db file scattered read

一：db file scattered read说明

二：db file scattered read解决思路

三：db file scattered read重现过程

四：db file scattered read官方文档

一：db file scattered read说明

Oracle在执行全表扫描(Full Table Scan,FTS)、索引快速全扫描(Index Fast Full Scan)时，为保障性能，尽量一次性读取多个块，这称为Multi Block I/O。
每次执行Multi Block I/O，都会等待物理I/O结束，此时等待db file scattered read事件。

https://docs.oracle.com/cd/E11882_01/server.112/e41573/instance_tune.htm#PFGRF94479

This event signifies that the user process is reading buffers into the SGA buffer cache and is waiting for a physical I/O call to return. Adb file scattered readissues a scattered read to read the data into multiple discontinuous memory locations. A scattered read is usually a multiblock read. It can occur for a fast full scan (of an index) in addition to a full table scan.

Thedbfilescatteredreadwait event identifies that a full scan is occurring. When performing a full scan into thebuffer cache, the blocks read are read into memory locations that are not physically adjacent to each other. Such reads are called scattered read calls, because the blocks are scattered throughout memory. This is why the corresponding wait event is called 'db file scattered read'. multiblock (up toDB_FILE_MULTIBLOCK_READ_COUNTblocks) reads due to full scans into the buffer cache show up as waits for 'db file scattered read'.

https://docs.oracle.com/cd/E11882_01/server.112/e40402/waitevents003.htm#BGGIBDJI

Similar todb file sequential read, except that the session is reading multiple data blocks.

二：db file scattered read解决思路

1 SQL优化

如果是某些SQL引起的，例如统计信息不准确，没有索引或使用低效的索引等，可以通过优化SQL，降低db file scattered read；

2 分区表

可以考虑将全表扫描优化成分区扫描；

3 增大BUFFER CACHE

如果db file scattered read出现特别频繁，Buffer HIT较低，可以考虑增大buffer cache；

4 使用更快的存储；

selectname,parameter1,parameter2,parameter3

fromv$event_name

wherename='db file scattered read';

---查看含有db file scattered read等待事件的session；

SELECTsid,total_waits,time_waited

FROMv$session_event

WHEREevent ='db file scattered read'

andtotal_waits >0

ORDERBY2desc;

selectsid,event,p1,p2,p3,wait_class

fromv$session_wait

whereevent ='db file scattered read';

Check the followingV$SESSION_WAITparameter columns:

P1: The absolute file number

P2: The block being read

P3: The number of blocks (should be greater than 1)

selectowner,segment_name,segment_type

fromdba_extents

wherefile_id =6

and37475betweenblock_id andblock_id +blocks -1;

SELECTrow_wait_obj# FROMV$SESSION WHEREEVENT ='db file scattered read';

SELECTowner,object_name,subobject_name,object_type

FROMDBA_OBJECTS

WHEREdata_object_id ='88605';

selectv.last_call_et,

v.username,

v.sid,

sql.sql_text,

sql.sql_id,

sql.disk_reads,

v.event

fromv$session v,v$sql sql

wherev.sql_address =sql.address

andv.last_call_et >0

andv.status ='ACTIVE'

andv.username isnotnull;

select*fromtable(dbms_xplan.display_cursor('d24df9xbujb75'));

---SELECTSQL_ADDRESS,SQL_HASH_VALUE

FROMV$SESSION

WHEREEVENT LIKE'db file%read';

三：db file scattered read重现过程

索引快速全扫描(Index Fast Full Scan)

SQL>createtablespacechenjch_tbs datafile'/u01/app/oracle/oradata/orcl/chenjch_tbs01a.dbf'size10M autoextendonmaxsize1G;

SQL>createuserchenjch identifiedbya defaulttablespacechenjch_tbs;

SQL>grantconnect,resource,dbatochenjch;

SQL>createtablet1 asselect*fromdba_objects;

SQL>insertintot1 select*fromt1;

SQL>commit;

SQL>insertintot1 select*fromt1;

SQL>commit;

SQL>insertintot1 select*fromt1;

SQL>commit;

......

