5 - Document to have patience when SD first starts.
6 - Document running a test version.
7 - Document saving MySQL databases, where to find code for shutting
8 down and saving other databases.
9 http://www.backupcentral.com/free-backup-software1.html
10 - Document -i option on FD
13 - that restore options work in FD.
14 - that mod of restore options works.
15 - that console command line options work
16 - blocksize recognition code.
17 - make sure that update of volume new parameters works
20 - drive MaxVolJobs, VolUseDuration from media record
25 - Check if both CatalogFiles and UseCatalog are set to SD.
26 - Check if we can bump Bacula FD priorty in Win2000
27 - Implement FileOptions.
28 - Make bcopy read through bad tape records.
29 - Need return status on read_cb() from read_records(). Need multiple
30 records -- one per Job, maybe a JCR or some other structure with
32 - Continue improving the restore process (handling
33 of tapes, efficiency improvements e.g. use FSF to
34 position the tape, ...)
35 - Work more on how to to a Bacula restore beginning with
36 just a Bacula tape and a boot floppy (bare metal recovery).
37 - Try bare metal Windows restore
38 - Fix read_record to handle multiple sessions.
39 - Program files (i.e. execute a program to read/write files).
40 Pass read date of last backup, size of file last time.
41 - Put system type returned by FD into catalog.
42 - Add VOLUME_CAT_INFO to the EOS tape record (as
43 well as to the EOD record).
44 - Add code to fast seek to proper place on tape/file
45 when doing Restore. If it doesn't work, try linear
47 - Add code to reject whole blocks if not wanted on restore.
48 - Possibly add email to Watchdog if drive is unmounted too
49 long and a job is waiting on the drive.
50 - Strip trailing slashes from Include directory names in the FD.
51 - Use read_record.c in SD code.
52 - Add EOM records ???????
53 - Why don't we get an error message from Win32 FD when bootstrap
54 file cannot be created for restore command?
55 - Put MaximumVolumeSize in Director (MaximumVolumeJobs, MaximumVolumeFiles,
57 - When Marking a file in Restore that is a hard link, also
58 mark the link so that the data will be reloaded.
59 - Restore program that errors in SD due to no tape reports
60 OK incorrectly in output.
61 - Make BSR accept count (total files to be restored).
62 - Make BSR return next_block when it knows record is not
63 in block, done when count is reached, and possibly other
64 optimizations. I.e. add a state word.
66 - Running multiple simultaneous jobs has a problem when one Job
67 must label the volume -- the others apparently do not wait.
69 > 13-Nov-2002 02:06 dump01-dir: Start Backup JobId 20, Job=save-rho.2002-11-13_02.06.00
70 > 13-Nov-2002 02:06 dump01-dir: Start Backup JobId 21, Job=save-beta.2002-11-13_02.06.01
71 > 13-Nov-2002 02:06 dump01-dir: Start Backup JobId 22, Job=save-dump01.2002-11-13_02.06.02
72 > 13-Nov-2002 02:06 dump01-sd: Created Volume label File0013 on device /dump7/bacula.
73 > 13-Nov-2002 02:06 dump01-sd: Wrote label to prelabeled Volume File0013 on device /dump7/bacula
74 > 13-Nov-2002 02:06 dump01-sd: save-beta.2002-11-13_02.06.01 Fatal error: Device /dump7/bacula is
75 > busy writing with another Volume.
76 > 13-Nov-2002 02:06 dump01-sd: save-dump01.2002-11-13_02.06.02 Fatal error: Device /dump7/bacula is
77 > busy writing with another Volume.
