Bug Review 11.2.0.2/11.2.0.3
1. 参考文档及关键Bug筛选规则1.1 主要的参考文档 Oracle会定期针对每个版本所发现的所有Bug进行跟踪，并根据影响程度进行分类。如果出现关键问题都会及时发布Alert，并将这些内容都合并到已知问题－Known Issues文档中。关注这些文档可以及时了解到各个版本的最新信息。目前两个版本的主文档为：ALERT: Oracle 11g Release 2 (11.2) SupportStatus and Alerts (Doc ID 880782.1)
它里面包含了对各个小版本的Known Issue列表，比如11.2.0.2/11.2.0.3等。
1.2 辅助的分析工具，RAT Patch Advisory 通常数据库的主要补丁或补丁集均在Known Issue中以PSU（PatchSet Update）的方式列出，但这却不适合10.2.0.3及以前版本，因为PSU尚未推出。同时，在某些情况下Known Issue的更新不如RAT的补丁分析工具来得及时，因为该工具是通过直接查询Bug库获取信息的，也就保证了它的结果是最新的。
1.3 Bug的筛选规则 Oracle数据库的Bug有许多的分类，其中按照影响程度可分为30多种，包含了从实例宕机到代码提高等许多种分类。而本次主要的关注点是关键Bug，即对系统影响较大的补丁，所以我们从所有的补丁分析中选择以下分类供局方参考：
影响程度 描述－Bug触发时导致的结果
Instance May Crash可能导致实例宕机
Hang(Process Hang/Spins)可能导致进程僵死
Corruption(MEM/Block/Dictionary/Index)可能导致损坏（数据，字典表或索引等）
Process May Dump (Failure)可能导致进程/会话异常退出
Wrong Results可能产生错误的返回结果
Performance Affected对实例性能产生影响
对于其它Bug影响程度描述请参考：245840.1 Introduction to "BugDescription" Articles
2. 11.2.0.2版本的关键Bug筛选
2.1 11.2.0.2版本的基本情况11.2.0.2发布于2010年9月，是11gR2第二个被大量采用的版本。
目前采用11.2.0.2版本，且未安装PSU补丁。就目前安装情况分析在11.2.0.2上有哪些软件BUG具有较高的危险性，并给出相关的介绍和建议。
PSU(PatchSet Updates)类型的补丁包是在两个PATCHSET之间以季度为周期发布的补丁，它一般包含了一些关键补丁和CPU补丁。通常建议用户周期性的应用PSU补丁来增加代码的代码的强壮性。
对于任何数据库版本而言，关键Bug以两种方式来区分：1. 通用平台的关键Bug：即与特定的硬件通用平台的，所有相同的数据库版本均会受影响；2. 平台相关的关键Bug：即与特定平台相关的环境,OS补丁有关的关键Bug，该Bug只在这个平台出现，而在其它平台不会触发；
接下来将按照两种分类来描述关键Bug。
2.2 11.2.0.2的关键Bug通用平台的关键Bug
参考Known Issue文档及RAT建议，目前该版本通用平台的关键Bug包含：
Bug/Doc DescriptionImpact
Note:1561271.1P*Solaris / HP: ORA-29701 raised in ASM i/o pathI/O ERROR
Note:1527740.1*ORA-600 ORA-600 . Corruption in Rollback with Clusterwide Global Transactions in RACMEM/DATA/Dictionary损坏
Note:1191474.1*Corrupt Undo. ORA-600 during rollback in undo block for COMPRESS table with SUPPLEMENTAL LOGGINGMEM/DATA/Dictionary损坏
Note:1260095.1P+AIX: Excess "work USLA heap" process memory use in 11.2 on AIXMemory Overhead
14540423+Memory corruption / wrong results / ORA-7445 for particular type of chained row with ExadataMEM/DATA/Dictionary损坏 only on Exadata
13645917+PMON block recovery loop - instance hangHANG
11799496+ORA-600 block corruption in buffer cache for 32k block size / ORA-7445 by cache protectMEM/DATA/Dictionary损坏
11666959+ORA-7445 / LPX-200 / wrong results etc.. from new XML parserWRONG RESULT
10425010+Stale data blocks may be returned by Exadata FlashCacheWRONG RESULT only on Exadata
10269193+Wrong results with outer join and CASE expression optimization (CASE need not to be present) or LIKE operatorWRONG RESULT
10194190P+Solaris: Process spin and/or ASM and DB crash if RAC instance up for > 248 daysInstance Crash
9965278+Assorted dumps and errors with function based indexes or virtual columns presentERROR
11807012+Agent gets timeout for check action against SVC, DB, DG, ASM resources / CRS-5818ERROR