SQL>createindexi_t1_001 ont1(object_id,object_name,object_type);

SQL>

begin

dbms_workload_repository.create_snapshot();

end;

SQL>altersystemflushbuffer_cache;

SQL>altersessionsettracefile_identifier='10046';

SQL>altersessionsetevents'10046 trace name context forever, level 12';

SQL>select/*+ index_ffs(t1 i_t1_001) */object_id,object_name,object_type fromt1 whereobject_id ='20';

SQL>altersessionsetevents'10046 trace name context off';

SQL>

begin

dbms_workload_repository.create_snapshot();

end;

SQL>

selectdistinct(m.sid),p.pid,p.tracefile

fromv$mystat m,v$session s,v$process p

wherem.sid =s.sid

ands.paddr =p.addr;

---/u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_57032_10046.trc

[oracle@dip ~]$ tkprof /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_57032_10046.trc /home/oracle/10046_3.trc

全表扫描(Full Table Scan,FTS)

http://f.dataguru.cn/thread-44033-1-1.html

Oracle 11g，在大表的全表扫描算法上有新的变化，根据表的大小、高速缓存的大小等信息，决定是否绕过SGA直接从磁盘读取数据。

而10g则是全部通过高速缓存读取数据。

Oracle 11g认为大表全表时使用直接路径读，可能比10g中的数据文件散列读(db file scattered reads)速度更快，效率更高，因此我们看到的大部分为direct path read等待事件。
Oracle 11g提供我们选择权，这个选择权可以通过隐含参数来控制，就是"_serial_direct_read"，此参数默认值为auto

SQL>createtablespacechenjch_tbs datafile'/u01/app/oracle/oradata/orcl/chenjch_tbs01a.dbf'size10M autoextendonmaxsize1G;

SQL>createuserchenjch identifiedbya defaulttablespacechenjch_tbs;

SQL>grantconnect,resource,dbatochenjch;

SQL>createtablet1 asselect*fromdba_objects;

SQL>insertintot1 select*fromt1;

SQL>commit;

SQL>insertintot1 select*fromt1;

SQL>commit;

SQL>insertintot1 select*fromt1;

SQL>commit;

......

SQL>

begin

dbms_workload_repository.create_snapshot();

end;

SQL>altersystemflushbuffer_cache;

SQL>altersessionsettracefile_identifier='10046';

SQL>altersessionsetevents'10046 trace name context forever, level 12';

SQL>selectobject_id,object_name,object_type fromt1 whereobject_id ='20';

SQL>altersessionsetevents'10046 trace name context off';

SQL>

begin

dbms_workload_repository.create_snapshot();

end;

SQL>

selectdistinct(m.sid),p.pid,p.tracefile

fromv$mystat m,v$session s,v$process p

wherem.sid =s.sid

ands.paddr =p.addr;

---/u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_55600_10046.trc

[oracle@dip ~]$ cp /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_55600_10046.trc .

[oracle@dip ~]$ tkprof orcl_ora_55600_10046.trc 10046_1.trc

关闭direct path read特效，全表扫描时等待事件变成db file scattered read

SQL>

selecta.ksppinm name,b.ksppstvl value,a.ksppdesc description

fromx$ksppi a,x$ksppcv b

wherea.indx =b.indx

anda.ksppinm like'_serial_direct_read';---auto

SQL>altersessionset"_serial_direct_read"=never;

SQL>

begin

dbms_workload_repository.create_snapshot();

end;

SQL>altersystemflushbuffer_cache;

SQL>altersessionsettracefile_identifier='10046';

SQL>altersessionsetevents'10046 trace name context forever, level 12';

SQL>selectobject_id,object_name,object_type fromt1 whereobject_id ='20';

SQL>altersessionsetevents'10046 trace name context off';

SQL>

begin

dbms_workload_repository.create_snapshot();

end;

SQL>

selectdistinct(m.sid),p.pid,p.tracefile

fromv$mystat m,v$session s,v$process p

wherem.sid =s.sid

ands.paddr =p.addr;

---/u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_55600_10046.trc

[oracle@dip ~]$ cp /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_55600_10046.trc 10046_2.trc

[oracle@dip ~]$ tkprof 10046_2.trc 10046_2a.trc

四：db file scattered read官方文档

https://docs.oracle.com/cd/E11882_01/server.112/e40402/waitevents003.htm#BGGIBDJI

db file scattered read

Similar todb file sequential read, except that the session is reading multiple data blocks.

Wait Time:The wait time is the actual time it takes to do all of the I/Os

https://docs.oracle.com/cd/E11882_01/server.112/e41573/instance_tune.htm#PFGRF94479

10.3.2db file scattered read

This event signifies that the user process is reading buffers into the SGA buffer cache and is waiting for a physical I/O call to return. Adb file scattered readissues a scattered read to read the data into multiple discontinuous memory locations. A scattered read is usually a multiblock read. It can occur for a fast full scan (of an index) in addition to a full table scan.

