格式：

通过:符号，可以连续设置多个事件，也可以通过连续使用alter session set events来设置多个事件。

格式说明：

eventnumber指触发dump的事件号，事件号可以是Oracle错误号（出现相应错误时跟踪指定的事件）或oralce内部事件号，内部事件号在10000到10999之间，不能与immediate关键字同用。

immediate关键字表示命令发出后，立即将指定的结构dump到跟踪文件中，这个关键字只用在alter session语句中，并且不能与eventnumber、forever关键字同用。

trace name是关键字

eventname指事件名称（见后面），即要进行dump的实际结构名。若eventname为context，则指根据内部事件号进行跟踪。

forever关键字表示事件在实例或会话的周期内保持有效状态，不能与immediate同用。

level为事件级别关键字。但在dump错误栈(errorstack)时不存在级别。

levelnumber表示事件级别号，一般从1到10，1表示只dump结构头部信息，10表示dump结构的所有信息。

常用转储命令：

一.Memory Dumps

1).Global Area 

Alter SESSION SET EVENTS 'immediate trace name global_area level n'; 

1 包含PGA 

2 包含SGA 

4 包含UGA 

8 包含indrect memory

2).Library Cache 

Alter SESSION SET EVENTS 'immediate trace name library_cache level n'; 

1 library cache统计信息 

2 包含hash table histogram 

3 包含object handle 

4 包含object结构（Heap 0）

3).Row Cache 

Alter SESSION SET EVENTS 'immediate trace name row_cache level n'; 

1 row cache统计信息 

2 包含hash table histogram 

8 包含object结构

4).Buffers 

Alter SESSION SET EVENTS 'immediate trace name buffers level n'; 

1 buffer header 

2 level 1 + block header 

3 level 2 + block contents 

4 level 1 + hash chain 

5 level 2 + hash chain 

6 level 3 + hash chain 

8 level 4 + users/waiters 

9 level 5 + users/waiters 

10 level 6 + users/waiters

5).Buffer 

Alter SESSION SET EVENTS 'immediate trace name buffer level n'; 

n为某个指定block的rdba，该命令可以转储某个block在buffer中的所有版本。

6).Heap 

Alter SESSION SET EVENTS 'immediate trace name heapdump level level'; 

1 PGA摘要 

2 SGA摘要 

4 UGA摘要 

8 Current call(CGA)摘要 

16 User call(CGA)摘要 

32 Large call(LGA)摘要 

1025 PGA内容 

2050 SGA内容 

4100 UGA内容 

8200 Current call内容 

16400 User call内容 

32800 Large call内容

7).Sub Heap 

oracle 9.0.1版本之前 

Alter SESSION SET EVENTS 'immediate trace name heapdump_addr level n'; 

若n为subheap的地址，转储的是subheap的摘要信息 

若n为subheap的地址+1，转储的则是subheap的内容 

oracle 9.2.0版本之后 

Alter SESSION SET EVENTS 'immediate trace name heapdump_addr level n, addr m'; 

其中m为subheap的地址 

n为1转储subheap的摘要，n为2转储subheap的内容

8).Process State 

Alter SESSION SET EVENTS 'immediate trace name processstate level n';

9).System State 

Alter SESSION SET EVENTS 'immediate trace name systemstate level n';

10).Error State 

Alter SESSION SET EVENTS 'immediate trace name errorstack level n'; 

0 Error stack 

1 level 0 + function call stack 

2 level 1 + process state 

3 level 2 + context area

11).Hang Analysis 

Alter SESSION SET EVENTS 'immediate trace name hanganalyze level n';

12).Work Area 

Alter SESSION SET EVENTS 'immediate trace name workareatab_dump level n'; 

1 SGA信息 

2 Workarea Table摘要信息 

3 Workarea Table详细信息

13).Latches 

Alter SESSION SET EVENTS 'immediate trace name latches level n'; 

