1<!--$Id: 4.7.html,v 1.38 2008/05/15 21:13:24 carol Exp $--> 2<html> 3<head> 4<title>The Berkeley DB Package: DB 4.7.25 Change Log</title> 5<meta name="description" content="Berkeley DB: A database programmatic toolkit."> 6<meta name="keywords" content="embedded,database,programmatic,toolkit,b+tree,btree,hash,hashing,transaction,transactions,locking,logging,access method,access methods"> 7</head> 8<body bgcolor=white> 9 10<h3 align=center>Berkeley DB 4.7.25 Change Log</h3> 11<h3>Database or Log File On-Disk Format Changes:</h3> 12 <ol> 13 <li>The log file format changed in 4.7. </li> 14 </ol> 15 16<h3>New Features:</h3> 17<ol> 18 19<li>The lock manager may now be fully partitioned, improving performance 20on some multi-CPU systems. [#15880] 21 <li>Replication groups are now architecture-neutral, supporting 22connections between differing architectures (big-endian or 23little-endian, independent of structure padding). [#15787] [#15840] </li> 24 <li>Java: A new Direct Persistence Layer adds a built-in Plain Old Java Object (POJO)-based 25 persistent object model, which provides support for complex object models without 26 compromises in performance. For an introduction to the Direct Persistence Layer 27 API, see Getting Started with Data Storage. [#15936] </li> 28 <li>Add the DB_ENV->set_intermediate_dir_mode method to support the 29creation of intermediate directories needed during recovery. [#15097] </li> 30<li>The DB_ENV->failchk method can now abort transactions for threads, 31which have failed while blocked on a concurrency lock. This 32significantly decreases the need for database environment recovery after 33thread of control failure. [#15626] 34<li>Replication Manager clients now can be configured to monitor the 35connection to the master using heartbeat messages, in order to promptly 36discover connection failures. [#15714] 37<li>The logging system may now be configured to pre-zero log files when 38they are created, improving performance on some systems. [#15758] 39 40</ol> 41<h3>Database Environment Changes:</h3> 42<ol> 43 44<li>Restructure aborted page allocation handling on systems without an 45ftruncate system call. This enables the Berkeley DB High Availability 46product on systems, which do not support ftruncate. [#15602] 47<li>Fix a bug where closing a database handle after aborting a 48transaction which included a failed open of that handle could result in 49application failure. [#15650] 50<li>Fix minor memory leaks when closing a private database environment. 51[#15663] 52<li>Fix a bug leading to a panic of "unpinned page returned" if a cursor 53was used for a delete multiple times and deadlocked during one of the 54deletes. [#15944] 55<li>Optionally signal processes still running in the environment before running 56recovery. [#15984] 57 58</ol> 59<h3>Concurrent Data Store Changes:</h3> 60 61None. 62 63<h3>General Access Method Changes:</h3> 64<ol> 65 66<li>Fix a bug where closing a database handle after aborting a 67transaction which included a failed open of that database handle could 68result in application failure. [#15650] 69<li>Fix a bug that could cause panic in a database environment 70configured with POSIX-style thread locking, if a database open failed. 71[#15662] 72<li>Fix bug in the DB->compact method which could cause a panic if a 73thread was about to release a page while another thread was truncating 74the database file. [#15671] 75 <li>Fix an obscure case of interaction between a cursor scan and delete that was prematurely 76 returning DB_NOTFOUND. [#15785]</li> 77<li>Fix a bug in the DB->compact method where if read-uncommitted was 78configured, a reader reading uncommitted data my see an inconsistent 79entry between when the compact method detects an error and when it 80aborts the enclosing transaction. [#15856] 81<li>Fix a bug in the DB->compact method where a thread of control 82mail fail if two threads are compacting the same section of a Recno 83database. [#15856] 84<li>Fix a bug in DB->compact method, avoid an assertion failure when zero pages 85can be freed. [#15965] 86<li>Fix a bug return a non-zero error when DB->truncate is called with open cursors. 