9 - Document running multiple Jobs
10 - Document that two Verifys at same time on same client do not work.
11 - Document how to recycle a tape in 7 days even if the backup takes a long time.
12 - Document default config file locations.
13 - Document better includes (does it cross file systems ?).
14 - Document specifically how to add new File daemon to config files.
15 - Document forcing a new tape to be used.
16 - Document "Error in message.c:500 Mail program terminated in error.
19 --bindir is wrong and does not reflect prefix= in the *_sqlite_* scripts
21 --top level configure options are not passed to the depkgs, particularly
23 --Also, it might be better to split the depkgs location from the --with-sqlite
25 --should be able to specify e.g. --with-sqlite=/opt/local and have it find
26 lib, bin, sbin for itself
27 I tried this and it didn't find sqlite.h
28 --sd.conf password does not match dir.conf storage password
31 - Write updated bootstrap after every Job.
32 - Correct Warning: Volume name mismatch. Wanted test2 got test1
33 when reading a tape and want the next one.
34 - Create all pools when Director starts
35 - Implement autochanger for restore. ARRRGGG! I forgot!
36 - Make BSR accept count (total files to be restored).
37 - Make BSR return next_block when it knows record is not
38 in block, done when count is reached, and possibly other
39 optimizations. I.e. add a state word.
40 - After unmount, if restore job started, ask to mount.
41 - Fix db_get_fileset in cats/sql_get.c for multiple records.
42 - Fix start/end blocks for File
43 - Add new code to scheduler.c and run_conf.c
44 - Volume Bytes shows bytes on last volume written in Job summary.
45 - Fix catalog filename truncation in sql_get and sql_create. Use
46 only a single filename split routine.
47 - Add command to reset VolFiles to a larger value (don't allow
48 a smaller number or print big warning).
49 - Make SD disallow writing on Volume with fewer files than in
51 - Make Restore report an error if FD or SD term codes are not OK.
52 - Convert all %x substitution variables, which are hard to remember
53 and read to %(variable-name). Idea from TMDA.
54 - Report compression % and other compression statistics if turned on.
55 - Add JobLevel in FD status (but make sure it is defined).
56 - Make Pool resource handle Counter resources.
57 - Remove NextId for SQLite. Optimize.
58 - Fix gethostbyname() to use gethostbyname_r()
59 - Implement ./configure --with-client-only
60 - Strip trailing / from Include
61 - Move all SQL statements into a single location.
62 - Cleanup db_update_media and db_update_pool
63 - Add UA rc and history files.
64 - put termcap (used by console) in ./configure and
65 allow -with-termcap-dir.
66 - Remove JobMediaId it is not used.
67 - Enhance time and size scanning routines.
68 - Fix Autoprune for Volumes to respect need for full save.
69 - DateWritten field on tape may be wrong.
70 - Fix Win32 config file definition name on /install
71 - No READLINE_SRC if found in alternate directory.
72 - Add Client FS/OS id (Linux, Win95/98, ...).
73 - Put Windows files in Windows stream?
74 - Ensure that everyone uses btime routines.
78 Bacula Projects Roadmap
81 Item 1: Multiple simultaneous Jobs. (done)
83 What: Permit multiple simultaneous jobs in Bacula.
85 Why: An enterprise level solution needs to go fast without the
86 need for the system administrator to carefully tweak
87 timing. Based on the benchmarks, during a full
88 backup, NetWorker typically hit 10 times the bandwidth to
89 the tape compared to Bacula--largely. This is probably due to
90 running parallel jobs and multi-threaded filling of buffers
91 and writing them to tape. This should also make things work
92 better when you have a mix of fast and slow machines backing
95 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
96 implementing this is a matter of some small cleanups and
100 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
103 What: If data is coming into the SD too fast, buffer it to
104 disk if the user has configured this option.