13384397+wrong results / OERI: with star transformationWRONG RESULT
10259620+Wrong results / ORA-7445 with DESC indexes and OR expansionWRONG RESULT
13550185+Hang / SGA memory corruption / ORA-7445 when using multiple shared pool subpoolsMEM/DATA/Dictionary损坏
Note:1472116.1*ORA-7445 / Assorted Errors on 11.2.0.2 if cursor sharing is enabled - Affects RMANMEM/DATA/Dictionary损坏
Note:1390632.1*Dictionary corruption / ORA-959 due to DROP TABLESPACEMEM/DATA/Dictionary损坏
11814891+ORA-600 [<lob block rdba>] / ORA-1555 double allocated LOB blockMEM/DATA/Dictionary损坏
12830339+Long shared pool latch waits / instance crash in 11.2 with ORA-240 / ORA-15064Instance Crash
Note:1318986.1*ORA-600 / corruption possible during shutdown in RACMEM/DATA/Dictionary损坏
9724970+Block Corruption with PDML UPDATE. ORA_600 OERI by block checkMEM/DATA/Dictionary损坏
14668670+Wrong results when execution plan shows nested loop batchingWRONG RESULT
10209232+ORA-1578 / ORA-600 Corruption. Misplaced Blocks and Lost Write in ASMMEM/DATA/Dictionary损坏
9735237+Dump / ORA-1722 as SQL uses child with mismatched BIND metadataERROR
以下是各个Bug的详细描述：Note: 1561271.1
影响程度 I/O 报错
有无补丁有
Bug描述仅仅发生在Solaris 或者 HP-ux平台，出现unrecoverable error ORA-29701、ORA-15032 raised in ASM I/O path; terminating process nnnnn
Workaround减轻负载压力
Note:1527740.1*
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述Database Block Corruption during ROLLBACK to <SAVEPOINT> when updating a RAC database with Clusterwide Global Transactions (database link or XA transaction).
WorkaroundSet _clusterwide_global_transactions=false (at the RAC database)
Note:1191474.1*
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述Supplemental Logging for COMPRESSED table may corrupt UNDO block.
The affected UNDO block results in the inability of tables from being read for transaction abort/recovery or consistent-read.
SMON may crash the database as it is unable to rollback the transaction associated to the affected UNDO segment.
Workaround1. Set _lmn_compression=FALSE to disable supplemental logging on compress tables.
2. Disable Supplemental Logging.
3. Disable COMPRESSION :
Note:1260095.1
影响程度 Memory Overhead
有无补丁可用
Bug描述仅仅在AIX平台上发生，每一个独立服务进程所消耗的内存对比9i,10g变多
Workaround linking the Oracle 11gR2 binary without -bexpful and -brtllib option.
BUG 14540423+
影响程度 MEM/DATA/Dictionary损坏 only on Exadata
有无补丁可用
Bug描述在Exadata上出现ORA-7445错误，或者查询结果集错误
WorkaroundTurn off cell offload eg: Set CELL_OFFLOAD_PROCESSING = false
BUG 13645917+
影响程度 Hang(Process Hang/Spins)
有无补丁可用
Bug描述由于PMON进程在线做block recover导致实例HANG
Workaroundshutdown abort then startup
BUG 11799496+
影响程度 MEM/DATA/Dictionary损坏
有无补丁可用
Bug描述32k的表空间上会出现ORA-600 数据块损坏
WorkaroundUse smaller block sizes such as 8k.