87[#15973] 88<li>Fix a bug add HANDLE_DEAD checking for DB cursors. [#15990] 89<li>Fix a bug to now generate errors when DB_SEQUENCE->stat is called without 90first opening the sequence. [#15995] 91 <li>Fix a bug to no longer dereference a pointer into a hash structure, when hash functionality 92 is disabled. [#16095]</li> 93 94</ol> 95<h3>Btree Access Method Changes:</h3> 96 97None. 98 99<h3>Hash Access Method Changes:</h3> 100<ol> 101 102<li>Fix a bug where a database store into a Hash database could 103self-deadlock in a database environment configured for the Berkeley DB 104Concurrent Data Store product, and with a free-threaded DB_ENV or DB 105handle. [#15718] 106 107</ol> 108<h3>Queue Access Method Changes:</h3> 109<ol> 110 111<li>Fix a bug that could cause a put or delete of a queue element to 112return a DB_NOTGRANTED error, if blocked. [#15933] 113 114</ol> 115<h3>Recno Access Method Changes:</h3> 116 117<ol> 118 119<li>Expose db_env_set_func_malloc, db_env_set_func_realloc, and 120db_env_set_func_free through the Windows API for the DB dll. [#16045] 121 122</ol> 123 124<h3>C-specific API Changes:</h3> 125 126None. 127 128<h3>Java-specific API Changes:</h3> 129<ol> 130 131<li>Fix a bug where enabling MVCC on a database through the Java API was 132ignored. [#15644] 133<li>Fixed memory leak bugs in error message buffering in the Java API. 134[#15843] 135<li>Fix a bug where Java SecondaryConfig was not setting SecondaryMultiKeyCreator from 136the underlying db handle [OTN FORUM} 137<li>Fix a bug so that getStartupComplete will now return a boolean instead of an int. 138[#16067] 139<li>Fix a bug in the Java API, where Berkeley DB would hang on exit when using replication. [#16142]</li> 140 141</ol> 142<h3> 143 Direct Persistence Layer (DPL), Bindings and Collections API:</h3> 144 <ol> 145 146<li>A new Direct Persistence Layer adds a built-in Plain Old Java Object (POJO)-based 147 persistent object model, which provides support for complex object models without 148 compromises in performance. For an introduction to the Direct Persistence Layer 149 API, see Getting Started with Data Storage. [#15936]</li> 150 <li> Fixed a bug in the remove method of the Iterator instances returned by the 151 StoredCollection.iterator method in the collections package. This bug caused ArrayIndexOutOfBoundsException 152 in some cases when calling next, previous, hasNext or hasPrevious after calling 153 remove. (Note that this issue does not apply to StoredIterator instances returned 154 by the StoredCollection.storedIterator method.) This bug was reported in this forum 155 thread: 156 <br /> 157 <a class="moz-txt-link-freetext" href="http://forums.oracle.com/forums/thread.jspa?messageID=2187896"> 158 http://forums.oracle.com/forums/thread.jspa?messageID=2187896</a> 159 <br /> 160 [#15858] </li> 161 <li>Fixed a bug in the remove method of the StoredIterator instances returned by StoredCollection.storedIterator 162 method in the collections package. If the sequence of methods next-remove-previous 163 was called, previous would sometimes return the removed record. If the sequence 164 of methods previous-remove-next was called, next would sometimes return the removed 165 record. (Note that this issue does not apply to Iterator instances returned by the 166 StoredCollection.iterator method.) [#15909] </li> 167 <li>Fixed a bug that causes a memory leak for applications where many Environment objects 168 are opened and closed and the CurrentTransaction or TransactionRunner class is used. 169 The problem was reported in this JE Forum thread: <a class="moz-txt-link-freetext" 170 href="http://forums.oracle.com/forums/thread.jspa?messageID=1782659">http://forums.oracle.com/forums/thread.jspa?messageID=1782659</a> 171 [#15444] </li> 172 <li>Added StoredContainer.areKeyRangesAllowed method. Key ranges and the methods 173 in SortedMap and SortedSet such as subMap and subSet are now explicitly disallowed 174 for RECNO and QUEUE databases -- they are only supported for BTREE databases. 