4 Documentation to do: (a little bit at a time)
5 - Document running a test version.
6 - Document query file format.
7 - Document static linking
8 - Document how to automatically backup all local partitions
9 - Document problems with Verify and pruning.
10 - Document how to use multiple databases.
11 - Add a section to the doc on Manual cycling of Volumes.
14 - Document default file backup
17 Testing to do: (painful)
18 - that ALL console command line options work and are always implemented
19 - blocksize recognition code.
20 - multiple simultaneous Volumes
23 - Do not ignore SIGCHLD
24 - Fix Bare Metal restore problem.
25 - Fix "access not allowed" for backup of files on WinXP.
26 - Document new transparent Console commands and wait command.
27 - Document new immediate File save capability.
28 - Document OPTIMIZE TABLE in MySQL
29 - Fix watchdog pthread crash on Win32
30 - Document labeling a whole magazine using "cat"
32 - Figure out how to use ssh or stunnel to protect Bacula communications.
35 - Implement restore "current system", but take all files without
36 doing selection tree -- so that jobs without File records can
38 - Implement scheduling of one time "run" jobs (i.e. instead of
39 starting immediately start at some specified time).
40 - Make | and < work on FD side.
41 - Pass prefix_links to FD.
42 - Implement a M_SECURITY message class.
43 - Implement disk spooling. Two parts: 1. Spool to disk then
44 immediately to tape to speed up tape operations. 2. Spool to
45 disk only when the tape is full, then when a tape is hung move
47 - From Phil Stracchino:
48 It would probably be a per-client option, and would be called
49 something like, say, "Automatically purge obsoleted jobs". What it
50 would do is, when you successfully complete a Differential backup of a
51 client, it would automatically purge all Incremental backups for that
52 client that are rendered redundant by that Differential. Likewise,
53 when a Full backup on a client completed, it would automatically purge
54 all Differential and Incremental jobs obsoleted by that Full backup.
55 This would let people minimize the number of tapes they're keeping on
56 hand without having to master the art of retention times.
58 - Allow multiple Storage specifications (or multiple names on
59 a single Storage specification) in the Job record. Thus a job
60 can be backed up to a number of storage devices.
61 - Implement dump label to UA
62 - Add prefixlinks to where or not where absolute links to FD.
63 - Look at Python for a Bacula scripting language -- www.python.org
64 - Issue message to mount a new tape before the rewind.
65 - Simplified client job initiation for portables.
66 - If SD cannot open a drive, make it periodically retry.
67 - Implement LabelTemplate (at least first cut).
68 - Add more of the config info to the tape label.
69 - Implement a Mount Command and an Unmount Command where
70 the users could specify a system command to be performed
71 to do the mount, after which Bacula could attempt to
72 read the device. This is for Removeable media such as a CDROM.
73 - Most likely, this mount command would be invoked explicitly
74 by the user using the current Console "mount" and "unmount"
75 commands -- the Storage Daemon would do the right thing
76 depending on the exact nature of the device.
77 - As with tape drives, when Bacula wanted a new removable
78 disk mounted, it would unmount the old one, and send a message
79 to the user, who would then use "mount" as described above
80 once he had actually inserted the disk.
82 - Make some way so that if a machine is skipped because it is not up
83 that Bacula will continue retrying for a specified period of time --
85 - If tape is marked read-only, then try opening it read-only rather than
86 failing, and remember that it cannot be written.
87 - Refine SD waiting output:
88 Device is being positioned
89 > Device is being positioned for append
90 > Device is being positioned to file x
92 - Figure out some way to estimate output size and to avoid splitting
93 a backup across two Volumes -- this could be useful for writing CDROMs
94 where you really prefer not to have it split -- not serious.
95 - Add RunBeforeJob and RunAfterJob to the Client program.
96 - Have SD compute MD5 or SHA1 and compare to what FD computes.
97 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
98 actual data on the Volume and compare it.
99 - Implement FileOptions (see end of this document)
100 - Implement Bacula plugins -- design API
101 - Make bcopy read through bad tape records.
102 - Fix read_record to handle multiple sessions.
103 - Program files (i.e. execute a program to read/write files).
