Oracle10g裸设备数据文件恢复，ORA-19501、ORA-27063

Oracle10g 裸设备数据文件恢复，ORA-19501、ORA-27063

环境说明：
OS:aix 6.1
DB:Oracle 10gR2 裸设备

使用rman备份整个数据库备份失败，，提示71数据文件失败，单独备份这个数据文件
$ rman target /

Recovery Manager: Release 10.2.0.1.0 - Production on Thu Nov 7 09:14:31 2013

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

connected to target database: ORCL (DBID=1237792397, not open)

RMAN> backup as copy datafile 71 format '/backup_dir/in71_%U';

Starting backup at 07-NOV-13
using target database control file instead of recovery catalog
allocated channel: ORA_DISK_1
channel ORA_DISK_1: sid=1086 devtype=DISK
channel ORA_DISK_1: starting datafile copy
input datafile fno=00071 name=/dev/rdata_03
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03009: failure of backup command on ORA_DISK_1 channel at 11/07/2013 09:18:10
ORA-19501: read error on file "/dev/rdata_03", blockno 178305 (blocksize=8192)
ORA-27063: number of bytes read/written is incorrect
IBM AIX RISC System/6000 Error: 5: I/O error
Additional information: -1
Additional information: 1048576

RMAN> backup check logical validate datafile 71;

Starting backup at 07-NOV-13
using channel ORA_DISK_1
channel ORA_DISK_1: starting full datafile backupset
channel ORA_DISK_1: specifying datafile(s) in backupset
input datafile fno=00071 name=/dev/rdata_03
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03009: failure of backup command on ORA_DISK_1 channel at 11/07/2013 09:47:03
ORA-19501: read error on file "/dev/rdata_03", blockno 178305 (blocksize=8192)
ORA-27063: number of bytes read/written is incorrect
IBM AIX RISC System/6000 Error: 5: I/O error
Additional information: -1
Additional information: 1048576

查询坏块个数
SQL> select * from v$database_block_corruption;

no rows selected

没有查询到逻辑坏块，猜测为物理坏块，数据库层面还认为是好的数据库造成了读的时候提示I/O error

检查71号文件的数据块个数
SQL> select bytes/8192 from v$datafile where file#=71;

BYTES/8192
----------
1280000


验证裸设备是否可以读
$ dbv file=/dev/rdata_03 blocksize=8192

DBVERIFY: Release 10.2.0.1.0 - Production on Thu Nov 7 09:48:08 2013

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

DBVERIFY - Verification starting : FILE = /dev/rdata_03

DBV-00102: File I/O error on FILE (/dev/rdata_03) during verification read operation (-2)

找出坏块，使用Oracle自带的工具
$ dbv file=/dev/rdata_03 blocksize=8192

DBVERIFY: Release 10.2.0.1.0 - Production on Thu Nov 7 09:48:08 2013

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

DBVERIFY - Verification starting : FILE = /dev/rdata_03

DBV-00102: File I/O error on FILE (/dev/rdata_03) during verification read operation (-2)
同时在另一个窗口中查看
$ ps -ef | grep dbv
oracle 598170 593980 38 09:57:42 pts/1 0:00 dbv file=/dev/rdata_03 blocksize=8192
oracle 1712282 581750 0 09:57:46 pts/0 0:00 grep dbv
$ truss -faieo /tmp/strace.out -p 598170

598170: klseek(5, 746045440, 0, 0x0FFFFFFFFFFF9530) = 0
...skipping...
598170: kread(5, " ( 0\011 ???D 5 ??.., 8192) = 8192
598170: klseek(5, 1461190656, 0, 0x0FFFFFFFFFFF9530) = 0
598170: kread(5, " ( 0\011 ???D 5 ??.., 8192) Err#5 EIO
598170: kwrite(2, "\n D B V - 0 0 1 0 2 : ", 12) = 12
598170: kwrite(2, " F i l e I / O e r r".., 24) = 24
598170: kwrite(2, " / d e v / r d a t a _ c".., 15) = 15
598170: kwrite(2, " ) d u r i n g ", 9) = 9
598170: kwrite(2, " v e r i f i c a t i o n".., 17) = 17
598170: kwrite(2, " o p e r a t i o n (", 12) = 12
598170: kwrite(2, " -", 1) = 1
598170: kwrite(2, " 2", 1) = 1
598170: kwrite(2, " )", 1) = 1
598170: kwrite(2, "\n", 1) = 1
598170: close(5) = 0
598170: close(4) = 0
598170: kfcntl(1, F_GETFL, 0x00000000101A2087) = 67110914
598170: kfcntl(2, F_GETFL, 0x00000000101A2087) = 67110914
598170: _exit(1)
...skipping...
598170: kread(5, " ( 0\011 ???D 5 ??.., 8192) = 8192
598170: klseek(5, 1461190656, 0, 0x0FFFFFFFFFFF9530) = 0
598170: kread(5, " ( 0\011 ???D 5 ??.., 8192) Err#5 EIO
598170: kwrite(2, "\n D B V - 0 0 1 0 2 : ", 12) = 12
598170: kwrite(2, " F i l e I / O e r r".., 24) = 24
598170: kwrite(2, " / d e v / r d a t a _ c".., 15) = 15
598170: kwrite(2, " ) d u r i n g ", 9) = 9
598170: kwrite(2, " v e r i f i c a t i o n".., 17) = 17
598170: kwrite(2, " o p e r a t i o n (", 12) = 12
598170: kwrite(2, " -", 1) = 1
598170: kwrite(2, " 2", 1) = 1
598170: kwrite(2, " )", 1) = 1
598170: kwrite(2, "\n", 1) = 1
598170: close(5) = 0
598170: close(4) = 0
598170: kfcntl(1, F_GETFL, 0x00000000101A2087) = 67110914
598170: kfcntl(2, F_GETFL, 0x00000000101A2087) = 67110914
598170: _exit(1)
找到开始的坏块
1461190656/8192=178368
说明第一个坏块为178368，验证最后的坏块，最大的数据块是1280000

$ dbv file=/dev/rdata_03 blocksize=8192 start=1461190656

DBVERIFY: Release 10.2.0.1.0 - Production on Thu Nov 7 10:10:58 2013

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

DBV-00102: File I/O error on FILE (/dev/rdata_03) during start read operation (-2)

$ dbv file=/dev/rdata_03 blocksize=8192 start=194719

DBVERIFY: Release 10.2.0.1.0 - Production on Thu Nov 7 10:52:56 2013

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

DBVERIFY - Verification starting : FILE = /dev/rdata_03

DBVERIFY - Verification complete