oracle hanganalyze工具的使用

hanganalyze是[u][b]ORACLE[/b][/u]的一款性能诊断工具，这个款工具是从oracle 8.0.6开始可用，在oracle[u][b]数据库[/b][/u]出现严重的性能问题的时候它可以帮助你定位问题所在。



1.首先说说hanganalyze工具的用法

对于单实例数据库语法如下

alter session set events 'immediate trace name hanganalyze level <level>';

或则使用oradebug进行hanganalyze

conn /as sysdba

SQLPLUS>oradebug hanganalyze <level>;



对于[u][b]RAC[/b][/u]数据的语法如下

con /as sysda

SQLPLUS> oradebug setmypid

SQLPLUS>oradebug setinst all

SQLPLUS>oradebug -g def hanganalyze <level>





关于level的说明：

10 Dump all processes (IGN state)

5 Level 4 + Dump all processes involved in wait chains (NLEAF state)

4 Level 3 + Dump leaf nodes (blockers) in wait chains (LEAF,LEAF_NW,IGN_DMP state)

3 Level 2 + Dump only processes thought to be in a hang (IN_HANG state)

1-2 Only HANGANALYZE output, no process dump at all

#############################

[level 4] : 23 node dumps -- [LEAF] [LEAF_NW] [IGN_DMP]

[level 5] : 36 node dumps -- [NLEAF]

[level 10] : 130 node dumps -- [IGN]







2.dump文件的分析



下面是一个例子：

[oracle@SHDBService01 ~]$ more /data/oracle/admin/ora10g/udump/ora10g_ora_28378.trc

/data/oracle/admin/ora10g/udump/ora10g_ora_28378.trc

Oracle Database[u][b]10g[/b][/u]Enterprise Edition Release 10.2.0.4.0 - 64bit Production

With the Partitioning, OLAP,[u][b]Data[/b][/u]Mining and Real Application Testing options

ORACLE_HOME = /data/oracle/product/10.2.1

System name:

Linux

Node name:

SHDBService01

Release:



2.6.9-67.ELlargesmp

Version:



#1 SMP Wed Nov 7 14:07:22 EST 2007

Machine:



x86_64

Instance name: ora10g

Redo thread mounted by this instance: 1

Oracle process number: 62

Unix process pid: 28378, image: oracle@SHDBService01 (TNS V1-V3)



*** ACTION NAME:() 2009-08-21 13:36:46.238

*** MODULE NAME:(sqlplus@SHDBService01 (TNS V1-V3)) 2009-08-21 13:36:46.238

*** SERVICE NAME:(SYS$USERS) 2009-08-21 13:36:46.238

*** SESSION ID:(532.3192) 2009-08-21 13:36:46.238

*** 2009-08-21 13:36:46.238

==============

HANG ANALYSIS:

==============

Open chains found:

Chain 1 : <cnode/sid/sess_srno/proc_ptr/ospid/wait_event> :

<0/542/2126/0xd1006f28/25642/SQL*Net message from client>

-- <0/1097/44386/0xd2001048/26064/enq: TX - row lock contention>

Other chains found:

Chain 2 : <cnode/sid/sess_srno/proc_ptr/ospid/wait_event> :

<0/532/3192/0xd1007710/28378/No Wait>

Chain 3 : <cnode/sid/sess_srno/proc_ptr/ospid/wait_event> :

<0/534/3/0xd10096b0/30838/Streams AQ: waiting for time man>

Chain 4 : <cnode/sid/sess_srno/proc_ptr/ospid/wait_event> :

<0/539/3/0xd1008ec8/30830/Streams AQ: qmn coordinator idle>

Chain 5 : <cnode/sid/sess_srno/proc_ptr/ospid/wait_event> :

<0/541/2497/0xd1005f58/16409/jobq slave wait>

Chain 6 : <cnode/sid/sess_srno/proc_ptr/ospid/wait_event> :

<0/1099/3/0xd2002fe8/30840/Streams AQ: qmn slave idle wait>

Extra information that will be dumped at higher levels:

[level 4] : 1 node dumps -- [REMOTE_WT] [LEAF] [LEAF_NW]

[level 5] : 5 node dumps -- [SINGLE_NODE] [SINGLE_NODE_NW] [IGN_DMP]

[level 6] : 1 node dumps -- [NLEAF]

[level 10] : 17 node dumps -- [IGN]



State of nodes

([nodenum]/cnode/sid/sess_srno/session/ospid/state/start/finish/[adjlist]/predecessor):