BUG 11666959+
Bug号 11666959
影响程度WRONG RESULT
有无补丁可用
Bug描述当解析XML时出现错误的结果集或者ORA-7445错误
Workaroundalter session set events= '31156 trace name context forever, level 0x400';
Bug 10425010
影响程度 WRONG RESULT only on Exadata
有无补丁可用
Bug描述Exadata Flashcache上的陈旧数据可能被标记为有效，导致返回错误的结果集合
Workaround无
Bug 10269193
影响程度 WRONG RESULT
有无补丁可用
Bug描述当存在以下情况时可能导致错误的结果集： 1. Wrong Results OR an ORA-600 in Parallel Query 2. Outer Join 3. CASE Expression 4. ROWID column is involved in CASE Expression
Workaround无
Bug 10194190P+
影响程度 Instance Crash
有无补丁可用
Bug描述仅仅在Solaris平台上发生，当实例运行超过248天，ASM和DB的进程可能spin导致实例CRASH
WorkaroundInstance needs to be restarted, but before doing that take pstack and truss on the spinning process for diagnostic purposes. This will help to determine whether this is same issue or different one.
To proactively avoid this issue schedule instance restarts to occur before 248 days of uptime.
Bug 9965278
影响程度 ERROR
有无补丁可用
Bug描述包含虚拟字段或者虚拟表达式字段的查询可能出现ORA-7445错误
WorkaroundThis workaround should help in all cases: alter session set "_replace_virtual_columns" = false;
These workarounds may work for specific variants of the issue: alter session set "_disable_function_based_index"=true; or alter session set "_fix_control"="6754080:off"; or alter session set "_fix_control"="8706652:off";
Bug 11807012
影响程度 ERROR
有无补丁可用
Bug描述Agents gets timeout for check action against SVC, DB, DG, ASM resources. eg: CRS-5818:Aborted command 'check for resource: ora.REDO1.dg orac11g04 1' for resource 'ora.REDO1.dg'. Details at (:CRSAGF00113:) {0:19:2} in /.../agent/crsd/oraagent_grid/oraagent_grid.log. Rediscovery Notes: If Agents gets timeout for check action against SVC, DB, DG, ASM resources in 11.2.0.2 even though the is database is healthy, you might be hitting this issue. If you can get call stack against agent threads before check action aborts, check if there are several threads waiting for mutex via ConnectionPool class. Example for such thread is #0 0xb7fd4402 in __kernel_vsyscall () #1 0x00c43379 in __lll_lock_wait () #2 0x00c3ea2f in _L_lock_885 #3 0x00c3e8f6 in pthread_mutex_lock () #4 0xb74b4a34 in sltsmna () #5 0x081d3fc2 in ConnectionPool::releaseConnection :
Workaround无
Bug 13384397
影响程度 WRONG RESULT
有无补丁可用
Bug描述星型转换时可能出现错误的结果集
Workaround_optimizer_use_cbqt_star_transformation=false
Bug 10259620+
影响程度 WRONG RESULT
有无补丁可用
Bug描述Wrong results may be produced by a query when the following conditions are all true for the query: - A column is included in a normal index and in a DESCending indexes - The same column is used in an OR predicate (or an INLIST) - OR-expansion is applied (CONCATENATE shown in Execution Plan) This is particularly likely for queries using a USE_CONCAT hint
WorkaroundAny of the following can avoid this problem: - alter session set "_fix_control" ='8352378:OFF'; - alter session set "_no_or_expansion" = true; - alter session set "_disable_function_based_index" = true; - Hint the query not to use OR expansion, or to avoid use of the function based index.
Bug 13550185
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述This problem is introduced in 11.2.0.2. SGA memory corruption and/or ORA-7445 can occur when the shared pool has multiple subpools. This is more likely to occur if literal replacement is used (CURSOR_SHARING=FORCE or SIMILAR) but it can also occur without cursor sharing explicitly enabled as described in the rediscovery notes below.
