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.
13 Testing to do: (painful)
14 - that console command line options work
15 - blocksize recognition code.
16 - multiple simultaneous Volumes
19 - Figure out some way to specify a retention period for files
20 that no longer exist on the machine -- so that we maintain
21 say backups for 30 days, but if the file is deleted, we maintain
22 the last copy for 1 year.
23 - Have Bacula "poll" the tape to see if it is there.
24 - Add more of the config info to the tape label.
25 - Implement bar code reader for autochangers
26 - Add a default File storage so that new users can do backup
27 and restores right away.
28 - Implement a Mount Command and an Unmount Command where
29 the users could specify a system command to be performed
30 to do the mount, after which Bacula could attempt to
31 read the device. This is for Removeable media such as a
33 - Most likely, this mount command would be invoked explicitly
34 by the user using the current Console "mount" and "unmount"
35 commands -- the Storage Daemon would do the right thing
36 depending on the exact nature of the device.
37 - As with tape drives, when Bacula wanted a new removable
38 disk mounted, it would unmount the old one, and send a message
39 to the user, who would then use "mount" as described above
40 once he had actually inserted the disk.
42 - Make some way so that if a machine is skipped because it is not up
43 that Bacula will continue retrying for a specified period of time --
45 - Figure out some way to estimate output size and to avoid splitting
46 a backup across two Volumes -- this could be useful for writing CDROMs
47 where you really prefer not to have it split -- not serious.
48 - Add chflags() code for FreeBSD file flags
49 - Add RunBeforeJob and RunAfterJob to the Client program.
50 - Have SD compute MD5 or SHA1 and compare to what FD computes.
51 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
52 actual data on the Volume and compare it.
53 - Implement FileOptions (see end of this document)
54 - Implement Bacula plugins -- design API
55 - Make bcopy read through bad tape records.
56 - Fix read_record to handle multiple sessions.
57 - Program files (i.e. execute a program to read/write files).
58 Pass read date of last backup, size of file last time.
59 - Add Signature type to File DB record.
60 - Make Restore report an error if FD or SD term codes are not OK.
61 - CD into subdirectory when open()ing files for backup to
62 speed up things. Test with testfind().
63 - Add prefixlinks to where or not where absolute links to FD.
64 - Look at handling <> in smtp doesn't work with exim.
65 - Priority job to go to top of list.
66 - Implement Bar code handling
67 - Why is catreq.c:111 Find vol called twice for a job?
68 - Find out why Full saves run slower and slower (hashing?)
69 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
70 - Implement some way for the Console to dynamically create a job.
71 - Restore to a particular time -- e.g. before date, after date.
72 - Implement disk spooling
73 - Implement finer multiprocessing options.
74 - Solaris -I on tar for include list
75 - Enable avoid backing up archive device (findlib/find_one.c:128)
76 - Need a verbose mode in restore, perhaps to bsr.
77 - bscan without -v is too quiet -- perhaps show jobs.
78 - Add code to reject whole blocks if not wanted on restore.
79 - Start working on Base jobs.
80 - Make sure the MaxVolFiles is fully implemented in SD
81 - Check if both CatalogFiles and UseCatalog are set to SD.
82 - Check if we can increase Bacula FD priorty in Win2000
83 - Need return status on read_cb() from read_records(). Need multiple
84 records -- one per Job, maybe a JCR or some other structure with
86 - Figure out how to do a bare metal Windows restore
87 - Put system type returned by FD into catalog.
88 - Need to specify MaximumConcurrentJobs in the Job resource.
89 - Possibly add email to Watchdog if drive is unmounted too
90 long and a job is waiting on the drive.
91 - Strip trailing slashes from Include directory names in the FD.
92 - Use read_record.c in SD code.
93 - Why don't we get an error message from Win32 FD when bootstrap
94 file cannot be created for restore command?
95 - When Marking a file in Restore that is a hard link, also
96 mark the link so that the data will be reloaded.