[531]/0/532/3192/0xd13bd850/28378/SINGLE_NODE_NW/1/2//none

[532]/0/533/655/0xd13bedb8/19321/IGN/3/4//none

[533]/0/534/3/0xd13c0320/30838/SINGLE_NODE/5/6//none

[538]/0/539/3/0xd13c6e28/30830/SINGLE_NODE/7/8//none

[540]/0/541/2497/0xd13c98f8/16409/SINGLE_NODE/9/10//none

[541]/0/542/2126/0xd13cae60/25642/LEAF/11/12//1096

[542]/0/543/1/0xd13cc3c8/30755/IGN/13/14//none

[543]/0/544/1/0xd13cd930/30753/IGN/15/16//none

[544]/0/545/1/0xd13cee98/30751/IGN/17/18//none

[545]/0/546/1/0xd13d0400/30749/IGN/19/20//none

[546]/0/547/1/0xd13d1968/30746/IGN/21/22//none

[547]/0/548/1/0xd13d2ed0/30744/IGN/23/24//none

[548]/0/549/1/0xd13d4438/30734/IGN/25/26//none

[549]/0/550/1/0xd13d59a0/30732/IGN/27/28//none

[550]/0/551/1/0xd13d6f08/30730/IGN/29/30//none

[551]/0/552/1/0xd13d8470/30728/IGN/31/32//none

[1096]/0/1097/44386/0xd23cee98/26064/NLEAF/33/34/[541]/none

[1098]/0/1099/3/0xd23d1968/30840/SINGLE_NODE/35/36//none

[1099]/0/1100/6/0xd23d2ed0/30861/IGN/37/38//none

[1101]/0/1102/1/0xd23d59a0/30742/IGN/39/40//none

[1102]/0/1103/1/0xd23d6f08/30736/IGN/41/42//none

[1651]/0/1652/3/0xd13d3c50/30858/IGN/43/44//none

[1653]/0/1654/1/0xd13d6720/30738/IGN/45/46//none

[2204]/0/2205/1/0xd03d6720/30740/IGN/47/48//none

====================

END OF HANG ANALYSIS

====================





open chains部分例子

Chain 1 : <cnode/sid/sess_srno/proc_ptr/ospid/wait_event> :

<0/542/2126/0xd1006f28/25642/SQL*Net message from client>

-- <0/1097/44386/0xd2001048/26064/enq: TX - row lock contention>

Other chains found:



sid = Session ID

sess_srno = Serial#

proc_ptr = Process Pointer

ospid = OS Process Id

wait_event = Waitevent

cnode = Node Id (Only available since Oracle9i)



State of nodes部分例子：

[nodenum]/cnode/sid/sess_srno/session/ospid/state/start/finish/[adjlist]/predecessor

[541]/0/542/2126/0xd13cae60/25642/LEAF/11/12//1096

[1096]/0/1097/44386/0xd23cee98/26064/NLEAF/33/34/[541]/none





Nodenum = This is secuencial number used by HANGANALYZE to identify each session

sid = Session ID

sess_srno = Serial#

ospid = OS Process Id

state = State of the node

adjlist = adjacent node (Usually represents a blocker node)

predecessor = predecessor node (Usually represents a waiter node)

cnode = Node number (Only available since Oracle9i)



IN_HANG:This might be considered as the most critical STATE. Basically a node in this state is involved in a deadlock, or is hung. Usually there will be another “adjacent node” in the same status. For example:



LEAF and LEAF_NW:Leaf nodes are considered on top of the wait chain (usually blockers). They are considered “Blockers” when there is another session waiting. This can be easily identified using the “predecesor” field. If there is a node referenced in the
‘prdecessor’ field, the node is considered as “blocker”, otherwise it is considered as a “slow” session waiting for some resource.

The difference between LEAF and LEAF_NW is that LEAF nodes are not waiting for something, while LEAF_NW are not waiting or may be using the CPU





可以看出上面例子中看出



session 542 2126阻塞了session 1097 44386.







3.在很多情况下如果数据库HANG住则无法登陆sqlplus，这时如果想要对系统进行hanganalyze可以加参数登录sqlplus

具体语法如下：



[oracle@SHDBService01 ~]$ sqlplus -prelim /nolog



SQL*Plus: Release 10.2.0.4.0 - Production on Fri Aug 21 15:42:23 2009



Copyright (c) 1982, 2007, Oracle. All Rights Reserved.



SQL> conn /as sysdba

Prelim connection established

SQL>





prelim参数只对10g以后的版本有效。



10g以前的版本不能登录SQLPLUS时可以使用dbx或则gdb。