4 Documentation to do: (a little bit at a time)
5 - Document running a test version.
6 - Make sure restore options are documented
7 - Document query file format.
9 Testing to do: (painful)
10 - that restore options work in FD.
11 - that console command line options work
12 - blocksize recognition code.
17 - Make BSR accept count (total files to be restored).
18 - Make BSR return next_block when it knows record is not
19 in block, done when count is reached, and possibly other
20 optimizations. I.e. add a state word.
21 - Continue improving the restore process (handling
22 of tapes, efficiency improvements e.g. use FSF to
23 position the tape, ...)
24 - Add code to fast seek to proper place on tape/file
25 when doing Restore. If it doesn't work, try linear
27 - Add code to reject whole blocks if not wanted on restore.
29 - Figure out how to allow multiple simultaneous file Volumes on
31 - Start working on Base jobs.
32 - Implement FileOptions (see end of this document)
33 - Replace popen() and pclose() -- fail safe and timeout, no SIG dep.
34 - Ensure that restore of differential jobs works (check SQL).
35 - Make sure the MaxVolFiles is fully implemented in SD
36 - Flush all the daemon messages at the end of every job.
37 - Check if both CatalogFiles and UseCatalog are set to SD.
38 - Check if we can bump Bacula FD priorty in Win2000
39 - Make bcopy read through bad tape records.
40 - Need return status on read_cb() from read_records(). Need multiple
41 records -- one per Job, maybe a JCR or some other structure with
43 - Think about how to make Bacula work better with File archives.
45 - Work more on how to to a Bacula restore beginning with
46 just a Bacula tape and a boot floppy (bare metal recovery).
47 - Try bare metal Windows restore
48 - Fix read_record to handle multiple sessions.
49 - Program files (i.e. execute a program to read/write files).
50 Pass read date of last backup, size of file last time.
51 - Put system type returned by FD into catalog.
52 - Possibly add email to Watchdog if drive is unmounted too
53 long and a job is waiting on the drive.
54 - Strip trailing slashes from Include directory names in the FD.
55 - Use read_record.c in SD code.
56 - Why don't we get an error message from Win32 FD when bootstrap
57 file cannot be created for restore command?
58 - When Marking a file in Restore that is a hard link, also
59 mark the link so that the data will be reloaded.
60 - Restore program that errors in SD due to no tape reports
61 OK incorrectly in output.
62 - After unmount, if restore job started, ask to mount.
63 - Fix db_get_fileset in cats/sql_get.c for multiple records.
64 - Fix catalog filename truncation in sql_get and sql_create. Use
65 only a single filename split routine.
66 - Make Restore report an error if FD or SD term codes are not OK.
67 - Convert all %x substitution variables, which are hard to remember
68 and read to %(variable-name). Idea from TMDA.
69 - Add JobLevel in FD status (but make sure it is defined).
70 - Make Pool resource handle Counter resources.
71 - Remove NextId for SQLite. Optimize.
72 - Fix gethostbyname() to use gethostbyname_r()
73 - Implement ./configure --with-client-only
74 - Strip trailing / from Include
75 - Move all SQL statements into a single location.
76 - Cleanup db_update_media and db_update_pool
77 - Add UA rc and history files.
78 - put termcap (used by console) in ./configure and
79 allow -with-termcap-dir.
80 - Enhance time and size scanning routines.
81 - Fix Autoprune for Volumes to respect need for full save.
82 - DateWritten field on tape may be wrong.
83 - Fix Win32 config file definition name on /install
84 - No READLINE_SRC if found in alternate directory.
85 - Add Client FS/OS id (Linux, Win95/98, ...).
86 - Test a second language e.g. french.
87 - Compare tape to Client files (attributes, or attributes and data)
88 - Restore options (overwrite, overwrite if older,
89 overwrite if newer, never overwrite, ...)
90 - Restore to a particular time -- e.g. before date, after date.
91 - Make all database Ids 64 bit.
92 - Write an applet for Linux.
93 - Add estimate to Console commands
94 - Find solution to blank filename (i.e. path only) problem.
95 - Implement new daemon communications protocol.
96 - Remove PoolId from Job table, it exists in Media.
97 - Allow console commands to detach or run in background.
