分类目录归档:ORA-xxxxx

物理备库在read only时报ORA-01552错误处理

物理备库在read only时报ORA-01552错误

Tue Jan 06 11:53:38 中国标准时间 2015
alter database open read only
Tue Jan 06 11:53:38 中国标准时间 2015
SMON: enabling cache recovery
Tue Jan 06 11:53:39 中国标准时间 2015
Database Characterset is ZHS16GBK
Opening with internal Resource Manager plan 
replication_dependency_tracking turned off (no async multimaster replication found)
Physical standby database opened for read only access.
Completed: alter database open read only
Tue Jan 06 11:54:04 中国标准时间 2015
Errors in file c:\oracle\product\10.2.0\admin\ntsy\udump\ntsy_ora_9080.trc:
ORA-00604: 递归 SQL 级别 1 出现错误
ORA-01552: 非系统表空间 'MY_SPACE' 不能使用系统回退段
ORA-06512: 在 line 2

分析trace文件

*** ACTION NAME:() 2015-01-06 11:54:04.828
*** MODULE NAME:(sqlplus.exe) 2015-01-06 11:54:04.828
*** SERVICE NAME:(SYS$USERS) 2015-01-06 11:54:04.828
*** SESSION ID:(1284.9) 2015-01-06 11:54:04.828
Error in executing triggers on connect internal
*** 2015-01-06 11:54:04.828
ksedmp: internal or fatal error
ORA-00604: 递归 SQL 级别 1 出现错误
ORA-01552: 非系统表空间 'MY_SPACE' 不能使用系统回退段
ORA-06512: 在 line 2
*** 2015-01-06 11:54:05.843
Process diagnostic dump for ORACLE.EXE (MMNL), OS id=10492,
pid: 13, proc_ser: 1, sid: <no session> 

这里可以看出来,是由于执行触发器导致该问题,根据经验第一感觉很可能是logon之类的触发器导致。

查询触发器

SQL> select trigger_name,trigger_type,OWNER from dba_triggers where owner='OP';

TRIGGER_NAME                   TRIGGER_TYPE     OWNER
------------------------------ ---------------- ------------------------------
LOGAD                          AFTER EVENT      OP
TR_TRACE_DDL                   AFTER EVENT      OP

只有这两个触发器是基于事件的,另外从名字和dba_source中确定

SQL> select text from dba_source where name='LOGAD';

TEXT
--------------------------------------------------------------------------------

TRIGGER "OP".logad after logon on database

begin
insert into logad values (SYS_CONTEXT('USERENV', 'SESSION_USER'), SYSDATE,SYS_CO

NTEXT('USERENV','IP_ADDRESS')) ;

end;


已选择6行。

SQL> select text from dba_source where name='TR_TRACE_DDL';

TEXT
--------------------------------------------------------------------------------

TRIGGER "OP".tr_trace_ddl
AFTER ddl
ON database

DECLARE
sql_text ora_name_list_t;
state_sql ddl$trace.ddl_sql%TYPE;
BEGIN
FOR i IN 1..ora_sql_txt(sql_text) LOOP
state_sql := state_sql||sql_text(i);
END LOOP;

TEXT
--------------------------------------------------------------------------------

INSERT INTO ddl$trace(login_user,audsid,machine,ipaddress,
schema_user,schema_object,ddl_time,ddl_sql)
VALUES(ora_login_user,userenv('SESSIONID'),sys_context('userenv','host'),
sys_context('userenv','ip_address'),ora_dict_obj_owner,ora_dict_obj_name,SYSDATE

,state_sql);

EXCEPTION
WHEN OTHERS THEN
-- sp_write_log('捕获DDL语句异常错误:'||SQLERRM);
null;
END tr_trace_ddl;

基本上确定LOGAD是登录触发器,tr_trace_ddl是记录ddl触发器,那现在问题应该出在LOGAD的触发器上.因为该触发器在备库上当有用户登录之时,他也会工作插入记录到logad表中,由于数据库是只读,因此就出现了类似ORA-01552错误

解决方法
在触发器中加判断数据库角色条件,当数据库为物理备库之时才执行dml操作

SQL> CREATE OR REPLACE TRIGGER "OP".logad
  2  AFTER LOGON on database
  3  declare
  4  db_role varchar2(30);
  5  begin
  6  select database_role into db_role from v$database;
  7  If db_role <> 'PHYSICAL STANDBY' then
  8  insert into op.logad values (SYS_CONTEXT('USERENV', 'SESSION_USER'), 
  9  SYSDATE,SYS_CONTEXT('USERENV','IP_ADDRESS')) ;
 10  end if;
 11  end;
 12  /

Warning: Trigger created with compilation errors.