78 > 13-Nov-2002 02:06 beta-fd: save-beta.2002-11-13_02.06.01 Error: bnet.c:292 Write error sending to
79 > Storage daemon:dump01:9103: ERR=Broken pipe
80 > 13-Nov-2002 02:06 dump01-fd: save-dump01.2002-11-13_02.06.02 Error: bnet.c:319 Write error sending
81 > to Storage daemon:dump01:9103: ERR=Broken pipe
82 > 13-Nov-2002 02:08 dump01-dir: 13-Nov-2002 02:08
85 - Add VolumeUseDuration and MaximumVolumeJobs to Pool db record and
87 - Figure out how compress everything except .gz,... files.
88 - Make bcopy copy with a single tape drive.
89 - Make sure catalog doesn't keep growing.
90 - Permit changing ownership during restore.
91 - After unmount, if restore job started, ask to mount.
92 - Fix db_get_fileset in cats/sql_get.c for multiple records.
93 - Fix start/end blocks for File devices
94 - Add new code to scheduler.c and run_conf.c
95 schedule options (1-sat, 2-sat, ...).
96 - Fix catalog filename truncation in sql_get and sql_create. Use
97 only a single filename split routine.
98 - Add command to reset VolFiles to a larger value (don't allow
99 a smaller number or print big warning).
100 - Make Restore report an error if FD or SD term codes are not OK.
101 - Convert all %x substitution variables, which are hard to remember
102 and read to %(variable-name). Idea from TMDA.
103 - Add JobLevel in FD status (but make sure it is defined).
104 - Make Pool resource handle Counter resources.
105 - Remove NextId for SQLite. Optimize.
106 - Fix gethostbyname() to use gethostbyname_r()
107 - Implement ./configure --with-client-only
108 - Strip trailing / from Include
109 - Move all SQL statements into a single location.
110 - Cleanup db_update_media and db_update_pool
111 - Add UA rc and history files.
112 - put termcap (used by console) in ./configure and
113 allow -with-termcap-dir.
114 - Enhance time and size scanning routines.
115 - Fix Autoprune for Volumes to respect need for full save.
116 - DateWritten field on tape may be wrong.
117 - Fix Win32 config file definition name on /install
118 - No READLINE_SRC if found in alternate directory.
119 - Add Client FS/OS id (Linux, Win95/98, ...).
123 Bacula Projects Roadmap
125 last update 7 October 2002
127 Item 1: Multiple simultaneous Jobs. (done)
130 What: Permit multiple simultaneous jobs in Bacula.
132 Why: An enterprise level solution needs to go fast without the
133 need for the system administrator to carefully tweak
134 timing. Based on the benchmarks, during a full
135 backup, NetWorker typically hit 10 times the bandwidth to
136 the tape compared to Bacula--largely. This is probably due to
137 running parallel jobs and multi-threaded filling of buffers
138 and writing them to tape. This should also make things work
139 better when you have a mix of fast and slow machines backing
142 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
143 implementing this is a matter of some small cleanups and
147 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
148 Deferred -- not necessary yet.
150 What: If data is coming into the SD too fast, buffer it to
151 disk if the user has configured this option.
153 Why: This would be nice, especially if it more or less falls out
154 when implementing (1) above. If not, it probably should not
155 be given a high priority because fundamentally the backup time
156 is limited by the tape bandwidth. Even though you may finish a
157 client job quicker by spilling to disk, you still have to
158 eventually get it onto tape. If intermediate disk buffering
159 allows us to improve write bandwidth to tape, it may make
162 Notes: Whether or not this is implemented will depend upon performance
163 testing after item 1 is implemented.
166 Item 3: Write the bscan program.
169 What: Write a program that reads a Bacula tape and puts all the
170 appropriate data into the catalog. This allows recovery
171 from a tape that is no longer in the database, or it allows
172 re-creation of a database if lost.
174 Why: This is a fundamental robustness and disaster recovery tool
175 which will increase the comfort level of a sysadmin
176 considering adopting Bacula.
178 Notes: A skeleton of this program already exists, but much work
179 needs to be done. Implementing this will also make apparent
180 any deficiencies in the current Bacula tape format.