106 Why: This would be nice, especially if it more or less falls out
107 when implementing (1) above. If not, it probably should not
108 be given a high priority because fundamentally the backup time
109 is limited by the tape bandwidth. Even though you may finish a
110 client job quicker by spilling to disk, you still have to
111 eventually get it onto tape. If intermediate disk buffering
112 allows us to improve write bandwidth to tape, it may make
115 Notes: Whether or not this is implemented will depend upon performance
116 testing after item 1 is implemented.
119 Item 3: Write the bscan program.
121 What: Write a program that reads a Bacula tape and puts all the
122 appropriate data into the catalog. This allows recovery
123 from a tape that is no longer in the database, or it allows
124 re-creation of a database if lost.
126 Why: This is a fundamental robustness and disaster recovery tool
127 which will increase the comfort level of a sysadmin
128 considering adopting Bacula.
130 Notes: A skeleton of this program already exists, but much work
131 needs to be done. Implementing this will also make apparent
132 any deficiencies in the current Bacula tape format.
135 Item 4: Implement Base jobs.
137 What: A base job is sort of like a Full save except that you
138 will want the FileSet to contain only files that are unlikely
139 to change in the future (i.e. a snapshot of most of your
140 system after installing it). After the base job has been run,
141 when you are doing a Full save, you can specify to exclude
142 all files saved by the base job that have not been modified.
144 Why: This is something none of the competition does, as far as we know
145 (except BackupPC, which is a Perl program that saves to disk
146 only). It is big win for the user, it makes Bacula stand out
147 as offering a unique optimization that immediately saves time
150 Notes: Big savings in tape usage. Will require more resources because
151 the e. DIR must send FD a list of files/attribs, and the FD must
152 search the list and compare it for each file to be saved.
155 Item 5: Implement Label templates
157 What: This is a mechanism whereby Bacula can automatically create
158 a tape label for new tapes according to a detailed specification
159 provided by the user.
161 Why: It is a major convenience item for folks who use automated label
164 Notes: Bacula already has a working form of automatic tape label
165 creation, but it is very crude. The design for the complete
166 tape labeling project is already documented in the manual.
169 Item 6: Write a regression script.
171 What: This is an automatic script that runs and tests as many features
172 of Bacula as possible. The output is compared to previous
173 versions of Bacula and any differences are reported.
175 Why: This is an enormous help in preventing introduction of new
176 errors in parts of the program that already work correctly.
178 Notes: This probably should be ranked higher, it's something the typical
179 user doesn't see. Depending on how it's implemented, it may
180 make sense to defer it until the archival tape format and
181 user interface mature.
184 Item 7: GUI for interactive restore
185 Item 8: GUI for interactive backup
187 What: The current interactive restore is implemented with a tty
188 interface. It would be much nicer to be able to "see" the
189 list of files backed up in typical GUI tree format.
190 The same mechanism could also be used for creating
191 ad-hoc backup FileSets (item 8).
193 Why: Ease of use -- especially for the end user.
195 Notes: Rather than implementing in Gtk, we probably should go directly
196 for a Browser implementation, even if doing so meant the
197 capability wouldn't be available until much later. Not only
198 is there the question of Windows sites, most
199 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
200 without installing lots of stuff admins are very wary about.
201 Real sysadmins will always use the command line anyway, and
202 the user who's doing an interactive restore or backup of his
203 own files will in most cases be on a Windows machine running
207 Item 9: Add SSL to daemon communications.
209 What: This provides for secure communications between the daemons.
211 Why: This would allow doing backup across the Internet without
212 privacy concerns (or with much less concern).
214 Notes: The vast majority of near term potential users will be backing up
215 a single site over a LAN and, correctly or not, they probably
216 won't be concerned with security, at least not enough to go to
217 the trouble to set up keys, etc. to screw things down. We suspect
218 that many users genuinely interested in multi-site backup
219 already run some form of VPN software in their internetwork
220 connections, and are willing to delegate security to that layer.
223 Item 10: Define definitive tape format.
225 What: Define that definitive tape format that will not change
226 for the next millennium.