SQL> show error;
Errors for TRIGGER "OP".logad:

LINE/COL ERROR
-------- -----------------------------------------------------------------
4/1      PL/SQL: SQL Statement ignored
4/40     PL/SQL: ORA-00942: table or view does not exist
SQL> conn / as sysdba
Connected.
SQL> grant select on v_$database to op;

Grant succeeded.

SQL> CREATE OR REPLACE TRIGGER "OP".logad
  2  AFTER LOGON on database
  3  declare
  4  db_role varchar2(30);
  5  begin
  6  select database_role into db_role from v$database;
  7  If db_role <> 'PHYSICAL STANDBY' then
  8  insert into op.logad values (SYS_CONTEXT('USERENV', 'SESSION_USER'),
  9 SYSDATE,SYS_CONTEXT('USERENV','IP_ADDRESS')) ;
 10  end if;
 12  end;
 12  /

Trigger created.

数据库open正常

Completed: ALTER DATABASE RECOVER MANAGED STANDBY DATABASE cancel
Tue Jan 06 13:51:20 中国标准时间 2015
alter database open read only
Tue Jan 06 13:51:21 中国标准时间 2015
SMON: enabling cache recovery
Tue Jan 06 13:51:21 中国标准时间 2015
Database Characterset is ZHS16GBK
Opening with internal Resource Manager plan 
replication_dependency_tracking turned off (no async multimaster replication found)
Physical standby database opened for read only access.
Tue Jan 06 13:51:23 中国标准时间 2015
db_recovery_file_dest_size of 102400 MB is 0.00% used. This is a
user-specified limit on the amount of space that will be used by this
database for recovery-related files, and does not reflect the amount of
space available in the underlying filesystem or ASM diskgroup.
Tue Jan 06 13:51:23 中国标准时间 2015
Completed: alter database open read only
发表在 Data Guard, ORA-xxxxx | 标签为 | 评论关闭

升级数据库到10.2.0.5遭遇ORA-00918: column ambiguously defined

一个数据库从10201升级到10205之后,出现ORA-00918错误,查询mos发现在以前版本中是bug,Oracle好像在10205中把它修复了,结果就是以前应用的sql无法正常执行.这次升级的结果就是客户晚上3点联系开发商紧急修改程序。再次提醒:再小的系统数据库升级都需要做,功能测试,SPA测试,确保升级后功能和性能都正常.

SQL> select * from v$version;

BANNER
----------------------------------------------------------------
Oracle Database 10g Enterprise Edition Release 10.2.0.5.0 - 64bi
PL/SQL Release 10.2.0.5.0 - Production
CORE    10.2.0.5.0      Production
TNS for 64-bit Windows: Version 10.2.0.5.0 - Production
NLSRTL Version 10.2.0.5.0 - Production

执行报错ORA-00918
多个表JOIN连接,由于在select中的列未指定表名,而且该列在多个表中有,因此在10205中报ORA-00918错误,Oracle认为在以前的版本中是 Bug 5368296: SQL NOT GENERATING ORA-918 WHEN USING JOIN. 升级到10.2.0.5, 11.1.0.7 and 11.2.0.2版本,需要注意此类问题。修复bug没事,但是修复了之后导致系统需要修改sql才能够运行,确实让人很无语

SQL> set autot trace
SQL> set lines 100
SQL> SELECT   yz_id, item_code, DECODE (yzlx, 0, '长期医嘱', '临时医嘱') yzlx,
  2             item_name, gg, sl || sldw sl, zyjs, yf, a.pc, zbj, zbh,
  3             TO_CHAR (dcl, 'fm9999990.009') || dcldw dcl, a.bz, lb, zyh,ch,xm,
  4             bq, cfh, lrysdm, lrysxm, lrrq, hdrdm, hdrxm, hdrq, sender_code,
  5             sender_name, send_date, tzysdm, tzysxm, tzrq, ksrq, zxfy,
  6             lb_yp_yl, zsq_code
  7        FROM op.yz a LEFT OUTER JOIN op.pc b
  8             ON NVL (TRIM (UPPER (a.pc)), ' ') = NVL (TRIM (UPPER (b.pc)), ' ')
  9             LEFT JOIN op.zy p ON a.zyh = p.zyh
 10       WHERE p.cy='在院' AND p.new_patient='1'
 11         AND upper(nvl(p.bj,1))<> 'Y'
 12         AND (state = '已核对')
 13         AND is_in_bill IS NULL
 14    ORDER BY ksrq, yz_id ;
           bq, cfh, lrysdm, lrysxm, lrrq, hdrdm, hdrxm, hdrq, sender_code,
           *
ERROR at line 4:
ORA-00918: column ambiguously defined


