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