5 - Document running a test version.
8 - that restore options work in FD.
9 - that mod of restore options works.
10 - that console command line options work
11 - blocksize recognition code.
16 - Ensure that restore of differential jobs works (check SQL).
17 - Enhance schedule to have 1stSat, ...
18 - Make sure the MaxVolFiles is fully implemented in SD
19 - Make Job err if WriteBootstrap fails.
20 - Flush all the daemon messages at the end of every job.
21 - Check if both CatalogFiles and UseCatalog are set to SD.
22 - Check if we can bump Bacula FD priorty in Win2000
23 - Implement FileOptions.
24 - Make bcopy read through bad tape records.
25 - Need return status on read_cb() from read_records(). Need multiple
26 records -- one per Job, maybe a JCR or some other structure with
28 - Continue improving the restore process (handling
29 of tapes, efficiency improvements e.g. use FSF to
30 position the tape, ...)
31 - Work more on how to to a Bacula restore beginning with
32 just a Bacula tape and a boot floppy (bare metal recovery).
33 - Try bare metal Windows restore
34 - Fix read_record to handle multiple sessions.
35 - Program files (i.e. execute a program to read/write files).
36 Pass read date of last backup, size of file last time.
37 - Put system type returned by FD into catalog.
38 - Add VOLUME_CAT_INFO to the EOS tape record (as
39 well as to the EOD record).
40 - Add code to fast seek to proper place on tape/file
41 when doing Restore. If it doesn't work, try linear
43 - Add code to reject whole blocks if not wanted on restore.
44 - Possibly add email to Watchdog if drive is unmounted too
45 long and a job is waiting on the drive.
46 - Strip trailing slashes from Include directory names in the FD.
47 - Use read_record.c in SD code.
48 - Add EOM records ???????
49 - Why don't we get an error message from Win32 FD when bootstrap
50 file cannot be created for restore command?
51 - Put MaximumVolumeSize in Director (MaximumVolumeJobs, MaximumVolumeFiles,
53 - When Marking a file in Restore that is a hard link, also
54 mark the link so that the data will be reloaded.
55 - Restore program that errors in SD due to no tape reports
56 OK incorrectly in output.
57 - Make BSR accept count (total files to be restored).
58 - Make BSR return next_block when it knows record is not
59 in block, done when count is reached, and possibly other
60 optimizations. I.e. add a state word.
62 - Running multiple simultaneous jobs has a problem when one Job
63 must label the volume -- the others apparently do not wait.
65 > 13-Nov-2002 02:06 dump01-dir: Start Backup JobId 20, Job=save-rho.2002-11-13_02.06.00
66 > 13-Nov-2002 02:06 dump01-dir: Start Backup JobId 21, Job=save-beta.2002-11-13_02.06.01
67 > 13-Nov-2002 02:06 dump01-dir: Start Backup JobId 22, Job=save-dump01.2002-11-13_02.06.02
68 > 13-Nov-2002 02:06 dump01-sd: Created Volume label File0013 on device /dump7/bacula.
69 > 13-Nov-2002 02:06 dump01-sd: Wrote label to prelabeled Volume File0013 on device /dump7/bacula
70 > 13-Nov-2002 02:06 dump01-sd: save-beta.2002-11-13_02.06.01 Fatal error: Device /dump7/bacula is
71 > busy writing with another Volume.
72 > 13-Nov-2002 02:06 dump01-sd: save-dump01.2002-11-13_02.06.02 Fatal error: Device /dump7/bacula is
73 > busy writing with another Volume.
74 > 13-Nov-2002 02:06 beta-fd: save-beta.2002-11-13_02.06.01 Error: bnet.c:292 Write error sending to
75 > Storage daemon:dump01:9103: ERR=Broken pipe
76 > 13-Nov-2002 02:06 dump01-fd: save-dump01.2002-11-13_02.06.02 Error: bnet.c:319 Write error sending
77 > to Storage daemon:dump01:9103: ERR=Broken pipe
78 > 13-Nov-2002 02:08 dump01-dir: 13-Nov-2002 02:08
81 - Add VolumeUseDuration and MaximumVolumeJobs to Pool db record and
83 - Figure out how compress everything except .gz,... files.