98 - Fix status delay on storage daemon during rewind.
99 - Add SD message variables to control operator wait time
100 - Maximum Operator Wait
101 - Minimum Message Interval
102 - Maximum Message Interval
103 - Send Operator message when cannot read tape label.
104 - Verify level=Volume (scan only), level=Data (compare of data to file).
105 Verify level=Catalog, level=InitCatalog
107 - Add keyword search to show command in Console.
108 - Fix Win2000 error with no messages during startup.
109 - Events : tape has more than xxx bytes.
110 - Restrict characters permitted in a Resource name.
111 - Complete code in Bacula Resources -- this will permit
112 reading a new config file at any time.
113 - Handle ctl-c in Console
114 - Implement LabelTemplate (at least first cut).
115 - Implement script driven addition of File daemon to config files.
117 - see setgroup and user for Bacula p4-5 of stunnel.c
118 - Implement new serialize subroutines
119 send(socket, "string", &Vol, "uint32", &i, NULL)
120 - Audit all UA commands to ensure that we always prompt where possible.
121 - If ./btape is called without /dev, assume argument is a Storage resource name.
122 - Put memory utilization in Status output of each daemon
123 if full status requested or if some level of debug on.
124 - Make database type selectable by .conf files i.e. at runtime
125 - gethostbyname failure in bnet_connect() continues
126 generating errors -- should stop.
127 - Add HOST to Volume label.
128 - Set flag for uname -a. Add to Volume label.
129 - Implement throttled work queue.
130 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
131 - Allow multiple Storage specifications (or multiple names on
132 a single Storage specification) in the Job record. Thus a job
133 can be backed up to a number of storage devices.
134 - Implement dump label to UA
135 - Copy volume using single drive.
136 - Concept of VolumeSet during restore which is a list
137 of Volume names needed.
138 - Restore files modified after date
139 - Restore file modified before date
140 - Emergency restore info:
142 - Backup working directory
144 - Restore -- do nothing but show what would happen
145 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
146 - Implement Restore FileSet=
147 - Create a protocol.h and protocol.c where all protocol messages
149 - If SD cannot open a drive, make it periodically retry.
150 - Remove duplicate fields from jcr (e.g. jcr.level and jcr.jr.Level, ...).
151 - Timout a job or terminate if link goes down, or reopen link and query.
152 - Fill all fields in Vol/Job Header -- ensure that everything
153 needed is written to tape. Think about restore to Catalog
154 from tape. Client record needs improving.
155 - Find general solution for sscanf size problems (as well
156 as sprintf. Do at run time?
157 - Concept of precious tapes (cannot be reused).
158 - Make bcopy copy with a single tape drive.
159 - Permit changing ownership during restore.
161 - Restore should get Device and Pool information from
162 job record rather than from config.
163 - Autolabel should be specified by DR instead of SD.
164 - Find out how to get the system tape block limits, e.g.:
165 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
166 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
169 - AutoScan (check checksum of tape)
170 - Format command = "format /dev/nst0"
174 - Seek resolution (usually corresponds to buffer size)
175 - EODErrorCode=ENOSPC or code
176 - Partial Read error code
177 - Partial write error code
178 - Nonformatted read error
179 - Nonformatted write error
180 - WriteProtected error
184 - IgnoreCloseErrors=yes
194 - FD sends unsaved file list to Director at end of job.
195 - Write a Storage daemon that uses pipes and
196 standard Unix programs to write to the tape.
198 - Need something that monitors the JCR queue and
199 times out jobs by asking the deamons where they are.
201 - Enhance Jmsg code to permit buffering and saving to disk.
202 - device driver = "xxxx" for drives.
203 - restart: paranoid: read label fsf to
204 eom read append block, and go
205 super-paranoid: read label, read all files
206 in between, read append block, and go
207 verify: backspace, read append block, and go
208 permissive: same as above but frees drive
209 if tape is not valid.
211 - Ensure that /dev/null works
212 - File daemon should build list of files skipped, and then
213 at end of save retry and report any errors.
214 - Need report class for messages. Perhaps
215 report resource where report=group of messages
216 - enhance scan_attrib and rename scan_jobtype, and
217 fill in code for "since" option
218 - Need to save contents of FileSet to tape?