WorkaroundUse a single shared pool subpool, but note that this can affect performance in highly concurrent environments which have high shared pool load. eg: Add init parameter _kghdsidx_count = 1 and restart database.
Note:1472116.1
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述当cursor_sharing = FORCE or SIMILAR时在11.2.0.2下RMAN可能报ORA-7445错误
WorkaroundPossible workaround is to flush the shared pool. SQL> alter system flush shared_pool; This will not work in all circumstances, but might allow you to connect with RMAN again.
Bug 1390632.1
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述If a segment is created using Deferred Segment Creation and the related tablespace is later dropped before the segment is populated with data, accessing the object fails with errors ORA-00604 ORA-00959.
WorkaroundTo prevent this problem from being introduced (if Patch 13326736 is not applied), always attempt DROP TABLESPACE with INCLUDING CONTENTS AND DATAFILES clause, e.g.: DROP TABLESPACE TEST2 INCLUDING CONTENTS AND DATAFILES ;
Bug 11814891
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述ORA-600 [<lob block rdba>] / ORA-1555 caused by a double allocated LOB block after a LOB column UPDATE. The second argument in the ORA-600 indicates ORA-1. The problem is introduced by an UPDATE producing an ORA-1551 (internal error not visible to user) of a LOB column based on a subquery. The statement is basically of the form:
Workaround无
Bug 12830339
影响程度 Instance Crash
有无补丁可用
Bug描述Long shared pool latch waits may be seen in the DB with shared pool contention in 11.2.
If the DB uses ASM the ASM may terminate the DB instance.
Workaround无
Note 1318986.1
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述The issue could happen under certain timing conditions during shutdown normal/ transactional/ immediate of RAC instances in a rolling fashion when the other instances are still under normal workload. Note that this issue does not happen with shutdown abort. Oracle ASM instance is not impacted by this issue.
Workaround1. Shutdown database as a whole using abort option: SQL> alter system checkpoint; srvctl stop database -d <db_uniqueue_name> -o abort -f The first command writes out dirty buffers for all instances to minimize instance recovery; The srvctl command shutdown all instances with abort option and its dependent resources. Shutdown abort completely by-passes the vulnerable code path of the bug. 2. If shutdown instances in rolling fashion is needed, instead of shutdown normal/transactional/immediate, one should shutdown each instance with the following 2 commands: SQL> alter system checkpoint local; SQL> shutdown abort; The first command writes out dirty buffers for this instance to minimize instance recovery. Shutdown abort completely by-passes the vulnerable code path of the bug. Instance can also be shutdown abort using "srvctl stop instance -d <db_unique_name> -i <instance_name> -o abort".
Note 9724970
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述A parallel DML UPDATE may cause block corruption with several dumps or ORA-600 errors on a large table if there are any row which has a NULL value in the last-N column(s). If db_block_checking is enabled, ORA-600 or ORA-600 can be produced with different error codes; e.g. 6135, 6110, 6255, etc.
Workaround使用串行更新
Note 14668670+
影响程度 WRONG RESULT
有无补丁可用
Bug描述A query may return wrong results if the execution plan shows Nested Loop Batching being used, usually during the first execution. Nested Loop Batching is being used if the execution plan shows access to a table's index that is being joined using a nested loop to its table. In the below example, T5 is being accessed by Nested Loop (Batching) using the result of I_T5. The rest of the tables are not using Nested Loop Batching.
WorkaroundThe best workaround for this issue is to set "_nlj_batching_misses_enabled"=0 This can be set at session or system level and will avoid the above bug scenario without impacting execution plans. Other possible workarounds are shown below but these may alter execution plans: Set _projection_pushdown=false or set _NLJ_BATCHING_ENABLED=0
Note 10209232+
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述Blocks can be misplaced in ASM after using a wrong extent map to write blocks during rebalance. The Blocks intended to write may not be written. Those blocks become stale blocks (Lost Write). Misplaced Blocks ================ The blocks are written to wrong locations and those blocks become misplaced with wrong rdba producing ORA-1578 and dbverify may report the misplaced blocks as follow:
Workaround无
Note 9735237
影响程度 ERROR
有无补丁可用
Bug描述It is possible for a session to end up using a mismatched child cursor where the bind metadata of the child cursor differs from the actual binds used. This can lead to a dump (typically under kxspoac), unexpected errors, such as ORA-1722, as the bind data is interpreted as a different type or in the case of DMLs logical data corruption.
