Oracle数据库数据恢复、性能优化

找回密码
注册
搜索
热搜: 活动 交友 discuz
发新帖

4

积分

0

好友

2

主题
1#
发表于 2012-6-11 17:12:43 | 查看: 20973| 回复: 3
下面是我测试库的awr报告:
DB NameDB IdInstanceInst numStartup TimeReleaseRAC
xxxx3032207194xxxx101-Nov-11 09:1111.2.0.1.0NO
Host NamePlatformCPUsCoresSocketsMemory (GB)
oraclevm02AIX-Based Systems (64-bit)168 4.00
Cache Sizes
BeginEnd
Buffer Cache:352M352MStd Block Size:8K
Shared Pool Size:624M624MLog Buffer:6,684K
Instance Efficiency Percentages (Target 100%)
Buffer Nowait %:100.00Redo NoWait %:100.00
Buffer Hit %:99.57In-memory Sort %:100.00
Library Hit %:99.97Soft Parse %:99.72
Execute to Parse %:36.64Latch Hit %:100.00
Parse CPU to Parse Elapsd %:6.04% Non-Parse CPU:99.55

Foreground Wait Events                 DB/Inst: BDLMS/bdlms  Snaps: 1417-1425
-> s  - second, ms - millisecond -    1000th of a second
-> Only events with Total Wait Time (s) >= .001 are shown
-> ordered by wait time desc, waits desc (idle events last)
-> %Timeouts: value of 0 indicates value was < .5%.  Value of null is truly 0

                                                             Avg
                                        %Time Total Wait    wait    Waits   % DB
Event                             Waits -outs   Time (s)    (ms)     /txn   time
-------------------------- ------------ ----- ---------- ------- -------- ------
OJVM: Generic                       270    30         80     296      0.1   11.0
log file sync                       917     0          1       1      0.3     .1
Disk file operations I/O          1,425     0          0       0      0.4     .0
db file sequential read              26     0          0       4      0.0     .0
library cache: mutex X                1     0          0      68      0.0     .0
buffer busy waits                     4     0          0      13      0.0     .0
enq: JS - queue lock                  1     0          0       7      0.0     .0
read by other session                 1     0          0       5      0.0     .0
SQL*Net message to client         1,370     0          0       0      0.4     .0
asynch descriptor resize          2,024   100          0       0      0.6     .0
control file sequential re          320     0          0       0      0.1     .0
jobq slave wait                  62,750   100     31,329     499     19.3
SQL*Net message from clien        1,370     0         17      13      0.4
          -------------------------------------------------------------


请问maclean 这个OJVM: Generic是什么?文档上没找到。谢谢


[ 本帖最后由 jimmy 于 2012-6-11 17:22 编辑 ]
2#
发表于 2012-6-11 17:27:26
请上传完整的 AWR

回复 只看该作者 道具 举报

3#
发表于 2012-6-11 17:32:34
原帖由 maclean 于 2012-6-11 17:27 发表 请上传完整的 AWR
awr完整报告在附件里面

awrrpt_1_1417_1425.html (387.45 KB, 下载次数: 932)

回复 只看该作者 道具 举报

4#
发表于 2012-6-11 17:42:23
11.2.0.1.0+ AIX-Based Systems (64-bit)


resource_manager_plan        SCHEDULER[0x3006]:DEFAULT_MAINTENANCE_PLAN

AWR 快照期间 遇到了 resource_manager_plan 被 激活为 DEFAULT_MAINTENANCE_PLAN


Java execution elapsed time        661.71        90.95

Java运行耗时 661s

耗费 最多 DB TIME的是一个JOB ,该JOB可能调用oracle内的JAVA ,引发了OJVM: Generic 等待事件, 该等待事件是 JVM java 虚拟机的常规操作等待

f8f7gbzbstcyx        DECLARE job BINARY_INTEGER := :job; next_date DATE := :mydate; broken BOOLEAN := FALSE; BEGIN wb_rti_service_job.check_service(2); :mydate := next_date; IF broken THEN :b := 1; ELSE :b := 0; END IF; END;

ODM FINDING

OJVM: Generic Wait Event in v$session_wait

Applies to:
Oracle Server - Enterprise Edition - Version: 11.2.0.2 and later   [Release: 11.2 and later ]
Information in this document applies to any platform.
Symptoms
Java stored procedure hangs with wait on OJVM: Generic. This occurs only if Resource Manager is active.

SQL> select event from v$session_wait where sid = 100;

EVENT
----------------------------------------------------------------
OJVM: Generic

The errorstack placed on the session indicates high waits on this event:

---------------------------------------------------
Sampled Session History Summary:
longest_non_idle_wait: 'OJVM: Generic'
[3 samples, 09:07:20 - 09:07:22]
time_waited: >= 2 sec (still in wait)
---------------------------------------------------
Cause
Bug 12701301 - WAIT CLASSE: OTHER OJVM : GENERIC was opened for this issue.
It was closed as duplicate of :
Bug 10325230 - RUNTIME.EXEC CAN HANG WHEN USED WITH RESOURCE MANAGER ACTIVE

Solution
Bug 10325230 will be fixed in 12.1, so currently the following options are available to solve this issue.


1. The workaround from bug 12701301:
'The issue only reproduces if resource_manager_plan ='DEFAULT_MAINTENANCE_PLAN'. If resource_manager_plan is not set, then the operation completes.'

2. The workaround from bug 10325230:
'Don't use user threads to read from Process streams, or don't use them
when RM is active, or use Thread.stop or other mechanisms to actively
stop/kill the threads once the related Process has exited.'

3. Apply a patch for bug 10325230. Patches can be requested only for 11.2.0.2.0 and later. Affected code did not exist before that.



BUG 12701301 当resource_manager_plan ='DEFAULT_MAINTENANCE_PLAN' 可能触发造成OJVM: Generic 等待的 BUG


建议

1.建议你至少升级到 11.2.0.3 , 11.2.0.1 是一个极不稳定的版本

2.  如果无法升级, 且没有出现 更严重的OJVM: Generic 等待,可以忽略该问题

回复 只看该作者 道具 举报

您需要登录后才可以回帖 登录 | 注册

QQ|手机版|Archiver|Oracle数据库数据恢复、性能优化

GMT+8, 2024-12-26 10:34 , Processed in 0.059671 second(s), 24 queries .

Powered by Discuz! X2.5

© 2001-2012 Comsenz Inc.

回顶部
TEL/電話+86 13764045638
Email service@parnassusdata.com
QQ 47079569