84 - Make bcopy copy with a single tape drive.
85 - Make sure catalog doesn't keep growing.
86 - Permit changing ownership during restore.
87 - After unmount, if restore job started, ask to mount.
88 - Fix db_get_fileset in cats/sql_get.c for multiple records.
89 - Fix start/end blocks for File devices
90 - Add new code to scheduler.c and run_conf.c
91 schedule options (1-sat, 2-sat, ...).
92 - Fix catalog filename truncation in sql_get and sql_create. Use
93 only a single filename split routine.
94 - Add command to reset VolFiles to a larger value (don't allow
95 a smaller number or print big warning).
96 - Make Restore report an error if FD or SD term codes are not OK.
97 - Convert all %x substitution variables, which are hard to remember
98 and read to %(variable-name). Idea from TMDA.
99 - Add JobLevel in FD status (but make sure it is defined).
100 - Make Pool resource handle Counter resources.
101 - Remove NextId for SQLite. Optimize.
102 - Fix gethostbyname() to use gethostbyname_r()
103 - Implement ./configure --with-client-only
104 - Strip trailing / from Include
105 - Move all SQL statements into a single location.
106 - Cleanup db_update_media and db_update_pool
107 - Add UA rc and history files.
108 - put termcap (used by console) in ./configure and
109 allow -with-termcap-dir.
110 - Enhance time and size scanning routines.
111 - Fix Autoprune for Volumes to respect need for full save.
112 - DateWritten field on tape may be wrong.
113 - Fix Win32 config file definition name on /install
114 - No READLINE_SRC if found in alternate directory.
115 - Add Client FS/OS id (Linux, Win95/98, ...).
119 Bacula Projects Roadmap
121 last update 7 October 2002
123 Item 1: Multiple simultaneous Jobs. (done)
126 What: Permit multiple simultaneous jobs in Bacula.
128 Why: An enterprise level solution needs to go fast without the
129 need for the system administrator to carefully tweak
130 timing. Based on the benchmarks, during a full
131 backup, NetWorker typically hit 10 times the bandwidth to
132 the tape compared to Bacula--largely. This is probably due to
133 running parallel jobs and multi-threaded filling of buffers
134 and writing them to tape. This should also make things work
135 better when you have a mix of fast and slow machines backing
138 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
139 implementing this is a matter of some small cleanups and
143 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
144 Deferred -- not necessary yet.
146 What: If data is coming into the SD too fast, buffer it to
147 disk if the user has configured this option.
149 Why: This would be nice, especially if it more or less falls out
150 when implementing (1) above. If not, it probably should not
151 be given a high priority because fundamentally the backup time
152 is limited by the tape bandwidth. Even though you may finish a
153 client job quicker by spilling to disk, you still have to
154 eventually get it onto tape. If intermediate disk buffering
155 allows us to improve write bandwidth to tape, it may make
158 Notes: Whether or not this is implemented will depend upon performance
159 testing after item 1 is implemented.
162 Item 3: Write the bscan program.
165 What: Write a program that reads a Bacula tape and puts all the
166 appropriate data into the catalog. This allows recovery
167 from a tape that is no longer in the database, or it allows
168 re-creation of a database if lost.
170 Why: This is a fundamental robustness and disaster recovery tool
171 which will increase the comfort level of a sysadmin
172 considering adopting Bacula.
174 Notes: A skeleton of this program already exists, but much work
175 needs to be done. Implementing this will also make apparent
176 any deficiencies in the current Bacula tape format.
179 Item 4: Implement Base jobs.
181 What: A base job is sort of like a Full save except that you
182 will want the FileSet to contain only files that are unlikely
183 to change in the future (i.e. a snapshot of most of your
184 system after installing it). After the base job has been run,
185 when you are doing a Full save, you can specify to exclude
186 all files saved by the base job that have not been modified.
