详述Redo日志中操作系统块头信息损坏的处理过程

墨墨导读：前几天某客户遇到这个问题：文件系统损坏导致Current redo log异常，最终恢复过程比较简单，这里不再累述。本文详述redo log  os header block损坏的处理过程，希望对大家有帮助。

如下是简单测试过程：

SQL> select * from v$Log;
GROUP#    THREAD#  SEQUENCE#      BYTES  BLOCKSIZE    MEMBERS ARCHIV STATUS                           FIRST_CHANGE# FIRST_TIME   NEXT_CHANGE# NEXT_TIME        CON_ID
---------- ---------- ---------- ---------- ---------- ---------- ------ -------------------------------- ------------- ------------ ------------ ------------ ----------
1          1         21  209715200        512          1 NO     INACTIVE                              13356440 12-MAY-20        13356445 12-MAY-20             0
2          1         22  209715200        512          1 NO     CURRENT                               13356445 12-MAY-20      9.2954E+18                       0
3          1         19  209715200        512          1 NO     INACTIVE                              13356426 12-MAY-20        13356435 12-MAY-20             0
4          1         20  104857600        512          1 NO     INACTIVE                              13356435 12-MAY-20        13356440 12-MAY-20             0
SQL> select member from v$Logfile;
MEMBER
--------------------------------------------------
/opt/oracle/oradata/ENMOTECH/redo03.log
/opt/oracle/oradata/ENMOTECH/redo02.log
/opt/oracle/oradata/ENMOTECH/redo01.log
/opt/oracle/oradata/ENMOTECH/redo04.log
SQL> shutdown abort;
ORACLE instance shut down.
SQL> host
[oracle@mysqldb1 ~]$ cp /opt/oracle/oradata/ENMOTECH/redo02.log /opt/oracle/oradata/ENMOTECH/redo02.log.bak

这里我们分别dump一下redo log的os block：

[oracle@mysqldb1 ~]$ dd if=/opt/oracle/oradata/ENMOTECH/redo02.log bs=512 count=1 | od -x
1+0 records in
1+0 records out
512 bytes (512 B) copied, 4.6621e-05 s, 11.0 MB/s
0000000 2200 0000 0000 ffc0 0000 0000 0000 0400
0000020 9dc3 0000 0200 0000 4000 0006 7c7d 7a7b
0000040 0003 0000 0000 0000 0000 0000 0000 0000
0000060 0000 0000 0000 0000 0000 0000 0000 0000
*
0001000
[oracle@mysqldb1 ~]$ dd if=/opt/oracle/oradata/ENMOTECH/redo01.log bs=512 count=1 | od -x
1+0 records in
1+0 records out
512 bytes (512 B) copied, 6.8299e-05 s, 7.5 MB/s
0000000 2200 0000 0000 ffc0 0000 0000 0000 0400
0000020 9dc3 0000 0200 0000 4000 0006 7c7d 7a7b
0000040 0003 0000 0000 0000 0000 0000 0000 0000
0000060 0000 0000 0000 0000 0000 0000 0000 0000
*
0001000

由于这2个redo log文件大小完全一致࿰