104 Pass read date of last backup, size of file last time.
105 - Add Signature type to File DB record.
106 - Make Restore report an error if FD or SD term codes are not OK.
107 - CD into subdirectory when open()ing files for backup to
108 speed up things. Test with testfind().
109 - Priority job to go to top of list.
110 - Find out why Full saves run slower and slower (hashing?)
111 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
112 - Implement some way for the Console to dynamically create a job.
113 - Restore to a particular time -- e.g. before date, after date.
114 - Solaris -I on tar for include list
115 - Prohibit backing up archive device (findlib/find_one.c:128)
116 - Need a verbose mode in restore, perhaps to bsr.
117 - bscan without -v is too quiet -- perhaps show jobs.
118 - Add code to reject whole blocks if not wanted on restore.
119 - Start working on Base jobs.
120 - Check if we can increase Bacula FD priorty in Win2000
121 - Make sure the MaxVolFiles is fully implemented in SD
122 - Check if both CatalogFiles and UseCatalog are set to SD.
123 - Need return status on read_cb() from read_records(). Need multiple
124 records -- one per Job, maybe a JCR or some other structure with
125 a block and a record.
126 - Figure out how to do a bare metal Windows restore
127 - Put system type returned by FD into catalog.
128 - Possibly add email to Watchdog if drive is unmounted too
129 long and a job is waiting on the drive.
130 - Use read_record.c in SD code.
131 - Why don't we get an error message from Win32 FD when bootstrap
132 file cannot be created for restore command?
133 - When Marking a file in Restore that is a hard link, also
134 mark the link so that the data will be reloaded.
135 - Restore program that errors in SD due to no tape reports
136 OK incorrectly in output.
137 - After unmount, if restore job started, ask to mount.
138 - Convert all %x substitution variables, which are hard to remember
139 and read to %(variable-name). Idea from TMDA.
140 - Add JobLevel in FD status (but make sure it is defined).
141 - Make Pool resource handle Counter resources.
142 - Remove NextId for SQLite. Optimize.
143 - Move all SQL statements into a single location.
144 - Add UA rc and history files.
145 - put termcap (used by console) in ./configure and
146 allow -with-termcap-dir.
147 - Enhance time and size scanning routines.
148 - Fix Autoprune for Volumes to respect need for full save.
149 - Fix Win32 config file definition name on /install
150 - No READLINE_SRC if found in alternate directory.
151 - Test a second language e.g. french.
152 - Compare tape to Client files (attributes, or attributes and data)
153 - Make all database Ids 64 bit.
154 - Write an applet for Linux.
155 - Add estimate to Console commands
156 - Find solution to blank filename (i.e. path only) problem.
157 - Implement new daemon communications protocol.
158 - Remove PoolId from Job table, it exists in Media.
159 - Allow console commands to detach or run in background.
160 - Fix status delay on storage daemon during rewind.
161 - Add SD message variables to control operator wait time
162 - Maximum Operator Wait
163 - Minimum Message Interval
164 - Maximum Message Interval
165 - Send Operator message when cannot read tape label.
166 - Verify level=Volume (scan only), level=Data (compare of data to file).
167 Verify level=Catalog, level=InitCatalog
169 - Add keyword search to show command in Console.
170 - Fix Win2000 error with no messages during startup.
171 - Events : tape has more than xxx bytes.
172 - Restrict characters permitted in a Resource name.
173 - Complete code in Bacula Resources -- this will permit
174 reading a new config file at any time.
175 - Handle ctl-c in Console
176 - Implement script driven addition of File daemon to config files.
177 - Think about how to make Bacula work better with File (non-tape) archives.
178 - Write Unix emulator for Windows.
179 - Implement new serialize subroutines
180 send(socket, "string", &Vol, "uint32", &i, NULL)
181 - Audit all UA commands to ensure that we always prompt where possible.
182 - If ./btape is called without /dev, assume argument is a Storage resource name.
183 - Put memory utilization in Status output of each daemon
184 if full status requested or if some level of debug on.
185 - Make database type selectable by .conf files i.e. at runtime
186 - Set flag for uname -a. Add to Volume label.
187 - Implement throttled work queue.
