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 - Pool requested twice in "relabel" command.
24 Must delete old volume from catalog.
25 - Make non-zero status from RunJobBefore/After error the job.
26 - Remove kern and kelvin from mysql_grant...
27 - Install grant_mysql...
28 - Look at Python for a Bacula scripting language -- www.python.org
29 - add #define ENABLE_NLS for Gnome compile on SuSE.
30 - Need define of int_least16_t in sha1.h for SuSE.
31 - Figure out how to use ssh to protect Bacula communications.
32 - Fix "access not allowed" for backup of files on WinXP.
33 - Issue message to mount a new tape before the rewind.
34 - Simplified client job initiation for portables.
35 - If SD cannot open a drive, make it periodically retry.
36 - Implement LabelTemplate (at least first cut).
37 - Add more of the config info to the tape label.
38 - Implement bar code reader for autochangers
39 - Add a default File storage so that new users can do backup
40 and restores right away.
41 - Implement a Mount Command and an Unmount Command where
42 the users could specify a system command to be performed
43 to do the mount, after which Bacula could attempt to
44 read the device. This is for Removeable media such as a CDROM.
45 - Most likely, this mount command would be invoked explicitly
46 by the user using the current Console "mount" and "unmount"
47 commands -- the Storage Daemon would do the right thing
48 depending on the exact nature of the device.
49 - As with tape drives, when Bacula wanted a new removable
50 disk mounted, it would unmount the old one, and send a message
51 to the user, who would then use "mount" as described above
52 once he had actually inserted the disk.
54 - Make some way so that if a machine is skipped because it is not up
55 that Bacula will continue retrying for a specified period of time --
57 - If tape is marked read-only, then try opening it read-only rather than
58 failing, and remember that it cannot be written.
59 - Refine SD waiting output:
60 Device is being positioned
61 > Device is being positioned for append
62 > Device is being positioned to file x
64 - Figure out some way to estimate output size and to avoid splitting
65 a backup across two Volumes -- this could be useful for writing CDROMs
66 where you really prefer not to have it split -- not serious.
67 - Add RunBeforeJob and RunAfterJob to the Client program.
68 - Have SD compute MD5 or SHA1 and compare to what FD computes.
69 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
70 actual data on the Volume and compare it.
71 - Implement FileOptions (see end of this document)
72 - Implement Bacula plugins -- design API
73 - Make bcopy read through bad tape records.
74 - Fix read_record to handle multiple sessions.
75 - Program files (i.e. execute a program to read/write files).
76 Pass read date of last backup, size of file last time.
77 - Add Signature type to File DB record.
78 - Make Restore report an error if FD or SD term codes are not OK.
79 - CD into subdirectory when open()ing files for backup to
80 speed up things. Test with testfind().
81 - Add prefixlinks to where or not where absolute links to FD.
82 - Priority job to go to top of list.
83 - Find out why Full saves run slower and slower (hashing?)
84 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
85 - Implement some way for the Console to dynamically create a job.
86 - Restore to a particular time -- e.g. before date, after date.
87 - Implement disk spooling
88 - Solaris -I on tar for include list
89 - Enable avoid backing up archive device (findlib/find_one.c:128)
90 - Need a verbose mode in restore, perhaps to bsr.
91 - bscan without -v is too quiet -- perhaps show jobs.
92 - Add code to reject whole blocks if not wanted on restore.
93 - Start working on Base jobs.
94 - Make sure the MaxVolFiles is fully implemented in SD
95 - Check if both CatalogFiles and UseCatalog are set to SD.
96 - Check if we can increase Bacula FD priorty in Win2000
97 - Need return status on read_cb() from read_records(). Need multiple
98 records -- one per Job, maybe a JCR or some other structure with
100 - Figure out how to do a bare metal Windows restore
101 - Put system type returned by FD into catalog.
102 - Possibly add email to Watchdog if drive is unmounted too
103 long and a job is waiting on the drive.
104 - Use read_record.c in SD code.
105 - Why don't we get an error message from Win32 FD when bootstrap
106 file cannot be created for restore command?
107 - When Marking a file in Restore that is a hard link, also
108 mark the link so that the data will be reloaded.
109 - Restore program that errors in SD due to no tape reports
110 OK incorrectly in output.
111 - After unmount, if restore job started, ask to mount.
