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.
17 - Fix db_get_job_volume_names() to return array of strings.
18 - Report compression % and other compression statistics if turned on.
19 - Add JobLevel in FD status (but make sure it is defined).
20 - Eliminate MySQL shared libraries from smtp and daemons not
22 - Pass "Catalog Files = no" to storage daemon to eliminate
24 - Implement alter_sqlite_tables
25 - Fix scheduler -- see "Hourly cycle". It doesn't do both each
26 hour, rather it alternates between 0:05 and 0:35.
27 - Create Counter DB records.
28 - Make Pool resource handle Counter resources.
29 - Remove NextId for SQLite. Optimize.
30 - Termination status in FD for Verify = C -- incorrect.
31 - Fix strerror() to use strerror_r()
32 - Fix gethostbyname() to use gethostbyname_r()
33 - Cleanup path/filename separation in sql_get.c and sql_create.c
34 - Implement ./configure --with-client-only
35 - Strip trailing / from Include
36 - Move all SQL statements into a single location.
37 - Cleanup db_update_media and db_update_pool
38 - Add UA rc and history files.
39 - put termcap (used by console) in ./configure and
40 allow -with-termcap-dir.
41 - Remove JobMediaId it is not used.
42 - Enhance time and size scanning routines.
43 - Fix Autoprune for Volumes to respect need for full save.
44 - DateWritten may be wrong.
45 - Fix Win32 config file definition name on /install
46 - When we are at EOM, we must ask each job to write JobMedia
47 record (update_volume_info).
48 - No READLINE_SRC if found in alternate directory.
49 - Add Client FS/OS id (Linux, Win95/98, ...).
50 - Put Windows files in Windows stream?
52 ====== 31 May 2002 ========
53 Now that Bacula 1.20 is released, virtually all the
54 basic features are implemented (some are still quite
55 primitive though). Over the next month or two, I'm
56 planning to focus on the following items:
59 - Fix any bugs I find or you report.
60 - Finish the implementation of automatic pruning
61 (add pruning of Restore and Verify jobs).
62 - Make sure pruning of Volumes won't prune the
63 only backup of a FileSet
66 - Improve the Restore capabilities of Bacula
67 * Restore to most recent system state (i.e.
68 figure out what tapes need to be mounted and
70 * Restore to a particular time (perhaps several
71 variations -- e.g. before date, after date).
72 * Interactive Restore where you get to select
73 what files are to be restored (much like the Unix
74 "restore" program permits). Now that we have a
75 catalog of all files saved, it would be nice to
77 * Restore options (overwrite, overwrite if older,
78 overwrite if newer, never overwrite, ...)
79 * Improve the standalone programs (bls and bextract)
80 to have pattern matching capabilities (e.g. restore
81 by FileSet, Job, JobType, JobLevel, ...).
82 * Ideally after each Job, Bacula could write out a
83 set of commands to a file that if later feed to
84 bextract would restore your system to the current
85 state (at least for the saved FileSet). This would
86 provide a simple disaster recovery that could be
87 initiated from a "floppy" and one simple ASCII control
88 file. I'm not exactly sure how to do this, but it
89 shouldn't be too hard and I'll
90 be trying to go in this direction.
93 - Implement tape verification to ensure that the data
94 written for a particular Job can really be read.
95 - Compare tape File attributes to Catalog.
96 (File attributes are size, dates, MD5, but not
98 - Compare tape to Client files (attributes, or
102 - With the current Bacula 1.21 (not yet in the CVS) I
103 expect there is about 95% chance that running multiple
104 simultaneous Jobs will actually work without stepping
105 on each other. I'm planning to try this sometime soon.
110 - Rework Storage daemon with new rwl_lock routines.
111 - Implement Label templates
112 - Pass JCR to database routines permitting better error printing.
115 - Verify against Full.
118 mysqldump -f --opt bacula >bacula
122 - Probably add End of Data tape records (this would make
123 the tape format incompatible with the previous version).
124 - I'll most likely enhance the current tape format
125 in the way that I previously described, which will make
126 some of the labels incompatible, but the change will
127 not affect the current restore code since it does not
128 look at the details of the labels.
129 - I may add a few more waiting conditions in the Storage
130 daemon where it will current immediately aborts a
131 Job if the necessary resources are not available (e.g.
132 tape is being written and a read request arrives).
133 - Write an applet for Linux.
135 - Remove PoolId from Job table, it exists in Media.
136 - Allow commands to detach or run in background.
137 - Write better dump of Messages resource.
138 - Fix status delay on storage daemon during rewind.
139 - Add VerNo to each Session label record.
140 - Add Job to Session records.
141 - Add VOLUME_CAT_INFO to the EOS tape record (as
142 well as to the EOD record).
143 - Add SD message variables to control operator wait time
144 - Maximum Operator Wait
145 - Minimum Message Interval
146 - Maximum Message Interval
147 - Add EOM handling variables
149 - Require EOD records
150 - Send Operator message when cannot read tape label.
