4 Documentation to do: (any release a little bit at a time)
5 - Document running a test version.
6 - Document query file format.
7 - Document static linking
8 - Document problems with Verify and pruning.
9 - Document how to use multiple databases.
10 - For FreeBSD typical /dev/nrsa0 and for mtx /dev/pass1
11 - VXA drives have a "cleaning required"
12 indicator, but Exabyte recommends preventive cleaning after every 75
14 - Lookup HP cleaning recommendations.
15 - Lookup HP tape replacement recommendations (see trouble shooting autochanger)
17 Testing to do: (painful)
18 - that ALL console command line options work and are always implemented
19 - blocksize recognition code.
20 - Test if rewind at end of tape waits for tape to rewind.
22 - Test not zeroing Autochanger slot when it is wrong.
23 - Figure out how to use ssh or stunnel to protect Bacula communications.
25 For 1.33 Testing/Documentation:
26 - Fix FreeBSD build with tcp_wrapper -- should not have -lnsl
27 - Document to start higher priorty jobs before lower ones.
28 - suppress "Do not forget to mount the drive!!!" if error
29 - Document new records in Director. SDAddress SDDeviceName, SDPassword.
30 FDPassword, FDAddress, DBAddress, DBPort, DBPassword.
31 - Document new Include/Exclude ...
32 - Add test of exclusion, test multiple Include {} statements.
33 - Add counter variable test.
36 - Either restrict the characters in a name, or fix the problem
37 emailing with names containing / (smtp command line breaks).
38 - Eliminate ua_retention.c (retentioncmd) if possible.
39 - Eliminate orphaned jobs: dbcheck, normal pruning, delete job command.
40 Hm. Well, there are the remaining orphaned job records:
42 | 105 | Llioness Save | 0000-00-00 00:00:00 | B | D | 0 | 0 | f |
43 | 110 | Llioness Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
44 | 115 | Llioness Save | 2003-09-10 02:22:03 | B | I | 0 | 0 | A |
45 | 128 | Catalog Save | 2003-09-11 03:53:32 | B | I | 0 | 0 | C |
46 | 131 | Catalog Save | 0000-00-00 00:00:00 | B | I | 0 | 0 | f |
48 As you can see, three of the five are failures. I already deleted the
49 one restore and one other failure using the by-client option. Deciding
50 what is an orphaned job is a tricky problem though, I agree. All these
51 records have or had 0 files/ 0 bytes, except for the restore. With no
52 files, of course, I don't know of the job ever actually becomes
53 associated with a Volume.
55 (I'm not sure if this is documented anywhere -- what are the meanings of
56 all the possible JobStatus codes?)
58 Looking at my database, it appears to me as though all the "orphaned"
59 jobs fit into one of two categories:
61 1) The Job record has a StartTime but no EndTime, and the job is not
64 2) The Job record has an EndTime, indicating that it completed, but
65 it has no associated JobMedia record.
68 This does suggest an approach. If failed jobs (or jobs that, for some
69 other reason, write no files) are associated with a volume via a
70 JobMedia record, then they should be purged when the associated volume
71 is purged. I see two ways to handle jobs that are NOT associated with a
74 1) purge them automatically whenever any volume is manually purged;
76 2) add an option to the purge command to manually purge all jobs with
79 I think Restore jobs also fall into category 2 above .... so one might
80 want to make that "The Job record has an EndTime,, but no associated
81 JobMedia record, and is not a Restore job."
82 - Implement RestoreJobRetention? Maybe better "JobRetention" in a Job,
83 which would take precidence over the Catalog "JobRetention".
84 - Implement Label Format in Add and Label console commands.
85 - make "btape /tmp" work.
86 - Make sure a rescheduled job is properly reported by status.
87 - Walk through the Pool records rather than the Job records
88 in dird.c to create/update pools.
89 - Figure out a way to move Volumes from one pool to another.
90 - What to do about "list files job=xxx".
91 - Implement delete Job.
92 - Document need to put LabelFormat in quotes.
93 - Implement scan: for every slot it finds, zero the slot of
94 Volume other volume having that slot.
95 - When job rescheduled, status gives is waiting for Client Rufus
96 to connect to Storage File. Dir needs to inform SD that job
98 - Fix get_storage_from_media_type (ua_restore) to use command line
100 - Enhance "update slots" to include a "scan" feature
101 scan 1; scan 1-5; scan 1,2,4 ... to update the catalog
102 - Allow a slot or range of slots on the label barcodes command.
103 - Don't print "Warning: Wrong Volume mounted ..." if mounting second volume.
104 - Make Dmsg look at global before calling subroutine.
105 - Enable trace output at runtime for Win32
106 - Make sure that Volumes are recycled based on "Least recently used"
107 rather than lowest MediaId.