112 - Convert all %x substitution variables, which are hard to remember
113 and read to %(variable-name). Idea from TMDA.
114 - Add JobLevel in FD status (but make sure it is defined).
115 - Make Pool resource handle Counter resources.
116 - Remove NextId for SQLite. Optimize.
117 - Strip trailing / from Include
118 - Move all SQL statements into a single location.
119 - Add UA rc and history files.
120 - put termcap (used by console) in ./configure and
121 allow -with-termcap-dir.
122 - Enhance time and size scanning routines.
123 - Fix Autoprune for Volumes to respect need for full save.
124 - Fix Win32 config file definition name on /install
125 - No READLINE_SRC if found in alternate directory.
126 - Add Client FS/OS id (Linux, Win95/98, ...).
127 - Test a second language e.g. french.
128 - Compare tape to Client files (attributes, or attributes and data)
129 - Make all database Ids 64 bit.
130 - Write an applet for Linux.
131 - Add estimate to Console commands
132 - Find solution to blank filename (i.e. path only) problem.
133 - Implement new daemon communications protocol.
134 - Remove PoolId from Job table, it exists in Media.
135 - Allow console commands to detach or run in background.
136 - Fix status delay on storage daemon during rewind.
137 - Add SD message variables to control operator wait time
138 - Maximum Operator Wait
139 - Minimum Message Interval
140 - Maximum Message Interval
141 - Send Operator message when cannot read tape label.
142 - Verify level=Volume (scan only), level=Data (compare of data to file).
143 Verify level=Catalog, level=InitCatalog
145 - Add keyword search to show command in Console.
146 - Fix Win2000 error with no messages during startup.
147 - Events : tape has more than xxx bytes.
148 - Restrict characters permitted in a Resource name.
149 - Complete code in Bacula Resources -- this will permit
150 reading a new config file at any time.
151 - Handle ctl-c in Console
152 - Implement script driven addition of File daemon to config files.
153 - Think about how to make Bacula work better with File (non-tape) archives.
154 - Write Unix emulator for Windows.
156 - Implement new serialize subroutines
157 send(socket, "string", &Vol, "uint32", &i, NULL)
158 - Audit all UA commands to ensure that we always prompt where possible.
159 - If ./btape is called without /dev, assume argument is a Storage resource name.
160 - Put memory utilization in Status output of each daemon
161 if full status requested or if some level of debug on.
162 - Make database type selectable by .conf files i.e. at runtime
163 - Set flag for uname -a. Add to Volume label.
164 - Implement throttled work queue.
165 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
166 - Allow multiple Storage specifications (or multiple names on
167 a single Storage specification) in the Job record. Thus a job
168 can be backed up to a number of storage devices.
169 - Implement dump label to UA
170 - Concept of VolumeSet during restore which is a list
171 of Volume names needed.
172 - Restore files modified after date
173 - Restore file modified before date
174 - Emergency restore info:
176 - Backup working directory
178 - Restore -- do nothing but show what would happen
179 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
180 - Implement Restore FileSet=
181 - Create a protocol.h and protocol.c where all protocol messages
183 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
184 - Timout a job or terminate if link goes down, or reopen link and query.
185 - Find general solution for sscanf size problems (as well
186 as sprintf. Do at run time?
187 - Concept of precious tapes (cannot be reused).
188 - Make bcopy copy with a single tape drive.
189 - Permit changing ownership during restore.
191 - Autolabel should be specified by DIR instead of SD.
192 - Find out how to get the system tape block limits, e.g.:
193 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
194 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
197 - AutoScan (check checksum of tape)
198 - Format command = "format /dev/nst0"
202 - Seek resolution (usually corresponds to buffer size)
203 - EODErrorCode=ENOSPC or code
204 - Partial Read error code
205 - Partial write error code
206 - Nonformatted read error
207 - Nonformatted write error
208 - WriteProtected error
212 - IgnoreCloseErrors=yes
222 - FD sends unsaved file list to Director at end of job (see
224 - File daemon should build list of files skipped, and then
225 at end of save retry and report any errors.
226 - Write a Storage daemon that uses pipes and
227 standard Unix programs to write to the tape.
229 - Need something that monitors the JCR queue and
230 times out jobs by asking the deamons where they are.
231 - Enhance Jmsg code to permit buffering and saving to disk.
232 - device driver = "xxxx" for drives.
