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 - Make BSR accept count (total files to be restored).
32 - Make BSR return next_block when it knows record is not
33 in block, done when count is reached, and possibly other
34 optimizations. I.e. add a state word.
35 - After unmount, if restore job started, ask to mount.
36 - Fix db_get_fileset in cats/sql_get.c for multiple records.
37 - Fix start/end blocks for File
38 - Add new code to scheduler.c and run_conf.c
39 - Volume Bytes shows bytes on last volume written in Job summary.
40 - Fix catalog filename truncation in sql_get and sql_create. Use
41 only a single filename split routine.
42 - Add command to reset VolFiles to a larger value (don't allow
43 a smaller number or print big warning).
44 - Make SD disallow writing on Volume with fewer files than in
46 - Make Restore report an error if FD or SD term codes are not OK.
47 - Convert all %x substitution variables, which are hard to remember
48 and read to %(variable-name). Idea from TMDA.
49 - Report compression % and other compression statistics if turned on.
50 - Add JobLevel in FD status (but make sure it is defined).
51 - Make Pool resource handle Counter resources.
52 - Remove NextId for SQLite. Optimize.
53 - Fix gethostbyname() to use gethostbyname_r()
54 - Implement ./configure --with-client-only
55 - Strip trailing / from Include
56 - Move all SQL statements into a single location.
57 - Cleanup db_update_media and db_update_pool
58 - Add UA rc and history files.
59 - put termcap (used by console) in ./configure and
60 allow -with-termcap-dir.
61 - Remove JobMediaId it is not used.
62 - Enhance time and size scanning routines.
63 - Fix Autoprune for Volumes to respect need for full save.
64 - DateWritten field on tape may be wrong.
65 - Fix Win32 config file definition name on /install
66 - No READLINE_SRC if found in alternate directory.
67 - Add Client FS/OS id (Linux, Win95/98, ...).
68 - Put Windows files in Windows stream?
69 - Ensure that everyone uses btime routines.
73 Bacula Projects Roadmap
76 Item 1: Multiple simultaneous Jobs. (done)
78 What: Permit multiple simultaneous jobs in Bacula.
80 Why: An enterprise level solution needs to go fast without the
81 need for the system administrator to carefully tweak
82 timing. Based on the benchmarks, during a full
83 backup, NetWorker typically hit 10 times the bandwidth to
84 the tape compared to Bacula--largely. This is probably due to
85 running parallel jobs and multi-threaded filling of buffers
86 and writing them to tape. This should also make things work
87 better when you have a mix of fast and slow machines backing
90 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
91 implementing this is a matter of some small cleanups and
95 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
98 What: If data is coming into the SD too fast, buffer it to
99 disk if the user has configured this option.
101 Why: This would be nice, especially if it more or less falls out
102 when implementing (1) above. If not, it probably should not
103 be given a high priority because fundamentally the backup time
104 is limited by the tape bandwidth. Even though you may finish a
105 client job quicker by spilling to disk, you still have to
106 eventually get it onto tape. If intermediate disk buffering
107 allows us to improve write bandwidth to tape, it may make
110 Notes: Whether or not this is implemented will depend upon performance
111 testing after item 1 is implemented.
114 Item 3: Write the bscan program.
116 What: Write a program that reads a Bacula tape and puts all the
117 appropriate data into the catalog. This allows recovery
118 from a tape that is no longer in the database, or it allows
119 re-creation of a database if lost.
121 Why: This is a fundamental robustness and disaster recovery tool
122 which will increase the comfort level of a sysadmin
123 considering adopting Bacula.
125 Notes: A skeleton of this program already exists, but much work
126 needs to be done. Implementing this will also make apparent
127 any deficiencies in the current Bacula tape format.
130 Item 4: Implement Base jobs.
132 What: A base job is sort of like a Full save except that you
133 will want the FileSet to contain only files that are unlikely
134 to change in the future (i.e. a snapshot of most of your
135 system after installing it). After the base job has been run,
136 when you are doing a Full save, you can specify to exclude
137 all files saved by the base job that have not been modified.
139 Why: This is something none of the competition does, as far as we know
140 (except BackupPC, which is a Perl program that saves to disk
141 only). It is big win for the user, it makes Bacula stand out
142 as offering a unique optimization that immediately saves time
145 Notes: Big savings in tape usage. Will require more resources because
146 the e. DIR must send FD a list of files/attribs, and the FD must
147 search the list and compare it for each file to be saved.
150 Item 5: Implement Label templates
152 What: This is a mechanism whereby Bacula can automatically create
153 a tape label for new tapes according to a detailed specification
154 provided by the user.
156 Why: It is a major convenience item for folks who use automated label
159 Notes: Bacula already has a working form of automatic tape label
160 creation, but it is very crude. The design for the complete
161 tape labeling project is already documented in the manual.
