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.
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
21 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
24 - Ability to backup to a file then later transfer to a tape.
25 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
26 set of volumes recycling the oldest volume when it is needed.
27 - Fix "access not allowed" for backup of files on WinXP.
28 - Figure out some way to specify a retention period for files
29 that no longer exist on the machine -- so that we maintain
30 say backups for 30 days, but if the file is deleted, we maintain
31 the last copy for 1 year.
32 - Have Bacula "poll" the tape to see if it is there.
33 - Add more of the config info to the tape label.
34 - Implement bar code reader for autochangers
35 - Add a default File storage so that new users can do backup
36 and restores right away.
37 - Implement a Mount Command and an Unmount Command where
38 the users could specify a system command to be performed
39 to do the mount, after which Bacula could attempt to
40 read the device. This is for Removeable media such as a CDROM.
41 - Most likely, this mount command would be invoked explicitly
42 by the user using the current Console "mount" and "unmount"
43 commands -- the Storage Daemon would do the right thing
44 depending on the exact nature of the device.
45 - As with tape drives, when Bacula wanted a new removable
46 disk mounted, it would unmount the old one, and send a message
47 to the user, who would then use "mount" as described above
48 once he had actually inserted the disk.
50 - Make some way so that if a machine is skipped because it is not up
51 that Bacula will continue retrying for a specified period of time --
53 - Figure out some way to estimate output size and to avoid splitting
54 a backup across two Volumes -- this could be useful for writing CDROMs
55 where you really prefer not to have it split -- not serious.
56 - Add chflags() code for FreeBSD file flags
57 - Add RunBeforeJob and RunAfterJob to the Client program.
58 - Have SD compute MD5 or SHA1 and compare to what FD computes.
59 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
60 actual data on the Volume and compare it.
61 - Implement FileOptions (see end of this document)
62 - Implement Bacula plugins -- design API
63 - Make bcopy read through bad tape records.
64 - Fix read_record to handle multiple sessions.
65 - Program files (i.e. execute a program to read/write files).
66 Pass read date of last backup, size of file last time.
67 - Add Signature type to File DB record.
68 - Make Restore report an error if FD or SD term codes are not OK.
69 - CD into subdirectory when open()ing files for backup to
70 speed up things. Test with testfind().
71 - Add prefixlinks to where or not where absolute links to FD.
72 - Look at handling <> in smtp doesn't work with exim.
73 - Priority job to go to top of list.
74 - Why is catreq.c:111 Find vol called twice for a job?
75 - Find out why Full saves run slower and slower (hashing?)
76 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
77 - Implement some way for the Console to dynamically create a job.
78 - Restore to a particular time -- e.g. before date, after date.
79 - Implement disk spooling
80 - Implement finer multiprocessing options.
81 - Solaris -I on tar for include list
82 - Enable avoid backing up archive device (findlib/find_one.c:128)
83 - Need a verbose mode in restore, perhaps to bsr.
84 - bscan without -v is too quiet -- perhaps show jobs.
85 - Add code to reject whole blocks if not wanted on restore.
86 - Start working on Base jobs.
87 - Make sure the MaxVolFiles is fully implemented in SD
88 - Check if both CatalogFiles and UseCatalog are set to SD.
89 - Check if we can increase Bacula FD priorty in Win2000
90 - Need return status on read_cb() from read_records(). Need multiple
91 records -- one per Job, maybe a JCR or some other structure with
93 - Figure out how to do a bare metal Windows restore
94 - Put system type returned by FD into catalog.
95 - Need to specify MaximumConcurrentJobs in the Job resource.
96 - Possibly add email to Watchdog if drive is unmounted too
97 long and a job is waiting on the drive.
98 - Use read_record.c in SD code.
99 - Why don't we get an error message from Win32 FD when bootstrap
100 file cannot be created for restore command?
101 - When Marking a file in Restore that is a hard link, also
102 mark the link so that the data will be reloaded.
103 - Restore program that errors in SD due to no tape reports
104 OK incorrectly in output.
105 - After unmount, if restore job started, ask to mount.
106 - Convert all %x substitution variables, which are hard to remember
107 and read to %(variable-name). Idea from TMDA.