108 - Available volumes for autochangers (see patrick@baanboard.com 3 Sep 03
109 and 4 Sep) scan slots.
110 - Upgrade to cygwin 1.5
112 - Get and test MySQL 4.0
113 - Do a complete audit of all pthreads_mutex, cond, ... to ensure that
114 any that are dynamically initialized are destroyed when no longer used.
115 - Write a mini-readline with history and editing.
116 - Look at how fuser works and /proc/PID/fd that is how Nic found the
117 file descriptor leak in Bacula.
118 - Implement WrapCounters in Counters.
119 - Turn on SIGHUP in dird.c and test.
120 - Use system dependent calls to get more precise info on tape errors.
121 - Add heartbeat from FD to SD if hb interval expires.
122 - Suppress read error on blank tape when doing a label.
123 - Can we dynamically change FileSets.
124 - If pool specified to label command and Label Format is specified,
125 automatically generate the Volume name.
126 - Take a careful look a the Basic recycling algorithm. When Bacula
127 chooses, the order should be:
129 - Look for Recycle or Purged
132 Instead of using lowest media Id, find the least recently used
135 When the tape is mounted and Bacula requests the status
136 - Do everything possible to use it.
138 Define a "available" status, which is the currently mounted
139 Volume and all volumes that are currently in the autochanger.
141 - Why can't SQL do the filename sort for restore?
142 - Is a pool specification really needed for a restore? Yes, and
143 you may want to exclude archive Pools.
144 - Look at libkse (man kse) for FreeBSD threading.
145 - Look into Microsoft Volume Shadowcopy Service VSS for backing
146 up system state components (Active Directory, System Volume, ...)
147 - Add ExhautiveRestoreSearch
148 - Look at the possibility of loading only the necessary
149 data into the restore tree (i.e. do it one directory at a
150 time as the user walks through the tree).
151 - Possibly use the hash code if the user selects all for a restore command.
152 - Orphaned Dir buffer at parse_conf.c:373 => store_dir
153 - Implement some way for the File daemon to contact the Director
154 to start a job or pass its DHCP obtained IP number.
155 - Implement multiple Consoles.
156 - Add Console usr permissions.
157 - Fix "restore all" to bypass building the tree.
158 - Fix restore to list errors if Invalid block found, and if # files
159 restored does not match # expected.
160 - Prohibit backing up archive device (findlib/find_one.c:128)
161 - Implement Release Device in the Job resource to unmount a drive.
162 - Implement Acquire Device in the Job resource to mount a drive,
163 be sure this works with admin jobs so that the user can get
164 prompted to insert the correct tape. Possibly some way to say to
165 run the job but don't save the files.
166 - Add JobName= to VerifyToCatalog so that all verifies can be done at the end.
167 - Implement FileOptions (see end of this document)
168 - Make things like list where a file is saved case independent for
170 - Edit the Client/Storage name into authentication failure messages.
171 - Implement job in VerifyToCatalog
173 - Implement a PostgreSQL driver.
174 - Bacula needs to propagate SD errors.
175 > > cluster-dir: Start Backup JobId 252, Job=REUTERS.2003-08-11_15.04.12
176 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Write error on device
177 /dev/nst0. ERR=Input/output error.
178 > > prod4-sd: REUTERS.2003-08-11_15.04.12 Error: Re-read of last block failed.
179 Last block=5162 Current block=5164.
180 > > prod4-sd: End of medium on Volume "REU007" Bytes=16,303,521,933
182 - Use autochanger to handle multiple devices.
183 - Fix packet too big problem.
184 - Add SuSE install doc to list.
185 - Check and rechedk "Invalid block number"
186 - Make bextract release the drive properly between tapes
187 so that an autochanger can be made to work.
188 - User wants to NOT backup up certain big files (email files).
189 - Maybe remove multiple simultaneous devices code in SD.
190 - On Windows with very long path names, it may be impossible to create
191 a file (and thus restore it) because the total length is too long.
192 We must cd into the directory then create the file without the
194 - lstat() is not going to work on Win32 for testing date.
195 - Something is not right in last block of fill command.
196 - Implement a Recycle command
197 - Add FileSet to command line arguments for restore.
198 - Allow multiple Storage specifications (or multiple names on
199 a single Storage specification) in the Job record. Thus a job
200 can be backed up to a number of storage devices.
201 - Add client name to cram-md5 challenge so Director can immediately
202 verify if it is the correct client.
203 - Implement ClientRunBeforeJob and ClientRunAfterJob.
204 - Add JobLevel in FD status (but make sure it is defined).
205 - Audit all UA commands to ensure that we always prompt where possible.
206 - Restrict characters permitted in a Resource name, and don't permit
208 - Check Jmsg in bnet, may not work, must dup bsock.