97 - Restore program that errors in SD due to no tape reports
98 OK incorrectly in output.
99 - After unmount, if restore job started, ask to mount.
100 - Convert all %x substitution variables, which are hard to remember
101 and read to %(variable-name). Idea from TMDA.
102 - Add JobLevel in FD status (but make sure it is defined).
103 - Make Pool resource handle Counter resources.
104 - Remove NextId for SQLite. Optimize.
105 - Strip trailing / from Include
106 - Move all SQL statements into a single location.
107 - Add UA rc and history files.
108 - put termcap (used by console) in ./configure and
109 allow -with-termcap-dir.
110 - Enhance time and size scanning routines.
111 - Fix Autoprune for Volumes to respect need for full save.
112 - Fix Win32 config file definition name on /install
113 - No READLINE_SRC if found in alternate directory.
114 - Add Client FS/OS id (Linux, Win95/98, ...).
115 - Test a second language e.g. french.
116 - Compare tape to Client files (attributes, or attributes and data)
117 - Make all database Ids 64 bit.
118 - Write an applet for Linux.
119 - Add estimate to Console commands
120 - Find solution to blank filename (i.e. path only) problem.
121 - Implement new daemon communications protocol.
122 - Remove PoolId from Job table, it exists in Media.
123 - Allow console commands to detach or run in background.
124 - Fix status delay on storage daemon during rewind.
125 - Add SD message variables to control operator wait time
126 - Maximum Operator Wait
127 - Minimum Message Interval
128 - Maximum Message Interval
129 - Send Operator message when cannot read tape label.
130 - Verify level=Volume (scan only), level=Data (compare of data to file).
131 Verify level=Catalog, level=InitCatalog
133 - Add keyword search to show command in Console.
134 - Fix Win2000 error with no messages during startup.
135 - Events : tape has more than xxx bytes.
136 - Restrict characters permitted in a Resource name.
137 - Complete code in Bacula Resources -- this will permit
138 reading a new config file at any time.
139 - Handle ctl-c in Console
140 - Implement LabelTemplate (at least first cut).
141 - Implement script driven addition of File daemon to config files.
142 - Think about how to make Bacula work better with File (non-tape) archives.
143 - Write Unix emulator for Windows.
145 - Implement new serialize subroutines
146 send(socket, "string", &Vol, "uint32", &i, NULL)
147 - Audit all UA commands to ensure that we always prompt where possible.
148 - If ./btape is called without /dev, assume argument is a Storage resource name.
149 - Put memory utilization in Status output of each daemon
150 if full status requested or if some level of debug on.
151 - Make database type selectable by .conf files i.e. at runtime
152 - gethostbyname failure in bnet_connect() continues
153 generating errors -- should stop.
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 - If SD cannot open a drive, make it periodically retry.
175 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
176 - Timout a job or terminate if link goes down, or reopen link and query.
177 - Find general solution for sscanf size problems (as well
178 as sprintf. Do at run time?
179 - Concept of precious tapes (cannot be reused).
180 - Make bcopy copy with a single tape drive.
181 - Permit changing ownership during restore.
183 - Autolabel should be specified by DIR instead of SD.
184 - Find out how to get the system tape block limits, e.g.:
185 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
186 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
189 - AutoScan (check checksum of tape)
190 - Format command = "format /dev/nst0"
194 - Seek resolution (usually corresponds to buffer size)
195 - EODErrorCode=ENOSPC or code
196 - Partial Read error code
197 - Partial write error code
198 - Nonformatted read error
199 - Nonformatted write error
200 - WriteProtected error
204 - IgnoreCloseErrors=yes
214 - FD sends unsaved file list to Director at end of job (see
216 - File daemon should build list of files skipped, and then
217 at end of save retry and report any errors.
218 - Write a Storage daemon that uses pipes and
219 standard Unix programs to write to the tape.
221 - Need something that monitors the JCR queue and
222 times out jobs by asking the deamons where they are.
223 - Enhance Jmsg code to permit buffering and saving to disk.
224 - device driver = "xxxx" for drives.