108 - Add JobLevel in FD status (but make sure it is defined).
109 - Make Pool resource handle Counter resources.
110 - Remove NextId for SQLite. Optimize.
111 - Strip trailing / from Include
112 - Move all SQL statements into a single location.
113 - Add UA rc and history files.
114 - put termcap (used by console) in ./configure and
115 allow -with-termcap-dir.
116 - Enhance time and size scanning routines.
117 - Fix Autoprune for Volumes to respect need for full save.
118 - Fix Win32 config file definition name on /install
119 - No READLINE_SRC if found in alternate directory.
120 - Add Client FS/OS id (Linux, Win95/98, ...).
121 - Test a second language e.g. french.
122 - Compare tape to Client files (attributes, or attributes and data)
123 - Make all database Ids 64 bit.
124 - Write an applet for Linux.
125 - Add estimate to Console commands
126 - Find solution to blank filename (i.e. path only) problem.
127 - Implement new daemon communications protocol.
128 - Remove PoolId from Job table, it exists in Media.
129 - Allow console commands to detach or run in background.
130 - Fix status delay on storage daemon during rewind.
131 - Add SD message variables to control operator wait time
132 - Maximum Operator Wait
133 - Minimum Message Interval
134 - Maximum Message Interval
135 - Send Operator message when cannot read tape label.
136 - Verify level=Volume (scan only), level=Data (compare of data to file).
137 Verify level=Catalog, level=InitCatalog
139 - Add keyword search to show command in Console.
140 - Fix Win2000 error with no messages during startup.
141 - Events : tape has more than xxx bytes.
142 - Restrict characters permitted in a Resource name.
143 - Complete code in Bacula Resources -- this will permit
144 reading a new config file at any time.
145 - Handle ctl-c in Console
146 - Implement LabelTemplate (at least first cut).
147 - Implement script driven addition of File daemon to config files.
148 - Think about how to make Bacula work better with File (non-tape) archives.
149 - Write Unix emulator for Windows.
151 - Implement new serialize subroutines
152 send(socket, "string", &Vol, "uint32", &i, NULL)
153 - Audit all UA commands to ensure that we always prompt where possible.
154 - If ./btape is called without /dev, assume argument is a Storage resource name.
155 - Put memory utilization in Status output of each daemon
156 if full status requested or if some level of debug on.
157 - Make database type selectable by .conf files i.e. at runtime
158 - gethostbyname failure in bnet_connect() continues
159 generating errors -- should stop.
160 - Set flag for uname -a. Add to Volume label.
161 - Implement throttled work queue.
162 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
163 - Allow multiple Storage specifications (or multiple names on
164 a single Storage specification) in the Job record. Thus a job
165 can be backed up to a number of storage devices.
166 - Implement dump label to UA
167 - Concept of VolumeSet during restore which is a list
168 of Volume names needed.
169 - Restore files modified after date
170 - Restore file modified before date
171 - Emergency restore info:
173 - Backup working directory
175 - Restore -- do nothing but show what would happen
176 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
177 - Implement Restore FileSet=
178 - Create a protocol.h and protocol.c where all protocol messages
180 - If SD cannot open a drive, make it periodically retry.
181 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
182 - Timout a job or terminate if link goes down, or reopen link and query.
183 - Find general solution for sscanf size problems (as well
184 as sprintf. Do at run time?
185 - Concept of precious tapes (cannot be reused).
186 - Make bcopy copy with a single tape drive.
187 - Permit changing ownership during restore.
189 - Autolabel should be specified by DIR instead of SD.
190 - Find out how to get the system tape block limits, e.g.:
191 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
192 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
195 - AutoScan (check checksum of tape)
196 - Format command = "format /dev/nst0"
200 - Seek resolution (usually corresponds to buffer size)
201 - EODErrorCode=ENOSPC or code
202 - Partial Read error code
203 - Partial write error code
204 - Nonformatted read error
205 - Nonformatted write error
206 - WriteProtected error
210 - IgnoreCloseErrors=yes
220 - FD sends unsaved file list to Director at end of job (see
222 - File daemon should build list of files skipped, and then
223 at end of save retry and report any errors.