188 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
189 - Restore files modified after date
190 - Restore file modified before date
191 - Emergency restore info:
193 - Backup working directory
195 - Restore -- do nothing but show what would happen
196 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
197 - Implement Restore FileSet=
198 - Create a protocol.h and protocol.c where all protocol messages
200 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
201 - Timout a job or terminate if link goes down, or reopen link and query.
202 - Find general solution for sscanf size problems (as well
203 as sprintf. Do at run time?
204 - Concept of precious tapes (cannot be reused).
205 - Make bcopy copy with a single tape drive.
206 - Permit changing ownership during restore.
208 - Autolabel should be specified by DIR instead of SD.
209 - Find out how to get the system tape block limits, e.g.:
210 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
211 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
214 - AutoScan (check checksum of tape)
215 - Format command = "format /dev/nst0"
219 - Seek resolution (usually corresponds to buffer size)
220 - EODErrorCode=ENOSPC or code
221 - Partial Read error code
222 - Partial write error code
223 - Nonformatted read error
224 - Nonformatted write error
225 - WriteProtected error
229 - IgnoreCloseErrors=yes
239 - FD sends unsaved file list to Director at end of job (see
241 - File daemon should build list of files skipped, and then
242 at end of save retry and report any errors.
243 - Write a Storage daemon that uses pipes and
244 standard Unix programs to write to the tape.
246 - Need something that monitors the JCR queue and
247 times out jobs by asking the deamons where they are.
248 - Enhance Jmsg code to permit buffering and saving to disk.
249 - device driver = "xxxx" for drives.
250 - restart: paranoid: read label fsf to
251 eom read append block, and go
252 super-paranoid: read label, read all files
253 in between, read append block, and go
254 verify: backspace, read append block, and go
255 permissive: same as above but frees drive
256 if tape is not valid.
258 - Ensure that /dev/null works
259 - Need report class for messages. Perhaps
260 report resource where report=group of messages
261 - enhance scan_attrib and rename scan_jobtype, and
262 fill in code for "since" option
263 - Director needs a time after which the report status is sent
264 anyway -- or better yet, a retry time for the job.
265 Don't reschedule a job if previous incarnation is still running.
266 - Some way to automatically backup everything is needed????
267 - Need a structure for pending actions:
269 - termination status (part of buffered msgs?)
271 Read, Write, Clean, Delete
272 - Login to Bacula; Bacula users with different permissions:
273 owner, group, user, quotas
274 - Store info on each file system type (probably in the job header on tape.
275 This could be the output of df; or perhaps some sort of /etc/mtab record.
278 - Design at hierarchial storage for Bacula. Migration and Clone.
279 - Implement FSM (File System Modules).
280 - Identify unchanged or "system" files and save them to a
281 special tape thus removing them from the standard
282 backup FileSet -- BASE backup.
283 - Heartbeat between daemons.
284 - Audit M_ error codes to ensure they are correct and consistent.
285 - Add variable break characters to lex analyzer.
286 Either a bit mask or a string of chars so that
287 the caller can change the break characters.
288 - Make a single T_BREAK to replace T_COMMA, etc.
289 - Ensure that File daemon and Storage daemon can
290 continue a save if the Director goes down (this
291 is NOT currently the case). Must detect socket error,
292 buffer messages for later.
293 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
294 - Add ability to backup to two Storage devices (two SD sessions) at
295 the same time -- e.g. onsite, offsite.
296 - Add the ability to consolidate old backup sets (basically do a restore
297 to tape and appropriately update the catalog). Compress Volume sets.
298 Might need to spool via file is only one drive is available.
299 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
300 someway to do so with every volume that has less than x% valid
305 Bacula Projects Roadmap
307 last update 5 January 2003
309 Item 1: Multiple simultaneous Jobs. (done)
310 Done -- Restore part needs better implementation to work correctly
312 What: Permit multiple simultaneous jobs in Bacula.
314 Why: An enterprise level solution needs to go fast without the
315 need for the system administrator to carefully tweak
316 timing. Based on the benchmarks, during a full
317 backup, NetWorker typically hit 10 times the bandwidth to
318 the tape compared to Bacula--largely. This is probably due to
319 running parallel jobs and multi-threaded filling of buffers
320 and writing them to tape. This should also make things work
321 better when you have a mix of fast and slow machines backing
324 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
325 implementing this is a matter of some small cleanups and
329 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
330 Deferred -- not necessary yet.
332 What: If data is coming into the SD too fast, buffer it to
333 disk if the user has configured this option.