209 - Suppress Job Name in Jmsg for console
210 - Create Pools that are referenced in a Run statement at startup if possible.
211 - Use runbeforejob to unload, then reload a volume previously used,
212 then the next job run gets an error reading the drive.
213 - Make bootstrap filename unique.
214 - Test a second language e.g. french.
215 - Start working on Base jobs.
216 - Make "make binary-release" work from any directory.
217 - Implement UnsavedFiles DB record.
218 - Implement argc/argv for daemon command line scanning using table driven
220 - Implement table driven single argc/argv scanner to pickup all arguments.
221 Much like xxx_conf.c scan table.
222 keyword, handler(store_routine), store_address, code, flags, default.
223 - From Phil Stracchino:
224 It would probably be a per-client option, and would be called
225 something like, say, "Automatically purge obsoleted jobs". What it
226 would do is, when you successfully complete a Differential backup of a
227 client, it would automatically purge all Incremental backups for that
228 client that are rendered redundant by that Differential. Likewise,
229 when a Full backup on a client completed, it would automatically purge
230 all Differential and Incremental jobs obsoleted by that Full backup.
231 This would let people minimize the number of tapes they're keeping on
232 hand without having to master the art of retention times.
233 - Implement new serialize subroutines
234 send(socket, "string", &Vol, "uint32", &i, NULL)
235 - Scratch Pool where the volumes can be re-assigned to any Pool.
236 - Implement a M_SECURITY message class.
237 - Implement forward spacing block/file: position_device(bsr) --
238 just before read_block_from_device();
239 - When doing a Backup send all attributes back to the Director, who
240 would then figure out what files have been deleted.
241 - Currently in mount.c:236 the SD simply creates a Volume. It should have
242 explicit permission to do so. It should also mark the tape in error
243 if there is an error.
244 - Make sure all restore counters are working correctly in the FD.
245 - SD Bytes Read is wrong.
246 - Look at ALL higher level routines that call block.c to be sure
247 they don't expect something in errmsg.
248 - Investigate doing RAW backup of Win32 partition.
249 - Add thread specific data to hold the jcr -- send error messages from
250 low level routines by accessing it and using Jmsg().
251 - Cancel waiting for Client connect in SD if FD goes away.
253 - Add Progress command that periodically reports the progress of
255 - One block was orphaned in the SD probably after cancel.
258 - Examine Bare Metal restore problem (a FD crash exists somewhere ...).
259 - Implement timeout in response() when it should come quickly.
260 - Implement console @echo command.
261 - Implement a Slot priority (loaded/not loaded).
262 - Implement "vacation" Incremental only saves.
263 - Implement single pane restore (much like the Gftp panes).
264 - Implement Automatic Mount even in operator wait.
265 - Implement create "FileSet"?
266 - Fix watchdog pthread crash on Win32 (this is pthread_kill() Cygwin bug)
267 - Implement "scratch pool" where tapes are defined and can be
268 taken by any pool that needs them.
269 - Implement restore "current system", but take all files without
270 doing selection tree -- so that jobs without File records can
272 - Implement disk spooling. Two parts: 1. Spool to disk then
273 immediately to tape to speed up tape operations. 2. Spool to
274 disk only when the tape is full, then when a tape is hung move
276 - Implement a relocatable bacula.spec
277 - Implement dump/print label to UA
278 - Add prefixlinks to where or not where absolute links to FD.
279 - Issue message to mount a new tape before the rewind.
280 - Simplified client job initiation for portables.
281 - If SD cannot open a drive, make it periodically retry.
282 - Implement LabelTemplate (at least first cut).
283 - Add more of the config info to the tape label.
284 - Implement a Mount Command and an Unmount Command where
285 the users could specify a system command to be performed
286 to do the mount, after which Bacula could attempt to
287 read the device. This is for Removeable media such as a CDROM.
288 - Most likely, this mount command would be invoked explicitly
289 by the user using the current Console "mount" and "unmount"
290 commands -- the Storage Daemon would do the right thing
291 depending on the exact nature of the device.
292 - As with tape drives, when Bacula wanted a new removable
293 disk mounted, it would unmount the old one, and send a message
294 to the user, who would then use "mount" as described above
295 once he had actually inserted the disk.
297 - If tape is marked read-only, then try opening it read-only rather than
298 failing, and remember that it cannot be written.
299 - Refine SD waiting output:
300 Device is being positioned
301 > Device is being positioned for append
302 > Device is being positioned to file x
304 - Figure out some way to estimate output size and to avoid splitting
305 a backup across two Volumes -- this could be useful for writing CDROMs
306 where you really prefer not to have it split -- not serious.
307 - Have SD compute MD5 or SHA1 and compare to what FD computes.