233 - restart: paranoid: read label fsf to
234 eom read append block, and go
235 super-paranoid: read label, read all files
236 in between, read append block, and go
237 verify: backspace, read append block, and go
238 permissive: same as above but frees drive
239 if tape is not valid.
241 - Ensure that /dev/null works
242 - Need report class for messages. Perhaps
243 report resource where report=group of messages
244 - enhance scan_attrib and rename scan_jobtype, and
245 fill in code for "since" option
246 - Director needs a time after which the report status is sent
247 anyway -- or better yet, a retry time for the job.
248 Don't reschedule a job if previous incarnation is still running.
249 - Some way to automatically backup everything is needed????
250 - Need a structure for pending actions:
252 - termination status (part of buffered msgs?)
253 - Concept of grouping Storage devices and job can use
254 any of a number of devices
256 Read, Write, Clean, Delete
257 - Login to Bacula; Bacula users with different permissions:
258 owner, group, user, quotas
259 - Store info on each file system type (probably in the job header on tape.
260 This could be the output of df; or perhaps some sort of /etc/mtab record.
263 - Design at hierarchial storage for Bacula. Migration and Clone.
264 - Implement FSM (File System Modules).
265 - Identify unchanged or "system" files and save them to a
266 special tape thus removing them from the standard
267 backup FileSet -- BASE backup.
268 - Turn virutally all sprintfs into snprintfs.
269 - Heartbeat between daemons.
270 - Audit M_ error codes to ensure they are correct and consistent.
271 - Add variable break characters to lex analyzer.
272 Either a bit mask or a string of chars so that
273 the caller can change the break characters.
274 - Make a single T_BREAK to replace T_COMMA, etc.
275 - Ensure that File daemon and Storage daemon can
276 continue a save if the Director goes down (this
277 is NOT currently the case). Must detect socket error,
278 buffer messages for later.
279 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
280 - Add ability to backup to two Storage devices (two SD sessions) at
281 the same time -- e.g. onsite, offsite.
282 - Add the ability to consolidate old backup sets (basically do a restore
283 to tape and appropriately update the catalog). Compress Volume sets.
284 Might need to spool via file is only one drive is available.
285 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
286 someway to do so with every volume that has less than x% valid
291 Bacula Projects Roadmap
293 last update 5 January 2003
295 Item 1: Multiple simultaneous Jobs. (done)
296 Done -- Restore part needs better implementation to work correctly
298 What: Permit multiple simultaneous jobs in Bacula.
300 Why: An enterprise level solution needs to go fast without the
301 need for the system administrator to carefully tweak
302 timing. Based on the benchmarks, during a full
303 backup, NetWorker typically hit 10 times the bandwidth to
304 the tape compared to Bacula--largely. This is probably due to
305 running parallel jobs and multi-threaded filling of buffers
306 and writing them to tape. This should also make things work
307 better when you have a mix of fast and slow machines backing
310 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
311 implementing this is a matter of some small cleanups and
315 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
316 Deferred -- not necessary yet.
318 What: If data is coming into the SD too fast, buffer it to
319 disk if the user has configured this option.
321 Why: This would be nice, especially if it more or less falls out
322 when implementing (1) above. If not, it probably should not
323 be given a high priority because fundamentally the backup time
324 is limited by the tape bandwidth. Even though you may finish a
325 client job quicker by spilling to disk, you still have to
326 eventually get it onto tape. If intermediate disk buffering
327 allows us to improve write bandwidth to tape, it may make
330 Notes: Whether or not this is implemented will depend upon performance
331 testing after item 1 is implemented.
334 Item 3: Write the bscan program -- also write a bcopy program.
337 What: Write a program that reads a Bacula tape and puts all the
338 appropriate data into the catalog. This allows recovery
339 from a tape that is no longer in the database, or it allows
340 re-creation of a database if lost.
342 Why: This is a fundamental robustness and disaster recovery tool
343 which will increase the comfort level of a sysadmin
344 considering adopting Bacula.
346 Notes: A skeleton of this program already exists, but much work
347 needs to be done. Implementing this will also make apparent
348 any deficiencies in the current Bacula tape format.
351 Item 4: Implement Base jobs.
353 What: A base job is sort of like a Full save except that you
354 will want the FileSet to contain only files that are unlikely
355 to change in the future (i.e. a snapshot of most of your
356 system after installing it). After the base job has been run,
357 when you are doing a Full save, you can specify to exclude
358 all files saved by the base job that have not been modified.