WorkaroundFor the problem statement ensure that bind types and metadata are always the same wherever that SQL is issued from. If bind types differ for calls from different locations then alter the SQL to include comment (or similar) so that the SQL text itself differs for the different bind types used.
11.2.0.2与Soaris平台相关的Bug
Note: 1561271.1
影响程度 I/O 报错
有无补丁有
Bug描述仅仅发生在Solaris 或者 HP-ux平台，出现unrecoverable error ORA-29701、ORA-15032 raised in ASM I/O path; terminating process nnnnn
Workaround减轻负载压力
Bug 10194190P+
影响程度 Instance Crash
有无补丁可用
Bug描述仅仅在Solaris平台上发生，当实例运行超过248天，ASM和DB的进程可能spin导致实例CRASH
WorkaroundInstance needs to be restarted, but before doing that take pstack and truss on the spinning process for diagnostic purposes. This will help to determine whether this is same issue or different one.
To proactively avoid this issue schedule instance restarts to occur before 248 days of uptime.
2.3 11.2.0.2版本的补丁建议
针对11.2.0.2如果不想升级到最新的11.2.0.4+LatestPSU的话，建议考虑安装上述这些关键的BUG的一次性补丁one-offPatch，特别是可能导致Instance Crash、InstanceHang和Corruption数据块损坏的bug。
Instance Crash Bug列表：10194190P+、12830339+
Instance Hang Bug列表：13645917+、13550185+
3. 11.2.0.3版本的关键Bug筛选
3.1 11.2.0.3版本的基本情况
本系统采用版本11.2.0.3 且安装了PSU 补丁"Database Patch Set Update :11.2.0.3.3 (13923374)"和"Grid InfrastructurePatch Set Update : 11.2.0.3.3 (13919095)"，我们在后续介绍11.2.0.3上的BUG时会排除那些已经在11.2.0.3.3 PSU上修复了的问题。
11.2.0.4是11gr2最新版本，第一次发布时间是2013年8月发布的。ORACLE 官方推荐用户升级到最新的PatchSet。
11.2.0.3是11gR2第二个PatchSet，第一次发布时间是2011年9月。 目前版本11.2.0.3仍是 ORACLE官方主要支持的版本。
Release PlatformRelease DatePatch EndPSUCPUBundle Patch
11.2.0.3Solaris29-OCT-2011Jan-2015Patch: 17478514Jan 2014N/A
一个数据库版本DeSupport就意味着：l 如果是已知的Bug，已经有相应Patch的，仍然可以下载使用；l 如果问题可能是新的Bug引发的，后台部门不再提供补丁开发支持；
对于任何数据库版本而言，关键Bug以两种方式来区分：1. 通用平台的关键Bug：即与特定的硬件通用平台的，所有相同的数据库版本均会受影响；2. 平台相关的关键Bug：即与特定平台相关的环境,OS补丁有关的关键Bug，该Bug只在这个平台出现，而在其它平台不会触发；
接下来将按照两种分类来描述关键Bug。3.2 11.2.0.3的关键Bug
通用平台的关键Bug
参考Known Issue文档及RAT建议，目前该版本通用平台的关键Bug包含：
Bug/Doc DescriptionFixed in PSU/Bundle
Note:1260095.1P+AIX: Excess "work USLA heap" process memory use in 11.2 on AIX AIX (ignore)
13605839+ORA-600 ORA-600 . Corruption in Rollback with Clusterwide Global Transactions in RAC 12.1
14668670+wrong results the first time a query is executed11.2.0.3.BP14
13786142+Cannot drop/replace trigger in editioning environment11.2.0.3.BP07
13384397+wrong results / OERI: with star transformation11.2.0.3.BP05
13645917+PMON block recovery loop - instance hang11.2.0.3.5
13550185+Hang / SGA memory corruption / ORA-7445 when using multiple shared pool subpools11.2.0.3.4
10194190P+Solaris: Process spin and/or ASM and DB crash if RAC instance up for > 248 days