1 latch信息 

2 统计信息

14).Events 

Alter SESSION SET EVENTS 'immediate trace name events level n'; 

1 session 

2 process 

3 system

15).Locks 

Alter SESSION SET EVENTS 'immediate trace name locks level n';

16).Shared Server Process 

Alter SESSION SET EVENTS 'immediate trace name shared_server_state level n'; 

n取值为1~14

17).Background Messages 

Alter SESSION SET EVENTS 'immediate trace name bg_messages level n'; 

n为pid+1

二.File Dumps

1).Block 

oracle 7之前 

Alter SESSION SET EVENTS 'immediate trace name blockdump level n'; 

n为block的rdba 

oracle8以后 

Alter SYSTEM DUMP DATAFILE file# BLOCK block#; 

Alter SYSTEM DUMP DATAFILE file# 

BLOCK MIN minimum_block# 

BLOCK MAX maximum_block#;

2).Tree Dump 

Alter SESSION SET EVENTS 'immediate trace name treedump level n'; 

n为object_id

3).Undo Segment Header 

Alter SYSTEM DUMP UNDO_HEADER 'segment_name';

4).Undo for a Transaction 

Alter SYSTEM DUMP UNDO BLOCK 'segment_name' XID xidusn xidslot xidsqn;

5).File Header 

Alter SESSION SET EVENTS 'immediate trace name file_hdrs level n'; 

1 控制文件中的文件头信息 

2 level 1 + 文件头信息 

3 level 2 + 数据文件头信息 

10 level 3

6).Control file 

Alter SESSION SET EVENTS 'immediate trace name controlf level n'; 

1 文件头信息 

2 level 1 + 数据库信息 + 检查点信息 

3 level 2 + 可重用节信息 

10 level 3

7).Redo log Header 

Alter SESSION SET EVENTS 'immediate trace name redohdr level n'; 

1 控制文件中的redo log信息 

2 level 1 + 文件头信息 

3 level 2 + 日志文件头信息 

10 level 3

8).Redo log 

Alter SYSTEM DUMP LOGFILE 'FileName'; 

Alter SYSTEM DUMP LOGFILE 'FileName' 

SCN MIN MinimumSCN 

SCN MAX MaximumSCN 

TIME MIN MinimumTime 

TIME MAX MaximumTime 

LAYER Layer 

OPCODE Opcode 

DBA MIN FileNumber . BlockNumber 

DBA MAX FileNumber . BlockNumber 

RBA MIN LogFileSequenceNumber . BlockNumber 

RBA MAX LogFileSequenceNumber . BlockNumber; 

其中time = (((((yyyy - 1988)) * 12 + mm - 1) * 31 + dd - 1) * 24 + hh) * 60 + mi) * 60 + ss; 

9).Loghist 

Alter SESSION SET EVENTS 'immediate trace name loghist level n'; 

