kccrsz: denied expansion of controlfile section *

前些时间，日志里频繁出现以下信息：

kccrsz: denied expansion of controlfile section 9 by 65535 record(s)
 the number of records is already at maximum value (65535)
krcpwnc: following controlfile record written over:
 RECID #520891 Recno 53663 Record timestamp
...
kccrsz: denied expansion of controlfile section 9 by 65535 record(s)
 the number of records is already at maximum value (65535)
krcpwnc: following controlfile record written over:
 RECID #520892 Recno 53664 Record timestamp

 

另外还有：

kccrsz: denied expansion of controlfile section 19 by 26767 record(s)
  the number of records is already at maximum value (65535)
krcpwnc: following controlfile record written over:
 RECID #768025 Recno 25227 Record timestamp
 12/18/05 00:30:54
  Object type=245  Object recid=7622  Object timestamp=
 10/22/05 22:11:19
kccrsz: denied expansion of controlfile section 19 by 26767 record(s)
  the number of records is already at maximum value (65535)
krcpwnc: following controlfile record written over:
 RECID #768026 Recno 25228 Record timestamp
 12/18/05 00:30:54
  Object type=245  Object recid=7625  Object timestamp=
 10/22/05 22:11:53

 

 

使用语句：

SQL> set pagesize 30 linesize 132
SQL> select rownum-1, crs.* from v$controlfile_record_section crs;

  ROWNUM-1 TYPE                 RECORD_SIZE RECORDS_TOTAL RECORDS_USED FIRST_INDEX LAST_INDEX LAST_RECID
---------- -------------------- ----------- ------------- ------------ ----------- ---------- ----------
         0 DATABASE                     316             1            1           0          0          0
         1 CKPT PROGRESS               4084            35            0           0          0          0
         2 REDO THREAD                  228            32            2           0          0          2
         3 REDO LOG                      72           192           10           0          0         24
         4 DATAFILE                     428          1024          387           0          0       2123
         5 FILENAME                     268          1601          412           0          0          0
         6 TABLESPACE                    68          1024           34           0          0        172
         7 TEMPORARY FILENAME            56          1024            5           0          0          5
         8 RMAN CONFIGURATION          1108            50            2           0          0          2
         9 LOG HISTORY                   36          3403         3403        2730       2729      29953
        10 OFFLINE RANGE                180          1043            0           0          0          0
        11 ARCHIVED LOG                 328          3212         3212        2797       2796      22068
        12 BACKUP SET                    40          1021         1021         180        179       5284
        13 BACKUP PIECE                 480          1003         1003         265        264       5279
        14 BACKUP DATAFILE              116          2886         2886         531        530      63959
        15 BACKUP REDOLOG                76           429            0           0          0          0
        16 DATAFILE COPY                404          1010            0           0          0          0
        17 BACKUP CORRUPTION             44          1113            0           0          0          0
        18 COPY CORRUPTION               40          1021            0           0          0          0
        19 DELETED OBJECT                20         65535        65535       14162      14161    2067659
        20 PROXY COPY                   596          1000            0           0          0          0
        21 BACKUP SPFILE                 36           226          226          63         62        288
        22 DATABASE INCARNATION          56           145            1           1          1          1

 

检查参数control_file_record_keep_time发现设置为7，

 

首先可以看到第9条记录的使用已经达到记录允许总大小的上限，也就是MAXLOGHISTORY在建库的时候只定义了3403个（control_file_record_keep_time不是0的时候，该参数自动扩展，但不会超过定义上限），用alter database backup controlfile to trace后从trace文件上可以看到该参数（不知道当时定义这么低）。其实这个信息并不影响数据库的使用，因为在该事件发生时控制文件中的记录将被覆盖重写

不让日志出现以上信息可以将control_file_record_keep_time设置为0，但会对rman的使用产生一定的危险，用户可能无法正常的restore备份。

 

第二部分信息主要发生在rman备份的时候，上面查询结果中第19条记录为DELETED OBJECT，这个参数主要是跟踪一些被删除的东西，例如：

archived logs
archived log file copies
controlfile backups
controlfile copies
datafile backups
datafile copies

使用语句

SQL> select count(*) from v$deleted_object ;

  COUNT(*)
----------
     65535

v$deleted_object视图显示的是一些控制文件中被删除的archived logs, datafile copies和backup pieces等信息，它唯一的目的就是使recovery catalog的同步操作得到优化，当archived log, datafile copy或者 backup piece被删除后，相应的记录就会标记为deleted

减少这样信息可以通过减少备份保留时间来解决，比如将recovery window 的天数减少。当备份历史有必要保留较长时间的时候，这个信息是可以忽略的。