360 Why: This is something none of the competition does, as far as we know
361 (except BackupPC, which is a Perl program that saves to disk
362 only). It is big win for the user, it makes Bacula stand out
363 as offering a unique optimization that immediately saves time
366 Notes: Big savings in tape usage. Will require more resources because
367 the e. DIR must send FD a list of files/attribs, and the FD must
368 search the list and compare it for each file to be saved.
371 Item 5: Implement Label templates
373 What: This is a mechanism whereby Bacula can automatically create
374 a tape label for new tapes according to a detailed specification
375 provided by the user.
377 Why: It is a major convenience item for folks who use automated label
380 Notes: Bacula already has a working form of automatic tape label
381 creation, but it is very crude. The design for the complete
382 tape labeling project is already documented in the manual.
385 Item 6: Write a regression script.
388 What: This is an automatic script that runs and tests as many features
389 of Bacula as possible. The output is compared to previous
390 versions of Bacula and any differences are reported.
392 Why: This is an enormous help in preventing introduction of new
393 errors in parts of the program that already work correctly.
395 Notes: This probably should be ranked higher, it's something the typical
396 user doesn't see. Depending on how it's implemented, it may
397 make sense to defer it until the archival tape format and
398 user interface mature.
401 Item 7: GUI for interactive restore
402 Item 8: GUI for interactive backup
404 What: The current interactive restore is implemented with a tty
405 interface. It would be much nicer to be able to "see" the
406 list of files backed up in typical GUI tree format.
407 The same mechanism could also be used for creating
408 ad-hoc backup FileSets (item 8).
410 Why: Ease of use -- especially for the end user.
412 Notes: Rather than implementing in Gtk, we probably should go directly
413 for a Browser implementation, even if doing so meant the
414 capability wouldn't be available until much later. Not only
415 is there the question of Windows sites, most
416 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
417 without installing lots of stuff admins are very wary about.
418 Real sysadmins will always use the command line anyway, and
419 the user who's doing an interactive restore or backup of his
420 own files will in most cases be on a Windows machine running
424 Item 9: Add SSL to daemon communications.
426 What: This provides for secure communications between the daemons.
428 Why: This would allow doing backup across the Internet without
429 privacy concerns (or with much less concern).
431 Notes: The vast majority of near term potential users will be backing up
432 a single site over a LAN and, correctly or not, they probably
433 won't be concerned with security, at least not enough to go to
434 the trouble to set up keys, etc. to screw things down. We suspect
435 that many users genuinely interested in multi-site backup
436 already run some form of VPN software in their internetwork
437 connections, and are willing to delegate security to that layer.
440 Item 10: Define definitive tape format.
443 What: Define that definitive tape format that will not change
444 for the next millennium.
446 Why: Stability, security.
448 Notes: See notes for item 11 below.
451 Item 11: New daemon communication protocol.
453 What: The current daemon to daemon protocol is basically an ASCII
454 printf() and sending the buffer. On the receiving end, the
455 buffer is sscanf()ed to unpack it. The new scheme would
456 be a binary format that allows quick packing and unpacking
457 of any data type with named fields.
459 Why: Using binary packing would be faster. Named fields will permit
460 error checking to ensure that what is sent is what the
461 receiver really wants.
463 Notes: These are internal improvements in the interest of the
464 long-term stability and evolution of the program. On the one
465 hand, the sooner they're done, the less code we have to rip
466 up when the time comes to install them. On the other hand, they
467 don't bring an immediately perceptible benefit to potential
468 users. Item 10 and possibly item 11 should be deferred until Bacula
469 is well established with a growing user community more or
470 less happy with the feature set. At that time, it will make a
471 good "next generation" upgrade in the interest of data
477 ======================================================
479 It is somewhat like a Full save becomes an incremental since
480 the Base job (or jobs) plus other non-base files.
482 - New BaseFile table that contains:
483 JobId, BaseJobId, FileId (from Base).
484 i.e. for each base file that exists but is not saved because
485 it has not changed, the File daemon sends the JobId, BaseId,
486 and FileId back to the Director who creates the DB entry.
487 - To initiate a Base save, the Director sends the FD
488 the FileId, and full filename for each file in the Base.