224 - Write a Storage daemon that uses pipes and
225 standard Unix programs to write to the tape.
227 - Need something that monitors the JCR queue and
228 times out jobs by asking the deamons where they are.
229 - Enhance Jmsg code to permit buffering and saving to disk.
230 - device driver = "xxxx" for drives.
231 - restart: paranoid: read label fsf to
232 eom read append block, and go
233 super-paranoid: read label, read all files
234 in between, read append block, and go
235 verify: backspace, read append block, and go
236 permissive: same as above but frees drive
237 if tape is not valid.
239 - Ensure that /dev/null works
240 - Need report class for messages. Perhaps
241 report resource where report=group of messages
242 - enhance scan_attrib and rename scan_jobtype, and
243 fill in code for "since" option
244 - Director needs a time after which the report status is sent
245 anyway -- or better yet, a retry time for the job.
246 Don't reschedule a job if previous incarnation is still running.
247 - Some way to automatically backup everything is needed????
248 - Need a structure for pending actions:
250 - termination status (part of buffered msgs?)
251 - Concept of grouping Storage devices and job can use
252 any of a number of devices
254 Read, Write, Clean, Delete
255 - Login to Bacula; Bacula users with different permissions:
256 owner, group, user, quotas
257 - Store info on each file system type (probably in the job header on tape.
258 This could be the output of df; or perhaps some sort of /etc/mtab record.
261 - Design at hierarchial storage for Bacula.
262 - Implement FSM (File System Modules).
263 - Identify unchanged or "system" files and save them to a
264 special tape thus removing them from the standard
265 backup FileSet -- BASE backup.
266 - Turn virutally all sprintfs into snprintfs.
267 - Heartbeat between daemons.
268 - Audit M_ error codes to ensure they are correct and consistent.
269 - Add variable break characters to lex analyzer.
270 Either a bit mask or a string of chars so that
271 the caller can change the break characters.
272 - Make a single T_BREAK to replace T_COMMA, etc.
273 - Ensure that File daemon and Storage daemon can
274 continue a save if the Director goes down (this
275 is NOT currently the case). Must detect socket error,
276 buffer messages for later.
277 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
278 - Add ability to backup to two Storage devices (two SD sessions) at
279 the same time -- e.g. onsite, offsite.
280 - Add the ability to consolidate old backup sets (basically do a restore
281 to tape and appropriately update the catalog). Compress Volume sets.
282 Might need to spool via file is only one drive is available.
283 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
284 someway to do so with every volume that has less than x% valid
289 Bacula Projects Roadmap
291 last update 5 January 2003
293 Item 1: Multiple simultaneous Jobs. (done)
294 Done -- Restore part needs better implementation to work correctly
296 What: Permit multiple simultaneous jobs in Bacula.
298 Why: An enterprise level solution needs to go fast without the
299 need for the system administrator to carefully tweak
300 timing. Based on the benchmarks, during a full
301 backup, NetWorker typically hit 10 times the bandwidth to
302 the tape compared to Bacula--largely. This is probably due to
303 running parallel jobs and multi-threaded filling of buffers
304 and writing them to tape. This should also make things work
305 better when you have a mix of fast and slow machines backing
308 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
309 implementing this is a matter of some small cleanups and
313 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
314 Deferred -- not necessary yet.
316 What: If data is coming into the SD too fast, buffer it to
317 disk if the user has configured this option.
319 Why: This would be nice, especially if it more or less falls out
320 when implementing (1) above. If not, it probably should not
321 be given a high priority because fundamentally the backup time
322 is limited by the tape bandwidth. Even though you may finish a
323 client job quicker by spilling to disk, you still have to
324 eventually get it onto tape. If intermediate disk buffering
325 allows us to improve write bandwidth to tape, it may make
328 Notes: Whether or not this is implemented will depend upon performance
329 testing after item 1 is implemented.
332 Item 3: Write the bscan program -- also write a bcopy program.
335 What: Write a program that reads a Bacula tape and puts all the
336 appropriate data into the catalog. This allows recovery
337 from a tape that is no longer in the database, or it allows
338 re-creation of a database if lost.