335 Why: This would be nice, especially if it more or less falls out
336 when implementing (1) above. If not, it probably should not
337 be given a high priority because fundamentally the backup time
338 is limited by the tape bandwidth. Even though you may finish a
339 client job quicker by spilling to disk, you still have to
340 eventually get it onto tape. If intermediate disk buffering
341 allows us to improve write bandwidth to tape, it may make
344 Notes: Whether or not this is implemented will depend upon performance
345 testing after item 1 is implemented.
348 Item 3: Write the bscan program -- also write a bcopy program.
351 What: Write a program that reads a Bacula tape and puts all the
352 appropriate data into the catalog. This allows recovery
353 from a tape that is no longer in the database, or it allows
354 re-creation of a database if lost.
356 Why: This is a fundamental robustness and disaster recovery tool
357 which will increase the comfort level of a sysadmin
358 considering adopting Bacula.
360 Notes: A skeleton of this program already exists, but much work
361 needs to be done. Implementing this will also make apparent
362 any deficiencies in the current Bacula tape format.
365 Item 4: Implement Base jobs.
367 What: A base job is sort of like a Full save except that you
368 will want the FileSet to contain only files that are unlikely
369 to change in the future (i.e. a snapshot of most of your
370 system after installing it). After the base job has been run,
371 when you are doing a Full save, you can specify to exclude
372 all files saved by the base job that have not been modified.
374 Why: This is something none of the competition does, as far as we know
375 (except BackupPC, which is a Perl program that saves to disk
376 only). It is big win for the user, it makes Bacula stand out
377 as offering a unique optimization that immediately saves time
380 Notes: Big savings in tape usage. Will require more resources because
381 the e. DIR must send FD a list of files/attribs, and the FD must
382 search the list and compare it for each file to be saved.
385 Item 5: Implement Label templates
387 What: This is a mechanism whereby Bacula can automatically create
388 a tape label for new tapes according to a detailed specification
389 provided by the user.
391 Why: It is a major convenience item for folks who use automated label
394 Notes: Bacula already has a working form of automatic tape label
395 creation, but it is very crude. The design for the complete
396 tape labeling project is already documented in the manual.
399 Item 6: Write a regression script.
400 Done -- Continue to expand its testing.
402 What: This is an automatic script that runs and tests as many features
403 of Bacula as possible. The output is compared to previous
404 versions of Bacula and any differences are reported.
406 Why: This is an enormous help in preventing introduction of new
407 errors in parts of the program that already work correctly.
409 Notes: This probably should be ranked higher, it's something the typical
410 user doesn't see. Depending on how it's implemented, it may
411 make sense to defer it until the archival tape format and
412 user interface mature.
415 Item 7: GUI for interactive restore
416 Item 8: GUI for interactive backup
418 What: The current interactive restore is implemented with a tty
419 interface. It would be much nicer to be able to "see" the
420 list of files backed up in typical GUI tree format.
421 The same mechanism could also be used for creating
422 ad-hoc backup FileSets (item 8).
424 Why: Ease of use -- especially for the end user.
426 Notes: Rather than implementing in Gtk, we probably should go directly
427 for a Browser implementation, even if doing so meant the
428 capability wouldn't be available until much later. Not only
429 is there the question of Windows sites, most
430 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
431 without installing lots of stuff admins are very wary about.
432 Real sysadmins will always use the command line anyway, and
433 the user who's doing an interactive restore or backup of his
434 own files will in most cases be on a Windows machine running
438 Item 9: Add SSL to daemon communications.
440 What: This provides for secure communications between the daemons.
442 Why: This would allow doing backup across the Internet without
443 privacy concerns (or with much less concern).
445 Notes: The vast majority of near term potential users will be backing up
446 a single site over a LAN and, correctly or not, they probably
447 won't be concerned with security, at least not enough to go to
448 the trouble to set up keys, etc. to screw things down. We suspect
449 that many users genuinely interested in multi-site backup
450 already run some form of VPN software in their internetwork
451 connections, and are willing to delegate security to that layer.