188 Why: This is something none of the competition does, as far as we know
189 (except BackupPC, which is a Perl program that saves to disk
190 only). It is big win for the user, it makes Bacula stand out
191 as offering a unique optimization that immediately saves time
194 Notes: Big savings in tape usage. Will require more resources because
195 the e. DIR must send FD a list of files/attribs, and the FD must
196 search the list and compare it for each file to be saved.
199 Item 5: Implement Label templates
201 What: This is a mechanism whereby Bacula can automatically create
202 a tape label for new tapes according to a detailed specification
203 provided by the user.
205 Why: It is a major convenience item for folks who use automated label
208 Notes: Bacula already has a working form of automatic tape label
209 creation, but it is very crude. The design for the complete
210 tape labeling project is already documented in the manual.
213 Item 6: Write a regression script.
216 What: This is an automatic script that runs and tests as many features
217 of Bacula as possible. The output is compared to previous
218 versions of Bacula and any differences are reported.
220 Why: This is an enormous help in preventing introduction of new
221 errors in parts of the program that already work correctly.
223 Notes: This probably should be ranked higher, it's something the typical
224 user doesn't see. Depending on how it's implemented, it may
225 make sense to defer it until the archival tape format and
226 user interface mature.
229 Item 7: GUI for interactive restore
230 Item 8: GUI for interactive backup
232 What: The current interactive restore is implemented with a tty
233 interface. It would be much nicer to be able to "see" the
234 list of files backed up in typical GUI tree format.
235 The same mechanism could also be used for creating
236 ad-hoc backup FileSets (item 8).
238 Why: Ease of use -- especially for the end user.
240 Notes: Rather than implementing in Gtk, we probably should go directly
241 for a Browser implementation, even if doing so meant the
242 capability wouldn't be available until much later. Not only
243 is there the question of Windows sites, most
244 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
245 without installing lots of stuff admins are very wary about.
246 Real sysadmins will always use the command line anyway, and
247 the user who's doing an interactive restore or backup of his
248 own files will in most cases be on a Windows machine running
252 Item 9: Add SSL to daemon communications.
254 What: This provides for secure communications between the daemons.
256 Why: This would allow doing backup across the Internet without
257 privacy concerns (or with much less concern).
259 Notes: The vast majority of near term potential users will be backing up
260 a single site over a LAN and, correctly or not, they probably
261 won't be concerned with security, at least not enough to go to
262 the trouble to set up keys, etc. to screw things down. We suspect
263 that many users genuinely interested in multi-site backup
264 already run some form of VPN software in their internetwork
265 connections, and are willing to delegate security to that layer.
268 Item 10: Define definitive tape format.
269 Mostly done (version 1.27)
271 What: Define that definitive tape format that will not change
272 for the next millennium.
274 Why: Stability, security.
276 Notes: See notes for item 11 below.
279 Item 11: New daemon communication protocol.
281 What: The current daemon to daemon protocol is basically an ASCII
282 printf() and sending the buffer. On the receiving end, the
283 buffer is sscanf()ed to unpack it. The new scheme would
284 be a binary format that allows quick packing and unpacking
285 of any data type with named fields.
287 Why: Using binary packing would be faster. Named fields will permit
288 error checking to ensure that what is sent is what the
289 receiver really wants.
291 Notes: These are internal improvements in the interest of the
292 long-term stability and evolution of the program. On the one
293 hand, the sooner they're done, the less code we have to rip
294 up when the time comes to install them. On the other hand, they
295 don't bring an immediately perceptible benefit to potential
296 users. Item 10 and possibly item 11 should be deferred until Bacula
297 is well established with a growing user community more or
298 less happy with the feature set. At that time, it will make a
299 good "next generation" upgrade in the interest of data
304 I haven't put these in any particular order.
307 - Rework Storage daemon with new rwl_lock routines.