228 Why: Stability, security.
230 Notes: See notes for item 11 below.
233 Item 11: New daemon communication protocol.
235 What: The current daemon to daemon protocol is basically an ASCII
236 printf() and sending the buffer. On the receiving end, the
237 buffer is sscanf()ed to unpack it. The new scheme would
238 be a binary format that allows quick packing and unpacking
239 of any data type with named fields.
241 Why: Using binary packing would be faster. Named fields will permit
242 error checking to ensure that what is sent is what the
243 receiver really wants.
245 Notes: These are internal improvements in the interest of the
246 long-term stability and evolution of the program. On the one
247 hand, the sooner they're done, the less code we have to rip
248 up when the time comes to install them. On the other hand, they
249 don't bring an immediately perceptible benefit to potential
250 users. Item 10 and possibly item 11 should be deferred until Bacula
251 is well established with a growing user community more or
252 less happy with the feature set. At that time, it will make a
253 good "next generation" upgrade in the interest of data
258 I haven't put these in any particular order.
261 - Rework Storage daemon with new rwl_lock routines.
262 - Compare tape to Client files (attributes, or attributes and data)
263 - Restore options (overwrite, overwrite if older,
264 overwrite if newer, never overwrite, ...)
265 - Restore to a particular time -- e.g. before date, after date.
266 - On command write out a bootstrap file (at end of job).
267 - Make all database Ids 64 bit.
268 - Pass JCR to database routines permitting better error printing.
269 - Make bls accept bootstrap record.
270 - Write an applet for Linux.
271 - Make SD reject writing on tape where Catalog and tape # files
272 don't agree (possibly OK if tape > catalog).
273 - Implement new daemon communications protocol.
274 - Add DIR config directive to spool attributes.
275 - Pass DIR config variable to SD for no attributes.
276 - Create JobMedia record for all running Jobs when Media changes.
277 - Send Volumes needed during restore to Console (just after
278 create_volume_list) -- also in restore command?
279 - Add estimate to Console commands
280 - Find solution to blank filename (i.e. path only) problem.
283 mysqldump -f --opt bacula >bacula
287 - Remove PoolId from Job table, it exists in Media.
288 - Allow console commands to detach or run in background.
289 - Fix status delay on storage daemon during rewind.
290 - Add VerNo to each Session label record.
291 - Add Job to Session records.
292 - Add VOLUME_CAT_INFO to the EOS tape record (as
293 well as to the EOD record).
294 - Add SD message variables to control operator wait time
295 - Maximum Operator Wait
296 - Minimum Message Interval
297 - Maximum Message Interval
298 - Add EOM handling variables
300 - Require EOD records
301 - Send Operator message when cannot read tape label.
302 - Think about how to handle I/O error on MTEOM.
303 - If Storage daemon aborts a job, ensure that this
304 is printed in the error message.
305 - Verify level=Volume (scan only), level=Data (compare of data to file).
306 Verify level=Catalog, level=InitCatalog
308 - Cold start full restore (restore catalog then
309 user selects what to restore). Write summary file containing only
310 Job, Media, and Catalog information. Store on another machine.
311 - Dump/Restore database
314 - Add keyword search to show command in Console.
315 - Fix Win2000 error with no messages during startup.
316 - Events : tape has more than xxx bytes.
317 - In Storage daemon, status should include job cancelled.
318 - Write general list maintenance subroutines.
319 - Implement immortal format with EDOs.
320 - Restrict characters permitted in a Resource name.
321 - Provide definitive identification of type in backup.
322 - Complete code in Bacula Resources -- this will permit
323 reading a new config file at any time.
324 - Document new Console
325 - Handle ctl-c in Console
326 - Test restore of Windows backup
327 - Implement LabelTemplate (at least first cut).
328 - Implement script driven addition of File daemon to
331 - Bug: anonymous Volumes requires mount in some cases.
332 - see setgroup and user for Bacula p4-5 of stunnel.c
333 - Implement new serialize subroutines
334 send(socket, "string", &Vol, "uint32", &i, NULL)
335 - Add save type to Session label.
336 - Correct date on Session label.
337 - On I/O error, write EOF, then try to write again.
