
576 record will be deprecated (i.e. removed in a later version a
577 year or so from now).
579 - The Exclude record will be deprecated as the same functionality
580 can be obtained by using an Exclude = yes in the FileOptions.
583 The following records can appear in the FileOptions resource. An
584 asterisk preceding the name indicates a feature not currently
588 - Compression= (GZIP, ...)
589 - Signature= (MD5, SHA1, ...)
591 - OneFs= (yes/no) - remain on one filesystem
592 - Recurse= (yes/no) - recurse into subdirectories
593 - Sparse= (yes/no) - do sparse file backup
594 - *Exclude= (yes/no) - exclude file from being saved
595 - *Reader= (filename) - external read (backup) program
596 - *Plugin= (filename) - read/write plugin module
599 - verify= (ipnougsamc5) - verify options
602 - replace= (always/ifnewer/ifolder/never) - replace options currently
604 - *Writer= (filename) - external write (restore) program
608 Currently options specifying compression, MD5 signatures, recursion,
609 ... of a FileSet are supplied on the Include record. These will now
610 all be collected into a FileOptions resource, which will be
611 specified on the Include in place of the options. Multiple FileOptions
612 may be specified. Since the FileOptions contain regular expressions
613 that are applied to the full filename, this will give the ability
614 to specify backup options on a file by file basis to whatever level
623 Include = compression=GZIP signature=MD5 {
632 Include = FileOptions=Opts {
643 That's a lot more to do the same thing, but it gives the ability to
644 apply options on a file by file basis. For example, suppose you
645 want to compress all files but not any file with extensions .gz or .Z.
646 You could do so as follows:
650 Include = FileOptions=NoCompress FileOptions=Opts {
658 ApplyTo = /*.?*/ # matches all files
663 # Note multiple ApplyTos are ORed
664 ApplyTo = /*.gz/ # matches .gz files */
665 ApplyTo = /*.Z/ # matches .Z files */
668 Now, since the NoCompress FileOptions is specified first on the
669 Include line, any *.gz or *.Z file will have an MD5 signature computed,
670 but will not be compressed. For all other files, the NoCompress will not
671 match, so the Opts options will be used which will include GZIP
675 - Is it necessary to provide some means of ANDing regular expressions
676 and negation? (not currently planned)
678 e.g. ApplyTo = /*.gz/ && !/big.gz/
680 - I see that Networker has a "null" module which, if specified, does not
681 backup the file, but does make an record of the file in the catalog
682 so that the catalog will reflect an exact picture of the filesystem.
683 The result is that the file can be "seen" when "browsing" the save
684 sets, but it cannot be restored.
686 Is this really useful? Should it be implemented in Bacula?
689 After implementing the above, the user will be able to specify
690 on a file by file basis (using regular expressions) what options are
691 applied for the backup.
692 ====================================
694 =========================================
695 Proposal by Bill Sellers
697 Return-Path: <w.a.sellers@larc.nasa.gov>
698 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
699 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)
700 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
701 X-Sender: w.a.sellers@pop.larc.nasa.gov
702 X-Mailer: QUALCOMM Windows Eudora Version 5.1
703 Date: Tue, 14 Jan 2003 16:30:18 -0500
704 To: Kern Sibbald <kern@sibbald.com>
705 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
706 Subject: Re: [Bacula-users] Bacula remote storage?
707 In-Reply-To: <1042565382.1845.177.camel@rufus>
708 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>
710 Content-Type: text/plain; charset="us-ascii"; format=flowed
711 X-Annoyance-Filter-Junk-Probability: 0
712 X-Annoyance-Filter-Classification: Mail
713 At 06:29 PM 1/14/2003 +0100, you wrote:
716 >Well, if you cannot put a Bacula client on the machine,
717 >then it is a big problem. If you know of some software
718 >that can do what you want, let me know, because I
719 >really just don't know how to do it -- at least not
725 We have been able to get Amanda to use the HSM as a storage
726 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
727 handle Windows systems very well (or at all without Samba). So I am
728 looking for a backup system that has a Windows client. I really like the
729 Windows integration of Bacula.
731 From the command line, its rather trivial to move the data around. We use
734 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
736 or if you use GNU tar:
738 tar czf hsm:path/file.tgz ./files
740 One idea for you to consider; Sendmail offers pipes in the aliases file;
741 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
742 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
743 pipe available, as a storage device? Then we could use any command that
744 handles stdin as a storage destination.
751 #Password is not used in rsh, but might be used in ftp.
752 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
759 Password = "foobar&-"
760 Device = "| ncftpput -c hsm /path/file.bacula"
764 >If you have some local storage available, you could
765 >use Bacula to backup to disk volumes, then use some
766 >other software (ftp, scp) to move them to the HSM
767 >machine. However, this is a bit kludgy.