183 Item 4: Implement Base jobs.
185 What: A base job is sort of like a Full save except that you
186 will want the FileSet to contain only files that are unlikely
187 to change in the future (i.e. a snapshot of most of your
188 system after installing it). After the base job has been run,
189 when you are doing a Full save, you can specify to exclude
190 all files saved by the base job that have not been modified.
192 Why: This is something none of the competition does, as far as we know
193 (except BackupPC, which is a Perl program that saves to disk
194 only). It is big win for the user, it makes Bacula stand out
195 as offering a unique optimization that immediately saves time
198 Notes: Big savings in tape usage. Will require more resources because
199 the e. DIR must send FD a list of files/attribs, and the FD must
200 search the list and compare it for each file to be saved.
203 Item 5: Implement Label templates
205 What: This is a mechanism whereby Bacula can automatically create
206 a tape label for new tapes according to a detailed specification
207 provided by the user.
209 Why: It is a major convenience item for folks who use automated label
212 Notes: Bacula already has a working form of automatic tape label
213 creation, but it is very crude. The design for the complete
214 tape labeling project is already documented in the manual.
217 Item 6: Write a regression script.
220 What: This is an automatic script that runs and tests as many features
221 of Bacula as possible. The output is compared to previous
222 versions of Bacula and any differences are reported.
224 Why: This is an enormous help in preventing introduction of new
225 errors in parts of the program that already work correctly.
227 Notes: This probably should be ranked higher, it's something the typical
228 user doesn't see. Depending on how it's implemented, it may
229 make sense to defer it until the archival tape format and
230 user interface mature.
233 Item 7: GUI for interactive restore
234 Item 8: GUI for interactive backup
236 What: The current interactive restore is implemented with a tty
237 interface. It would be much nicer to be able to "see" the
238 list of files backed up in typical GUI tree format.
239 The same mechanism could also be used for creating
240 ad-hoc backup FileSets (item 8).
242 Why: Ease of use -- especially for the end user.
244 Notes: Rather than implementing in Gtk, we probably should go directly
245 for a Browser implementation, even if doing so meant the
246 capability wouldn't be available until much later. Not only
247 is there the question of Windows sites, most
248 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
249 without installing lots of stuff admins are very wary about.
250 Real sysadmins will always use the command line anyway, and
251 the user who's doing an interactive restore or backup of his
252 own files will in most cases be on a Windows machine running
256 Item 9: Add SSL to daemon communications.
258 What: This provides for secure communications between the daemons.
260 Why: This would allow doing backup across the Internet without
261 privacy concerns (or with much less concern).
263 Notes: The vast majority of near term potential users will be backing up
264 a single site over a LAN and, correctly or not, they probably
265 won't be concerned with security, at least not enough to go to
266 the trouble to set up keys, etc. to screw things down. We suspect
267 that many users genuinely interested in multi-site backup
268 already run some form of VPN software in their internetwork
269 connections, and are willing to delegate security to that layer.
272 Item 10: Define definitive tape format.
273 Mostly done (version 1.27)
275 What: Define that definitive tape format that will not change
276 for the next millennium.
278 Why: Stability, security.
280 Notes: See notes for item 11 below.
283 Item 11: New daemon communication protocol.
285 What: The current daemon to daemon protocol is basically an ASCII
286 printf() and sending the buffer. On the receiving end, the
287 buffer is sscanf()ed to unpack it. The new scheme would
288 be a binary format that allows quick packing and unpacking
289 of any data type with named fields.
291 Why: Using binary packing would be faster. Named fields will permit
292 error checking to ensure that what is sent is what the
293 receiver really wants.
295 Notes: These are internal improvements in the interest of the
296 long-term stability and evolution of the program. On the one
297 hand, the sooner they're done, the less code we have to rip
298 up when the time comes to install them. On the other hand, they
299 don't bring an immediately perceptible benefit to potential
300 users. Item 10 and possibly item 11 should be deferred until Bacula
301 is well established with a growing user community more or
302 less happy with the feature set. At that time, it will make a
303 good "next generation" upgrade in the interest of data
308 I haven't put these in any particular order.
311 - Rework Storage daemon with new rwl_lock routines.