308 - Compare tape to Client files (attributes, or attributes and data)
309 - Restore options (overwrite, overwrite if older,
310 overwrite if newer, never overwrite, ...)
311 - Restore to a particular time -- e.g. before date, after date.
312 - Make all database Ids 64 bit.
313 - Write an applet for Linux.
314 - Add estimate to Console commands
315 - Find solution to blank filename (i.e. path only) problem.
316 - Implement new daemon communications protocol.
319 mysqldump -f --opt bacula >bacula
323 - Remove PoolId from Job table, it exists in Media.
324 - Allow console commands to detach or run in background.
325 - Fix status delay on storage daemon during rewind.
326 - Add SD message variables to control operator wait time
327 - Maximum Operator Wait
328 - Minimum Message Interval
329 - Maximum Message Interval
330 - Add EOM handling variables
332 - Require EOD records
333 - Send Operator message when cannot read tape label.
334 - Think about how to handle I/O error on MTEOM.
335 - Verify level=Volume (scan only), level=Data (compare of data to file).
336 Verify level=Catalog, level=InitCatalog
338 - Add keyword search to show command in Console.
339 - Fix Win2000 error with no messages during startup.
340 - Events : tape has more than xxx bytes.
341 - Write general list maintenance subroutines.
342 - Restrict characters permitted in a Resource name.
343 - Complete code in Bacula Resources -- this will permit
344 reading a new config file at any time.
345 - Handle ctl-c in Console
346 - Implement LabelTemplate (at least first cut).
347 - Implement script driven addition of File daemon to config files.
349 - see setgroup and user for Bacula p4-5 of stunnel.c
350 - Implement new serialize subroutines
351 send(socket, "string", &Vol, "uint32", &i, NULL)
352 - On I/O error, write EOF, then try to write again ????
353 - Audit all UA commands to ensure that we always prompt where possible.
354 - If ./btape is called without /dev, assume argument is a Storage resource name.
355 - Put memory utilization in Status output of each daemon
356 if full status requested or if some level of debug on.
357 - Make database type selectable by .conf files i.e. at runtime
358 - gethostbyname failure in bnet_connect() continues
359 generating errors -- should stop.
360 - Add HOST to Volume label.
361 - Set flag for uname -a. Add to Volume label.
362 - Implement throttled work queue.
363 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
364 - Allow multiple Storage specifications (or multiple names on
365 a single Storage specification) in the Job record. Thus a job
366 can be backed up to a number of storage devices.
367 - Implement full MediaLabel code.
368 - Implement dump label to UA
369 - Copy volume using single drive.
370 - Copy volume with multiple driven (same or different block size).
371 - Add block size (min, max) to Vol label.
372 - Concept of VolumeSet during restore which is a list
373 of Volume names needed.
374 - Restore files modified after date
375 - Restore file modified before date
376 - Emergency restore info:
378 - Backup working directory
380 - Restore options (do not overwrite)
381 - Restore -- do nothing but show what would happen
382 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
383 - Implement Restore FileSet=
384 - Write a scanner for the UA (keyword, scan-routine, result, prompt).
385 - Create a protocol.h and protocol.c where all protocol messages
387 - If SD cannot open a drive, make it periodically retry.
388 - Put Bacula version somewhere in Job stream, probably Start Session
390 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
391 - Timout a job or terminate if link goes down, or reopen link and query.
392 - Fill all fields in Vol/Job Header -- ensure that everything
393 needed is written to tape. Think about restore to Catalog
394 from tape. Client record needs improving.
395 - Find general solution for sscanf size problems (as well
396 as sprintf. Do at run time?
397 - Concept of precious tapes (cannot be reused).
398 - Allow FD to run from inetd ???
400 - Restore should get Device and Pool information from
401 job record rather than from config.
402 - Autolabel should be specified by DR instead of SD.