454 Item 10: Define definitive tape format.
457 What: Define that definitive tape format that will not change
458 for the next millennium.
460 Why: Stability, security.
462 Notes: See notes for item 11 below.
465 Item 11: New daemon communication protocol.
467 What: The current daemon to daemon protocol is basically an ASCII
468 printf() and sending the buffer. On the receiving end, the
469 buffer is sscanf()ed to unpack it. The new scheme would
470 be a binary format that allows quick packing and unpacking
471 of any data type with named fields.
473 Why: Using binary packing would be faster. Named fields will permit
474 error checking to ensure that what is sent is what the
475 receiver really wants.
477 Notes: These are internal improvements in the interest of the
478 long-term stability and evolution of the program. On the one
479 hand, the sooner they're done, the less code we have to rip
480 up when the time comes to install them. On the other hand, they
481 don't bring an immediately perceptible benefit to potential
482 users. Item 10 and possibly item 11 should be deferred until Bacula
483 is well established with a growing user community more or
484 less happy with the feature set. At that time, it will make a
485 good "next generation" upgrade in the interest of data
491 ======================================================
493 It is somewhat like a Full save becomes an incremental since
494 the Base job (or jobs) plus other non-base files.
496 - New BaseFile table that contains:
497 JobId, BaseJobId, FileId (from Base).
498 i.e. for each base file that exists but is not saved because
499 it has not changed, the File daemon sends the JobId, BaseId,
500 and FileId back to the Director who creates the DB entry.
501 - To initiate a Base save, the Director sends the FD
502 the FileId, and full filename for each file in the Base.
503 - When the FD finds a Base file, he requests the Director to
504 send him the full File entry (stat packet plus MD5), or
505 conversely, the FD sends it to the Director and the Director
506 says yes or no. This can be quite rapid if the FileId is kept
507 by the FD for each Base Filename.
508 - It is probably better to have the comparison done by the FD
509 despite the fact that the File entry must be sent across the
511 - An alternative would be to send the FD the whole File entry
512 from the start. The disadvantage is that it requires a lot of
513 space. The advantage is that it requires less communications
515 - The Job record must be updated to indicate that one or more
517 - At end of Job, FD returns:
518 1. Count of base files/bytes not written to tape (i.e. matches)
519 2. Count of base file that were saved i.e. had changed.
520 - No tape record would be written for a Base file that matches, in the
521 same way that no tape record is written for Incremental jobs where
522 the file is not saved because it is unchanged.
523 - On a restore, all the Base file records must explicitly be
524 found from the BaseFile tape. I.e. for each Full save that is marked
525 to have one or more Base Jobs, search the BaseFile for all occurrences
527 - An optimization might be to make the BaseFile have:
533 This would avoid the need to explicitly fetch each File record for
534 the Base job. The Base Job record will be fetched to get the
535 VolSessionId and VolSessionTime.
536 =========================================================
538 ==========================================================
540 For each Incremental job that is run, there may be files that
541 were found but not saved because they were locked (this applies
542 only to Windows). Such a system could send back to the Director
543 a list of Unsaved files.
545 - New UnSavedFiles table that contains:
549 - Then in the next Incremental job, the list of Unsaved Files will be
550 feed to the FD, who will ensure that they are explicitly chosen even
551 if standard date/time check would not have selected them.
552 =============================================================
556 =============================================================
558 Request For Comments For File Backup Options
561 Subject: File Backup Options
564 A few days ago, a Bacula user who is backing up to file volumes and
565 using compression asked if it was possible to suppress compressing
566 all .gz files since it was a waste of CPU time. Although Bacula
567 currently permits using different options (compression, ...) on
568 a directory by directory basis, it cannot do it on a file by
569 file basis, which is clearly what was desired.
571 Proposed Implementation:
572 To solve this problem, I propose the following:
574 - Add a new Director resource type called FileOptions.
576 - The FileOptions resource will have records for all
577 options that can currently be specified on the Include record
578 (in a FileSet). Examples below.