312 - Compare tape to Client files (attributes, or attributes and data)
313 - Restore options (overwrite, overwrite if older,
314 overwrite if newer, never overwrite, ...)
315 - Restore to a particular time -- e.g. before date, after date.
316 - Make all database Ids 64 bit.
317 - Write an applet for Linux.
318 - Add estimate to Console commands
319 - Find solution to blank filename (i.e. path only) problem.
320 - Implement new daemon communications protocol.
323 mysqldump -f --opt bacula >bacula
327 - Remove PoolId from Job table, it exists in Media.
328 - Allow console commands to detach or run in background.
329 - Fix status delay on storage daemon during rewind.
330 - Add SD message variables to control operator wait time
331 - Maximum Operator Wait
332 - Minimum Message Interval
333 - Maximum Message Interval
334 - Add EOM handling variables
336 - Require EOD records
337 - Send Operator message when cannot read tape label.
338 - Think about how to handle I/O error on MTEOM.
339 - Verify level=Volume (scan only), level=Data (compare of data to file).
340 Verify level=Catalog, level=InitCatalog
342 - Add keyword search to show command in Console.
343 - Fix Win2000 error with no messages during startup.
344 - Events : tape has more than xxx bytes.
345 - Write general list maintenance subroutines.
346 - Restrict characters permitted in a Resource name.
347 - Complete code in Bacula Resources -- this will permit
348 reading a new config file at any time.
349 - Handle ctl-c in Console
350 - Implement LabelTemplate (at least first cut).
351 - Implement script driven addition of File daemon to config files.
353 - see setgroup and user for Bacula p4-5 of stunnel.c
354 - Implement new serialize subroutines
355 send(socket, "string", &Vol, "uint32", &i, NULL)
356 - On I/O error, write EOF, then try to write again ????
357 - Audit all UA commands to ensure that we always prompt where possible.
358 - If ./btape is called without /dev, assume argument is a Storage resource name.
359 - Put memory utilization in Status output of each daemon
360 if full status requested or if some level of debug on.
361 - Make database type selectable by .conf files i.e. at runtime
362 - gethostbyname failure in bnet_connect() continues
363 generating errors -- should stop.
364 - Add HOST to Volume label.
365 - Set flag for uname -a. Add to Volume label.
366 - Implement throttled work queue.
367 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
368 - Allow multiple Storage specifications (or multiple names on
369 a single Storage specification) in the Job record. Thus a job
370 can be backed up to a number of storage devices.
371 - Implement full MediaLabel code.
372 - Implement dump label to UA
373 - Copy volume using single drive.
374 - Copy volume with multiple driven (same or different block size).
375 - Add block size (min, max) to Vol label.
376 - Concept of VolumeSet during restore which is a list
377 of Volume names needed.
378 - Restore files modified after date
379 - Restore file modified before date
380 - Emergency restore info:
382 - Backup working directory
384 - Restore options (do not overwrite)
385 - Restore -- do nothing but show what would happen
386 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
387 - Implement Restore FileSet=
388 - Write a scanner for the UA (keyword, scan-routine, result, prompt).
389 - Create a protocol.h and protocol.c where all protocol messages
391 - If SD cannot open a drive, make it periodically retry.
392 - Put Bacula version somewhere in Job stream, probably Start Session
394 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
395 - Timout a job or terminate if link goes down, or reopen link and query.
396 - Fill all fields in Vol/Job Header -- ensure that everything
397 needed is written to tape. Think about restore to Catalog
398 from tape. Client record needs improving.
399 - Find general solution for sscanf size problems (as well
400 as sprintf. Do at run time?
401 - Concept of precious tapes (cannot be reused).