308 - Make VolumeToCatalog calculate an MD5 or SHA1 from the
309 actual data on the Volume and compare it.
310 - Implement Bacula plugins -- design API
311 - Make bcopy read through bad tape records.
312 - Program files (i.e. execute a program to read/write files).
313 Pass read date of last backup, size of file last time.
314 - Add Signature type to File DB record.
315 - CD into subdirectory when open()ing files for backup to
316 speed up things. Test with testfind().
317 - Priority job to go to top of list.
318 - Why are save/restore of device different sizes (sparse?) Yup! Fix it.
319 - Implement some way for the Console to dynamically create a job.
320 - Restore to a particular time -- e.g. before date, after date.
321 - Solaris -I on tar for include list
322 - Need a verbose mode in restore, perhaps to bsr.
323 - bscan without -v is too quiet -- perhaps show jobs.
324 - Add code to reject whole blocks if not wanted on restore.
325 - Check if we can increase Bacula FD priorty in Win2000
326 - Make sure the MaxVolFiles is fully implemented in SD
327 - Check if both CatalogFiles and UseCatalog are set to SD.
328 - Figure out how to do a bare metal Windows restore
329 - Possibly add email to Watchdog if drive is unmounted too
330 long and a job is waiting on the drive.
331 - Restore program that errs in SD due to no tape, reports
332 OK incorrectly in output.
333 - After unmount, if restore job started, ask to mount.
334 - Convert all %x substitution variables, which are hard to remember
335 and read to %(variable-name). Idea from TMDA.
336 - Remove NextId for SQLite. Optimize.
337 - Move all SQL statements into a single location.
338 - Add UA rc and history files.
339 - put termcap (used by console) in ./configure and
340 allow -with-termcap-dir.
341 - Enhance time and size scanning routines.
342 - Fix Autoprune for Volumes to respect need for full save.
343 - Fix Win32 config file definition name on /install
344 - Compare tape to Client files (attributes, or attributes and data)
345 - Make all database Ids 64 bit.
346 - Write an applet for Linux.
347 - Implement new inter-daemon communications protocol.
348 - Allow console commands to detach or run in background.
349 - Fix status delay on storage daemon during rewind.
350 - Add SD message variables to control operator wait time
351 - Maximum Operator Wait
352 - Minimum Message Interval
353 - Maximum Message Interval
354 - Send Operator message when cannot read tape label.
355 - Verify level=Volume (scan only), level=Data (compare of data to file).
356 Verify level=Catalog, level=InitCatalog
358 - Add keyword search to show command in Console.
359 - Events : tape has more than xxx bytes.
360 - Complete code in Bacula Resources -- this will permit
361 reading a new config file at any time.
362 - Handle ctl-c in Console
363 - Implement script driven addition of File daemon to config files.
364 - Think about how to make Bacula work better with File (non-tape) archives.
365 - Write Unix emulator for Windows.
366 - Put memory utilization in Status output of each daemon
367 if full status requested or if some level of debug on.
368 - Make database type selectable by .conf files i.e. at runtime
369 - Set flag for uname -a. Add to Volume label.
370 - Implement throttled work queue.
371 - Restore files modified after date
372 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
373 - Implement Restore FileSet=
374 - Create a protocol.h and protocol.c where all protocol messages
376 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
377 - Timout a job or terminate if link goes down, or reopen link and query.
378 - Concept of precious tapes (cannot be reused).
379 - Make bcopy copy with a single tape drive.
380 - Permit changing ownership during restore.
383 > My suggestion: Add a feature on the systray menu-icon menu to request
384 > an immediate backup now. This would be useful for laptop users who may
385 > not be on the network when the regular scheduled backup is run.
387 > My wife's suggestion: Add a setting to the win32 client to allow it to
388 > shut down the machine after backup is complete (after, of course,
389 > displaying a "System will shut down in one minute, click here to cancel"
390 > warning dialog). This would be useful for sites that want user
391 > woorkstations to be shut down overnight to save power.
394 - From Terry Manderson <terry@apnic.net>
395 jobset { # new structure
396 name = "monthlyUnixBoxen"
399 jobs = "wakame;durian;soy;wasabi;miso" #new!
409 fileset = "genericUnixSet"
415 fileset = "genericUnixSet"
421 fileset = "UnixDevelBoxSet"
426 - Autolabel should be specified by DIR instead of SD.
429 - AutoScan (check checksum of tape)
430 - Format command = "format /dev/nst0"
434 - Seek resolution (usually corresponds to buffer size)
435 - EODErrorCode=ENOSPC or code
436 - Partial Read error code
437 - Partial write error code
438 - Nonformatted read error
439 - Nonformatted write error
440 - WriteProtected error
444 - IgnoreCloseErrors=yes
454 - FD sends unsaved file list to Director at end of job (see
456 - File daemon should build list of files skipped, and then
457 at end of save retry and report any errors.