Thedbfilescatteredreadwait event identifies that a full scan is occurring. When performing a full scan into thebuffer cache, the blocks read are read into memory locations that are not physically adjacent to each other. Such reads are called scattered read calls, because the blocks are scattered throughout memory. This is why the corresponding wait event is called 'db file scattered read'. multiblock (up toDB_FILE_MULTIBLOCK_READ_COUNTblocks) reads due to full scans into the buffer cache show up as waits for 'db file scattered read'.

Check the followingV$SESSION_WAITparameter columns:

P1: The absolute file number

P2: The block being read

P3: The number of blocks (should be greater than 1)

10.3.2.2Managing Excessive I/O

There are several ways to handle excessive I/O waits. In the order of effectiveness, these are as follows:

Reduce the I/O activity by SQL tuning.

Reduce the need to do I/O by managing the workload.

Gather system statistics withDBMS_STATSpackage, allowing the query optimizer to accurately cost possible access paths that use full scans.

Use Automatic Storage Management.

Add more disks to reduce the number of I/Os for each disk.

Alleviate I/O hot spots by redistributing I/O across existing disks.

The first course of action should be to find opportunities to reduce I/O. Examine the SQL statements being run by sessions waiting for these events and statements causing high physical I/Os fromV$SQLAREA. Factors that can adversely affect the execution plans causing excessive I/O include the following:

Improperly optimized SQL

Missing indexes

High degree of parallelism for the table (skewing the optimizer toward scans)

Lack of accurate statistics for the optimizer

Setting the value forDB_FILE_MULTIBLOCK_READ_COUNTinitialization parameter too high which favors full scans

10.3.2.3Inadequate I/O Distribution

Besides reducing I/O, also examine the I/O distribution of files across the disks. Is I/O distributed uniformly across the disks, or are there hot spots on some disks? Are the number of disks sufficient to meet the I/O needs of the database?

See the total I/O operations (reads and writes) by the database, and compare those with the number of disks used. Remember to include the I/O activity of LGWR and ARCH processes.

10.3.2.4Finding the SQL Statement executed by Sessions Waiting for I/O

Use the following query to determine, at a point in time, which sessions are waiting for I/O:

SELECT SQL_ADDRESS, SQL_HASH_VALUE

FROM V$SESSION

WHERE EVENT LIKE 'db file%read';

10.3.2.5Finding the Object Requiring I/O

To determine the possible causes, first queryV$SESSIONto identify the value ofROW_WAIT_OBJ#when the session waits fordbfilescatteredread. For example:

SELECT row_wait_obj#

FROM V$SESSION

WHERE EVENT = 'db file scattered read';

To identify the object and object type contended for, queryDBA_OBJECTSusing the value forROW_WAIT_OBJ#that is returned fromV$SESSION. For example:

SELECT owner, object_name, subobject_name, object_type

FROM DBA_OBJECTS

WHERE data_object_id = &row_wait_obj;

https://support.oracle.com/epmos/faces/DocContentDisplay?_afrLoop=162624307195503&id=34558.1&_afrWindowMode=0&_adf.ctrl-state=1a6zu7mmvj_153

WAITEVENT: "db file scattered read" Reference Note (文档 ID 34558.1)

修改时间:

2018-1-17

类型:

REFERENCE

***Checked for relevance on 05-Aug-2015***

"db file scattered read" Reference Note

This is a reference note for the wait event"db file scattered read"which includes the following subsections:

·Brief definition

·Individual wait details(eg: For waits seen inV$SESSION_WAIT)

·Systemwide wait details(eg: For waits seen inV$SYSTEM_EVENT)

·Reducing waits / wait times

·Troubleshooting

·Known Bugs

SeeNote:61998.1for an introduction to Wait Events.

Definition:

·Versions:7.0 - 12.1Documentation:12.111.211.110.210.1

·This wait happens when a session is waiting for a multiblock IO to complete. This typically occurs during FULL TABLE SCANs or INDEX FAST FULL SCANs. Oracle reads up to DB_FILE_MULTIBLOCK_READ_COUNT consecutive blocks at a time and scatters them into buffers in the buffer cache. How this is done depends on the platform and the release of Oracle you are running. It may also vary depending on the device type being read from and the number of blocks requested.