402 - Allow FD to run from inetd ???
404 - Restore should get Device and Pool information from
405 job record rather than from config.
406 - Autolabel should be specified by DR instead of SD.
407 - Find out how to get the system tape block limits, e.g.:
408 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
409 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
412 - AutoScan (check checksum of tape)
413 - Format command = "format /dev/nst0"
417 - Seek resolution (usually corresponds to buffer size)
418 - EODErrorCode=ENOSPC or code
419 - Partial Read error code
420 - Partial write error code
421 - Nonformatted read error
422 - Nonformatted write error
423 - WriteProtected error
427 - IgnoreCloseErrors=yes
437 - FD sends unsaved file list to Director at end of job.
438 - Write a Storage daemon that uses pipes and
439 standard Unix programs to write to the tape.
441 - Need something that monitors the JCR queue and
442 times out jobs by asking the deamons where they are.
444 - Enhance Jmsg code to permit buffering and saving to disk.
445 - device driver = "xxxx" for drives.
446 - restart: paranoid: read label fsf to
447 eom read append block, and go
448 super-paranoid: read label, read all files
449 in between, read append block, and go
450 verify: backspace, read append block, and go
451 permissive: same as above but frees drive
452 if tape is not valid.
454 - Ensure that /dev/null works
455 - File daemon should build list of files skipped, and then
456 at end of save retry and report any errors.
457 - Need report class for messages. Perhaps
458 report resource where report=group of messages
459 - enhance scan_attrib and rename scan_jobtype, and
460 fill in code for "since" option
461 - To buffer messages, we need associated jobid and Director name.
462 - Need to save contents of FileSet to tape?
463 - Director needs a time after which the report status is sent
464 anyway -- or better yet, a retry time for the job.
465 Don't reschedule a job if previous incarnation is still running.
466 - Figure out how to save the catalog (possibly a special FileSet).
467 - Figure out how to restore the catalog.
468 - Some way to automatically backup everything is needed????
469 - Need a structure for pending actions:
471 - termination status (part of buffered msgs?)
472 - Concept of grouping Storage devices and job can use
473 any of a number of devices
475 Read, Write, Clean, Delete
476 - Login to Bacula; Bacula users with different permissions:
477 owner, group, user, quotas
478 - Tape recycle destination
479 - Job Schedule Status
483 - Store info on each file system type (probably in the job header on tape.
484 This could be the output of df; or perhaps some sort of /etc/mtab record.
487 - Use media 1 time (so that we can do 6 days of incremental
488 backups before switching to another tape) (already)
489 specify # times (jobs)
490 specify bytes (already)
491 specify time (seconds, hours, days)
492 - Implement FSM (File System Modules).
493 - Identify unchanged or "system" files and save them to a
494 special tape thus removing them from the standard
495 backup FileSet -- BASE backup.
496 - Turn virutally all sprintfs into snprintfs.
497 - Heartbeat between daemons.
498 - Audit M_ error codes to ensure they are correct and consistent.
499 - Add variable break characters to lex analyzer.
500 Either a bit mask or a string of chars so that
501 the caller can change the break characters.
502 - Make a single T_BREAK to replace T_COMMA, etc.
503 - Ensure that File daemon and Storage daemon can
504 continue a save if the Director goes down (this
505 is NOT currently the case). Must detect socket error,
506 buffer messages for later.
509 Done: (see kernsdone for more)
512 - Check if GZIP1 is working -- check speed.
513 - Document forcing a new tape to be used.
514 - Ensure that AcceptAnyVolume works.
515 - Document running multiple Jobs
516 - Preserve block number when EOT and writing on next tape.
517 - Document how to cancel a job that is waiting on a Volume.
518 Must "cancel" then "mount".
519 - Document Volume Bytes shows bytes on last volume written in Job summary.
520 - Restore all Windows attributes. Leave hooks for ACLs and security.