219 - Director needs a time after which the report status is sent
220 anyway -- or better yet, a retry time for the job.
221 Don't reschedule a job if previous incarnation is still running.
222 - Figure out how to save the catalog (possibly a special FileSet).
223 - Figure out how to restore the catalog.
224 - Some way to automatically backup everything is needed????
225 - Need a structure for pending actions:
227 - termination status (part of buffered msgs?)
228 - Concept of grouping Storage devices and job can use
229 any of a number of devices
231 Read, Write, Clean, Delete
232 - Login to Bacula; Bacula users with different permissions:
233 owner, group, user, quotas
234 - Store info on each file system type (probably in the job header on tape.
235 This could be the output of df; or perhaps some sort of /etc/mtab record.
238 - Implement FSM (File System Modules).
239 - Identify unchanged or "system" files and save them to a
240 special tape thus removing them from the standard
241 backup FileSet -- BASE backup.
242 - Turn virutally all sprintfs into snprintfs.
243 - Heartbeat between daemons.
244 - Audit M_ error codes to ensure they are correct and consistent.
245 - Add variable break characters to lex analyzer.
246 Either a bit mask or a string of chars so that
247 the caller can change the break characters.
248 - Make a single T_BREAK to replace T_COMMA, etc.
249 - Ensure that File daemon and Storage daemon can
250 continue a save if the Director goes down (this
251 is NOT currently the case). Must detect socket error,
252 buffer messages for later.
253 - Enhance time/duration input to allow multiple qualifiers e.g. 3d2h
257 Bacula Projects Roadmap
259 last update 27 November 2002
261 Item 1: Multiple simultaneous Jobs. (done)
264 What: Permit multiple simultaneous jobs in Bacula.
266 Why: An enterprise level solution needs to go fast without the
267 need for the system administrator to carefully tweak
268 timing. Based on the benchmarks, during a full
269 backup, NetWorker typically hit 10 times the bandwidth to
270 the tape compared to Bacula--largely. This is probably due to
271 running parallel jobs and multi-threaded filling of buffers
272 and writing them to tape. This should also make things work
273 better when you have a mix of fast and slow machines backing
276 Notes: Bacula was designed to run multiple simultaneous jobs. Thus
277 implementing this is a matter of some small cleanups and
281 Item 2: Make the Storage daemon use intermediate file storage to buffer data.
282 Deferred -- not necessary yet.
284 What: If data is coming into the SD too fast, buffer it to
285 disk if the user has configured this option.
287 Why: This would be nice, especially if it more or less falls out
288 when implementing (1) above. If not, it probably should not
289 be given a high priority because fundamentally the backup time
290 is limited by the tape bandwidth. Even though you may finish a
291 client job quicker by spilling to disk, you still have to
292 eventually get it onto tape. If intermediate disk buffering
293 allows us to improve write bandwidth to tape, it may make
296 Notes: Whether or not this is implemented will depend upon performance
297 testing after item 1 is implemented.
300 Item 3: Write the bscan program -- also write a bcopy program.
303 What: Write a program that reads a Bacula tape and puts all the
304 appropriate data into the catalog. This allows recovery
305 from a tape that is no longer in the database, or it allows
306 re-creation of a database if lost.
308 Why: This is a fundamental robustness and disaster recovery tool
309 which will increase the comfort level of a sysadmin
310 considering adopting Bacula.
312 Notes: A skeleton of this program already exists, but much work
313 needs to be done. Implementing this will also make apparent
314 any deficiencies in the current Bacula tape format.
317 Item 4: Implement Base jobs.
319 What: A base job is sort of like a Full save except that you
320 will want the FileSet to contain only files that are unlikely
321 to change in the future (i.e. a snapshot of most of your
322 system after installing it). After the base job has been run,
323 when you are doing a Full save, you can specify to exclude
324 all files saved by the base job that have not been modified.
326 Why: This is something none of the competition does, as far as we know
327 (except BackupPC, which is a Perl program that saves to disk
328 only). It is big win for the user, it makes Bacula stand out
329 as offering a unique optimization that immediately saves time
332 Notes: Big savings in tape usage. Will require more resources because
333 the e. DIR must send FD a list of files/attribs, and the FD must
334 search the list and compare it for each file to be saved.