Note:1561271.1P*Solaris / HP: ORA-29701 raised in ASM i/o path11.2.0.3.GIPSU08
16299065+ORA-1206 in SOURCE database after RMAN duplicate11.2.0.3.BP21
13384397+wrong results / OERI: with star transformation11.2.0.3.BP05
17761775+ORA-600 ORA-600 or ORA-600 Join of temp and permanent table in RAC might lead to corruption11.2.0.3.9
17752121+ORA-600 ORA-600 RAC diagnostic/fix to avoid a block being modified in Shared Mode and prevent corruption11.2.0.3.9
Note:1527740.1*ORA-600 ORA-600 . Corruption in Rollback with Clusterwide Global Transactions in RAC11.2.0.3.8
14668670+Wrong results when execution plan shows nested loop batching11.2.0.3.7
Bug 13605839
影响程度 Corruption (Logical)
有无补丁无
Bug描述ORA-600 , ORA-600 , ORA-600 with ktcrsp1 on stack, ORA-600 with check codes or when db_block_checking is enabled. Corruption during Rollback with Clusterwide Global Transactions when using savepoints and updating a RAC database through a database link.
WorkaroundSet _clusterwide_global_transactions=false (at the RAC database). Setting this parameter to false does not affect performance.
Bug 16299065
影响程度 Corruption (datafile)
有无补丁有
Bug描述当使用RMAN duplicate命令时可能损坏数据文件头
WorkaroundRestore the affected datafile from backup
Bug 14668670
影响程度 Wrong Results
有无补丁无
Bug描述A query may return wrong results if the execution plan shows Nested Loop Batching being used. Nested Loop Batching is being used if the execution plan shows access to a table's index that is being joined using a nested loop to its table.
Workaround_projection_pushdown=false or _NLJ_BATCHING_ENABLED=0
BUG 13645917+
影响程度 Hang(Process Hang/Spins)
有无补丁可用
Bug描述由于PMON进程在线做block recover导致实例HANG
Workaroundshutdown abort then startup
Bug 13550185
影响程度 Performance Affected
有无补丁可用
Bug描述当两个会话同时初始化某个游标结构时可能会造成SGA内存的损坏。往往这个游标具备以下特点： 1. 系统启用了游标共享即Cursor_sharing=similar/force; 2. 许多会话以不同的字符串来执行该游标； 3. 系统的共享池子池超过1；
Workaround将共享池子池数设置为1，即alter system set "_kghdsidx_count"=1 scope=spfile;
Bug 13384397
影响程度 Wrong Result
有无补丁可用
Bug描述使用星形转换时产生错误结果或内部错误。
Workaround_optimizer_use_cbqt_star_transformation=false
Note: 1561271.1
影响程度 I/O 报错
有无补丁有
Bug描述仅仅发生在Solaris 或者 HP-ux平台，出现unrecoverable error ORA-29701、ORA-15032 raised in ASM I/O path; terminating process nnnnn
Workaround减轻负载压力
Bug 10194190P+
影响程度 Instance Crash
有无补丁可用
Bug描述仅仅在Solaris平台上发生，当实例运行超过248天，ASM和DB的进程可能spin导致实例CRASH
WorkaroundInstance needs to be restarted, but before doing that take pstack and truss on the spinning process for diagnostic purposes. This will help to determine whether this is same issue or different one.
To proactively avoid this issue schedule instance restarts to occur before 248 days of uptime.