340 Why: This is a fundamental robustness and disaster recovery tool
341 which will increase the comfort level of a sysadmin
342 considering adopting Bacula.
344 Notes: A skeleton of this program already exists, but much work
345 needs to be done. Implementing this will also make apparent
346 any deficiencies in the current Bacula tape format.
349 Item 4: Implement Base jobs.
351 What: A base job is sort of like a Full save except that you
352 will want the FileSet to contain only files that are unlikely
353 to change in the future (i.e. a snapshot of most of your
354 system after installing it). After the base job has been run,
355 when you are doing a Full save, you can specify to exclude
356 all files saved by the base job that have not been modified.
358 Why: This is something none of the competition does, as far as we know
359 (except BackupPC, which is a Perl program that saves to disk
360 only). It is big win for the user, it makes Bacula stand out
361 as offering a unique optimization that immediately saves time
364 Notes: Big savings in tape usage. Will require more resources because
365 the e. DIR must send FD a list of files/attribs, and the FD must
366 search the list and compare it for each file to be saved.
369 Item 5: Implement Label templates
371 What: This is a mechanism whereby Bacula can automatically create
372 a tape label for new tapes according to a detailed specification
373 provided by the user.
375 Why: It is a major convenience item for folks who use automated label
378 Notes: Bacula already has a working form of automatic tape label
379 creation, but it is very crude. The design for the complete
380 tape labeling project is already documented in the manual.
383 Item 6: Write a regression script.
386 What: This is an automatic script that runs and tests as many features
387 of Bacula as possible. The output is compared to previous
388 versions of Bacula and any differences are reported.
390 Why: This is an enormous help in preventing introduction of new
391 errors in parts of the program that already work correctly.
393 Notes: This probably should be ranked higher, it's something the typical
394 user doesn't see. Depending on how it's implemented, it may
395 make sense to defer it until the archival tape format and
396 user interface mature.
399 Item 7: GUI for interactive restore
400 Item 8: GUI for interactive backup
402 What: The current interactive restore is implemented with a tty
403 interface. It would be much nicer to be able to "see" the
404 list of files backed up in typical GUI tree format.
405 The same mechanism could also be used for creating
406 ad-hoc backup FileSets (item 8).
408 Why: Ease of use -- especially for the end user.
410 Notes: Rather than implementing in Gtk, we probably should go directly
411 for a Browser implementation, even if doing so meant the
412 capability wouldn't be available until much later. Not only
413 is there the question of Windows sites, most
414 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
415 without installing lots of stuff admins are very wary about.
416 Real sysadmins will always use the command line anyway, and
417 the user who's doing an interactive restore or backup of his
418 own files will in most cases be on a Windows machine running
422 Item 9: Add SSL to daemon communications.
424 What: This provides for secure communications between the daemons.
426 Why: This would allow doing backup across the Internet without
427 privacy concerns (or with much less concern).
429 Notes: The vast majority of near term potential users will be backing up
430 a single site over a LAN and, correctly or not, they probably
431 won't be concerned with security, at least not enough to go to
432 the trouble to set up keys, etc. to screw things down. We suspect
433 that many users genuinely interested in multi-site backup
434 already run some form of VPN software in their internetwork
435 connections, and are willing to delegate security to that layer.
438 Item 10: Define definitive tape format.
441 What: Define that definitive tape format that will not change
442 for the next millennium.
444 Why: Stability, security.
446 Notes: See notes for item 11 below.
449 Item 11: New daemon communication protocol.
451 What: The current daemon to daemon protocol is basically an ASCII
452 printf() and sending the buffer. On the receiving end, the
453 buffer is sscanf()ed to unpack it. The new scheme would
454 be a binary format that allows quick packing and unpacking
455 of any data type with named fields.
457 Why: Using binary packing would be faster. Named fields will permit
458 error checking to ensure that what is sent is what the
459 receiver really wants.