337 Item 5: Implement Label templates
339 What: This is a mechanism whereby Bacula can automatically create
340 a tape label for new tapes according to a detailed specification
341 provided by the user.
343 Why: It is a major convenience item for folks who use automated label
346 Notes: Bacula already has a working form of automatic tape label
347 creation, but it is very crude. The design for the complete
348 tape labeling project is already documented in the manual.
351 Item 6: Write a regression script.
354 What: This is an automatic script that runs and tests as many features
355 of Bacula as possible. The output is compared to previous
356 versions of Bacula and any differences are reported.
358 Why: This is an enormous help in preventing introduction of new
359 errors in parts of the program that already work correctly.
361 Notes: This probably should be ranked higher, it's something the typical
362 user doesn't see. Depending on how it's implemented, it may
363 make sense to defer it until the archival tape format and
364 user interface mature.
367 Item 7: GUI for interactive restore
368 Item 8: GUI for interactive backup
370 What: The current interactive restore is implemented with a tty
371 interface. It would be much nicer to be able to "see" the
372 list of files backed up in typical GUI tree format.
373 The same mechanism could also be used for creating
374 ad-hoc backup FileSets (item 8).
376 Why: Ease of use -- especially for the end user.
378 Notes: Rather than implementing in Gtk, we probably should go directly
379 for a Browser implementation, even if doing so meant the
380 capability wouldn't be available until much later. Not only
381 is there the question of Windows sites, most
382 Solaris/HP/IRIX, etc, shops can't currently run Gtk programs
383 without installing lots of stuff admins are very wary about.
384 Real sysadmins will always use the command line anyway, and
385 the user who's doing an interactive restore or backup of his
386 own files will in most cases be on a Windows machine running
390 Item 9: Add SSL to daemon communications.
392 What: This provides for secure communications between the daemons.
394 Why: This would allow doing backup across the Internet without
395 privacy concerns (or with much less concern).
397 Notes: The vast majority of near term potential users will be backing up
398 a single site over a LAN and, correctly or not, they probably
399 won't be concerned with security, at least not enough to go to
400 the trouble to set up keys, etc. to screw things down. We suspect
401 that many users genuinely interested in multi-site backup
402 already run some form of VPN software in their internetwork
403 connections, and are willing to delegate security to that layer.
406 Item 10: Define definitive tape format.
409 What: Define that definitive tape format that will not change
410 for the next millennium.
412 Why: Stability, security.
414 Notes: See notes for item 11 below.
417 Item 11: New daemon communication protocol.
419 What: The current daemon to daemon protocol is basically an ASCII
420 printf() and sending the buffer. On the receiving end, the
421 buffer is sscanf()ed to unpack it. The new scheme would
422 be a binary format that allows quick packing and unpacking
423 of any data type with named fields.
425 Why: Using binary packing would be faster. Named fields will permit
426 error checking to ensure that what is sent is what the
427 receiver really wants.
429 Notes: These are internal improvements in the interest of the
430 long-term stability and evolution of the program. On the one
431 hand, the sooner they're done, the less code we have to rip
432 up when the time comes to install them. On the other hand, they
433 don't bring an immediately perceptible benefit to potential
434 users. Item 10 and possibly item 11 should be deferred until Bacula
435 is well established with a growing user community more or
436 less happy with the feature set. At that time, it will make a
437 good "next generation" upgrade in the interest of data
443 ====================================
448 Subject: File Backup Options
451 A few days ago, a Bacula user who is backing up to file volumes and
452 using compression asked if it was possible to suppress compressing
453 all .gz files since it was a waste of CPU time. Although Bacula
454 currently permits using different options (compression, ...) on
455 a directory by directory basis, it cannot do it on a file by
456 file basis, which is clearly what was desired.
458 Proposed Implementation:
459 To solve this problem, I propose the following:
461 - Add a new Director resource type called FileOptions.
463 - The FileOptions resource will have records for all
464 options that can currently be specified on the Include record
465 (in a FileSet). Examples below.
467 - The FileOptions resource will permit an exclude option as well
468 as a number of additional options.