458 - Write a Storage daemon that uses pipes and
459 standard Unix programs to write to the tape.
461 - Need something that monitors the JCR queue and
462 times out jobs by asking the deamons where they are.
463 - Enhance Jmsg code to permit buffering and saving to disk.
464 - device driver = "xxxx" for drives.
466 - Ensure that /dev/null works
467 - Need report class for messages. Perhaps
468 report resource where report=group of messages
469 - enhance scan_attrib and rename scan_jobtype, and
470 fill in code for "since" option
471 - Director needs a time after which the report status is sent
472 anyway -- or better yet, a retry time for the job.
473 - Don't reschedule a job if previous incarnation is still running.
474 - Some way to automatically backup everything is needed????
475 - Need a structure for pending actions:
477 - termination status (part of buffered msgs?)
479 Read, Write, Clean, Delete
480 - Login to Bacula; Bacula users with different permissions:
481 owner, group, user, quotas
482 - Store info on each file system type (probably in the job header on tape.
483 This could be the output of df; or perhaps some sort of /etc/mtab record.
486 - Design at hierarchial storage for Bacula. Migration and Clone.
487 - Implement FSM (File System Modules).
488 - Audit M_ error codes to ensure they are correct and consistent.
489 - Add variable break characters to lex analyzer.
490 Either a bit mask or a string of chars so that
491 the caller can change the break characters.
492 - Make a single T_BREAK to replace T_COMMA, etc.
493 - Ensure that File daemon and Storage daemon can
494 continue a save if the Director goes down (this
495 is NOT currently the case). Must detect socket error,
496 buffer messages for later.
497 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
498 - Add ability to backup to two Storage devices (two SD sessions) at
499 the same time -- e.g. onsite, offsite.
500 - Add the ability to consolidate old backup sets (basically do a restore
501 to tape and appropriately update the catalog). Compress Volume sets.
502 Might need to spool via file is only one drive is available.
503 - Compress or consolidate Volumes of old possibly deleted files. Perhaps
504 someway to do so with every volume that has less than x% valid
508 Migration: Move a backup from one Volume to another
509 Clone: Copy a backup -- two Volumes
511 Bacula Migration is based on Jobs (apparently Networker is file by file).
513 Migration triggered by:
517 Highwater mark (keep total size)
521 Bacula Projects Roadmap
523 last update 8 May 2003
525 Item 1: Multiple simultaneous Jobs. (done)
526 Done -- Restore part needs better implementation to work correctly
527 Also, it needs considerable testing
529 What: Permit multiple simultaneous jobs in Bacula.
531 Why: An enterprise level solution needs to go fast without the
532 need for the system administrator to carefully tweak
533 timing. Based on the benchmarks, during a full
534 backup, NetWorker typically hit 10 times the bandwidth to
535 the tape compared to Bacula--largely. This is probably due to
536 running parallel jobs and multi-threaded filling of buffers
537 and writing them to tape. This should also make things work
538 better when you have a mix of fast and slow machines backing
541 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
542 implementing this is a matter of some small cleanups and
546 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
547 Deferred -- not necessary yet -- possibly implement with Migration.
549 What: If data is coming into the SD too fast, buffer it to
550 disk if the user has configured this option.
552 Why: This would be nice, especially if it more or less falls out
553 when implementing (1) above. If not, it probably should not
554 be given a high priority because fundamentally the backup time
555 is limited by the tape bandwidth. Even though you may finish a
556 client job quicker by spilling to disk, you still have to
557 eventually get it onto tape. If intermediate disk buffering
558 allows us to improve write bandwidth to tape, it may make
561 Notes: Whether or not this is implemented will depend upon performance
562 testing after item 1 is implemented.
565 Item 3: Write the bscan program -- also write a bcopy program.
568 What: Write a program that reads a Bacula tape and puts all the
569 appropriate data into the catalog. This allows recovery
570 from a tape that is no longer in the database, or it allows
571 re-creation of a database if lost.
573 Why: This is a fundamental robustness and disaster recovery tool
574 which will increase the comfort level of a sysadmin
575 considering adopting Bacula.
577 Notes: A skeleton of this program already exists, but much work
578 needs to be done. Implementing this will also make apparent
579 any deficiencies in the current Bacula tape format.
582 Item 4: Implement Base jobs.
584 What: A base job is sort of like a Full save except that you
585 will want the FileSet to contain only files that are unlikely
586 to change in the future (i.e. a snapshot of most of your
587 system after installing it). After the base job has been run,
588 when you are doing a Full save, you can specify to exclude
589 all files saved by the base job that have not been modified.