770 It is, but maybe worth a try. Is there some function in Bacula to put
771 variables in filenames? i.e. backup.2003-01-15.root
778 w.a.sellers@larc.nasa.gov
780 ==============================================
781 The Project for the above
783 I finally realized that this is not at all
784 the same as reader/writer programs or plugins,
785 which are alternate ways of accessing the
786 files to be backed up. Rather, it is an alternate
787 form of storage device, and I have always planned
788 that Bacula should be able to handle all sorts
791 So, I propose the following phases:
793 1. OK from you to invest some time in testing
794 this as I implement it (requires that you
795 know how to download from the SourceForge
796 cvs -- which I imagine is a piece of cake
799 2. Dumb implementation by allowing a device to
800 be a fifo for write only.
801 Reason: easy to implement, proof of concept.
803 3. Try reading from fifo but with fixed block
805 Reason: proof of concept, easy to implement.
807 4. Extend reading from fifo (restores) to handle
809 Reason: requires some delicate low level coding
810 which could destabilize all of Bacula.
812 5. Implementation of above but to a program. E.g.
813 Device = "|program" (not full pipeline).
814 Reason: routines already exist, and program can
815 be a shell script which contains anything.
817 6. Add full pipeline as a possibility. E.g.
818 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
819 Reason: needs additional coding to implement full
820 pipeline (must fire off either a shell or all
821 programs and connect their pipes).
823 There are a good number of details in each step
824 that I have left out, but I will specify them at
825 every stage, and there may be a few changes as things
826 evolve. I expect that to get to stage 5 will take a
827 few weeks, and at that point, you will have
828 everything you need (just inside a script).
829 Stage 6 will probably take longer, but if this
830 project pleases you, what we do for 5 should
831 be adequate for some time.
835 =============================================
837 Done: (see kernsdone for more)
838 - Look into Pruning/purging problems or why there seem to
839 be so many files listed each night.
840 - Fix cancel in find_one -- need jcr.
841 - Cancel does not work for restore in FD.
842 - Write SetJobStatus() function so cancel status not lost.
843 - Add include list to end of chain in findlib
844 - Zap sd_auth_key after use
845 - Add Bar code reading capabilities (new mtx-changer)
846 - Figure out some way to automatically backup all local partitions
847 - Make hash table for linked files in findlib/find_one.c:161
849 - Rewrite find_one.c to use only pool_memory instead of
850 alloca and malloc (probably not necessary).
851 - Make sure btraceback goes into /sbin not sysconf directory.
852 - InitVerify is getting pruned and it shouldn't (document it)
853 - Make 1.28c release ??? NO do 1.29 directly
854 - Set timeout on opening fifo for save or restore (findlib)
855 - Document FIFO storage device.
856 - Document fifo and | and <
859 - Get correct error status from run_program or open_bpipe().
860 - Restrict permissions on File Volumes (now 0640).
861 - Umasked 022 daemons
862 - Fix restore of hard linked file.
863 - Figure out how to allow multiple simultaneous file Volumes on a single device.
864 - Implement multiple simultaneous file Volumes on a single device.
865 - Cleanup db_update_media and db_update_pool
866 - Flush all the daemon messages at the end of every job.
867 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
868 run_program -- bpipe.c
869 - Apparently cancel does not work for jobs waiting to be
871 - Implement TCP/IP connection for MySQL
872 - Pull a canceled job from the Scheduling queue.
873 - Implement max_file_size in block.c (already done, just tweaked).
874 - Look at purge jobs volume (at least document it, and see if it is
876 - Add list volumes does all pools. list volumes pool=xxx now works.
877 - Add pool= to "list media" in ua_output.c
878 - Strip trailing slashes from Include directory names in the FD.
879 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
880 ERR=Operation not permited loop.
881 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
882 - Produce better error messages in when error/eof writing block.
883 - Cancelling of a queued job does NOT work!!!!!!
885 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
886 rufus-sd: Recycled volume File0003 on device /home/kern/bacula/working, all previous data lost.
887 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
888 - Ability to backup to a file then later transfer to a tape -- Migration.
889 Migration based on MaxJobs(MinJobs),MaxVols(MinVols),AgeJobs,MaxBytes(MinBytes)
890 (i.e. HighwaterMark, LowwaterMark).
891 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
892 set of volumes recycling the oldest volume when it is needed.
893 - gethostbyname failure in bnet_connect() continues
894 generating errors -- should stop.
895 - Add chflags() code for FreeBSD file flags
896 - Bevan Anderson suggests having a run queue for each device
897 so that multiple simultaneous jobs can run but each writing
898 to a different Volume.
899 - Look at handling <> in smtp doesn't work with exim.
900 - Need to specify MaximumConcurrentJobs in the Job resource.
901 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
902 - Implement finer multiprocessing options.
903 - Implement | and < in Exclude statements.
904 - Figure out some way to specify a retention period for files
905 that no longer exist on the machine -- so that we maintain
906 say backups for 30 days, but if the file is deleted, we maintain
907 the last copy for 1 year. -- answer Volume retention.