225 - restart: paranoid: read label fsf to
226 eom read append block, and go
227 super-paranoid: read label, read all files
228 in between, read append block, and go
229 verify: backspace, read append block, and go
230 permissive: same as above but frees drive
231 if tape is not valid.
233 - Ensure that /dev/null works
234 - Need report class for messages. Perhaps
235 report resource where report=group of messages
236 - enhance scan_attrib and rename scan_jobtype, and
237 fill in code for "since" option
238 - Director needs a time after which the report status is sent
239 anyway -- or better yet, a retry time for the job.
240 Don't reschedule a job if previous incarnation is still running.
241 - Some way to automatically backup everything is needed????
242 - Need a structure for pending actions:
244 - termination status (part of buffered msgs?)
245 - Concept of grouping Storage devices and job can use
246 any of a number of devices
248 Read, Write, Clean, Delete
249 - Login to Bacula; Bacula users with different permissions:
250 owner, group, user, quotas
251 - Store info on each file system type (probably in the job header on tape.
252 This could be the output of df; or perhaps some sort of /etc/mtab record.
255 - Design at hierarchial storage for Bacula.
256 - Implement FSM (File System Modules).
257 - Identify unchanged or "system" files and save them to a
258 special tape thus removing them from the standard
259 backup FileSet -- BASE backup.
260 - Turn virutally all sprintfs into snprintfs.
261 - Heartbeat between daemons.
262 - Audit M_ error codes to ensure they are correct and consistent.
263 - Add variable break characters to lex analyzer.
264 Either a bit mask or a string of chars so that
265 the caller can change the break characters.
266 - Make a single T_BREAK to replace T_COMMA, etc.
267 - Ensure that File daemon and Storage daemon can
268 continue a save if the Director goes down (this
269 is NOT currently the case). Must detect socket error,
270 buffer messages for later.
271 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
272 - Add ability to backup to two Storage devices (two SD sessions) at
273 the same time -- e.g. onsite, offsite.
274 - Add the ability to consolidate old backup sets (basically do a restore
275 to tape and appropriately update the catalog). Compress Volume sets.
276 Might need to spool via file is only one drive is available.
277 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
278 someway to do so with every volume that has less than x% valid
283 Bacula Projects Roadmap
285 last update 5 January 2003
287 Item 1: Multiple simultaneous Jobs. (done)
288 Done -- Restore part needs better implementation to work correctly
290 What: Permit multiple simultaneous jobs in Bacula.
292 Why: An enterprise level solution needs to go fast without the
293 need for the system administrator to carefully tweak
294 timing. Based on the benchmarks, during a full
295 backup, NetWorker typically hit 10 times the bandwidth to
296 the tape compared to Bacula--largely. This is probably due to
297 running parallel jobs and multi-threaded filling of buffers
298 and writing them to tape. This should also make things work
299 better when you have a mix of fast and slow machines backing
302 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
303 implementing this is a matter of some small cleanups and
307 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
308 Deferred -- not necessary yet.
310 What: If data is coming into the SD too fast, buffer it to
311 disk if the user has configured this option.
313 Why: This would be nice, especially if it more or less falls out
314 when implementing (1) above. If not, it probably should not
315 be given a high priority because fundamentally the backup time
316 is limited by the tape bandwidth. Even though you may finish a
317 client job quicker by spilling to disk, you still have to
318 eventually get it onto tape. If intermediate disk buffering
319 allows us to improve write bandwidth to tape, it may make
322 Notes: Whether or not this is implemented will depend upon performance
323 testing after item 1 is implemented.
326 Item 3: Write the bscan program -- also write a bcopy program.
329 What: Write a program that reads a Bacula tape and puts all the
330 appropriate data into the catalog. This allows recovery
331 from a tape that is no longer in the database, or it allows
332 re-creation of a database if lost.
334 Why: This is a fundamental robustness and disaster recovery tool
335 which will increase the comfort level of a sysadmin
336 considering adopting Bacula.