461 Notes: These are internal improvements in the interest of the
462 long-term stability and evolution of the program. On the one
463 hand, the sooner they're done, the less code we have to rip
464 up when the time comes to install them. On the other hand, they
465 don't bring an immediately perceptible benefit to potential
466 users. Item 10 and possibly item 11 should be deferred until Bacula
467 is well established with a growing user community more or
468 less happy with the feature set. At that time, it will make a
469 good "next generation" upgrade in the interest of data
475 ======================================================
477 It is somewhat like a Full save becomes an incremental since
478 the Base job (or jobs) plus other non-base files.
480 - New BaseFile table that contains:
481 JobId, BaseJobId, FileId (from Base).
482 i.e. for each base file that exists but is not saved because
483 it has not changed, the File daemon sends the JobId, BaseId,
484 and FileId back to the Director who creates the DB entry.
485 - To initiate a Base save, the Director sends the FD
486 the FileId, and full filename for each file in the Base.
487 - When the FD finds a Base file, he requests the Director to
488 send him the full File entry (stat packet plus MD5), or
489 conversely, the FD sends it to the Director and the Director
490 says yes or no. This can be quite rapid if the FileId is kept
491 by the FD for each Base Filename.
492 - It is probably better to have the comparison done by the FD
493 despite the fact that the File entry must be sent across the
495 - An alternative would be to send the FD the whole File entry
496 from the start. The disadvantage is that it requires a lot of
497 space. The advantage is that it requires less communications
499 - The Job record must be updated to indicate that one or more
501 - At end of Job, FD returns:
502 1. Count of base files/bytes not written to tape (i.e. matches)
503 2. Count of base file that were saved i.e. had changed.
504 - No tape record would be written for a Base file that matches, in the
505 same way that no tape record is written for Incremental jobs where
506 the file is not saved because it is unchanged.
507 - On a restore, all the Base file records must explicitly be
508 found from the BaseFile tape. I.e. for each Full save that is marked
509 to have one or more Base Jobs, search the BaseFile for all occurrences
511 - An optimization might be to make the BaseFile have:
517 This would avoid the need to explicitly fetch each File record for
518 the Base job. The Base Job record will be fetched to get the
519 VolSessionId and VolSessionTime.
520 =========================================================
522 ==========================================================
524 For each Incremental job that is run, there may be files that
525 were found but not saved because they were locked (this applies
526 only to Windows). Such a system could send back to the Director
527 a list of Unsaved files.
529 - New UnSavedFiles table that contains:
533 - Then in the next Incremental job, the list of Unsaved Files will be
534 feed to the FD, who will ensure that they are explicitly chosen even
535 if standard date/time check would not have selected them.
536 =============================================================
540 =============================================================
542 Request For Comments For File Backup Options
545 Subject: File Backup Options
548 A few days ago, a Bacula user who is backing up to file volumes and
549 using compression asked if it was possible to suppress compressing
550 all .gz files since it was a waste of CPU time. Although Bacula
551 currently permits using different options (compression, ...) on
552 a directory by directory basis, it cannot do it on a file by
553 file basis, which is clearly what was desired.
555 Proposed Implementation:
556 To solve this problem, I propose the following:
558 - Add a new Director resource type called FileOptions.
560 - The FileOptions resource will have records for all
561 options that can currently be specified on the Include record
562 (in a FileSet). Examples below.
564 - The FileOptions resource will permit an exclude option as well
565 as a number of additional options.
567 - The heart of the FileOptions resource is the ability to
568 supply any number of ApplyTo records which specify POSIX
569 regular expressions. These ApplyTo regular expressions are
570 applied to the fully qualified filename (path and all). If
571 one matches, then the FileOptions will be used.
573 - When an ApplyTo specification matches an included file, the
574 options specified in the FileOptions resource will override
575 the default options specified on the Include record.
577 - Include records will be modified to permit referencing one or
578 more FileOptions resources. The FileOptions will be used
579 in the order listed on the Include record and the first
580 one that matches will be applied.
582 - Options (or specifications) currently supplied on the Include
583 record will be deprecated (i.e. removed in a later version a
584 year or so from now).
586 - The Exclude record will be deprecated as the same functionality
587 can be obtained by using an Exclude = yes in the FileOptions.
590 The following records can appear in the FileOptions resource. An
591 asterisk preceding the name indicates a feature not currently
595 - Compression= (GZIP, ...)