580 - The FileOptions resource will permit an exclude option as well
581 as a number of additional options.
583 - The heart of the FileOptions resource is the ability to
584 supply any number of ApplyTo records which specify POSIX
585 regular expressions. These ApplyTo regular expressions are
586 applied to the fully qualified filename (path and all). If
587 one matches, then the FileOptions will be used.
589 - When an ApplyTo specification matches an included file, the
590 options specified in the FileOptions resource will override
591 the default options specified on the Include record.
593 - Include records will be modified to permit referencing one or
594 more FileOptions resources. The FileOptions will be used
595 in the order listed on the Include record and the first
596 one that matches will be applied.
598 - Options (or specifications) currently supplied on the Include
599 record will be deprecated (i.e. removed in a later version a
600 year or so from now).
602 - The Exclude record will be deprecated as the same functionality
603 can be obtained by using an Exclude = yes in the FileOptions.
606 The following records can appear in the FileOptions resource. An
607 asterisk preceding the name indicates a feature not currently
611 - Compression= (GZIP, ...)
612 - Signature= (MD5, SHA1, ...)
614 - OneFs= (yes/no) - remain on one filesystem
615 - Recurse= (yes/no) - recurse into subdirectories
616 - Sparse= (yes/no) - do sparse file backup
617 - *Exclude= (yes/no) - exclude file from being saved
618 - *Reader= (filename) - external read (backup) program
619 - *Plugin= (filename) - read/write plugin module
622 - verify= (ipnougsamc5) - verify options
625 - replace= (always/ifnewer/ifolder/never) - replace options currently
627 - *Writer= (filename) - external write (restore) program
631 Currently options specifying compression, MD5 signatures, recursion,
632 ... of a FileSet are supplied on the Include record. These will now
633 all be collected into a FileOptions resource, which will be
634 specified on the Include in place of the options. Multiple FileOptions
635 may be specified. Since the FileOptions contain regular expressions
636 that are applied to the full filename, this will give the ability
637 to specify backup options on a file by file basis to whatever level
646 Include = compression=GZIP signature=MD5 {
655 Include = FileOptions=Opts {
666 That's a lot more to do the same thing, but it gives the ability to
667 apply options on a file by file basis. For example, suppose you
668 want to compress all files but not any file with extensions .gz or .Z.
669 You could do so as follows:
673 Include = FileOptions=NoCompress FileOptions=Opts {
681 ApplyTo = /*.?*/ # matches all files
686 # Note multiple ApplyTos are ORed
687 ApplyTo = /*.gz/ # matches .gz files */
688 ApplyTo = /*.Z/ # matches .Z files */
691 Now, since the NoCompress FileOptions is specified first on the
692 Include line, any *.gz or *.Z file will have an MD5 signature computed,
693 but will not be compressed. For all other files, the NoCompress will not
694 match, so the Opts options will be used which will include GZIP
698 - Is it necessary to provide some means of ANDing regular expressions
699 and negation? (not currently planned)
701 e.g. ApplyTo = /*.gz/ && !/big.gz/
703 - I see that Networker has a "null" module which, if specified, does not
704 backup the file, but does make an record of the file in the catalog
705 so that the catalog will reflect an exact picture of the filesystem.
706 The result is that the file can be "seen" when "browsing" the save
707 sets, but it cannot be restored.
709 Is this really useful? Should it be implemented in Bacula?
712 After implementing the above, the user will be able to specify
713 on a file by file basis (using regular expressions) what options are
714 applied for the backup.
715 ====================================
717 =========================================
718 Proposal by Bill Sellers
720 Return-Path: <w.a.sellers@larc.nasa.gov>
721 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
722 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)