SQL> select COLUMN_NAME,TABLE_NAME from DBA_tab_columns where column_name='BQ'
  2  AND TABLE_NAME IN('YZ','ZY','PC');

COLUMN_NAME                    TABLE_NAME
------------------------------ ------------------------------
BQ                             ZY
BQ                             YZ

10.2.0.1中执行正常

E:\>sqlplus / as sysdba

SQL*Plus: Release 10.2.0.1.0 - Production on 星期六 1月 3 14:09:51 2015

Copyright (c) 1982, 2005, Oracle.  All rights reserved.


连接到:
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - 64bit Production
With the Partitioning, OLAP and Data Mining options

SQL> set autot trace
SQL> set lines 100
SQL> SELECT   yz_id, item_code, DECODE (yzlx, 0, '长期医嘱', '临时医嘱') yzlx,
  2             item_name, gg, sl || sldw sl, zyjs, yf, a.pc, zbj, zbh,
  3             TO_CHAR (dcl, 'fm9999990.009') || dcldw dcl, a.bz, lb, zyh,ch,xm
,
  4             bq, cfh, lrysdm, lrysxm, lrrq, hdrdm, hdrxm, hdrq, sender_code,
  5             sender_name, send_date, tzysdm, tzysxm, tzrq, ksrq, zxfy,
  6             lb_yp_yl, zsq_code
  7        FROM op.yz a LEFT OUTER JOIN op.pc b
  8             ON NVL (TRIM (UPPER (a.pc)), ' ') = NVL (TRIM (UPPER (b.pc)), '
')
  9             LEFT JOIN op.zy p ON a.zyh = p.zyh
 10       WHERE p.cy='在院' AND p.new_patient='1'
 11         AND upper(nvl(p.bj,1))<> 'Y'
 12         AND (state = '已核对')
 13         AND is_in_bill IS NULL
 14    ORDER BY ksrq, yz_id ;

已选择19804行。


执行计划
----------------------------------------------------------
ERROR:
ORA-00604: 递归 SQL 级别 2 出现错误
ORA-16000: 打开数据库以进行只读访问


SP2-0612: 生成 AUTOTRACE EXPLAIN 报告时出错

统计信息
----------------------------------------------------------
          1  recursive calls
          0  db block gets
      41945  consistent gets
          0  physical reads
          0  redo size
    2075973  bytes sent via SQL*Net to client
      14989  bytes received via SQL*Net from client
       1322  SQL*Net roundtrips to/from client
          1  sorts (memory)
          0  sorts (disk)
      19804  rows processed

10.2.0.5库中同名列增加表名前缀执行OK

1
SQL> set autot trace
SQL> set lines 100
SQL> SELECT   yz_id, item_code, DECODE (yzlx, 0, '长期医嘱', '临时医嘱') yzlx,
  2             item_name, gg, sl || sldw sl, zyjs, yf, a.pc, zbj, zbh,
  3             TO_CHAR (dcl, 'fm9999990.009') || dcldw dcl, a.bz, lb,zyh,ch,xm,
  4             a.bq, cfh, lrysdm, lrysxm, lrrq, hdrdm, hdrxm, hdrq, sender_code,
  5             sender_name, send_date, tzysdm, tzysxm, tzrq, ksrq, zxfy,
  6             lb_yp_yl, zsq_code
  7        FROM op.yz a LEFT OUTER JOIN op.pc b
  8             ON NVL (TRIM (UPPER (a.pc)), ' ') = NVL (TRIM (UPPER (b.pc)), ' ')
  9             LEFT JOIN op.zy p ON a.zyh = p.zyh
 10       WHERE p.cy='在院' AND p.new_patient='1'
 11         AND upper(nvl(p.bj,1))<> 'Y'
 12         AND (state = '已核对')
 13         AND is_in_bill IS NULL
 14    ORDER BY ksrq, yz_id ;

20629 rows selected.