591 Why: This is something none of the competition does, as far as we know
592 (except BackupPC, which is a Perl program that saves to disk
593 only). It is big win for the user, it makes Bacula stand out
594 as offering a unique optimization that immediately saves time
597 Notes: Big savings in tape usage. Will require more resources because
598 the e. DIR must send FD a list of files/attribs, and the FD must
599 search the list and compare it for each file to be saved.
602 Item 5: Implement Label templates
605 What: This is a mechanism whereby Bacula can automatically create
606 a tape label for new tapes according to a detailed specification
607 provided by the user.
609 Why: It is a major convenience item for folks who use automated label
612 Notes: Bacula already has a working form of automatic tape label
613 creation, but it is very crude. The design for the complete
614 tape labeling project is already documented in the manual.
617 Item 6: Write a regression script.
618 Done -- Continue to expand its testing.
620 What: This is an automatic script that runs and tests as many features
621 of Bacula as possible. The output is compared to previous
622 versions of Bacula and any differences are reported.
624 Why: This is an enormous help in preventing introduction of new
625 errors in parts of the program that already work correctly.
627 Notes: This probably should be ranked higher, it's something the typical
628 user doesn't see. Depending on how it's implemented, it may
629 make sense to defer it until the archival tape format and
630 user interface mature.
633 Item 7: GUI for interactive restore
634 Item 8: GUI for interactive backup
636 What: The current interactive restore is implemented with a tty
637 interface. It would be much nicer to be able to "see" the
638 list of files backed up in typical GUI tree format.
639 The same mechanism could also be used for creating
640 ad-hoc backup FileSets (item 8).
642 Why: Ease of use -- especially for the end user.
644 Notes: Rather than implementing in Gtk, we probably should go directly
645 for a Browser implementation, even if doing so meant the
646 capability wouldn't be available until much later. Not only
647 is there the question of Windows sites, most
648 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
649 without installing lots of stuff admins are very wary about.
650 Real sysadmins will always use the command line anyway, and
651 the user who's doing an interactive restore or backup of his
652 own files will in most cases be on a Windows machine running
656 Item 9: Add SSL to daemon communications.
658 What: This provides for secure communications between the daemons.
660 Why: This would allow doing backup across the Internet without
661 privacy concerns (or with much less concern).
663 Notes: The vast majority of near term potential users will be backing up
664 a single site over a LAN and, correctly or not, they probably
665 won't be concerned with security, at least not enough to go to
666 the trouble to set up keys, etc. to screw things down. We suspect
667 that many users genuinely interested in multi-site backup
668 already run some form of VPN software in their internetwork
669 connections, and are willing to delegate security to that layer.
672 Item 10: Define definitive tape format.
675 What: Define that definitive tape format that will not change
676 for the next millennium.
678 Why: Stability, security.
680 Notes: See notes for item 11 below.
683 Item 11: New daemon communication protocol.
685 What: The current daemon to daemon protocol is basically an ASCII
686 printf() and sending the buffer. On the receiving end, the
687 buffer is sscanf()ed to unpack it. The new scheme would
688 be a binary format that allows quick packing and unpacking
689 of any data type with named fields.
691 Why: Using binary packing would be faster. Named fields will permit
692 error checking to ensure that what is sent is what the
693 receiver really wants.
695 Notes: These are internal improvements in the interest of the
696 long-term stability and evolution of the program. On the one
697 hand, the sooner they're done, the less code we have to rip
698 up when the time comes to install them. On the other hand, they
699 don't bring an immediately perceptible benefit to potential
700 users. Item 10 and possibly item 11 should be deferred until Bacula
701 is well established with a growing user community more or
702 less happy with the feature set. At that time, it will make a
703 good "next generation" upgrade in the interest of data
709 ======================================================
711 It is somewhat like a Full save becomes an incremental since
712 the Base job (or jobs) plus other non-base files.
714 - New BaseFile table that contains:
715 JobId, BaseJobId, FileId (from Base).
716 i.e. for each base file that exists but is not saved because
717 it has not changed, the File daemon sends the JobId, BaseId,
718 and FileId back to the Director who creates the DB entry.
719 - To initiate a Base save, the Director sends the FD
720 the FileId, and full filename for each file in the Base.
721 - When the FD finds a Base file, he requests the Director to
722 send him the full File entry (stat packet plus MD5), or
723 conversely, the FD sends it to the Director and the Director
724 says yes or no. This can be quite rapid if the FileId is kept
725 by the FD for each Base Filename.