338 Notes: A skeleton of this program already exists, but much work
339 needs to be done. Implementing this will also make apparent
340 any deficiencies in the current Bacula tape format.
343 Item 4: Implement Base jobs.
345 What: A base job is sort of like a Full save except that you
346 will want the FileSet to contain only files that are unlikely
347 to change in the future (i.e. a snapshot of most of your
348 system after installing it). After the base job has been run,
349 when you are doing a Full save, you can specify to exclude
350 all files saved by the base job that have not been modified.
352 Why: This is something none of the competition does, as far as we know
353 (except BackupPC, which is a Perl program that saves to disk
354 only). It is big win for the user, it makes Bacula stand out
355 as offering a unique optimization that immediately saves time
358 Notes: Big savings in tape usage. Will require more resources because
359 the e. DIR must send FD a list of files/attribs, and the FD must
360 search the list and compare it for each file to be saved.
363 Item 5: Implement Label templates
365 What: This is a mechanism whereby Bacula can automatically create
366 a tape label for new tapes according to a detailed specification
367 provided by the user.
369 Why: It is a major convenience item for folks who use automated label
372 Notes: Bacula already has a working form of automatic tape label
373 creation, but it is very crude. The design for the complete
374 tape labeling project is already documented in the manual.
377 Item 6: Write a regression script.
380 What: This is an automatic script that runs and tests as many features
381 of Bacula as possible. The output is compared to previous
382 versions of Bacula and any differences are reported.
384 Why: This is an enormous help in preventing introduction of new
385 errors in parts of the program that already work correctly.
387 Notes: This probably should be ranked higher, it's something the typical
388 user doesn't see. Depending on how it's implemented, it may
389 make sense to defer it until the archival tape format and
390 user interface mature.
393 Item 7: GUI for interactive restore
394 Item 8: GUI for interactive backup
396 What: The current interactive restore is implemented with a tty
397 interface. It would be much nicer to be able to "see" the
398 list of files backed up in typical GUI tree format.
399 The same mechanism could also be used for creating
400 ad-hoc backup FileSets (item 8).
402 Why: Ease of use -- especially for the end user.
404 Notes: Rather than implementing in Gtk, we probably should go directly
405 for a Browser implementation, even if doing so meant the
406 capability wouldn't be available until much later. Not only
407 is there the question of Windows sites, most
408 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
409 without installing lots of stuff admins are very wary about.
410 Real sysadmins will always use the command line anyway, and
411 the user who's doing an interactive restore or backup of his
412 own files will in most cases be on a Windows machine running
416 Item 9: Add SSL to daemon communications.
418 What: This provides for secure communications between the daemons.
420 Why: This would allow doing backup across the Internet without
421 privacy concerns (or with much less concern).
423 Notes: The vast majority of near term potential users will be backing up
424 a single site over a LAN and, correctly or not, they probably
425 won't be concerned with security, at least not enough to go to
426 the trouble to set up keys, etc. to screw things down. We suspect
427 that many users genuinely interested in multi-site backup
428 already run some form of VPN software in their internetwork
429 connections, and are willing to delegate security to that layer.
432 Item 10: Define definitive tape format.
435 What: Define that definitive tape format that will not change
436 for the next millennium.
438 Why: Stability, security.
440 Notes: See notes for item 11 below.
443 Item 11: New daemon communication protocol.
445 What: The current daemon to daemon protocol is basically an ASCII
446 printf() and sending the buffer. On the receiving end, the
447 buffer is sscanf()ed to unpack it. The new scheme would
448 be a binary format that allows quick packing and unpacking
449 of any data type with named fields.
451 Why: Using binary packing would be faster. Named fields will permit
452 error checking to ensure that what is sent is what the
453 receiver really wants.