164 Item 6: Write a regression script.
166 What: This is an automatic script that runs and tests as many features
167 of Bacula as possible. The output is compared to previous
168 versions of Bacula and any differences are reported.
170 Why: This is an enormous help in preventing introduction of new
171 errors in parts of the program that already work correctly.
173 Notes: This probably should be ranked higher, it's something the typical
174 user doesn't see. Depending on how it's implemented, it may
175 make sense to defer it until the archival tape format and
176 user interface mature.
179 Item 7: GUI for interactive restore
180 Item 8: GUI for interactive backup
182 What: The current interactive restore is implemented with a tty
183 interface. It would be much nicer to be able to "see" the
184 list of files backed up in typical GUI tree format.
185 The same mechanism could also be used for creating
186 ad-hoc backup FileSets (item 8).
188 Why: Ease of use -- especially for the end user.
190 Notes: Rather than implementing in Gtk, we probably should go directly
191 for a Browser implementation, even if doing so meant the
192 capability wouldn't be available until much later. Not only
193 is there the question of Windows sites, most
194 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
195 without installing lots of stuff admins are very wary about.
196 Real sysadmins will always use the command line anyway, and
197 the user who's doing an interactive restore or backup of his
198 own files will in most cases be on a Windows machine running
202 Item 9: Add SSL to daemon communications.
204 What: This provides for secure communications between the daemons.
206 Why: This would allow doing backup across the Internet without
207 privacy concerns (or with much less concern).
209 Notes: The vast majority of near term potential users will be backing up
210 a single site over a LAN and, correctly or not, they probably
211 won't be concerned with security, at least not enough to go to
212 the trouble to set up keys, etc. to screw things down. We suspect
213 that many users genuinely interested in multi-site backup
214 already run some form of VPN software in their internetwork
215 connections, and are willing to delegate security to that layer.
218 Item 10: Define definitive tape format.
220 What: Define that definitive tape format that will not change
221 for the next millennium.
223 Why: Stability, security.
225 Notes: See notes for item 11 below.
228 Item 11: New daemon communication protocol.
230 What: The current daemon to daemon protocol is basically an ASCII
231 printf() and sending the buffer. On the receiving end, the
232 buffer is sscanf()ed to unpack it. The new scheme would
233 be a binary format that allows quick packing and unpacking
234 of any data type with named fields.
236 Why: Using binary packing would be faster. Named fields will permit
237 error checking to ensure that what is sent is what the
238 receiver really wants.
240 Notes: These are internal improvements in the interest of the
241 long-term stability and evolution of the program. On the one
242 hand, the sooner they're done, the less code we have to rip
243 up when the time comes to install them. On the other hand, they
244 don't bring an immediately perceptible benefit to potential
245 users. Item 10 and possibly item 11 should be deferred until Bacula
246 is well established with a growing user community more or
247 less happy with the feature set. At that time, it will make a
248 good "next generation" upgrade in the interest of data
253 I haven't put these in any particular order.
256 - Rework Storage daemon with new rwl_lock routines.
257 - Compare tape to Client files (attributes, or attributes and data)
258 - Restore options (overwrite, overwrite if older,
259 overwrite if newer, never overwrite, ...)
260 - Restore to a particular time -- e.g. before date, after date.
261 - On command write out a bootstrap file (at end of job).
262 - Make all database Ids 64 bit.
263 - Pass JCR to database routines permitting better error printing.
264 - Make bls accept bootstrap record.
265 - Write an applet for Linux.
266 - Make SD reject writing on tape where Catalog and tape # files
267 don't agree (possibly OK if tape > catalog).
268 - Implement new daemon communications protocol.
269 - Add DIR config directive to spool attributes.
270 - Pass DIR config variable to SD for no attributes.
271 - Create JobMedia record for all running Jobs when Media changes.
272 - Send Volumes needed during restore to Console (just after
273 create_volume_list) -- also in restore command?
274 - Add estimate to Console commands
275 - Find solution to blank filename (i.e. path only) problem.
278 mysqldump -f --opt bacula >bacula
282 - Remove PoolId from Job table, it exists in Media.
283 - Allow console commands to detach or run in background.
284 - Fix status delay on storage daemon during rewind.
285 - Add VerNo to each Session label record.
286 - Add Job to Session records.
287 - Add VOLUME_CAT_INFO to the EOS tape record (as
288 well as to the EOD record).
289 - Add SD message variables to control operator wait time
290 - Maximum Operator Wait
291 - Minimum Message Interval
292 - Maximum Message Interval
293 - Add EOM handling variables
295 - Require EOD records
296 - Send Operator message when cannot read tape label.
297 - Think about how to handle I/O error on MTEOM.
298 - If Storage daemon aborts a job, ensure that this
299 is printed in the error message.