726 - It is probably better to have the comparison done by the FD
727 despite the fact that the File entry must be sent across the
729 - An alternative would be to send the FD the whole File entry
730 from the start. The disadvantage is that it requires a lot of
731 space. The advantage is that it requires less communications
733 - The Job record must be updated to indicate that one or more
735 - At end of Job, FD returns:
736 1. Count of base files/bytes not written to tape (i.e. matches)
737 2. Count of base file that were saved i.e. had changed.
738 - No tape record would be written for a Base file that matches, in the
739 same way that no tape record is written for Incremental jobs where
740 the file is not saved because it is unchanged.
741 - On a restore, all the Base file records must explicitly be
742 found from the BaseFile tape. I.e. for each Full save that is marked
743 to have one or more Base Jobs, search the BaseFile for all occurrences
745 - An optimization might be to make the BaseFile have:
751 This would avoid the need to explicitly fetch each File record for
752 the Base job. The Base Job record will be fetched to get the
753 VolSessionId and VolSessionTime.
754 =========================================================
758 =============================================================
760 Request For Comments For File Backup Options
763 Subject: File Backup Options
766 A few days ago, a Bacula user who is backing up to file volumes and
767 using compression asked if it was possible to suppress compressing
768 all .gz files since it was a waste of CPU time. Although Bacula
769 currently permits using different options (compression, ...) on
770 a directory by directory basis, it cannot do it on a file by
771 file basis, which is clearly what was desired.
773 Proposed Implementation:
774 To solve this problem, I propose the following:
776 - Add a new Director resource type called Options.
778 - The Options resource will have records for all
779 options that can currently be specified on the Include record
780 (in a FileSet). Examples below.
782 - The Options resource will permit an exclude option as well
783 as a number of additional options.
785 - The heart of the Options resource is the ability to
786 supply any number of Match records which specify POSIX
787 regular expressions. These Match regular expressions are
788 applied to the fully qualified filename (path and all). If
789 one matches, then the Options will be used.
791 - When an Match specification matches an included file, the
792 options specified in the Options resource will override
793 the default options specified on the Include record.
795 - Include records will be modified to permit referencing one or
796 more Options resources. The Options will be used
797 in the order listed on the Include record and the first
798 one that matches will be applied.
800 - Options (or specifications) currently supplied on the Include
801 record will be deprecated (i.e. removed in a later version a
802 year or so from now).
804 - The Exclude record will be deprecated as the same functionality
805 can be obtained by using an Exclude = yes in the Options.
808 The following records can appear in the Options resource. An
809 asterisk preceding the name indicates a feature not currently
813 - Compression= (GZIP, ...)
814 - Signature= (MD5, SHA1, ...)
816 - OneFs= (yes/no) - remain on one filesystem
817 - Recurse= (yes/no) - recurse into subdirectories
818 - Sparse= (yes/no) - do sparse file backup
819 - *Exclude= (yes/no) - exclude file from being saved
820 - *Reader= (filename) - external read (backup) program
821 - *Plugin= (filename) - read/write plugin module
824 - verify= (ipnougsamc5) - verify options
827 - replace= (always/ifnewer/ifolder/never) - replace options currently
829 - *Writer= (filename) - external write (restore) program
833 Currently options specifying compression, MD5 signatures, recursion,
834 ... of a FileSet are supplied on the Include record. These will now
835 all be collected into a Options resource, which will be
836 specified in the Include in place of the options. Multiple Options
837 may be specified. Since the Options may contain regular expressions
838 that are applied to the full filename, this will give the ability
839 to specify backup options on a file by file basis to whatever level
848 Include = compression=GZIP signature=MD5 {
860 Match = /*.?*/ # matches all files.
865 That's a lot more to do the same thing, but it gives the ability to
866 apply options on a file by file basis. For example, suppose you
867 want to compress all files but not any file with extensions .gz or .Z.
868 In that case, you will need to group two sets of options using
869 the Options resource as follows:
877 # Note multiple Matches are ORed
878 Match = /*.gz/ # matches .gz files */
879 Match = /*.Z/ # matches .Z files */
884 Match = /*.?*/ # matches all files
890 Now, since the no Compression option is specified in the
891 first group of Options, *.gz or *.Z file will have an MD5 signature computed,
892 but will not be compressed. For all other files, the *.gz *.Z will not
893 match, so the second group of options will be used which will include GZIP
897 - Is it necessary to provide some means of ANDing regular expressions
898 and negation? (not currently planned)
900 e.g. Match = /*.gz/ && !/big.gz/
902 - I see that Networker has a "null" module which, if specified, does not
903 backup the file, but does make an record of the file in the catalog
904 so that the catalog will reflect an exact picture of the filesystem.
905 The result is that the file can be "seen" when "browsing" the save
906 sets, but it cannot be restored.
908 Is this really useful? Should it be implemented in Bacula?
911 After implementing the above, the user will be able to specify
912 on a file by file basis (using regular expressions) what options are
913 applied for the backup.