455 Notes: These are internal improvements in the interest of the
456 long-term stability and evolution of the program. On the one
457 hand, the sooner they're done, the less code we have to rip
458 up when the time comes to install them. On the other hand, they
459 don't bring an immediately perceptible benefit to potential
460 users. Item 10 and possibly item 11 should be deferred until Bacula
461 is well established with a growing user community more or
462 less happy with the feature set. At that time, it will make a
463 good "next generation" upgrade in the interest of data
469 ======================================================
471 It is somewhat like a Full save becomes an incremental since
472 the Base job (or jobs) plus other non-base files.
474 - New BaseFile table that contains:
475 JobId, BaseJobId, FileId (from Base).
476 i.e. for each base file that exists but is not saved because
477 it has not changed, the File daemon sends the JobId, BaseId,
478 and FileId back to the Director who creates the DB entry.
479 - To initiate a Base save, the Director sends the FD
480 the FileId, and full filename for each file in the Base.
481 - When the FD finds a Base file, he requests the Director to
482 send him the full File entry (stat packet plus MD5), or
483 conversely, the FD sends it to the Director and the Director
484 says yes or no. This can be quite rapid if the FileId is kept
485 by the FD for each Base Filename.
486 - It is probably better to have the comparison done by the FD
487 despite the fact that the File entry must be sent across the
489 - An alternative would be to send the FD the whole File entry
490 from the start. The disadvantage is that it requires a lot of
491 space. The advantage is that it requires less communications
493 - The Job record must be updated to indicate that one or more
495 - At end of Job, FD returns:
496 1. Count of base files/bytes not written to tape (i.e. matches)
497 2. Count of base file that were saved i.e. had changed.
498 - No tape record would be written for a Base file that matches, in the
499 same way that no tape record is written for Incremental jobs where
500 the file is not saved because it is unchanged.
501 - On a restore, all the Base file records must explicitly be
502 found from the BaseFile tape. I.e. for each Full save that is marked
503 to have one or more Base Jobs, search the BaseFile for all occurrences
505 - An optimization might be to make the BaseFile have:
511 This would avoid the need to explicitly fetch each File record for
512 the Base job. The Base Job record will be fetched to get the
513 VolSessionId and VolSessionTime.
514 =========================================================
516 ==========================================================
518 For each Incremental job that is run, there may be files that
519 were found but not saved because they were locked (this applies
520 only to Windows). Such a system could send back to the Director
521 a list of Unsaved files.
523 - New UnSavedFiles table that contains:
527 - Then in the next Incremental job, the list of Unsaved Files will be
528 feed to the FD, who will ensure that they are explicitly chosen even
529 if standard date/time check would not have selected them.
530 =============================================================
534 =============================================================
536 Request For Comments For File Backup Options
539 Subject: File Backup Options
542 A few days ago, a Bacula user who is backing up to file volumes and
543 using compression asked if it was possible to suppress compressing
544 all .gz files since it was a waste of CPU time. Although Bacula
545 currently permits using different options (compression, ...) on
546 a directory by directory basis, it cannot do it on a file by
547 file basis, which is clearly what was desired.
549 Proposed Implementation:
550 To solve this problem, I propose the following:
552 - Add a new Director resource type called FileOptions.
554 - The FileOptions resource will have records for all
555 options that can currently be specified on the Include record
556 (in a FileSet). Examples below.
558 - The FileOptions resource will permit an exclude option as well
559 as a number of additional options.
561 - The heart of the FileOptions resource is the ability to
562 supply any number of ApplyTo records which specify POSIX
563 regular expressions. These ApplyTo regular expressions are
564 applied to the fully qualified filename (path and all). If
565 one matches, then the FileOptions will be used.
567 - When an ApplyTo specification matches an included file, the
568 options specified in the FileOptions resource will override
569 the default options specified on the Include record.
571 - Include records will be modified to permit referencing one or
572 more FileOptions resources. The FileOptions will be used
573 in the order listed on the Include record and the first
574 one that matches will be applied.
576 - Options (or specifications) currently supplied on the Include
577 record will be deprecated (i.e. removed in a later version a
578 year or so from now).
580 - The Exclude record will be deprecated as the same functionality
581 can be obtained by using an Exclude = yes in the FileOptions.
