1 Technical notes on version 1.38
8 - Add periodic (every 24 hours) garbage collection of memory
9 pool by releasing free buffers.
11 - Correct bug counting sized (for display only) in smartall.c
12 - Print FD mempool stats if debug > 0 rather than 5.
14 - Make db_lock() mutex error fail the job rather than abort
15 Bacula. Canceling the job caused the mutex to fail.
16 - Correct bug in alist.c that re-allocated the list if the
17 number of items goes to zero.
18 - Move the reservation system thread locking to the top level
19 so that one job at a time tries all possible drives before
21 - Implement a reservation 'fail' message queue that is built
22 and destroyed on each pass through the reservation system.
23 These messages are displayed in a 'Jobs waiting to reserve
24 a drive' list during a 'status storage='. Note, multiple
25 messages will generally print for each JobId because they
26 represent the different problems with either the same drive
27 or different drives. If this output proves too confusing
28 of voluminous, I will display it only when debug level 1
29 or greater is enabled in the SD.
31 - Add enable/disable job=<job-name>. This command prevents
32 the specified job from being scheduled. Even when disabled,
33 the job can be manually started from the console.
34 - During 'update slots' clear all InChanger flags where the
35 StorageId is zero (old Media records).
39 - Fix autochanger code to strip leading spaces from returned
40 slots number. Remove bc from chio-changer.
41 - Back port a bit of 1.39 crypto code to reduce diffs.
42 - Fix first call to autochanger that missed close()ing the
43 drive. Put close() just before each run_program(). Fixes
46 - Add PoolId to Job record when updating it at job start time.
48 - Pull in more code from 1.39 so that there are fewer file
49 differences (the new ua_dotcmds.c, base64.h, crypto.h
50 hmac.c jcr.c (dird and lib) lib.h md5.h parse_conf.c
51 util.c. Aside from ua_dotcmds.c these are mostly crypto
53 - Implement new method of walking the jcr chain. The
54 incr/dec of the use_count is done within the walking
55 routines. This should prevent a jcr from being freed
56 from under the walk routines.
58 Release 1.38.3 05Jan06:
60 - Move the suitable_drive flag to a better place to prevent
61 premature termination of the reservation if all drives
62 are busy -- should fix Arno's diff/inc pool failures.
64 - Add mutex to single thread VSS code in Win32.
68 - Add OPENSSL_INC to console dependencies, lib dependencies, and
69 wx-console dependencies in Makefile.in
70 - Add OPENSSL INC/LIB to gnome2_console Makefile.in.
71 - Simplify code in askdir.c that waits for creating an appendable
72 volume so that it can handle multiple returns from the wait
74 - Modify the wait code to permit multiple returns.
75 - Return a zero when 'autochanger drives' is called and
76 it is not an autochanger.
77 - Make rewind_dev() a method taking a DCR as an argument.
78 This permits closing and reopening the drive if the
79 rewind fails as happens if the drive was loaded while the
80 file descriptor was open. This refreshes the file descriptor.
81 - Remove the ST_OPENED flag and always rely on fd < 0 for knowing
82 if the device is open or not. This should eliminate
84 - Return error if reserve cannot find at least one suitable device.
85 - Make wait_for_sysop() return correct state information.
86 - Fix Win32 state file problem. write was not using compat
87 code. This should fix bug #500.
89 - Modify gui on command to set only GUI mode and not batch.
90 - Modify .messages command to always print messages regardless
92 - If GUI mode is on, suppress automatic printing of
94 - Delete old bnet packet code.
95 - Ignore new BNET_START_SELECT and BNET_END_SELECT signals in
97 - Modify restore command in wx-console to set gui on and to use
98 only .messages instead of messages. Hopefully this fixes bug
100 Beta release 20Dec05:
102 - Fix seg fault in exit of acquire when canceling a job --
103 reported by Wolfgang Denk
105 - Implement load balancing code.
106 - Pull a few files from 1.39 where we can maintain compatibility.