916 =============================================
918 ==========================================================
920 For each Incremental job that is run, there may be files that
921 were found but not saved because they were locked (this applies
922 only to Windows). Such a system could send back to the Director
923 a list of Unsaved files.
925 - New UnSavedFiles table that contains:
929 - Then in the next Incremental job, the list of Unsaved Files will be
930 feed to the FD, who will ensure that they are explicitly chosen even
931 if standard date/time check would not have selected them.
932 =============================================================
935 Done: (see kernsdone for more)
936 - Implement new alist in FileSet scanning.
937 - bls should continue reading even if it finds Win32 data on the tape.
938 The error should be Warning rather the Error.
939 - Add user configurable timeout for connecting to SD.
940 - Unsaved Flag in Job record (use JobMissingFiles).
941 - Base Flag in Job record.
942 - Configure mtx-changer to have correct path to mtx.
943 - Add all command line arguments to "update", e.g. slot=nn volStatus=append, ...
944 - Make some way so that if a machine is skipped because it is not up
945 that Bacula will continue retrying for a specified period of time --
947 - Implement all command line args on run.
948 - Implement command line "restore" args.
949 - Implement "restore current select=no"
950 - Restore file modified before date
951 - Restore -- do nothing but show what would happen
952 - Add estimate to Console commands
953 - Use read_record.c in SD code.
954 - Fix read_record to handle multiple sessions.
955 - Tip from Steve Allam
956 mt -f /dev/nst0 defblksize 0
957 - Document "status" in the console.
958 - Document driving console from shell script.
959 - Write JobMedia records with max file size is reached on tape.
960 - Handle the case of multiple JobMedia records pending (i.e. the
961 thread is slow and multiple situations requiring a JobMedia
963 - Do performance analysis on the restore tree routines.
964 - Fix maximum file size (block.c) to generate JobMedia records.
965 - Make the default file size 1GB on the tape.
966 - Implement forward spacing between files.
967 - Add Machine type (Linux/Windows) to Status report for daemons.
969 - Use repositioning at the beginning of the tape.
970 - Do full check the command line args in update (e.g. VolStatus ...).
971 - Specify list of files to restore
972 - Implement ClientRunBeforeJob and ClientRunAfterJob.
973 - Make | and < work on FD side.
974 - Check to see if "blocked" is set during restore.
975 - Figure out what is interrupting sql command in console.
976 - Make new job print warning User Unmounted Tape.
977 - Test recycling and purging (code changed in db_find_next_volume and
979 - Document SDConnectTimeout (in FD).
980 - Add restore by filename test.
981 - Document restore by files.
982 - Make variable expansion work correctly.
983 - Implement List Volume Job=xxx or List scheduled volumes or Status Director
984 - Copy static programs into install directory.
985 - Think about changing Storage resource Device record to be
987 - Add RunBeforeJob and RunAfterJob to the Client program.
988 - Need return status on read_cb() from read_records(). Need multiple
989 records -- one per Job, maybe a JCR or some other structure with
990 a block and a record.
991 - LabelFormat on tape volume apparently creates the db record but
992 never actually labels the volume.
993 - Recycling a volume when two jobs are using it is going to break. Fixed.
994 - Document list nextvol and new format status dir.
995 - Client files in Win32 with Unix eol conventions doesn't work.
996 - Either fix or document that fill command in btape can be
997 compressed enormously by the hardware - a 36GB tape wrote 750GB!
998 - Add multiple character duration qualifiers.
999 - Require some modifer.
1000 - Restrict characters permitted in a Resource name, and don't permit
1002 - Figure out some way to ignore or get past checksum errors in
1004 - The SD spooling file gets created even if it is not used.
1005 - Look at Cleaning tape in ua_label.c for media create/update
1006 - Add regression testing to the manual
1007 - End time: in job output of rescheduled job is time of first run.
1008 - Document list nextvol and status output.
1009 - Separate Dir heartbeat in FD from the SD heartbeat.
1010 - Fix sparse file handeling so that it always reads a multiple
1011 of 512. Currently, it subtracts 8 bytes (for faddr).
1012 Kludged with #ifdef for FreeBSD.
1013 - Document that Volume pruning can delete last Full backup and
1014 hence you will not have a valid backup.
1015 - Clarify the fact that having the Bacula cygwin1.dll loaded
1016 is not the same as having cygwin installed.
1017 - Document that it is safe to use the drive when the lights stop flashing.
1018 - Document all the status codes JobLevel, JobType, JobStatus.
1019 - Add GUI interface to manual
1020 - Combine the 3 places that search run records for the next
1021 job. Use find_job_pool() modified in ua_output.c
1022 - Test connect timeouts.