Execution Plan
----------------------------------------------------------
Plan hash value: 3468887510

--------------------------------------------------------------------------------------------
| Id  | Operation                     | Name       | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT              |            |    10 |  2580 |  2968   (2)| 00:00:36 |
|   1 |  SORT ORDER BY                |            |    10 |  2580 |  2968   (2)| 00:00:36 |
|*  2 |   HASH JOIN OUTER             |            |    10 |  2580 |  2967   (2)| 00:00:36 |
|*  3 |    TABLE ACCESS BY INDEX ROWID| YZ         |     3 |   672 |    42   (0)| 00:00:01 |
|   4 |     NESTED LOOPS              |            |    10 |  2390 |  2963   (2)| 00:00:36 |
|*  5 |      TABLE ACCESS FULL        | ZY         |     3 |    45 |  2917   (2)| 00:00:36 |
|*  6 |      INDEX RANGE SCAN         | DZBLYZ_ZYH |   118 |       |     2   (0)| 00:00:01 |
|   7 |    TABLE ACCESS FULL          | PC         |    33 |   627 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   2 - access(NVL(TRIM(UPPER("A"."PC")),' ')=NVL(TRIM(UPPER("B"."PC"(+))),' '))
   3 - filter("A"."STATE"='已核对' AND "A"."IS_IN_BILL" IS NULL)
   5 - filter("P"."CY"='在院' AND UPPER(NVL("P"."BJ",'1'))<>'Y' AND
              "P"."NEW_PATIENT"='1')
   6 - access("A"."ZYH"="P"."ZYH")


Statistics
----------------------------------------------------------
          0  recursive calls
          0  db block gets
      42121  consistent gets
          0  physical reads
          0  redo size
    2181383  bytes sent via SQL*Net to client
      15617  bytes received via SQL*Net from client
       1377  SQL*Net roundtrips to/from client
          1  sorts (memory)
          0  sorts (disk)
      20629  rows processed

Bug 5368296: SQL NOT GENERATING ORA-918 WHEN USING JOIN
Bug 12388159 : SQL REPORTING ORA00918 AFTER UPGRADE TO 10.2.0.5.0
再次提醒:再小的系统数据库升级都需要做,功能测试,SPA测试,确保升级后功能和性能都正常.

发表在 ORA-xxxxx | 标签为 | 评论关闭

查询v$session报ORA-04031错误

客户的数据库在出账期间有工具登录Oracle数据库偶尔性报ORA-04031,经过分析是因为该工具需要查询v$session,经过分析确定是Bug 12808696 – Shared pool memory leak of “hng: All sessi” memory (Doc ID 12808696.8),重现错误如下
节点1进行查询报ORA-4031

SQL> select count(*) from v$session; 

COUNT(*) 
---------- 
1536 

SQL> select count(*) from gv$session; 

COUNT(*) 
---------- 
2089 

SQL> select /*+ full(t) */ count(*) from gv$session t; 

COUNT(*) 
---------- 
2053 

SQL> select * from gv$session; 
select * from gv$session 
* 
ERROR at line 1: 
ORA-12801: error signaled in parallel query server PZ93, instance 
ocs_db_2:zjocs2 (2) 
ORA-04031: unable to allocate 308448 bytes of shared memory ("shared 
pool","unknown object","sga heap(1,0)","hng: All sessions data for API.")

节点2进行查询报ORA-04031

SQL> select * from gv$session; 
select * from gv$session 
* 
ERROR at line 1: 
ORA-12801: error signaled in parallel query server PZ95, instance 
ocs_db_2:zjocs2 (2) 
ORA-04031: unable to allocate 308448 bytes of shared memory ("shared 
pool","unknown object","sga heap(6,0)","hng: All sessions data for API.") 

SQL> select * from v$session; 
select * from v$session 
* 
ERROR at line 2: 
ORA-04031: unable to allocate 308448 bytes of shared memory ("shared 
pool","unknown object","sga heap(7,0)","hng: All sessions data for API.")

通过上述分析:确认是节点2的v$session遭遇到Bug 12808696,导致在该节点中中查询v$session和Gv$session报ORA-04031,而在节点1中查询v$session正常,查询Gv$session报ORA-04031.
bug-12808696
该bug在11.1.0.6中修复,所有的10g版本中未修复,只能通过临时重启来暂时避免,注意该bug通过flash shared_pool无法解决
如果您有权限可以进步一查询SR 3-7670890781: 查询v$session的BLOCKING_SESSION字段时,出现ora-04031错误

发表在 ORA-xxxxx | 标签为 , | 评论关闭