300 - Verify level=Volume (scan only), level=Data (compare of data to file).
301 Verify level=Catalog, level=InitCatalog
303 - Cold start full restore (restore catalog then
304 user selects what to restore). Write summary file containing only
305 Job, Media, and Catalog information. Store on another machine.
306 - Dump/Restore database
309 - Add keyword search to show command in Console.
310 - Fix Win2000 error with no messages during startup.
311 - Events : tape has more than xxx bytes.
312 - In Storage daemon, status should include job cancelled.
313 - Write general list maintenance subroutines.
314 - Implement immortal format with EDOs.
315 - Restrict characters permitted in a Resource name.
316 - Provide definitive identification of type in backup.
317 - Complete code in Bacula Resources -- this will permit
318 reading a new config file at any time.
319 - Document new Console
320 - Handle ctl-c in Console
321 - Test restore of Windows backup
322 - Implement LabelTemplate (at least first cut).
323 - Implement script driven addition of File daemon to
326 - Bug: anonymous Volumes requires mount in some cases.
327 - see setgroup and user for Bacula p4-5 of stunnel.c
328 - Implement new serialize subroutines
329 send(socket, "string", &Vol, "uint32", &i, NULL)
330 - Add save type to Session label.
331 - Correct date on Session label.
332 - On I/O error, write EOF, then try to write again.
333 - Audit all UA commands to ensure that we always prompt where
335 - If ./btape is called without /dev, assume argument is
336 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 - Don't create a volume that is already written. I.e. create only once.
343 - If error at end of tape, implement some way to kill waiting processes.
344 - Add HOST to Volume label.
345 - Set flag for uname -a. Add to Volume label.
346 - Implement throttled work queue.
347 - Write bscan program that will syncronize the DB Media record with
348 the contents of the Volume -- for use after a crash.
349 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
350 - Allow multiple Storage specifications (or multiple names on
351 a single Storage specification) in the Job record. Thus a job
352 can be backed up to a number of storage devices.
353 - Implement full MediaLabel code.
354 - Implement dump label to UA
355 - Copy volume using single drive.
356 - Copy volume with multiple driven (same or different block size).
357 - Add block size (min, max) to Vol label.
358 - Concept of VolumeSet during restore which is a list
359 of Volume names needed.
360 - Restore files modified after date
361 - Restore file modified before date
362 - Emergency restore info:
364 - Backup working directory
366 - Restore options (do not overwrite)
367 - Restore -- do nothing but show what would happen
368 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
369 - Put Job statistics in End Session Label (files saved,
370 total bytes, start time, ...).
371 - Put FileSet name in the SOS label.
372 - Implement Restore FileSet=
373 - Write a scanner for the UA (keyword, scan-routine, result, prompt).
374 - Create a protocol.h and protocol.c where all protocol messages
376 - If SD cannot open a drive, make it periodically retry.
377 - Put Bacula version somewhere in Job stream, probably Start Session
379 - Remove duplicate fields from jcr (e.g. jcr.level and
381 - Timout a job or terminate if link goes down, or reopen link and query.
382 - Define how we handle times to avoid problem with Unix dates (2049 ?).
383 - Fill all fields in Vol/Job Header -- ensure that everything
384 needed is written to tape. Think about restore to Catalog
385 from tape. Client record needs improving.
386 - Find general solution for sscanf size problems (as well
387 as sprintf. Do at run time?
389 - Concept of precious tapes (cannot be reused).
390 - Allow FD to run from inetd ???
391 - Preprocessing command per file.
392 - Postprocessing command per file (when restoring).
394 - Restore should get Device and Pool information from
395 job record rather than from config.
396 - Autolabel should be specified by DR instead of SD.
397 - Ability to recreate the catalog from a tape.
398 - Find out how to get the system tape block limits, e.g.:
399 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
400 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
403 - AutoScan (check checksum of tape)
404 - Format command = "format /dev/nst0"
408 - Seek resolution (usually corresponds to buffer size)
409 - EODErrorCode=ENOSPC or code
410 - Partial Read error code
411 - Partial write error code
412 - Nonformatted read error
413 - Nonformatted write error
414 - WriteProtected error
418 - IgnoreCloseErrors=yes
428 - Eliminate duplicate File records to shrink database.
429 - FD sends unsaved file list to Director at end of job.
430 - Write a Storage daemon that uses pipes and
431 standard Unix programs to write to the tape.
433 - Need something that monitors the JCR queue and
434 times out jobs by asking the deamons where they are.
435 - Add daemon JCR JobId=0 to have a daemon context
439 - Client (list of clients to force client)
440 - Devices (list of devices to force device)
445 - Recycle from other pools: Yes, No
446 - Recycle to other pools: Yes, no
449 - Optional MediaType to force media?