403 - Find out how to get the system tape block limits, e.g.:
404 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
405 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
408 - AutoScan (check checksum of tape)
409 - Format command = "format /dev/nst0"
413 - Seek resolution (usually corresponds to buffer size)
414 - EODErrorCode=ENOSPC or code
415 - Partial Read error code
416 - Partial write error code
417 - Nonformatted read error
418 - Nonformatted write error
419 - WriteProtected error
423 - IgnoreCloseErrors=yes
433 - FD sends unsaved file list to Director at end of job.
434 - Write a Storage daemon that uses pipes and
435 standard Unix programs to write to the tape.
437 - Need something that monitors the JCR queue and
438 times out jobs by asking the deamons where they are.
440 - Enhance Jmsg code to permit buffering and saving to disk.
441 - device driver = "xxxx" for drives.
442 - restart: paranoid: read label fsf to
443 eom read append block, and go
444 super-paranoid: read label, read all files
445 in between, read append block, and go
446 verify: backspace, read append block, and go
447 permissive: same as above but frees drive
448 if tape is not valid.
450 - Ensure that /dev/null works
451 - File daemon should build list of files skipped, and then
452 at end of save retry and report any errors.
453 - Need report class for messages. Perhaps
454 report resource where report=group of messages
455 - enhance scan_attrib and rename scan_jobtype, and
456 fill in code for "since" option
457 - To buffer messages, we need associated jobid and Director name.
458 - Need to save contents of FileSet to tape?
459 - Director needs a time after which the report status is sent
460 anyway -- or better yet, a retry time for the job.
461 Don't reschedule a job if previous incarnation is still running.
462 - Figure out how to save the catalog (possibly a special FileSet).
463 - Figure out how to restore the catalog.
464 - Some way to automatically backup everything is needed????
465 - Need a structure for pending actions:
467 - termination status (part of buffered msgs?)
468 - Concept of grouping Storage devices and job can use
469 any of a number of devices
471 Read, Write, Clean, Delete
472 - Login to Bacula; Bacula users with different permissions:
473 owner, group, user, quotas
474 - Tape recycle destination
475 - Job Schedule Status
479 - Store info on each file system type (probably in the job header on tape.
480 This could be the output of df; or perhaps some sort of /etc/mtab record.
483 - Use media 1 time (so that we can do 6 days of incremental
484 backups before switching to another tape) (already)
485 specify # times (jobs)
486 specify bytes (already)
487 specify time (seconds, hours, days)
488 - Implement FSM (File System Modules).
489 - Identify unchanged or "system" files and save them to a
490 special tape thus removing them from the standard
491 backup FileSet -- BASE backup.
492 - Turn virutally all sprintfs into snprintfs.
493 - Heartbeat between daemons.
494 - Audit M_ error codes to ensure they are correct and consistent.
495 - Add variable break characters to lex analyzer.
496 Either a bit mask or a string of chars so that
497 the caller can change the break characters.
498 - Make a single T_BREAK to replace T_COMMA, etc.
499 - Ensure that File daemon and Storage daemon can
500 continue a save if the Director goes down (this
501 is NOT currently the case). Must detect socket error,
502 buffer messages for later.
505 Done: (see kernsdone for more)
508 - Check if GZIP1 is working -- check speed.
509 - Document forcing a new tape to be used.
510 - Ensure that AcceptAnyVolume works.
511 - Document running multiple Jobs
512 - Preserve block number when EOT and writing on next tape.
513 - Document how to cancel a job that is waiting on a Volume.
514 Must "cancel" then "mount".
515 - Document Volume Bytes shows bytes on last volume written in Job summary.
516 - Restore all Windows attributes. Leave hooks for ACLs and security.
517 (Handle x = (HANDLE)get_osfhandle(fd);
518 - Test Windows restore.
520 - Implement sparse files.
521 - Document sparse files.
522 - Document better Include (does it cross file systems ?).
523 - Document default config file locations.
524 - Document specifically how to add new File daemon to config files.
525 - Add VerNo to each Session label record.
526 - Add Job to Session records.