151 - Think about how to handle I/O error on MTEOM.
152 - If Storage daemon aborts a job, ensure that this
153 is printed in the error message.
154 - Verify level=Volume (scan only), level=Data (compare of data to file).
155 Verify level=Catalog, level=InitCatalog
156 - Scan tape contents into database.
158 - Cold start full restore (restore catalog then
159 user selects what to restore). Write summary file containing only
160 Job, Media, and Catalog information. Store on another machine.
161 - Dump/Restore database
164 - Implement first cut of Catalog Retention period (remove old
165 entries from database).
166 - Add SessionTime/Id filters to bextract.
168 - Ensure that Start/End File/Block are correct.
169 - Add keyword search to show command in Console.
170 - If MySQL database is not running, job terminates with
171 wierd type and wierd error code.
172 - Write a regression script
173 - Report bad status from smtp or mail program.
174 - Fix Win2000 error with no messages during startup.
175 - Add estimate to Console
176 - Events : tape has more than xxx bytes.
177 - In Storage daemon, status should include job cancelled.
178 - Write general list maintenance subroutines.
179 - Implement immortal format with EDOs.
180 - Restrict characters permitted in a Resource name.
181 - Restore file xx or files xx, yy to their most recent values.
182 - Provide definitive identification of type in backup.
183 - Complete code in Bacula Resources -- this will permit
184 reading a new config file at any time.
185 - Document new Console
186 - Handle ctl-c in Console
187 - Test restore of Windows backup
188 - Implement LabelTemplate (at least first cut).
189 - Implement script driven addition of File daemon to
192 - Bug: anonymous Volumes requires mount in some cases.
193 - see setgroup and user for Bacula p4-5 of stunnel.c
194 - Implement new serialize subroutines
195 send(socket, "string", &Vol, "uint32", &i, NULL)
196 - Add save type to Session label.
197 - Correct date on Session label.
198 - On I/O error, write EOF, then try to write again.
199 - Audit all UA commands to ensure that we always prompt where
201 - If ./btape is called without /dev, assume argument is
202 a Storage resource name.
203 - Put memory utilization in Status output of each daemon
204 if full status requested or if some level of debug on.
205 - Make database type selectable by .conf files i.e. at runtime
206 - gethostbyname failure in bnet_connect() continues
207 generating errors -- should stop.
208 - Don't create a volume that is already written. I.e. create only once.
209 - If error at end of tape, implement some way to kill waiting processes.
210 - Get correct block/file information in Catalog, pay attention
212 - Add HOST to Volume label.
213 - Set flag for uname -a. Add to Volume label.
214 - Implement throttled work queue.
215 - Write bscan program that will syncronize the DB Media record with
216 the contents of the Volume -- for use after a crash.
217 - Check for EOT at ENOSPC or EIO or ENXIO (unix Pc)
218 - Allow multiple Storage specifications (or multiple names on
219 a single Storage specification) in the Job record. Thus a job
220 can be backed up to a number of storage devices.
221 - Implement full MediaLabel code.
222 - Implement dump label to UA
223 - Copy volume using single drive.
224 - Copy volume with multiple driven (same or different block size).
225 - Add block size (min, max) to Vol label.
226 - Concept of VolumeSet during restore which is a list
227 of Volume names needed.
228 - Restore files modified after date
229 - Restore file modified before date
230 - Emergency restore info:
232 - Backup working directory
234 - Restore options (do not overwrite)
235 - Restore -- do nothing but show what would happend
236 - Authentication between SD and FD
237 - SET LD_RUN_PATH=$HOME/mysql/lib/mysql
238 - Send Volumes needed during restore to Console
239 - Put Job statistics in End Session Label (files saved,
240 total bytes, start time, ...).
241 - Put FileSet name in the SOS label.
242 - Implement Restore FileSet=
243 - Write a scanner for the UA (keyword, scan-routine, result, prompt).
244 - Create a protocol.h and protocol.c where all protocol messages
246 - If SD cannot open a drive, make it periodically retry.
247 - Put Bacula version somewhere in Job stream, probably Start Session
249 - Remove duplicate fields from jcr (e.g. jcr.level and
251 - Timout a job or terminate if link goes down, or reopen link and query.
252 - Define how we handle times to avoid problem with Unix dates (2049 ?).
253 - The daemons should know when one is already
254 running and refuse to run a second copy.
255 - Fill all fields in Vol/Job Header -- ensure that everything
256 needed is written to tape. Think about restore to Catalog
257 from tape. Client record needs improving.
258 - Find general solution for sscanf size problems (as well
259 as sprintf. Do at run time?
261 - Concept of precious tapes (cannot be reused).
262 - Allow FD to run from inetd ???
263 - Preprocessing command per file.
264 - Postprocessing command per file (when restoring).
266 - Restore should get Device and Pool information from
267 job record rather than from config.
268 - Make SD send attribute stream to DR but first
269 buffering to file, then sending only when the
270 files are written to tape.