Individual Waits:

Parameters:

P1 = file#

P2 = block#

P3 = blocks

file#This is the file# of the file that Oracle is trying to read

from. In Oracle8 onwards it is the ABSOLUTE file number (AFN).

block#This is the starting block number in the file from where

Oracle starts reading the blocks.

See Note:181306.1to determine the tablespace, filename

and object for this file#,block#pair.

blocksThis parameter specifies the number of blocks that Oracle is

trying to read from the file# starting at block#.

The upper limit is DB_FILE_MULTIBLOCK_READ_COUNT, which is self

tuned from Oracle 10.2 onwards.

Wait Time:

The wait blocks until all blocks in the IO request have been read.

Note than an Oracle read request to the OS may be satisfied from an

OS file system cache and so may not incur a disk IO.

Systemwide Waits:

IO is a normal activity so you are really interested in unnecessary or slow IO activity.

If the TIME spent waiting for multiblock reads is significant then determine which segments/objects Oracle is performing the reads against. See the "Tablespace IO", and "File IO" sections of the AWR (or STATSPACK) reports, along with ADDM and ASH output. These should show which tablespaces / files are servicing the most IO requests, and give an indication of the speed of the IO subsystem. Tablespaces / files involved in "db file scattered read" waits will have"Av Blks/Rd" > 1.

The files where the reads are occuring can also be found by looking atV$FILESTATwhereBLKS_READ / READS > 1. (A ratio greater than 1 indicates there are some multiblock reads occuring).

See the "Top SQL by Disk Reads" sections of AWR reports for clues about any SQL causing high I/O. If statistics gathering is enabled then V$SQL_PLAN can also give clues about SQL statements using FULL scans.

It can sometimes be useful to see which sessions are performing scans and trace them to see if the scans are expected or not. This statement can be used to see which sessions are incurring waits:

SELECT sid, total_waits, time_waited

FROM v$session_event

WHERE event='db file scattered read'

and total_waits>0

ORDER BY 3,2

;

One can also look at:

·Statements with highDISK_READSinV$SQL- shown in the "Top SQL by Disk Reads" section of AWR.

·Sessions with high"table scans blocks gotten"inV$SESSTAT

Reducing Waits / Wait times:

Ideally you do not want to repeatedly perform full scans in online portions of an application when there is a faster more selective way to get at the data - in this case query tuning should be used to optimize the online SQL.In non online portions of an application table scanning is much more likely to be required. The main steps for tuning IO waits are described inNote:223117.1. Some specific points for "db file scattered read" waits include:

·Tuning of SQL usually gives the largest gains

·Consider partitioning to reduce the amount of data you need to scan

·Are affected objects sparsely populated? If so consider shrinking them

·Consider Advanced Compression to reduce the number of blocks that need to be visited

·Careful use of multiple buffer pools and the CACHE option might help.

Troubleshooting

See the following documents for help troubleshooting issues relating to "db file scattered read" waits:

Document:1475785.1Resolving Issues Where Application Queries are Waiting Too Often for 'db file scattered read' OperationsDocument:1476092.1Resolving Issues Where 'db file scattered read' Waits are Seen Due to IO Performance ProblemsDocument:223117.1Troubleshooting I/O Related WaitsDocument:1275596.1How to Tell if the I/O of the Database is Slow

Known Issues / Bugs:

You can restrict the list below to issues likely to affect one of the following versions by clicking the relevant button:

NB

Prob

Bug

Fixed

Description

II

25150925

12.1.0.2.180116, 18.1

Extreme Performance Degradation on Cursor-Duration Temporary Table (WITH Clause)

P

II

13400729

11.2.0.4, 12.1.0.1

increased elapsed time on "db file scattered read" in IBM AIX

I

6452766

10.2.0.4.3, 10.2.0.5, 11.2.0.1

10046 trace does not always show the correct "obj#" value in the trace

II

5376783

10.2.0.4, 11.1.0.6

DBMS_SPACE.OBJECT_GROWTH_TREND can perform excessive IO

·'*'indicates that an alert exists for that issue.

·'+'indicates a particularly notable issue / bug.

·SeeNote:1944526.1for details of other symbols used

Related:

Tracing User sessionsNote:62160.1<<Parameter:DB_FILE_MULTIBLOCK_READ_COUNT>>

欢迎关注我的微信公众号"IT小Chen"，共同学习，共同成长！！！