521 (Handle x = (HANDLE)get_osfhandle(fd);
522 - Test Windows restore.
524 - Implement sparse files.
525 - Document sparse files.
526 - Document better Include (does it cross file systems ?).
527 - Document default config file locations.
528 - Document specifically how to add new File daemon to config files.
529 - Add VerNo to each Session label record.
530 - Add Job to Session records.
531 - Cold start full restore (restore catalog then
532 user selects what to restore). Write summary file containing only
533 Job, Media, and Catalog information. Store on another machine.
534 - Dump/Restore database
535 - Write bscan program that will syncronize the DB Media record with
536 the contents of the Volume -- for use after a crash.
537 - Figure out how to put a Volume into the catalog (from the tape)
538 - Figure out how to do a restore from a Volume
539 - Report compression % and other compression statistics if turned on.
540 - Put Windows files in Windows stream?
541 - Ensure that everyone uses btime routines (mostly done).
542 - Put Job statistics in End Session Label (files saved,
543 total bytes, start time, ...).
544 - Put FileSet name in the SOS label.
545 - Eliminate duplicate File records to shrink database.
546 - If Storage daemon aborts a job, ensure that this
547 is printed in the error message.
548 - Add save type to Session label.
549 - Correct date on Session label.
550 - Test restore of Windows backup
551 - Ability to recreate the catalog from a tape.
552 - Bug: anonymous Volumes requires mount in some cases.
553 - Define how we handle times to avoid problem with Unix dates (2049 ?).
554 - Add daemon JCR JobId=0 to have a daemon context
555 - Implement full restoration of all Windows attributes
556 (such as Hidden, System, creation dates, ...)
557 - Handle sparse files (i.e. files with holes in them)
558 - Enhance testing for Bacula compatibility with
560 - Turn on new BB02 tape format implemented in 1.26 but
562 - More testing of restoring on Unix systems.
563 - Implement additional tape format enhancements to better
564 support Windows and other non-Unix systems e.g.
566 - Upgrade to latest version of cygwin. (not possible)
567 - Add configure for gettimeofday.
568 - At line 51 of ua_input.c, why is = 0 necessary. Previously without
569 it, if cancel gnome-console during sql command, DIR crashed. However,
570 with it, blank line input for Where: is not possible.
571 - Make SD disallow writing on Volume with fewer files than in
573 - Label (asks for slot, return and it stops).
574 - Make SD reject writing on tape where Catalog and tape # files
575 don't agree (possibly OK if tape > catalog).
576 - Document all daemon tools MUST have a config file.
577 - Why does btape error when pointed to a file?
578 - Disallow compile if long long not 64 bits.
579 - Send Volumes needed during restore to Console (just after
580 create_volume_list) -- also in restore command?
581 - File system type from File daemon
582 - Move block size code from block.c to init_dev().
583 - Add FileSet MD5 to bscan.
584 - Finish implementation of restore "replace" options, and document.
586 - Document buffer size considerations with Sparse files --
587 - that restore prints volumes.
588 - Document that two Verifys at same time on same client do not work.
590 - MD5 in bscan is set in FileSet.
591 - that Bacula won't write on tape where tape/catalog files differ.
592 - Document how to recycle a tape in 7 days even if the backup takes a long time.
593 - Document tape cycling
594 - Implement VolumeUseDuration (maximum duration/period) a volume can be used).
595 Possibly VolumeUseTimes (MaximumVolumeJobs) This permits a volume to be
596 recycled even if it is not full.
597 - Decide what to do with JobTDate in catalog (make real utime_t?)
598 - Implement VolMaxJob and Duration check in next_volume().
599 - Make sure pruning of Jobs removes JobMediaId
600 - Write bcopy program -- recovery of bad tape.
601 - Make gethostbyname() thread safe in bnet.c
602 - Add ORDER BY JobId to list of Jobs in query.sql, and in
603 ua_output.c (list command).