460 if(connect(sockfd, (struct sockaddr * ) (& addr), sizeof(addr)) .lt. 0){
466 linger.l_linger = 60;
467 i = setsockopt(sockfd, SOL_SOCKET, SO_LINGER, (char *) &linger,
470 fl = fcntl(sockfd, F_GETFL);
471 fcntl(sockfd, F_SETFL, fl & (~ O_NONBLOCK) & (~ O_NDELAY));
473 - Enhance Jmsg code to permit buffering and saving to disk.
474 - device driver = "xxxx" for drives.
475 - restart: paranoid: read label fsf to
476 eom read append block, and go
477 super-paranoid: read label, read all files
478 in between, read append block, and go
479 verify: backspace, read append block, and go
480 permissive: same as above but frees drive
481 if tape is not valid.
483 - Ensure that /dev/null works
484 - File daemon should build list of files skipped, and then
485 at end of save retry and report any errors.
486 - Need report class for messages. Perhaps
487 report resource where report=group of messages
489 - enhance scan_attrib and rename scan_jobtype, and
490 fill in code for "since" option
491 - To buffer messages, we need associated jobid and Director name.
492 - Need to save contents of FileSet to tape?
493 - Director needs a time after which the report status is sent
494 anyway -- or better yet, a retry time for the job.
495 Don't reschedule a job if previous incarnation is still running.
496 - Figure out how to save the catalog (possibly a special FileSet).
497 - Figure out how to restore the catalog.
498 - Figure out how to put a Volume into the catalog (from the tape)
499 - Figure out how to do a restore from a Volume
500 - Some way to automatically backup everything is needed????
501 - Need a structure for pending actions:
503 - termination status (part of buffered msgs?)
504 - Concept of grouping Storage devices and job can use
505 any of a number of devices
507 Read, Write, Clean, Delete
508 - Login to Bacula; Bacula users with different permissions:
510 - Tape recycle destination
511 - Job Schedule Status
515 - File daemon should pass Director the operating system info
516 to be stored in the Client Record (or verified that it has
518 - Store info on each file system type (probably in the job header on tape.
519 This could be the output of df; or perhaps some sort of /etc/mtab record.
522 - Use media 1 time (so that we can do 6 days of incremental
523 backups before switching to another tape) (already)
524 specify # times (jobs)
525 specify bytes (already)
526 specify time (seconds, hours, days)
527 - Implement FSM (File System Modules).
528 - Identify unchanged or "system" files and save them to a
529 special tape thus removing them from the standard
530 backup FileSet -- BASE backup.
531 - Turn virutally all sprintfs into snprintfs.
532 - Heartbeat between daemons.
533 - Audit M_ error codes to ensure they are correct and consistent.
534 - Add variable break characters to lex analyzer.
535 Either a bit mask or a string of chars so that
536 the caller can change the break characters.
537 - Make a single T_BREAK to replace T_COMMA, etc.
538 - Ensure that File daemon and Storage daemon can
539 continue a save if the Director goes down (this
540 is NOT currently the case). Must detect socket error,
541 buffer messages for later.
544 Done: (see kernsdone for more)
545 --the console script is broken as installed and has to be hand-massaged with
546 paths, config files etc.
547 - Termination status in FD for Verify = C -- incorrect.
548 - Implement alter_sqlite_tables
549 - Fix scheduler -- see "Hourly cycle". It doesn't do both each
550 hour, rather it alternates between 0:05 and 0:35.
551 - Create Counter DB records.
552 - Fix db_get_job_volume_names() to return array of strings (now works
554 - Eliminate MySQL shared libraries from smtp and daemons not using MySQL.
555 - Compare tape File attributes to Catalog.
556 (File attributes are size, dates, MD5, but not
558 - Report bad status from smtp or mail program.
559 - Ensure that Start/End File/Block are correct.
560 - If MySQL database is not running, job terminates with
561 wierd type and wierd error code.
562 - Probably create a jcr with JobId=0 as a master
563 catchall if jcr not found or if operation involves
565 - The daemons should know when one is already
566 running and refuse to run a second copy.
567 - Figure out how to do a "full" restore from catalog
568 - Make SD send attribute stream to DR but first
569 buffering to file, then sending only when the
570 files are written to tape.
571 - Restore file xx or files xx, yy to their most recent values.
572 - Get correct block/file information in Catalog, pay attention to change of media.
573 - Write better dump of Messages resource.
574 - Authentication between SD and FD
575 - In restore job, print some summary information at end, such
576 as rate, ... job status, ...
577 - Problem with len at 362 in tree.c
578 - Report volume write rate.
579 - Pass "Catalog Files = no" to storage daemon to eliminate network traffic.
580 - When we are at EOM, we must ask each job to write JobMedia
581 record (update_volume_info).