723 Message-Id: <5.1.0.14.2.20030114153452.028dbae8@pop.larc.nasa.gov>
724 X-Sender: w.a.sellers@pop.larc.nasa.gov
725 X-Mailer: QUALCOMM Windows Eudora Version 5.1
726 Date: Tue, 14 Jan 2003 16:30:18 -0500
727 To: Kern Sibbald <kern@sibbald.com>
728 From: Bill Sellers <w.a.sellers@larc.nasa.gov>
729 Subject: Re: [Bacula-users] Bacula remote storage?
730 In-Reply-To: <1042565382.1845.177.camel@rufus>
731 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>
733 Content-Type: text/plain; charset="us-ascii"; format=flowed
734 X-Annoyance-Filter-Junk-Probability: 0
735 X-Annoyance-Filter-Classification: Mail
736 At 06:29 PM 1/14/2003 +0100, you wrote:
739 >Well, if you cannot put a Bacula client on the machine,
740 >then it is a big problem. If you know of some software
741 >that can do what you want, let me know, because I
742 >really just don't know how to do it -- at least not
748 We have been able to get Amanda to use the HSM as a storage
749 device. Someone here wrote a driver for Amanda. BUT, Amanda doesn't
750 handle Windows systems very well (or at all without Samba). So I am
751 looking for a backup system that has a Windows client. I really like the
752 Windows integration of Bacula.
754 From the command line, its rather trivial to move the data around. We use
757 tar cf - ./files | gzip -c | rsh hsm dd of=path/file.tgz
759 or if you use GNU tar:
761 tar czf hsm:path/file.tgz ./files
763 One idea for you to consider; Sendmail offers pipes in the aliases file;
764 (mailpipe: "|/usr/bin/vacation root") and Perl supports pipes in the
765 "open" statement (open FILE, "|/bin/nroff -man";) Could you could make a
766 pipe available, as a storage device? Then we could use any command that
767 handles stdin as a storage destination.
774 #Password is not used in rsh, but might be used in ftp.
775 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
782 Password = "foobar&-"
783 Device = "| ncftpput -c hsm /path/file.bacula"
787 >If you have some local storage available, you could
788 >use Bacula to backup to disk volumes, then use some
789 >other software (ftp, scp) to move them to the HSM
790 >machine. However, this is a bit kludgy.
793 It is, but maybe worth a try. Is there some function in Bacula to put
794 variables in filenames? i.e. backup.2003-01-15.root
801 w.a.sellers@larc.nasa.gov
803 ==============================================
804 The Project for the above
806 I finally realized that this is not at all
807 the same as reader/writer programs or plugins,
808 which are alternate ways of accessing the
809 files to be backed up. Rather, it is an alternate
810 form of storage device, and I have always planned
811 that Bacula should be able to handle all sorts
814 So, I propose the following phases:
816 1. OK from you to invest some time in testing
817 this as I implement it (requires that you
818 know how to download from the SourceForge
819 cvs -- which I imagine is a piece of cake
822 2. Dumb implementation by allowing a device to
823 be a fifo for write only.
824 Reason: easy to implement, proof of concept.
826 3. Try reading from fifo but with fixed block
828 Reason: proof of concept, easy to implement.
830 4. Extend reading from fifo (restores) to handle
832 Reason: requires some delicate low level coding
833 which could destabilize all of Bacula.
835 5. Implementation of above but to a program. E.g.
836 Device = "|program" (not full pipeline).
837 Reason: routines already exist, and program can
838 be a shell script which contains anything.
840 6. Add full pipeline as a possibility. E.g.
841 Device = "| gzip -c | rsh hsm dd of=path/file.tgz"
842 Reason: needs additional coding to implement full
843 pipeline (must fire off either a shell or all
844 programs and connect their pipes).
846 There are a good number of details in each step
847 that I have left out, but I will specify them at
848 every stage, and there may be a few changes as things
849 evolve. I expect that to get to stage 5 will take a
850 few weeks, and at that point, you will have
851 everything you need (just inside a script).
852 Stage 6 will probably take longer, but if this
853 project pleases you, what we do for 5 should
854 be adequate for some time.
858 =============================================
860 Done: (see kernsdone for more)
861 - Look into Pruning/purging problems or why there seem to
862 be so many files listed each night.
863 - Fix cancel in find_one -- need jcr.
864 - Cancel does not work for restore in FD.
865 - Write SetJobStatus() function so cancel status not lost.
866 - Add include list to end of chain in findlib
867 - Zap sd_auth_key after use
868 - Add Bar code reading capabilities (new mtx-changer)
869 - Figure out some way to automatically backup all local partitions
870 - Make hash table for linked files in findlib/find_one.c:161
872 - Rewrite find_one.c to use only pool_memory instead of
873 alloca and malloc (probably not necessary).