489 - When the FD finds a Base file, he requests the Director to
490 send him the full File entry (stat packet plus MD5), or
491 conversely, the FD sends it to the Director and the Director
492 says yes or no. This can be quite rapid if the FileId is kept
493 by the FD for each Base Filename.
494 - It is probably better to have the comparison done by the FD
495 despite the fact that the File entry must be sent across the
497 - An alternative would be to send the FD the whole File entry
498 from the start. The disadvantage is that it requires a lot of
499 space. The advantage is that it requires less communications
501 - The Job record must be updated to indicate that one or more
503 - At end of Job, FD returns:
504 1. Count of base files/bytes not written to tape (i.e. matches)
505 2. Count of base file that were saved i.e. had changed.
506 - No tape record would be written for a Base file that matches, in the
507 same way that no tape record is written for Incremental jobs where
508 the file is not saved because it is unchanged.
509 - On a restore, all the Base file records must explicitly be
510 found from the BaseFile tape. I.e. for each Full save that is marked
511 to have one or more Base Jobs, search the BaseFile for all occurrences
513 - An optimization might be to make the BaseFile have:
519 This would avoid the need to explicitly fetch each File record for
520 the Base job. The Base Job record will be fetched to get the
521 VolSessionId and VolSessionTime.
522 =========================================================
524 ==========================================================
526 For each Incremental job that is run, there may be files that
527 were found but not saved because they were locked (this applies
528 only to Windows). Such a system could send back to the Director
529 a list of Unsaved files.
531 - New UnSavedFiles table that contains:
535 - Then in the next Incremental job, the list of Unsaved Files will be
536 feed to the FD, who will ensure that they are explicitly chosen even
537 if standard date/time check would not have selected them.
538 =============================================================
542 =============================================================
544 Request For Comments For File Backup Options
547 Subject: File Backup Options
550 A few days ago, a Bacula user who is backing up to file volumes and
551 using compression asked if it was possible to suppress compressing
552 all .gz files since it was a waste of CPU time. Although Bacula
553 currently permits using different options (compression, ...) on
554 a directory by directory basis, it cannot do it on a file by
555 file basis, which is clearly what was desired.
557 Proposed Implementation:
558 To solve this problem, I propose the following:
560 - Add a new Director resource type called FileOptions.
562 - The FileOptions resource will have records for all
563 options that can currently be specified on the Include record
564 (in a FileSet). Examples below.
566 - The FileOptions resource will permit an exclude option as well
567 as a number of additional options.
569 - The heart of the FileOptions resource is the ability to
570 supply any number of ApplyTo records which specify POSIX
571 regular expressions. These ApplyTo regular expressions are
572 applied to the fully qualified filename (path and all). If
573 one matches, then the FileOptions will be used.
575 - When an ApplyTo specification matches an included file, the
576 options specified in the FileOptions resource will override
577 the default options specified on the Include record.
579 - Include records will be modified to permit referencing one or
580 more FileOptions resources. The FileOptions will be used
581 in the order listed on the Include record and the first
582 one that matches will be applied.
584 - Options (or specifications) currently supplied on the Include
585 record will be deprecated (i.e. removed in a later version a
586 year or so from now).
588 - The Exclude record will be deprecated as the same functionality
589 can be obtained by using an Exclude = yes in the FileOptions.
592 The following records can appear in the FileOptions resource. An
593 asterisk preceding the name indicates a feature not currently
597 - Compression= (GZIP, ...)
598 - Signature= (MD5, SHA1, ...)
600 - OneFs= (yes/no) - remain on one filesystem
601 - Recurse= (yes/no) - recurse into subdirectories
602 - Sparse= (yes/no) - do sparse file backup
603 - *Exclude= (yes/no) - exclude file from being saved
604 - *Reader= (filename) - external read (backup) program
605 - *Plugin= (filename) - read/write plugin module
608 - verify= (ipnougsamc5) - verify options
611 - replace= (always/ifnewer/ifolder/never) - replace options currently
613 - *Writer= (filename) - external write (restore) program
617 Currently options specifying compression, MD5 signatures, recursion,
618 ... of a FileSet are supplied on the Include record. These will now
619 all be collected into a FileOptions resource, which will be
620 specified on the Include in place of the options. Multiple FileOptions
621 may be specified. Since the FileOptions contain regular expressions
622 that are applied to the full filename, this will give the ability
623 to specify backup options on a file by file basis to whatever level
632 Include = compression=GZIP signature=MD5 {
641 Include = FileOptions=Opts {
652 That's a lot more to do the same thing, but it gives the ability to
653 apply options on a file by file basis. For example, suppose you
654 want to compress all files but not any file with extensions .gz or .Z.