271 - Autolabel should be specified by DR instead of SD.
272 - Ability to recreate the catalog from a tape.
273 - Find out how to get the system tape block limits, e.g.:
274 Apr 22 21:22:10 polymatou kernel: st1: Block limits 1 - 245760 bytes.
275 Apr 22 21:22:10 polymatou kernel: st0: Block limits 2 - 16777214 bytes.
278 - AutoScan (check checksum of tape)
279 - Format command = "format /dev/nst0"
283 - Seek resolution (usually corresponds to buffer size)
284 - EODErrorCode=ENOSPC or code
285 - Partial Read error code
286 - Partial write error code
287 - Nonformatted read error
288 - Nonformatted write error
289 - WriteProtected error
293 - IgnoreCloseErrors=yes
303 - Eliminate duplicate File records to shrink database.
304 - FD sends unsaved file list to Director at end of job.
305 - Implement InsertUniqueDB.
306 - Write a Storage daemon that uses pipes and
307 standard Unix programs to write to the tape.
309 - Need something that monitors the JCR queue and
310 times out jobs by asking the deamons where they are.
311 - Add daemon JCR JobId=0 to have a daemon context
315 - Client (list of clients to force client)
316 - Devices (list of devices to force device)
321 - Recycle from other pools: Yes, No
322 - Recycle to other pools: Yes, no
325 - Optional MediaType to force media?
336 if(connect(sockfd, (struct sockaddr * ) (& addr), sizeof(addr)) .lt. 0){
342 linger.l_linger = 60;
343 i = setsockopt(sockfd, SOL_SOCKET, SO_LINGER, (char *) &linger,
346 fl = fcntl(sockfd, F_GETFL);
347 fcntl(sockfd, F_SETFL, fl & (~ O_NONBLOCK) & (~ O_NDELAY));
349 - Add "0nnn" in front of all sscanf %s fields
350 to prevent field overflow.
352 What: jobid or file list
353 From: tape, file, ...
354 Where: original location, another path
355 How: Always replace, Replace if newer, Never replace
356 Report: files restored; files not restored; errors; warnings
358 - Enhance Jmsg code to permit buffering and saving to disk.
359 - Probably create a jcr with JobId=0 as a master
360 catchall if jcr not found or if operation involves
362 - device driver = "xxxx" for drives.
363 - restart: paranoid: read label fsf to
364 eom read append block, and go
365 super-paranoid: read label, read all files
366 in between, read append block, and go
367 verify: backspace, read append block, and go
368 permissive: same as above but frees drive
369 if tape is not valid.
371 - Ensure that /dev/null works
372 - File daemon should build list of files skipped, and then
373 at end of save retry and report any errors.
374 - Need report class for messages. Perhaps
375 report resource where report=group of messages
376 - Extract what=(session_id|file_list); where
378 - enhance scan_attrib and rename scan_jobtype, and
379 fill in code for "since" option
380 - dir_config: get rid of all printfs
381 - To buffer messages, we need associated jobid and Director name.
382 - Need to save contents of FileSet to tape?
383 - Director needs a time after which the report status is sent
384 anyway -- or better yet, a retry time for the job.
385 Don't reschedule a job if previous incarnation is still running.
386 - Figure out how to do a "full" restore from catalog
387 - Figure out how to save the catalog (possibly a special FileSet).
388 - Figure out how to restore the catalog.
389 - Figure out how to put a Volume into the catalog (from the tape)
390 - Figure out how to do a restore from a Volume
391 - Some way to automatically backup everything is needed????
392 - Need a structure for pending actions:
394 - termination status (part of buffered msgs?)
395 - Concept of grouping Storage devices and job can use
396 any of a number of devices
398 Read, Write, Clean, Delete
399 - Login to Bacula; Bacula users with different permissions:
401 - Tape recycle destination
402 - Job Schedule Status
406 - File daemon should pass Director the operating system info
407 to be stored in the Client Record (or verified that it has
409 - Store info on each file system type (probably in the job header on tape.
410 This could be the output of df; or perhaps some sort of /etc/mtab record.
413 - Use media 1 time (so that we can do 6 days of incremental
414 backups before switching to another tape) (already)
415 specify # times (jobs)
416 specify bytes (already)
417 specify time (seconds, hours, days)
418 - Implement FSM (File System Modules).
419 - Identify unchanged or "system" files and save them to a
420 special tape thus removing them from the standard
421 backup FileSet -- BASE backup.
422 - Turn virutally all sprintfs into snprintfs.
423 - Heartbeat between daemons.
424 - Audit M_ error codes to ensure they are correct and
426 - Add variable break characters to lex analyzer.
427 Either a bit mask or a string of chars so that
428 the caller can change the break characters.
429 - Make a single T_BREAK to replace T_COMMA, etc.
430 - Ensure that File daemon and Storage daemon can
431 continue a save if the Director goes down (this
432 is NOT currently the case). Must detect socket error,
433 buffer messages for later.
436 Done: (see kernsdone for more)