584 The following records can appear in the FileOptions resource. An
585 asterisk preceding the name indicates a feature not currently
589 - Compression= (GZIP, ...)
590 - Signature= (MD5, SHA1, ...)
592 - OneFs= (yes/no) - remain on one filesystem
593 - Recurse= (yes/no) - recurse into subdirectories
594 - Sparse= (yes/no) - do sparse file backup
595 - *Exclude= (yes/no) - exclude file from being saved
596 - *Reader= (filename) - external read (backup) program
597 - *Plugin= (filename) - read/write plugin module
600 - verify= (ipnougsamc5) - verify options
603 - replace= (always/ifnewer/ifolder/never) - replace options currently
605 - *Writer= (filename) - external write (restore) program
609 Currently options specifying compression, MD5 signatures, recursion,
610 ... of a FileSet are supplied on the Include record. These will now
611 all be collected into a FileOptions resource, which will be
612 specified on the Include in place of the options. Multiple FileOptions
613 may be specified. Since the FileOptions contain regular expressions
614 that are applied to the full filename, this will give the ability
615 to specify backup options on a file by file basis to whatever level
624 Include = compression=GZIP signature=MD5 {
633 Include = FileOptions=Opts {
644 That's a lot more to do the same thing, but it gives the ability to
645 apply options on a file by file basis. For example, suppose you
646 want to compress all files but not any file with extensions .gz or .Z.
647 You could do so as follows:
651 Include = FileOptions=NoCompress FileOptions=Opts {
659 ApplyTo = /*.?*/ # matches all files
664 # Note multiple ApplyTos are ORed
665 ApplyTo = /*.gz/ # matches .gz files */
666 ApplyTo = /*.Z/ # matches .Z files */
669 Now, since the NoCompress FileOptions is specified first on the
670 Include line, any *.gz or *.Z file will have an MD5 signature computed,
671 but will not be compressed. For all other files, the NoCompress will not
672 match, so the Opts options will be used which will include GZIP
676 - Is it necessary to provide some means of ANDing regular expressions
677 and negation? (not currently planned)
679 e.g. ApplyTo = /*.gz/ && !/big.gz/
681 - I see that Networker has a "null" module which, if specified, does not
682 backup the file, but does make an record of the file in the catalog
683 so that the catalog will reflect an exact picture of the filesystem.
684 The result is that the file can be "seen" when "browsing" the save
685 sets, but it cannot be restored.
687 Is this really useful? Should it be implemented in Bacula?
690 After implementing the above, the user will be able to specify
691 on a file by file basis (using regular expressions) what options are
692 applied for the backup.
693 ====================================
695 =========================================
696 Proposal by Bill Sellers
698 Return-Path: <w.a.sellers@larc.nasa.gov>
699 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
700 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)
701 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
702 X-Sender: w.a.sellers@pop.larc.nasa.gov
703 X-Mailer: QUALCOMM Windows Eudora Version 5.1
704 Date: Tue, 14 Jan 2003 16:30:18 -0500
705 To: Kern Sibbald <kern@sibbald.com>
706 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
707 Subject: Re: [Bacula-users] Bacula remote storage?
708 In-Reply-To: <1042565382.1845.177.camel@rufus>
709 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>
711 Content-Type: text/plain; charset="us-ascii"; format=flowed
712 X-Annoyance-Filter-Junk-Probability: 0
713 X-Annoyance-Filter-Classification: Mail
714 At 06:29 PM 1/14/2003 +0100, you wrote:
717 >Well, if you cannot put a Bacula client on the machine,
718 >then it is a big problem. If you know of some software
719 >that can do what you want, let me know, because I
720 >really just don't know how to do it -- at least not
726 We have been able to get Amanda to use the HSM as a storage
727 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
728 handle Windows systems very well (or at all without Samba). So I am
729 looking for a backup system that has a Windows client. I really like the
730 Windows integration of Bacula.
732 From the command line, its rather trivial to move the data around. We use
735 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
737 or if you use GNU tar:
739 tar czf hsm:path/file.tgz ./files
741 One idea for you to consider; Sendmail offers pipes in the aliases file;
742 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
743 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
744 pipe available, as a storage device? Then we could use any command that
745 handles stdin as a storage destination.