338 - Audit all UA commands to ensure that we always prompt where
340 - If ./btape is called without /dev, assume argument is
341 a Storage resource name.
342 - Put memory utilization in Status output of each daemon
343 if full status requested or if some level of debug on.
344 - Make database type selectable by .conf files i.e. at runtime
345 - gethostbyname failure in bnet_connect() continues
346 generating errors -- should stop.
347 - Don't create a volume that is already written. I.e. create only once.
348 - If error at end of tape, implement some way to kill waiting processes.
349 - Add HOST to Volume label.
350 - Set flag for uname -a. Add to Volume label.
351 - Implement throttled work queue.
352 - Write bscan program that will syncronize the DB Media record with
353 the contents of the Volume -- for use after a crash.
354 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
355 - Allow multiple Storage specifications (or multiple names on
356 a single Storage specification) in the Job record. Thus a job
357 can be backed up to a number of storage devices.
358 - Implement full MediaLabel code.
359 - Implement dump label to UA
360 - Copy volume using single drive.
361 - Copy volume with multiple driven (same or different block size).
362 - Add block size (min, max) to Vol label.
363 - Concept of VolumeSet during restore which is a list
364 of Volume names needed.
365 - Restore files modified after date
366 - Restore file modified before date
367 - Emergency restore info:
369 - Backup working directory
371 - Restore options (do not overwrite)
372 - Restore -- do nothing but show what would happen
373 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
374 - Put Job statistics in End Session Label (files saved,
375 total bytes, start time, ...).
376 - Put FileSet name in the SOS label.
377 - Implement Restore FileSet=
378 - Write a scanner for the UA (keyword, scan-routine, result, prompt).
379 - Create a protocol.h and protocol.c where all protocol messages
381 - If SD cannot open a drive, make it periodically retry.
382 - Put Bacula version somewhere in Job stream, probably Start Session
384 - Remove duplicate fields from jcr (e.g. jcr.level and
386 - Timout a job or terminate if link goes down, or reopen link and query.
387 - Define how we handle times to avoid problem with Unix dates (2049 ?).
388 - Fill all fields in Vol/Job Header -- ensure that everything
389 needed is written to tape. Think about restore to Catalog
390 from tape. Client record needs improving.
391 - Find general solution for sscanf size problems (as well
392 as sprintf. Do at run time?
394 - Concept of precious tapes (cannot be reused).
395 - Allow FD to run from inetd ???
396 - Preprocessing command per file.
397 - Postprocessing command per file (when restoring).
399 - Restore should get Device and Pool information from
400 job record rather than from config.
401 - Autolabel should be specified by DR instead of SD.
402 - Ability to recreate the catalog from a tape.
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 - Eliminate duplicate File records to shrink database.
434 - FD sends unsaved file list to Director at end of job.
435 - Write a Storage daemon that uses pipes and
436 standard Unix programs to write to the tape.
438 - Need something that monitors the JCR queue and
439 times out jobs by asking the deamons where they are.
440 - Add daemon JCR JobId=0 to have a daemon context
444 - Client (list of clients to force client)
445 - Devices (list of devices to force device)
450 - Recycle from other pools: Yes, No
451 - Recycle to other pools: Yes, no
454 - Optional MediaType to force media?
465 if(connect(sockfd, (struct sockaddr * ) (& addr), sizeof(addr)) .lt. 0){
471 linger.l_linger = 60;
472 i = setsockopt(sockfd, SOL_SOCKET, SO_LINGER, (char *) &linger,
475 fl = fcntl(sockfd, F_GETFL);
476 fcntl(sockfd, F_SETFL, fl & (~ O_NONBLOCK) & (~ O_NDELAY));
478 - Enhance Jmsg code to permit buffering and saving to disk.
479 - device driver = "xxxx" for drives.
480 - restart: paranoid: read label fsf to
481 eom read append block, and go
482 super-paranoid: read label, read all files
483 in between, read append block, and go
484 verify: backspace, read append block, and go
485 permissive: same as above but frees drive
486 if tape is not valid.