655 You could do so as follows:
659 Include = FileOptions=NoCompress FileOptions=Opts {
667 ApplyTo = /*.?*/ # matches all files
672 # Note multiple ApplyTos are ORed
673 ApplyTo = /*.gz/ # matches .gz files */
674 ApplyTo = /*.Z/ # matches .Z files */
677 Now, since the NoCompress FileOptions is specified first on the
678 Include line, any *.gz or *.Z file will have an MD5 signature computed,
679 but will not be compressed. For all other files, the NoCompress will not
680 match, so the Opts options will be used which will include GZIP
684 - Is it necessary to provide some means of ANDing regular expressions
685 and negation? (not currently planned)
687 e.g. ApplyTo = /*.gz/ && !/big.gz/
689 - I see that Networker has a "null" module which, if specified, does not
690 backup the file, but does make an record of the file in the catalog
691 so that the catalog will reflect an exact picture of the filesystem.
692 The result is that the file can be "seen" when "browsing" the save
693 sets, but it cannot be restored.
695 Is this really useful? Should it be implemented in Bacula?
698 After implementing the above, the user will be able to specify
699 on a file by file basis (using regular expressions) what options are
700 applied for the backup.
701 ====================================
703 =========================================
704 Proposal by Bill Sellers
706 Return-Path: <w.a.sellers@larc.nasa.gov>
707 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
708 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)
709 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
710 X-Sender: w.a.sellers@pop.larc.nasa.gov
711 X-Mailer: QUALCOMM Windows Eudora Version 5.1
712 Date: Tue, 14 Jan 2003 16:30:18 -0500
713 To: Kern Sibbald <kern@sibbald.com>
714 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
715 Subject: Re: [Bacula-users] Bacula remote storage?
716 In-Reply-To: <1042565382.1845.177.camel@rufus>
717 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>
719 Content-Type: text/plain; charset="us-ascii"; format=flowed
720 X-Annoyance-Filter-Junk-Probability: 0
721 X-Annoyance-Filter-Classification: Mail
722 At 06:29 PM 1/14/2003 +0100, you wrote:
725 >Well, if you cannot put a Bacula client on the machine,
726 >then it is a big problem. If you know of some software
727 >that can do what you want, let me know, because I
728 >really just don't know how to do it -- at least not
734 We have been able to get Amanda to use the HSM as a storage
735 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
736 handle Windows systems very well (or at all without Samba). So I am
737 looking for a backup system that has a Windows client. I really like the
738 Windows integration of Bacula.
740 From the command line, its rather trivial to move the data around. We use
743 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
745 or if you use GNU tar:
747 tar czf hsm:path/file.tgz ./files
749 One idea for you to consider; Sendmail offers pipes in the aliases file;
750 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
751 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
752 pipe available, as a storage device? Then we could use any command that
753 handles stdin as a storage destination.
760 #Password is not used in rsh, but might be used in ftp.
761 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
768 Password = "foobar&-"
769 Device = "| ncftpput -c hsm /path/file.bacula"
773 >If you have some local storage available, you could
774 >use Bacula to backup to disk volumes, then use some
775 >other software (ftp, scp) to move them to the HSM
776 >machine. However, this is a bit kludgy.
779 It is, but maybe worth a try. Is there some function in Bacula to put
780 variables in filenames? i.e. backup.2003-01-15.root
787 w.a.sellers@larc.nasa.gov
789 ==============================================
790 The Project for the above
792 I finally realized that this is not at all
793 the same as reader/writer programs or plugins,
794 which are alternate ways of accessing the
795 files to be backed up. Rather, it is an alternate
796 form of storage device, and I have always planned
797 that Bacula should be able to handle all sorts
800 So, I propose the following phases:
802 1. OK from you to invest some time in testing
803 this as I implement it (requires that you
804 know how to download from the SourceForge
805 cvs -- which I imagine is a piece of cake
808 2. Dumb implementation by allowing a device to
809 be a fifo for write only.