107 - Rewrite reservation algorithm again. Rename variables to be
108 more logical, add HEARTBEAT with Director, allow cancel of
109 jobs stuck in reservation, add last resourt any_drive.
111 - Remove quotes from Version table name -- it breaks things.
112 - Fix seg fault if user labels a drive directly bug #513
113 - Remove quotes around Version as it breaks things.
115 - Merge in Aleksandar Milivojevic's mods to the spec file.
116 - Apply sparse code fix for raw drives and fifos. Bug 506
117 - Thorsten fixed Unicode cd problem with wx-console bug 505.
118 Beta release 14Dec05:
120 - Correct reservation system to do a last ditch try
121 for any mounted volume, then anyone anywhere.
122 - Add quotes around table Version because of
123 error in MySQL 4.1.15 -- bug report submitted.
124 - Correct some minor problems with btape in the fill
126 - Updates to ssh-tunnel from Joshua Kugler.
127 - Added a report.pl program from Jonas Bjorklund.
128 - Simplify the O_NONBLOCK open() code for tape drives,
129 and always open nonblocking.
130 - Do not wait for open() if EIO returned (shouldn't happen).
131 - Eliminate 3 argument to tape open().
132 - Correct the slot # edited in the 3995 Bad autochanger unload
134 - With -S on bscan (show progress) do not divide by zero.
137 - Make cancel pthread_cond_signal() pthread_cond_broadcast().
138 - When dcr is freed, also broadcast dev->wait_next_vol signal.
139 - Remove unused code in wait_for_device.
140 - Make wait_for_device() always return after 60 seconds of wait.
142 - Use localhost if no network configured
144 - Eliminated duplicate MaxVolBytes in cat update -- bug 509.
145 - Remove debug print.
146 - Add bail_out in error during state file reading.
147 Beta release 10Dec05:
149 - Merge updates into 1.38 branch
150 - Update specs to include mysql4 define.
151 - Stop read_record() if status not ok in second loop.
152 - Return rec->FileIndex in dcr->VolLastIndex for normal
153 and partial records in read_record(). This allows bscan
154 to get FileIndex at EOT correct.
155 - Fix butil.c to correctly set dcr -- fixes seg fault in bls.
157 - Apply patch supplied by user (slightly modified) to fix
158 correct detection of holes in block devices and FIFOs.
160 - Apply patch supplied by user (slightly modified)
161 to fix SD hang with multiple pools and bad client
164 - Add nagios plugin to the examples directory. Submitted by
166 - Remove warning message about multiple saves of hardlinked files
167 from find_one.c as it can generate too many warning messages.
169 - Reset timeout values before select() per patch from
170 Frank Sweetser for problems with non-blocking sockets.
171 - Unlink the state file if either reading or writing it gets
172 errors. Hopefully this will fix Win32 exit problems.
173 - Add sanity check in append.c to ensure that dcr is not NULL.
174 This can happen if multiple drive autochanger SCSI control
175 channel and drive indicies do not correspond.
177 - Get next volume from Scratch pool before creating a volume.
178 - Set new Pool defaults in Vol when moved from Scratch Pool.
179 - Remove argument from create_bacula_database for SQLite as it
181 - Add back next_vol index code so that two drive autochangers can get
183 - Change a bunch of debug levels to aid debugging autochangers.
184 - Fix reservation so that mutexes are properly applied.
185 - Rework reservation algorithm so that two drives can be used
188 - Apply days keyword patch from Alexander.Bergolth at wu-wien.ac.at
189 If this patch is applied, the number of days can be specified with
190 'list nextvol days=xx'
193 My use case is to be able to preview the next scheduled job (and the
194 next tape to be used) on fridays if there are no scheduled jobs during
197 - Fix font code in gnome2 console user patch. Fixes bug #501.
198 - Fix malformatted bnet error message that caused seg fault
200 - Applied user patch to improve README.vc8 in src/win32.
202 - Correct some more editing of JobId's (for 64 bit compatibility).
203 - Ensure that StorageId is stored in Media record when ever possible.