596 - Signature= (MD5, SHA1, ...)
598 - OneFs= (yes/no) - remain on one filesystem
599 - Recurse= (yes/no) - recurse into subdirectories
600 - Sparse= (yes/no) - do sparse file backup
601 - *Exclude= (yes/no) - exclude file from being saved
602 - *Reader= (filename) - external read (backup) program
603 - *Plugin= (filename) - read/write plugin module
606 - verify= (ipnougsamc5) - verify options
609 - replace= (always/ifnewer/ifolder/never) - replace options currently
611 - *Writer= (filename) - external write (restore) program
615 Currently options specifying compression, MD5 signatures, recursion,
616 ... of a FileSet are supplied on the Include record. These will now
617 all be collected into a FileOptions resource, which will be
618 specified on the Include in place of the options. Multiple FileOptions
619 may be specified. Since the FileOptions contain regular expressions
620 that are applied to the full filename, this will give the ability
621 to specify backup options on a file by file basis to whatever level
630 Include = compression=GZIP signature=MD5 {
639 Include = FileOptions=Opts {
650 That's a lot more to do the same thing, but it gives the ability to
651 apply options on a file by file basis. For example, suppose you
652 want to compress all files but not any file with extensions .gz or .Z.
653 You could do so as follows:
657 Include = FileOptions=NoCompress FileOptions=Opts {
665 ApplyTo = /*.?*/ # matches all files
670 # Note multiple ApplyTos are ORed
671 ApplyTo = /*.gz/ # matches .gz files */
672 ApplyTo = /*.Z/ # matches .Z files */
675 Now, since the NoCompress FileOptions is specified first on the
676 Include line, any *.gz or *.Z file will have an MD5 signature computed,
677 but will not be compressed. For all other files, the NoCompress will not
678 match, so the Opts options will be used which will include GZIP
682 - Is it necessary to provide some means of ANDing regular expressions
683 and negation? (not currently planned)
685 e.g. ApplyTo = /*.gz/ && !/big.gz/
687 - I see that Networker has a "null" module which, if specified, does not
688 backup the file, but does make an record of the file in the catalog
689 so that the catalog will reflect an exact picture of the filesystem.
690 The result is that the file can be "seen" when "browsing" the save
691 sets, but it cannot be restored.
693 Is this really useful? Should it be implemented in Bacula?
696 After implementing the above, the user will be able to specify
697 on a file by file basis (using regular expressions) what options are
698 applied for the backup.
699 ====================================
701 =========================================
702 Proposal by Bill Sellers
704 Return-Path: <w.a.sellers@larc.nasa.gov>
705 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
706 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)
707 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
708 X-Sender: w.a.sellers@pop.larc.nasa.gov
709 X-Mailer: QUALCOMM Windows Eudora Version 5.1
710 Date: Tue, 14 Jan 2003 16:30:18 -0500
711 To: Kern Sibbald <kern@sibbald.com>
712 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
713 Subject: Re: [Bacula-users] Bacula remote storage?
714 In-Reply-To: <1042565382.1845.177.camel@rufus>
715 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>
717 Content-Type: text/plain; charset="us-ascii"; format=flowed
718 X-Annoyance-Filter-Junk-Probability: 0
719 X-Annoyance-Filter-Classification: Mail
720 At 06:29 PM 1/14/2003 +0100, you wrote:
723 >Well, if you cannot put a Bacula client on the machine,
724 >then it is a big problem. If you know of some software
725 >that can do what you want, let me know, because I
726 >really just don't know how to do it -- at least not
732 We have been able to get Amanda to use the HSM as a storage
733 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
734 handle Windows systems very well (or at all without Samba). So I am
735 looking for a backup system that has a Windows client. I really like the
736 Windows integration of Bacula.
738 From the command line, its rather trivial to move the data around. We use
741 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
743 or if you use GNU tar:
745 tar czf hsm:path/file.tgz ./files
747 One idea for you to consider; Sendmail offers pipes in the aliases file;
748 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
749 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
750 pipe available, as a storage device? Then we could use any command that
751 handles stdin as a storage destination.
758 #Password is not used in rsh, but might be used in ftp.