Bug 13467683
影响程度 MEM/DATA/Dictionary损坏
有无补丁可用
Bug描述在RAC环境，在两个不同的实例中，相同的块可能被错误的修改，临时表和永久表作join时可能引起永久表损坏。 数据库在恢复时可能产生ORA-600 错误
Workaround无
Bug 17761775
影响程度 MEM/DATA/Dictionary损坏
有无补丁可用
Bug描述this bug is only relevant when using Real Application Clusters (RAC) Joining temporary and permanent tables in RAC might lead to corruption causing ORA-600 and/or ORA-600 during recovery. If the diagnostic/fix of bug 17752121 is in place, it may cause ORA-600 preventing further corruptions. This bug may commonly affect INDEX blocks associated to the permanent TABLE rather than affecting the TABLE blocks but it also may affect TABLE blocks.
Workaround无
Bug 17752121
影响程度 MEM/DATA/Dictionary损坏
有无补丁可用
Bug描述This bug is only relevant when using Real Application Clusters (RAC) RAC specific diagnostic fix to avoid a block buffer being modified in Shared Mode instead of Exclusive mode and prevent further errors or corruptions caused by stale blocks in buffer cache. The inconsistency/corruption can be also manifested by TABLE/INDEX mismatch reported with ORA-1499 by "analyze table validate structure cascade" or ORA-8102.
Workaround无
Note:1527740.1*
影响程度 Corruption(MEM/Block/Dictionary/Index)
有无补丁可用
Bug描述Database Block Corruption during ROLLBACK to <SAVEPOINT> when updating a RAC database with Clusterwide Global Transactions (database link or XA transaction).
WorkaroundSet _clusterwide_global_transactions=false (at the RAC database)
Note: 13786142
影响程度 ERROR
有无补丁可用
Bug描述if a trigger fails creation on an editioning environment, the trigger will not be able to be dropped or replaced. Attempts to do so will raise ora-4045 and ora-4098.
Workaround无
3.3 11.2.0.3版本的补丁建议
针对11.2.0.3如果不想升级到最新的11.2.0.4+Latest PSU的话，建议考虑安装上述这些关键的BUG的一次性补丁one-off Patch，特别是可能导致Instance Crash、Instance Hang和Corruption数据块损坏的bug。
目前已知的11.2.0.3上的Instance Crash Bug列表：10194190P
目前已知的11.2.0.3上的Instance Hang Bug列表：13645917 13550185
当然，RAT的补丁建议是基于现有的补丁库信息，如果在若干个月以后搭建11gRAC的话，就需要在搭建前重新获取这个版本的关键补丁分析，这样可以获取知道最新的补丁情况。另外，11.2.0.3开始每个Patchset均为一个完整的软件发布包，而不是一个增量的修订包。数据库的升级基本是一个软件的重新安装过程。
4. Bug分析及Patch建议总结
4.1 补丁冲突及解决建议以上无论是通用平台的关键补丁，还是平台相关的关键补丁，在应用补丁前需要考虑以下情况：l 是否与已应用的补丁冲突；l 是否在多个关键补丁之间存在冲突；
考虑到10.2.0.3/11.2.0.2已经终止支持，如果存在以上冲突，则需要选择关键级别最高的补丁来应用，除非有其它Merge Patch可用。
有关补丁冲突检测的方法因为不在本次现场支持的范围内，详细情况请参考文档：459360.1 How To Do ThePrerequisite/Conflicts Checks Using OUI(Oracle Universal Installer) And OpatchBefore Applying/Rolling Back A Patch
4.2 PSU补丁及安装步骤从11.2.0.2开始，Oracle推出的PSU补丁成为大的PATCHSET之后最为常用的小补丁集，如果暂时不能从以上关键补丁中确认的，也可以通过应用每个版本最新的PSU补丁来增强代码强壮性。
各个版本的最新PSU补丁已经在上面的各部分列出来了，局方可以根据情况选择合适的PSU补丁。
补丁安装过程：在PSU补丁包里都有详细的安装说明，局方可以参考这些说明自行安装补丁，或者向ACS部门申请人天现场应用补丁。
刘大，10g 的bug review 也可以贴下么？
页:
[1]