752 #Password is not used in rsh, but might be used in ftp.
753 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
760 Password = "foobar&-"
761 Device = "| ncftpput -c hsm /path/file.bacula"
765 >If you have some local storage available, you could
766 >use Bacula to backup to disk volumes, then use some
767 >other software (ftp, scp) to move them to the HSM
768 >machine. However, this is a bit kludgy.
771 It is, but maybe worth a try. Is there some function in Bacula to put
772 variables in filenames? i.e. backup.2003-01-15.root
779 w.a.sellers@larc.nasa.gov
781 ==============================================
782 The Project for the above
784 I finally realized that this is not at all
785 the same as reader/writer programs or plugins,
786 which are alternate ways of accessing the
787 files to be backed up. Rather, it is an alternate
788 form of storage device, and I have always planned
789 that Bacula should be able to handle all sorts
792 So, I propose the following phases:
794 1. OK from you to invest some time in testing
795 this as I implement it (requires that you
796 know how to download from the SourceForge
797 cvs -- which I imagine is a piece of cake
800 2. Dumb implementation by allowing a device to
801 be a fifo for write only.
802 Reason: easy to implement, proof of concept.
804 3. Try reading from fifo but with fixed block
806 Reason: proof of concept, easy to implement.
808 4. Extend reading from fifo (restores) to handle
810 Reason: requires some delicate low level coding
811 which could destabilize all of Bacula.
813 5. Implementation of above but to a program. E.g.
814 Device = "|program" (not full pipeline).
815 Reason: routines already exist, and program can
816 be a shell script which contains anything.
818 6. Add full pipeline as a possibility. E.g.
819 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
820 Reason: needs additional coding to implement full
821 pipeline (must fire off either a shell or all
822 programs and connect their pipes).
824 There are a good number of details in each step
825 that I have left out, but I will specify them at
826 every stage, and there may be a few changes as things
827 evolve. I expect that to get to stage 5 will take a
828 few weeks, and at that point, you will have
829 everything you need (just inside a script).
830 Stage 6 will probably take longer, but if this
831 project pleases you, what we do for 5 should
832 be adequate for some time.
836 =============================================
838 Done: (see kernsdone for more)
839 - Look into Pruning/purging problems or why there seem to
840 be so many files listed each night.
841 - Fix cancel in find_one -- need jcr.
842 - Cancel does not work for restore in FD.
843 - Write SetJobStatus() function so cancel status not lost.
844 - Add include list to end of chain in findlib
845 - Zap sd_auth_key after use
846 - Add Bar code reading capabilities (new mtx-changer)
847 - Figure out some way to automatically backup all local partitions
848 - Make hash table for linked files in findlib/find_one.c:161
850 - Rewrite find_one.c to use only pool_memory instead of
851 alloca and malloc (probably not necessary).
852 - Make sure btraceback goes into /sbin not sysconf directory.
853 - InitVerify is getting pruned and it shouldn't (document it)
854 - Make 1.28c release ??? NO do 1.29 directly
855 - Set timeout on opening fifo for save or restore (findlib)
856 - Document FIFO storage device.
857 - Document fifo and | and <
860 - Get correct error status from run_program or open_bpipe().
861 - Restrict permissions on File Volumes (now 0640).
862 - Umasked 022 daemons
863 - Fix restore of hard linked file.
864 - Figure out how to allow multiple simultaneous file Volumes on a single device.
865 - Implement multiple simultaneous file Volumes on a single device.
866 - Cleanup db_update_media and db_update_pool
867 - Flush all the daemon messages at the end of every job.
868 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
869 run_program -- bpipe.c
870 - Apparently cancel does not work for jobs waiting to be
872 - Implement TCP/IP connection for MySQL
873 - Pull a canceled job from the Scheduling queue.