759 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
766 Password = "foobar&-"
767 Device = "| ncftpput -c hsm /path/file.bacula"
771 >If you have some local storage available, you could
772 >use Bacula to backup to disk volumes, then use some
773 >other software (ftp, scp) to move them to the HSM
774 >machine. However, this is a bit kludgy.
777 It is, but maybe worth a try. Is there some function in Bacula to put
778 variables in filenames? i.e. backup.2003-01-15.root
785 w.a.sellers@larc.nasa.gov
787 ==============================================
788 The Project for the above
790 I finally realized that this is not at all
791 the same as reader/writer programs or plugins,
792 which are alternate ways of accessing the
793 files to be backed up. Rather, it is an alternate
794 form of storage device, and I have always planned
795 that Bacula should be able to handle all sorts
798 So, I propose the following phases:
800 1. OK from you to invest some time in testing
801 this as I implement it (requires that you
802 know how to download from the SourceForge
803 cvs -- which I imagine is a piece of cake
806 2. Dumb implementation by allowing a device to
807 be a fifo for write only.
808 Reason: easy to implement, proof of concept.
810 3. Try reading from fifo but with fixed block
812 Reason: proof of concept, easy to implement.
814 4. Extend reading from fifo (restores) to handle
816 Reason: requires some delicate low level coding
817 which could destabilize all of Bacula.
819 5. Implementation of above but to a program. E.g.
820 Device = "|program" (not full pipeline).
821 Reason: routines already exist, and program can
822 be a shell script which contains anything.
824 6. Add full pipeline as a possibility. E.g.
825 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
826 Reason: needs additional coding to implement full
827 pipeline (must fire off either a shell or all
828 programs and connect their pipes).
830 There are a good number of details in each step
831 that I have left out, but I will specify them at
832 every stage, and there may be a few changes as things
833 evolve. I expect that to get to stage 5 will take a
834 few weeks, and at that point, you will have
835 everything you need (just inside a script).
836 Stage 6 will probably take longer, but if this
837 project pleases you, what we do for 5 should
838 be adequate for some time.
842 =============================================
844 Done: (see kernsdone for more)
845 - Look into Pruning/purging problems or why there seem to
846 be so many files listed each night.
847 - Fix cancel in find_one -- need jcr.
848 - Cancel does not work for restore in FD.
849 - Write SetJobStatus() function so cancel status not lost.
850 - Add include list to end of chain in findlib
851 - Zap sd_auth_key after use
852 - Add Bar code reading capabilities (new mtx-changer)
853 - Figure out some way to automatically backup all local partitions
854 - Make hash table for linked files in findlib/find_one.c:161
856 - Rewrite find_one.c to use only pool_memory instead of
857 alloca and malloc (probably not necessary).
858 - Make sure btraceback goes into /sbin not sysconf directory.
859 - InitVerify is getting pruned and it shouldn't (document it)
860 - Make 1.28c release ??? NO do 1.29 directly
861 - Set timeout on opening fifo for save or restore (findlib)
862 - Document FIFO storage device.
863 - Document fifo and | and <
866 - Get correct error status from run_program or open_bpipe().
867 - Restrict permissions on File Volumes (now 0640).
868 - Umasked 022 daemons
869 - Fix restore of hard linked file.
870 - Figure out how to allow multiple simultaneous file Volumes on a single device.
871 - Implement multiple simultaneous file Volumes on a single device.
872 - Cleanup db_update_media and db_update_pool
873 - Flush all the daemon messages at the end of every job.
874 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
875 run_program -- bpipe.c
876 - Apparently cancel does not work for jobs waiting to be
878 - Implement TCP/IP connection for MySQL
879 - Pull a canceled job from the Scheduling queue.
880 - Implement max_file_size in block.c (already done, just tweaked).
881 - Look at purge jobs volume (at least document it, and see if it is
883 - Add list volumes does all pools. list volumes pool=xxx now works.
884 - Add pool= to "list media" in ua_output.c
885 - Strip trailing slashes from Include directory names in the FD.
886 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
887 ERR=Operation not permited loop.
888 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
889 - Produce better error messages in when error/eof writing block.