527 - Cold start full restore (restore catalog then
528 user selects what to restore). Write summary file containing only
529 Job, Media, and Catalog information. Store on another machine.
530 - Dump/Restore database
531 - Write bscan program that will syncronize the DB Media record with
532 the contents of the Volume -- for use after a crash.
533 - Figure out how to put a Volume into the catalog (from the tape)
534 - Figure out how to do a restore from a Volume
535 - Report compression % and other compression statistics if turned on.
536 - Put Windows files in Windows stream?
537 - Ensure that everyone uses btime routines (mostly done).
538 - Put Job statistics in End Session Label (files saved,
539 total bytes, start time, ...).
540 - Put FileSet name in the SOS label.
541 - Eliminate duplicate File records to shrink database.
542 - If Storage daemon aborts a job, ensure that this
543 is printed in the error message.
544 - Add save type to Session label.
545 - Correct date on Session label.
546 - Test restore of Windows backup
547 - Ability to recreate the catalog from a tape.
548 - Bug: anonymous Volumes requires mount in some cases.
549 - Define how we handle times to avoid problem with Unix dates (2049 ?).
550 - Add daemon JCR JobId=0 to have a daemon context
551 - Implement full restoration of all Windows attributes
552 (such as Hidden, System, creation dates, ...)
553 - Handle sparse files (i.e. files with holes in them)
554 - Enhance testing for Bacula compatibility with
556 - Turn on new BB02 tape format implemented in 1.26 but
558 - More testing of restoring on Unix systems.
559 - Implement additional tape format enhancements to better
560 support Windows and other non-Unix systems e.g.
562 - Upgrade to latest version of cygwin. (not possible)
563 - Add configure for gettimeofday.
564 - At line 51 of ua_input.c, why is = 0 necessary. Previously without
565 it, if cancel gnome-console during sql command, DIR crashed. However,
566 with it, blank line input for Where: is not possible.
567 - Make SD disallow writing on Volume with fewer files than in
569 - Label (asks for slot, return and it stops).
570 - Make SD reject writing on tape where Catalog and tape # files
571 don't agree (possibly OK if tape > catalog).
572 - Document all daemon tools MUST have a config file.
573 - Why does btape error when pointed to a file?
574 - Disallow compile if long long not 64 bits.
575 - Send Volumes needed during restore to Console (just after
576 create_volume_list) -- also in restore command?
577 - File system type from File daemon
578 - Move block size code from block.c to init_dev().
579 - Add FileSet MD5 to bscan.
580 - Finish implementation of restore "replace" options, and document.
582 - Document buffer size considerations with Sparse files --
583 - that restore prints volumes.
584 - Document that two Verifys at same time on same client do not work.
586 - MD5 in bscan is set in FileSet.
587 - that Bacula won't write on tape where tape/catalog files differ.
588 - Document how to recycle a tape in 7 days even if the backup takes a long time.
589 - Document tape cycling
590 - Implement VolumeUseDuration (maximum duration/period) a volume can be used).
591 Possibly VolumeUseTimes (MaximumVolumeJobs) This permits a volume to be
592 recycled even if it is not full.
593 - Decide what to do with JobTDate in catalog (make real utime_t?)
594 - Implement VolMaxJob and Duration check in next_volume().
595 - Make sure pruning of Jobs removes JobMediaId
596 - Write bcopy program -- recovery of bad tape.
597 - Make gethostbyname() thread safe in bnet.c
598 - Add ORDER BY JobId to list of Jobs in query.sql, and in
599 ua_output.c (list command).
600 - drive MaxVolJobs, VolUseDuration from media record
601 rather than Resource.
602 - Fix intmax_t on FreeBSD.
603 - make sure that update of volume new parameters works
604 - Document -i option on FD
605 - Document to have patience when SD first starts.
606 - Document saving MySQL databases, where to find code for shutting
607 down and saving other databases.
608 http://www.backupcentral.com/free-backup-software1.html
609 mysqldump -f --opt bacula >bacula