175 Before, using key ranges in a RECNO or QUEUE database did not work, but was not 176 explicitly prohibited in the Collections API. [#15936] </li> 177 </ol> 178 <h3> 179 Tcl-specific API Changes:</h3> 180<ol> 181 182<li>The Berkeley DB Tcl API does not attempt to avoid evaluating input 183as Tcl commands. For this reason, it may be dangerous to pass unreviewed 184user input through the Berkeley DB Tcl API, as the input may 185subsequently be evaluated as a Tcl command. To minimize the 186effectiveness of a Tcl injection attack, the Berkeley DB Tcl API in the 1874.7 release routine resets process' effective user and group IDs to the 188real user and group IDs. [#15597] 189 190</ol> 191<h3>RPC-specific Client/Server Changes:</h3> 192 193None. 194 195<h3>Replication Changes:</h3> 196<ol> 197 198<li>Fix a bug where a master failure resulted in multiple attempts to 199perform a "fast election"; subsequent elections, when necessary, now use 200the normal nsites value. [#15099] 201<li>Replication performance enhancements to speed up failover. [#15490] 202<li>Fix a bug where replication could self-block in a database environment 203configured for in-memory logging. [#15503] 204<li>Fix a bug where replication would attempt to read log file version 205numbers in a database configured for in-memory logging. [#15503] 206<li>Fix a bug where log files were not removed during client 207initialization in a database configured for in-memory logging. [#15503] 208<li>The 4.7 release no longer supports live replication upgrade from the 2094.2 or 4.3 releases, only from the 4.4 and later releases. [#15602] 210<li>Fix a bug where replication could re-request missing records on 211every arriving record. [#15629] 212<li>Change the DB_ENV->rep_set_request method to use time, not the 213number of messages, when re-requesting missed messages on a replication 214client. [#15629] 215<li>Fix a minor memory leak on the master when updating a client during 216internal initialization. [#15634] 217<li>Fix a bug where a client error when syncing with a new replication 218group master could result in an inability to ever re-join the group. [#15648] 219 <li>Change dbenv->rep_set_request to use time-based values instead of counters. [#15682]</li> 220<li>Fix a bug where a LOCK_NOTGRANTED error could be returned from the 221DB_ENV->rep_process_message method, instead of being handled 222internally by replication. [#15685] 223<li>Fix a bug where the Replication Manager would reject a fresh 224connection from a remote site that had crashed and restarted, displaying 225the message: "redundant incoming connection will be ignored". [#15731] 226<li>The Replication Manager now supports dynamic negotiation of the best 227available wire protocol version, on a per-connection basis. [#15783] 228<li>Fix a bug, which could lead to slow performance of internal 229initialization under the Replication Manager, as evidenced by "queue limit 230exceeded" messages in verbose replication diagnostic output. [#15788] 231<li>Fix a bug where replication control message were not portable between 232replication clients with different endian architectures. [#15793] 233<li>Add a configuration option to turn off Replication Manager's special 234handling of elections in 2-site groups. [#15873] 235<li>Fix a bug making it impossible to call replicationManagerAddRemoteSite 236in the Java API after having called replicationManagerStart. [#15875] 237<li>Fix a bug where the DB_EVENT_REP_STARTUPDONE event could be 238triggered too early. [#15887] 239 <li>Fix a bug where the rcvd_ts timestamp is reset when the user just changes the threshold. 240 [#15895]</li> 241<li>Fix a bug where the master in a 2-site replication group might wait 242for client acknowledgement, even when there was no client connected. 243[#15927] 244<li>Fix a bug, clean up and restart internal init if master log is gone. 245 [#16006] 246 <li>Fix a bug, ignore page messages that are from an old internal init. 247 [#16075] [#16059] 248 <li>Fix a bug where checkpoint records do not indicate a database was a named in-memory 249 database. [#16076]</li> 250 <li>Fix a bug with in-memory replication, where we returned with the log region mutex 251 held in an error path, leading to self-deadlock. [#16088]</li> 252 <li>Fix a bug which causes the DB_REP_CHECKPOINT_DELAY setting in rep_set_timeout() 253 to be interpreted in seconds, rather than microseconds. [#16153]</li> 254 255</ol> 256<h3>XA Resource Manager Changes:</h3> 257<ol> 258 259<li>Fix a bug where the DB_ENV->failchk method and replication in 260general could fail in database environments configured for XA. [#15654] 261 262</ol> 263<h3>Locking Subsystem Changes:</h3> 264<ol> 265 266<li>Fix a bug causing a lock or transaction timeout to not be set 267properly after the first timeout triggers on a particular lock id. 268[#15847] 269 270<li>Fix