874 - Make sure btraceback goes into /sbin not sysconf directory.
875 - InitVerify is getting pruned and it shouldn't (document it)
876 - Make 1.28c release ??? NO do 1.29 directly
877 - Set timeout on opening fifo for save or restore (findlib)
878 - Document FIFO storage device.
879 - Document fifo and | and <
882 - Get correct error status from run_program or open_bpipe().
883 - Restrict permissions on File Volumes (now 0640).
884 - Umasked 022 daemons
885 - Fix restore of hard linked file.
886 - Figure out how to allow multiple simultaneous file Volumes on a single device.
887 - Implement multiple simultaneous file Volumes on a single device.
888 - Cleanup db_update_media and db_update_pool
889 - Flush all the daemon messages at the end of every job.
890 - Change stat1= fgets()!=NULL to stat1=fgest()==NULL; in
891 run_program -- bpipe.c
892 - Apparently cancel does not work for jobs waiting to be
894 - Implement TCP/IP connection for MySQL
895 - Pull a canceled job from the Scheduling queue.
896 - Implement max_file_size in block.c (already done, just tweaked).
897 - Look at purge jobs volume (at least document it, and see if it is
899 - Add list volumes does all pools. list volumes pool=xxx now works.
900 - Add pool= to "list media" in ua_output.c
901 - Strip trailing slashes from Include directory names in the FD.
902 - Fix Error: bnet.c:408 gethostbyname() for lpmatou failed:
903 ERR=Operation not permited loop.
904 - Add code if there is no mtio.h (cannot do -- too many ioctl defines needed)
905 - Produce better error messages in when error/eof writing block.
906 - Cancelling of a queued job does NOT work!!!!!!
908 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
909 rufus-sd: Recycled volume File0003 on device /home/kern/bacula/working, all previous data lost.
910 rufus-dir: Volume used once. Marking Volume "File0003" as Used.
911 - Ability to backup to a file then later transfer to a tape -- Migration.
912 Migration based on MaxJobs(MinJobs),MaxVols(MinVols),AgeJobs,MaxBytes(MinBytes)
913 (i.e. HighwaterMark, LowwaterMark).
914 - Eugeny Fisher <efischer@vip-rus.com> wants to cycle through a
915 set of volumes recycling the oldest volume when it is needed.
916 - gethostbyname failure in bnet_connect() continues
917 generating errors -- should stop.
918 - Add chflags() code for FreeBSD file flags
919 - Bevan Anderson suggests having a run queue for each device
920 so that multiple simultaneous jobs can run but each writing
921 to a different Volume.
922 - Look at handling <> in smtp doesn't work with exim.
923 - Need to specify MaximumConcurrentJobs in the Job resource.
924 - ***test GetFileAttributexEx, and remove MessageBox at 335 of winservice.cpp ****
925 - Implement finer multiprocessing options.
926 - Implement | and < in Exclude statements.
927 - Figure out some way to specify a retention period for files
928 that no longer exist on the machine -- so that we maintain
929 say backups for 30 days, but if the file is deleted, we maintain
930 the last copy for 1 year. -- answer Volume retention.
931 - Make non-zero status from RunJobBefore/After error the job.
932 - Need define of int_least16_t in sha1.h for SuSE.
933 - Implement bar code reader for autochangers
934 - Document new MaximumConcurrentJob records (Job, Client, Storage)
935 - Write up how to use/manage disk Volume Storage. ******
936 - Remove kern and kelvin from mysql_grant...
937 - Install grant_mysql...
938 - Strip trailing / from Include
939 - add #define ENABLE_NLS for Gnome compile on SuSE.
940 - Add Client FS/OS id (Linux, Win95/98, ...).
941 - Concept of VolumeSet during restore which is a list
942 of Volume names needed.
943 - Turn virutally all sprintfs into snprintfs.
944 - Update volume=Test01 requests pool, then lists volumes.
945 **** Test select_pool_and_media ...
947 - Add IP address to authentication failures.
948 - Add a default File storage so that new users can do backup
949 and restores right away.