810 Reason: easy to implement, proof of concept.
812 3. Try reading from fifo but with fixed block
814 Reason: proof of concept, easy to implement.
816 4. Extend reading from fifo (restores) to handle
818 Reason: requires some delicate low level coding
819 which could destabilize all of Bacula.
821 5. Implementation of above but to a program. E.g.
822 Device = "|program" (not full pipeline).
823 Reason: routines already exist, and program can
824 be a shell script which contains anything.
826 6. Add full pipeline as a possibility. E.g.
827 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
828 Reason: needs additional coding to implement full
829 pipeline (must fire off either a shell or all
830 programs and connect their pipes).
832 There are a good number of details in each step
833 that I have left out, but I will specify them at
834 every stage, and there may be a few changes as things
835 evolve. I expect that to get to stage 5 will take a
836 few weeks, and at that point, you will have
837 everything you need (just inside a script).
838 Stage 6 will probably take longer, but if this
839 project pleases you, what we do for 5 should
840 be adequate for some time.
844 =============================================
846 Done: (see kernsdone for more)
847 - Look into Pruning/purging problems or why there seem to
848 be so many files listed each night.
849 - Fix cancel in find_one -- need jcr.
850 - Cancel does not work for restore in FD.
851 - Write SetJobStatus() function so cancel status not lost.
852 - Add include list to end of chain in findlib
853 - Zap sd_auth_key after use
854 - Add Bar code reading capabilities (new mtx-changer)
855 - Figure out some way to automatically backup all local partitions
856 - Make hash table for linked files in findlib/find_one.c:161
858 - Rewrite find_one.c to use only pool_memory instead of
859 alloca and malloc (probably not necessary).
860 - Make sure btraceback goes into /sbin not sysconf directory.
861 - InitVerify is getting pruned and it shouldn't (document it)
862 - Make 1.28c release ??? NO do 1.29 directly
863 - Set timeout on opening fifo for save or restore (findlib)
864 - Document FIFO storage device.
865 - Document fifo and | and <
868 - Get correct error status from run_program or open_bpipe().
869 - Restrict permissions on File Volumes (now 0640).
870 - Umasked 022 daemons
871 - Fix restore of hard linked file.
872 - Figure out how to allow multiple simultaneous file Volumes on a single device.
873 - Implement multiple simultaneous file Volumes on a single device.
874 - Cleanup db_update_media and db_update_pool
875 - Flush all the daemon messages at the end of every job.
876 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
877 run_program -- bpipe.c
878 - Apparently cancel does not work for jobs waiting to be
880 - Implement TCP/IP connection for MySQL
881 - Pull a canceled job from the Scheduling queue.
882 - Implement max_file_size in block.c (already done, just tweaked).
883 - Look at purge jobs volume (at least document it, and see if it is
885 - Add list volumes does all pools. list volumes pool=xxx now works.
886 - Add pool= to "list media" in ua_output.c
887 - Strip trailing slashes from Include directory names in the FD.
888 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
889 ERR=Operation not permited loop.
890 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
891 - Produce better error messages in when error/eof writing block.
892 - Cancelling of a queued job does NOT work!!!!!!
894 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
895 rufus-sd: Recycled volume File0003 on device /home/kern/bacula/working, all previous data lost.
896 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
897 - Ability to backup to a file then later transfer to a tape -- Migration.
898 Migration based on MaxJobs(MinJobs),MaxVols(MinVols),AgeJobs,MaxBytes(MinBytes)
899 (i.e. HighwaterMark, LowwaterMark).
900 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
901 set of volumes recycling the oldest volume when it is needed.
902 - gethostbyname failure in bnet_connect() continues
903 generating errors -- should stop.
904 - Add chflags() code for FreeBSD file flags
905 - Bevan Anderson suggests having a run queue for each device
906 so that multiple simultaneous jobs can run but each writing
907 to a different Volume.
908 - Look at handling <> in smtp doesn't work with exim.
909 - Need to specify MaximumConcurrentJobs in the Job resource.
910 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
911 - Implement finer multiprocessing options.
912 - Implement | and < in Exclude statements.
913 - Figure out some way to specify a retention period for files
914 that no longer exist on the machine -- so that we maintain
915 say backups for 30 days, but if the file is deleted, we maintain
916 the last copy for 1 year. -- answer Volume retention.