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