470 - The heart of the FileOptions resource is the ability to
471 supply any number of ApplyTo records which specify POSIX
472 regular expressions. These ApplyTo regular expressions are
473 applied to the fully qualified filename (path and all). If
474 one matches, then the FileOptions will be used.
476 - When an ApplyTo specification matches an included file, the
477 options specified in the FileOptions resource will override
478 the default options specified on the Include record.
480 - Include records will be modified to permit referencing one or
481 more FileOptions resources. The FileOptions will be used
482 in the order listed on the Include record and the first
483 one that matches will be applied.
485 - Options (or specifications) currently supplied on the Include
486 record will be deprecated (i.e. removed in a later version a
487 year or so from now).
489 - The Exclude record will be deprecated as the same functionality
490 can be obtained by using an Exclude = yes in the FileOptions.
493 The following records can appear in the FileOptions resource. An
494 asterisk preceding the name indicates a feature not currently
498 - Compression= (GZIP, ...)
499 - Signature= (MD5, SHA1, ...)
501 - OneFs= (yes/no) - remain on one filesystem
502 - Recurse= (yes/no) - recurse into subdirectories
503 - Sparse= (yes/no) - do sparse file backup
504 - *Exclude= (yes/no) - exclude file from being saved
505 - *Reader= (filename) - external read (backup) program
508 - verify= (ipnougsamc5) - verify options
511 - replace= (always/ifnewer/ifolder/never) - replace options currently
513 - *Writer= (filename) - external write (restore) program
517 Currently options specifying compression, MD5 signatures, recursion,
518 ... of a FileSet are supplied on the Include record. These will now
519 all be collected into a FileOptions resource, which will be
520 specified on the Include in place of the options. Multiple FileOptions
521 may be specified. Since the FileOptions contain regular expressions
522 that are applied to the full filename, this will give the ability
523 to specify backup options on a file by file basis to whatever level
532 Include = compression=GZIP signature=MD5 {
541 Include = FileOptions=Opts {
552 That's a lot more to do the same thing, but it gives the ability to
553 apply options on a file by file basis. For example, suppose you
554 want to compress all files but not any file with extensions .gz or .Z.
555 You could do so as follows:
559 Include = FileOptions=NoCompress FileOptions=Opts {
567 ApplyTo = /*.?*/ # matches all files
572 # Note multiple ApplyTos are ORed
573 ApplyTo = /*.gz/ # matches .gz files */
574 ApplyTo = /*.Z/ # matches .Z files */
577 Now, since the NoCompress FileOptions is specified first on the
578 Include line, any *.gz or *.Z file will have an MD5 signature computed,
579 but will not be compressed. For all other files, the NoCompress will not
580 match, so the Opts options will be used which will include GZIP
584 - Is it necessary to provide some means of ANDing regular expressions
585 and negation? (not currently planned)
587 e.g. ApplyTo = /*.gz/ && !/big.gz/
589 - I see that Networker has a "null" module which, if specified, does not
590 backup the file, but does make an record of the file in the catalog
591 so that the catalog will reflect an exact picture of the filesystem.
592 The result is that the file can be "seen" when "browsing" the save
593 sets, but it cannot be restored.
595 Is this really useful? Should it be implemented in Bacula?
598 After implementing the above, the user will be able to specify
599 on a file by file basis (using regular expressions) what options are
600 applied for the backup.
601 ====================================
603 Done: (see kernsdone for more)
604 - Add EOM records? No, not at this time. The current system works and
606 - Add VolumeUseDuration and MaximumVolumeJobs to Pool db record and
608 - Add VOLUME_CAT_INFO to the EOS tape record (as
609 well as to the EOD record). -- No, not at this time.
610 - Put MaximumVolumeSize in Director (MaximumVolumeJobs, MaximumVolumeFiles,
612 - Enhance schedule to have 1stSat, ...
613 - Make sure catalog doesn't keep growing.
614 - On I/O error, write EOF, then try to write again ? No, keep it simple.
615 - Figure out how compress everything except .gz,... files.
616 Implement FileOptions.
617 - Put Bacula version somewhere in Job stream, probably Start Session Labels.
618 - Fix start/end blocks for File devices
619 - Make Job err if WriteBootstrap fails.
620 - Test that mod of restore options works.
621 - Test that week position schedule code works.