204 - Add Migration Job to Job.
205 - Change Start Storage daemon job to require read and write storage
207 - Pass read storage data to SD as well as write storage data.
208 - Remove old code from winservice.cpp
209 - Break on error in scan.
210 - Fix typo in signal.c
211 - Separate read/write DCR in SD. Add jcr->read_dcr.
212 - Cleanup how find_device() works.
213 - Add read output to Status in SD.
215 - Remove abs() in bfile.c so that it compiles on Solaris.
218 Changes to 1.38.2: 22 November 2005
220 - Fix crash in tray-monitor when daemon disconnects. Bug #479.
221 - Fix bnet-server bug found on OpenBSD. Bug #486 (bug originator
222 says this does not fix *his* bug).
223 - Fix cancel failure bug. Bug #481
224 - Fix failure when Pool name has spaces. Bug #487
225 - Fix SD crash in autochanger code. Mutex failure. Bug #488
226 - Fix a couple of free()s in src/filed/acl.c
227 - Fix memory overrun in bfile.c in building OS X resource
228 fork filename. Bug #489
229 - Add Pool name to SD status output.
230 - Add Python install dir for Solaris to configure. Bug #492
232 Changes to 1.38.1: 15 November 2005
234 - Apply SunOS patch for ACLs submitted by David Duchscher.
235 - Make sure to set storage before trying to set drive.
236 - Add bacula_mail_summary.sh to examples directory. It makes
237 a single email summary of any number of jobs. Submitted
239 - Make sure when we do a mount to unblock the device even
240 if the drive could not be opened.
242 - Merge Scott's new spec files.
243 - Add doc on setting up Win32 environment variable supplied
244 by Kees van den Broek.
245 - Turn off API debug output unless debug set to avoid confusing
247 - Add Solaris ACL detection in configure.in as supplied by
250 - Implement 'autochanger drives' protocol so that Dir knows
251 how many drives an autochanger has.
252 - Do not request drive number in label, ... if only one drive.
253 - Turn off debug code.
254 - Fix update slots to clear slot number of every slot before
255 setting it. This fixes (I believe) bug #471
256 - Make unmount unload the autochanger slot.
257 - Modify open() on mount to be read-only and non-blocking,
258 otherwise the mount can block for a long time.
259 - Make a few error message numbers unique.
260 - Make a few error messages more correct.
261 - Apply patch from Thorsten to fix Win98 stat() command.
263 - Remove delete of CVS from all makefiles.
264 - Fix seg fault when clicking on Add button in wx-console
265 restore panel. Bug #470.
266 - Fix copyright date and URL typo -- bug #468.
267 - Change autostart install for FreeBSD to look for rc.conf
268 rather than rc.local as suggested fix for bug #466.
269 - Apply patch supplied by Eric Bollinger to fix PostgreSQL
270 grant on status. Bug #465
271 - Apply patch supplied by Eric Bollinger to fix PostgreSQL
272 update script. Bug #464
273 - Fix off by one for last Slot of autochanger for label.
274 - Update release date.
275 - Tweak an authentication error message in dir.
276 - Fix autoloader so that mutex is set and released around
277 each run_program(). There was a missing set.
278 - Remove an unnecessary drive release in autochanger.
279 - Modify configure.in to add execute option to sqlite3 catalog
281 - Create update_xxx_table_8_to_9 scripts for updatedb
282 - Fix wrong variable in bpipe.c debug output reported by user.
283 - Fix improper placement of encode_and_send_attributes() in
284 FD backup.c causing first file of non-portable Win32 backup
285 to have wrong stream. Reported by Thorsten.
286 - Move the -lcrypt for PostgreSQL after the PostgreSQL libs in
287 autoconf/bacula-macros/db.m4 as suggested by user. Fixes bug #457.
288 - Remove @STATIC_CONS@ from tray-monitor Makefile as suggested
289 by user. Fixes bug #456.
291 Released 1.38.0 (28Oct05): 31 October 2005