488 - Ensure that /dev/null works
489 - File daemon should build list of files skipped, and then
490 at end of save retry and report any errors.
491 - Need report class for messages. Perhaps
492 report resource where report=group of messages
494 - enhance scan_attrib and rename scan_jobtype, and
495 fill in code for "since" option
496 - To buffer messages, we need associated jobid and Director name.
497 - Need to save contents of FileSet to tape?
498 - Director needs a time after which the report status is sent
499 anyway -- or better yet, a retry time for the job.
500 Don't reschedule a job if previous incarnation is still running.
501 - Figure out how to save the catalog (possibly a special FileSet).
502 - Figure out how to restore the catalog.
503 - Figure out how to put a Volume into the catalog (from the tape)
504 - Figure out how to do a restore from a Volume
505 - Some way to automatically backup everything is needed????
506 - Need a structure for pending actions:
508 - termination status (part of buffered msgs?)
509 - Concept of grouping Storage devices and job can use
510 any of a number of devices
512 Read, Write, Clean, Delete
513 - Login to Bacula; Bacula users with different permissions:
515 - Tape recycle destination
516 - Job Schedule Status
520 - File daemon should pass Director the operating system info
521 to be stored in the Client Record (or verified that it has
523 - Store info on each file system type (probably in the job header on tape.
524 This could be the output of df; or perhaps some sort of /etc/mtab record.
527 - Use media 1 time (so that we can do 6 days of incremental
528 backups before switching to another tape) (already)
529 specify # times (jobs)
530 specify bytes (already)
531 specify time (seconds, hours, days)
532 - Implement FSM (File System Modules).
533 - Identify unchanged or "system" files and save them to a
534 special tape thus removing them from the standard
535 backup FileSet -- BASE backup.
536 - Turn virutally all sprintfs into snprintfs.
537 - Heartbeat between daemons.
538 - Audit M_ error codes to ensure they are correct and consistent.
539 - Add variable break characters to lex analyzer.
540 Either a bit mask or a string of chars so that
541 the caller can change the break characters.
542 - Make a single T_BREAK to replace T_COMMA, etc.
543 - Ensure that File daemon and Storage daemon can
544 continue a save if the Director goes down (this
545 is NOT currently the case). Must detect socket error,
546 buffer messages for later.
549 Done: (see kernsdone for more)
550 --the console script is broken as installed and has to be hand-massaged with
551 paths, config files etc.
552 - Termination status in FD for Verify = C -- incorrect.
553 - Implement alter_sqlite_tables
554 - Fix scheduler -- see "Hourly cycle". It doesn't do both each
555 hour, rather it alternates between 0:05 and 0:35.
556 - Create Counter DB records.
557 - Fix db_get_job_volume_names() to return array of strings (now works
559 - Eliminate MySQL shared libraries from smtp and daemons not using MySQL.
560 - Compare tape File attributes to Catalog.
561 (File attributes are size, dates, MD5, but not
563 - Report bad status from smtp or mail program.
564 - Ensure that Start/End File/Block are correct.
565 - If MySQL database is not running, job terminates with
566 wierd type and wierd error code.
567 - Probably create a jcr with JobId=0 as a master
568 catchall if jcr not found or if operation involves
570 - The daemons should know when one is already
571 running and refuse to run a second copy.
572 - Figure out how to do a "full" restore from catalog
573 - Make SD send attribute stream to DR but first
574 buffering to file, then sending only when the
575 files are written to tape.
576 - Restore file xx or files xx, yy to their most recent values.
577 - Get correct block/file information in Catalog, pay attention to change of media.
578 - Write better dump of Messages resource.
579 - Authentication between SD and FD
580 - In restore job, print some summary information at end, such
581 as rate, ... job status, ...
582 - Problem with len at 362 in tree.c
583 - Report volume write rate.
584 - Pass "Catalog Files = no" to storage daemon to eliminate network traffic.
585 - When we are at EOM, we must ask each job to write JobMedia
586 record (update_volume_info).