a bug that would cause a trap if DB_ENV->lock_id_free was passed an 271invalid locker id. [#16005] 272 273<li>Fix a bug when thread tracking is enabled where an attempt is made to 274release a mutex that is not lock. [#16011] 275</ol> 276<h3>Logging Subsystem Changes:</h3> 277<ol> 278<li>Fix a bug, handle zero-length log records doing HA sync with 279in-memory logs. [#15838] 280<li>Fix a bug that could cause DB_ENV->failcheck to leak log region 281memory. [#15925] 282<li>Fix a bug where the abort of a transaction that opened a database 283could leak log region memory. [#15953] 284<li>Fix a bug that could leak memory in the DB_ENV->log_archive interface if 285a log file was not found. [#16013] 286 287</ol> 288<h3>Memory Pool Subsystem Changes:</h3> 289<ol> 290 <li>Fix multiple MVCC bugs including a race, which could <b>result in 291incorrect data being returned</b> to the application. [#15653] </li> 292<li>Fixed a bug that left an active file in the buffer pool after a 293database create was aborted. [#15918] 294<li>Fix a bug where there could be uneven distribution of pages if a single database 295and multiple cache regions are configured. [#16015] 296<li>Fix a bug where DB_MPOOLFILE->set_maxsize was dropping the wrong 297mutex after open. [#16050] 298 299</ol> 300<h3>Mutex Subsystem Changes:</h3> 301<ol> 302 303<li>Fix a bug where mutex contention in database environments configured 304for hybrid mutex support could result in performance degradation. 305[#15646] 306<li>Set the DB_MUTEX_PROCESS_ONLY flag on all mutexes in private environments, 307they can't be shared and so we can use the faster, intra-process only mutex 308implementations [#16025] 309<li>Fix a bug so that mutexes are now removed from the environment signature if 310mutexes are disabled. [#16042] 311 312</ol> 313<h3>Transaction Subsystem Changes:</h3> 314<ol> 315 316<li>Fix a bug that could cause a checkpoint to selfblock attempting to 317flush a file, when the file handle was closed by another thread during 318the flush. [#15692] 319<li>Fix a bug that could cause DB_ENV->failcheck to hang if there 320were pending prepared transactions in the environment. [#15925] 321 <li>Prepared transactions will now use the sync setting from the environment. 322 Default to flushing the log on commit (was nosync). [#15995]</li> 323 <li>If __txn_getactive fails, we now return with the log region mutex held. This 324 is not a bus since __txn_getactive cannot really fail. [#16088]</li> 325 326</ol> 327<h3>Utility Changes:</h3> 328 329<ol> 330 331<li>Update db_stat with -x option for mutex stats 332 <li>Fix an incorrect assumption about buffer size when getting an overflow page in db_verify. 333 [#16064]</li> 334</ol> 335 336<h3>Configuration, Documentation, Sample Application, Portability and Build Changes:</h3> 337<ol> 338 339<li>Fix an installation bug where the Berkeley DB PHP header file was 340not installed in the correct place. 341<li>Merge the run-time configuration sleep and yield functions. [#15037] 342<li>Fix Handle_DEAD and other expected replication errors in the 343C++ sample application ReqQuoteExample.cpp. [15568] 344<li>Add support for monotonic timers. [#15670] 345<li>Fix bugs where applications using the db_env_func_map and 346db_env_func_unmap run-time configuration functions could not join 347existing database environments, or open multiple DB_ENV handles for a 348single environment. [#15930] 349<li> Add documentation about building Berkeley DB for VxWorks 6.x. 350<li>Remove the HAVE_FINE_GRAINED_LOCK_MANAGER flag, it is obsolete in 4.7. 351<li>Fix a bug in ex_rep, add a missing break which could cause a segment fault.</li> 352 <li>Fix build warnings from 64 bit Windows build. [#16029]</li> 353<li>Fix an alignment bug on ARM Linux. Force the assignment to use memcpy. 354 [#16125]</li> 355 <li>Fix a bug in the Windows specific code of ex_sequence.c, where there was an invalide 356 printf specifier. [#16131]</li> 357 <li>Improve the timer in ex_tpcb to use high resolution timers. [#16154]</li> 358 <li>Mention in the documentation that env->open() requires DB_THREAD to be specified when using repmgr. 359 [#16163]</li> 360 <li>Disable support for mmap on Windows CE. The only affect is that we do not 361 attempt to mmap small read only databases into the mpool. [#16169]</li> 362 363</ol> 364</body> 365</html> 366