解決 oracle 11G 出現的ORA-01555 問題

oracle我很不熟, 但在公司使用exp指令備份oracle 11g資料時, 遇到以下錯誤訊息

EXP-00056: ORACLE error 1555 encountered
ORA-01555: snapshot too old: rollback segment number  with name "" too small
ORA-22924: snapshot too old

爬了文, 解法有三, 我從最簡單的解法開始, 第一種,第二種都失敗, 直到第三種方式才成功.
我建議從第三種方式開始解,該方法不需修改系統設定, 影響較小.

• 方法一
  修改 undo_retention秒數, 把時間拉大,
  **我是用備份的方式測試, 發現改了之後還是一樣出錯

SQL>  show parameter undo;

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
undo_management                      string      AUTO
undo_retention                       integer     21600
undo_tablespace                      string      UNDOTBS1
SQL> ALTER SYSTEM SET UNDO_RETENTION = 43200;
SQL>
System altered.

SQL>

• 方法二
  直接加大undo tablespace ,
  這要小心點, 有很多眉角, 似乎牽涉到小心舊undo tablesapce 是否已呈現”OFFLINE”, 且新的undo tablespace 呈現”ONLINE’

# 新增undo tablespace 
create undo tablespace UNDOTBS2 datafile '/.../oradata/undotbs02.dbf' size 30000M autoextend on next 100m maxsize unlimited;
#生效
alter system set undo_tablespace = UNDOTBS2 scope=both;

#查看狀態
select tablespace_name, status, count(*) from dba_rollback_segs group by tablespace_name, status; 
select status,segment_name from dba_rollback_segs where status not in ('OFFLINE') and tablespace_name='UNDOTBS1';
select tablespace_name, status, count(*) from dba_rollback_segs group by tablespace_name, status;
#確認舊的已是offline , 就可以刪掉舊的
Drop tablespace UNDOTBS1 including contents and datafiles;
#再次查看是否已變成新的undo tablespace
show parameter undo
select tablespace_name tablespace, status, sum(bytes)/1024/1024 sum_in_mb, count(*) counts
from dba_undo_extents
group by tablespace_name, status order by 1,2;

• 方法三

# 先建立一個資料表, 用來儲存有問題的資料ID(ROWID)
SQL> CREATE TABLE CORRUPTED_ROWS (CORRUPTED_ROWID ROWID, ERROR_NUMBER NUMBER);

Table created.

接下來將有錯誤資料表table schema , 欄位類型為clob , blob 欄位列出來, 問題一定在這幾個欄位之中

# 以下紅色部分請改為有問題資料表與欄位名稱, 若沒有出現錯誤, 就換下一個欄位(fieldname)試試看
SET TIMING ON
DECLARE 
   ERROR_1578  EXCEPTION;
   ERROR_1555  EXCEPTION;
   ERROR_22922 EXCEPTION;
   PRAGMA EXCEPTION_INIT(ERROR_1578,  -1578);
   PRAGMA EXCEPTION_INIT(ERROR_1555,  -1555);
   PRAGMA EXCEPTION_INIT(ERROR_22922, -22922);
   N NUMBER;
BEGIN
   FOR ROW IN (SELECT ROWID, fieldname FROM user.table)
   LOOP
      BEGIN
         N:=DBMS_LOB.INSTR(ROW.fieldname, HEXTORAW('889911'));
	  EXCEPTION
         WHEN ERROR_1578  THEN 
              INSERT INTO CORRUPTED_ROWS VALUES (ROW.ROWID, 1578);
			  COMMIT;
         WHEN ERROR_1555  THEN 
              INSERT INTO CORRUPTED_ROWS VALUES (ROW.ROWID, 1555);
			  COMMIT;
         WHEN ERROR_22922 THEN 
              INSERT INTO CORRUPTED_ROWS VALUES (ROW.ROWID, 22922);
			  COMMIT;
       END;
   END LOOP;
END;
/

接下來查詢是否真有問題資料

SELECT * FROM CORRUPTED_ROWS;
#執行以下這行指令應該要出錯
SELECT fieldname FROM user.table WHERE ROWID IN (SELECT CORRUPTED_ROWID FROM CORRUPTED_ROWS);

清空有問題的資料欄位

#若欄位型態是clob就用empty_clob(), 若為blob,就改成empty_blob(), 有幾個欄位出問題就清空幾個
update xxx.table set fieldname = empty_clob()where ROWID IN (SELECT CORRUPTED_ROWID FROM CORRUPTED_ROWS);

若不敢清空, 可以跳過有問題的資料備份

exp system@yourinstance BUFFER=81920 file=/tmp/backup.dmp  tables=user.table QUERY=\"WHERE rowid NOT IN \(SELECT CORRUPTED_ROWID FROM CORRUPTED_ROWS\)\"

參考資料

ORA-01555: SNAPSHOT TOO OLD ERROR BECAUSE OF LOB CORRUPTION

https://www.modb.pro/db/47021