1 dump控制文件中最早和最迟的日志历史项 

>1 dump 2^n个日志历史项

eventnumber

1、例举

Code   Description 

10000 Controlfile debug event, name 'control_file' 

10001 Controlfile crash event1 

10002 Controlfile crash event2 

10003 Controlfile crash event3 

10004 Controlfile crash event4 

10005 Trace latch operations for debugging 

10006 Testing - block recovery forced 

10007 Log switch debug crash after new log select, thread %s 

10008 Log switch debug crash after new log header write, thread %s 

10009 Log switch debug crash after old log header write, thread %s 

10010 Begin Transaction 

10011 End Transaction 

10012 Abort Transaction 

10013 Instance Recovery 

10014 Roll Back to Save Point 

10015 Undo Segment Recovery 

10016 Undo Segment extend 

10017 Undo Segment Wrap 

10018 Data Segment Create 

10019 Data Segment Recovery 

10020 Partial link restored to linked list (KSG) 

10021 Latch cleanup for state objects (KSS) 

10022 Trace ktsgsp 

10023 Create Save Undo Segment 

10024 Write to Save Undo 

10025 Extend Save Undo Segment 

10026 Apply Save Undo 

10027 Latch cleanup for enqueue locks (KSQ) 

10028 Latch cleanup for enqueue resources (KSQ) 

10029 Session logon (KSU) 

10030 Session logoff (KSU) 

10031 Sort debug event (S*) 

10032 Sort statistics (SOR*) 

10033 Sort run information (SRD*/SRS*) 

10034 Access path analysis (APA*) 

10035 Parse SQL statement (OPIPRS) 

10036 Create remote row source (QKANET) 

10037 Allocate remote row source (QKARWS) 

10038 Dump row source tree (QBADRV) 

10039 Type checking (OPITCA) 

10040 Dirty cache list 

10041 Dump undo records skipped 

10042 Trap error during undo application 

10043 Check consistency of owner/waiter/converter lists in KSQ 

10044 Free list undo operations 

10045 Free list update operations - ktsrsp, ktsunl 

10046 Enable SQL statement timing 

10047 Trace switching of sessions 

10048 Undo segment shrink 

10049 Protect library cache memory heaps 

10050 Sniper trace 

10051 Trace OPI calls 

10052 Don't clean up obj$ 

10053 CBO Enable optimizer trace 

10054 Trace UNDO handling in MLS 

10055 Trace UNDO handing 

10056 Dump analyze stats (kdg) 

10057 Suppress file names in error messages 

10058 Use table scan cost in tab$.spare1 

10059 Simulate error in logfile create/clear 

10060 CBO Enable predicate dump 

10061 Disable SMON from cleaning temp segment 

10062 Disable usage of OS Roles in osds 

10063 Disable usage of DBA and OPER privileges in osds 

10064 Thread enable debug crash level %s, thread %s 

10065 Limit library cache dump information for state object dump 

10066 Simulate failure to verify file 

10067 Force redo log checksum errors - block number 

10068 Force redo log checksum errors - file number 

10069 Trusted Oracle test event 

10070 Force datafile checksum errors - block number 

10071 Force datafile checksum errors - file number 

10072 Protect latch recovery memory 

10073 Have PMON dump info before latch cleanup 

10074 Default trace function mask for kst 

10075 CBO Disable outer-join to regular join conversion 

10076 CBO Enable cartesian product join costing 

10077 CBO Disable view-merging optimization for outer-joins 

10078 CBO Disable constant predicate elimination optimization 

10079 Trace data sent/received via SQL*Net 

10080 Dump a block on a segment list which cannot be exchanged 

10081 Segment High Water Mark has been advanced 

10082 Free list head block is the same as the last block 

10083 A brand new block has been requested from space management 

10084 Free list becomes empty 

10085 Free lists have been merged 

10086 CBO Enable error if kko and qka disagree on oby sort 

10087 Disable repair of media corrupt data blocks 

10088 CBO Disable new NOT IN optimization 

10089 CBO Disable index sorting 

10090 Invoke other events before crash recovery 

10091 CBO Disable constant predicate merging 

10092 CBO Disable hash join 

10093 CBO Enable force hash joins 

10094 Before resizing a data file 

10095 Dump debugger commands to trace file 

10096 After the cross instance call when resizing a data file 

10097 After generating redo when resizing a data file 

10098 After the OS has increased the size of a data file 

10099 After updating the file header with the new file size 

10100 After the OS has decreased the size of a data file 

10101 Atomic redo write recovery 

10102 Switch off anti-joins 

10103 CBO Disable hash join swapping 

10104 Dump hash join statistics to trace file 

10105 CBO Enable constant pred trans and MPs w WHERE-clause 

10106 CBO Disable evaluating correlation pred last for NOT IN 

10107 CBO Always use bitmap index 

10108 CBO Don't use bitmap index 

10109 CBO Disable move of negated predicates 

10110 CBO Try index rowid range scans 

10111 Bitmap index creation switch