之前有同学在我的Oracle Allstarts群里讨论关于丢失包含ocr和votedisk的ASM diskgroup导致11gR2 RAC cluster无法正常启动的问题，最早我在
《在11gR2 RAC中修改ASM DISK Path磁盘路径》一文中介绍了
，如何在不启动CRS的情况下启动11.2中的ASM实例并实施操作， 这里也需要用到同样的命令"crsctl start crs -excl -nocrs "; 注意以下的恢复手段，针对ASM中单独的ocr或者单独的votedisk丢失也有效，因为11.2中普遍把ocr和votedisk存放在ASM中，而ASM的启动又依赖于ocr和votedisk，所以在丢失ocr或votedisk仍一一者都会导致cluter无法正常启动；这里我们仅仅讨论如何让CRS正常启动，如果丢失的diskgroup中还存放有数据库的话，数据的恢复不属于本篇文章的讨论范畴。 前提：恢复的前提是你仍有和故障前一样多的ASM LUN DISK，且你有OCR的自动备份，注意默认情况下每4个小时会自动备份一次，只要你没有删除$GI_HOME，一般都会有备份可用；不要求有votedisk备份 恢复场景： 利用dd命令清空ocr和votedisk所在diskgroup header，模拟diskgroup corruption:
1. 检查votedisk和 ocr备份
[root@vrh1 ~]# crsctl query css votedisk
## STATE File Universal Id File Name Disk group
-- ----- ----------------- --------- ---------
1. ONLINE a853d6204bbc4feabfd8c73d4c3b3001 (/dev/asm-diskh) [SYSTEMDG]
2. ONLINE a5b37704c3574f0fbf21d1d9f58c4a6b (/dev/asm-diskg) [SYSTEMDG]
3. ONLINE 36e5c51ff0294fc3bf2a042266650331 (/dev/asm-diski) [SYSTEMDG]
4. ONLINE af337d1512824fe4bf6ad45283517aaa (/dev/asm-diskj) [SYSTEMDG]
5. ONLINE 3c4a349e2e304ff6bf64b2b1c9d9cf5d (/dev/asm-diskk) [SYSTEMDG]
Located 5 voting disk(s).
su - grid
[grid@vrh1 ~]$ ocrconfig -showbackup
PROT-26: Oracle Cluster Registry backup locations were retrieved from a local copy
vrh1 2012/08/09 01:59:56 /g01/11.2.0/maclean/grid/cdata/vrh-cluster/backup00.ocr
vrh1 2012/08/08 21:59:56 /g01/11.2.0/maclean/grid/cdata/vrh-cluster/backup01.ocr
vrh1 2012/08/08 17:59:55 /g01/11.2.0/maclean/grid/cdata/vrh-cluster/backup02.ocr
vrh1 2012/08/08 05:59:54 /g01/11.2.0/grid/cdata/vrh-cluster/day.ocr
vrh1 2012/08/08 05:59:54 /g01/11.2.0/grid/cdata/vrh-cluster/week.ocr
PROT-25: Manual backups for the Oracle Cluster Registry are not available
2. 彻底关闭所有节点上的clusterware ,OHASD
crsctl stop has -f
3. GetAsmDH.sh ==> GetAsmDH.sh是ASM disk header的备份脚本
请养成良好的习惯，做危险操作前备份asm header
[grid@vrh1 ~]$ ./GetAsmDH.sh
############################################
1) Collecting Information About the Disks:
############################################
SQL*Plus: Release 11.2.0.3.0 Production on Thu Aug 9 03:28:13 2012
Copyright (c) 1982, 2011, Oracle. All rights reserved.
SQL> Connected.
SQL> SQL> SQL> SQL> SQL> SQL> SQL>
1 0 /dev/asm-diske
1 1 /dev/asm-diskd
2 0 /dev/asm-diskb
2 1 /dev/asm-diskc
2 2 /dev/asm-diskf
3 0 /dev/asm-diskh
3 1 /dev/asm-diskg
3 2 /dev/asm-diski
3 3 /dev/asm-diskj
3 4 /dev/asm-diskk
SQL> SQL> Disconnected from Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
With the Real Application Clusters and Automatic Storage Management options
-rw-r--r-- 1 grid oinstall 1048 Aug 9 03:28 /tmp/HC/asmdisks.lst
############################################
2) Generating asm_diskh.sh script.
############################################
-rwx------ 1 grid oinstall 666 Aug 9 03:28 /tmp/HC/asm_diskh.sh
############################################
3) Executing asm_diskh.sh script to
generate dd dumps.
############################################
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_1_0.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_1_1.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_2_0.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_2_1.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_2_2.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_3_0.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_3_1.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_3_2.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_3_3.dd
-rw-r--r-- 1 grid oinstall 1048576 Aug 9 03:28 /tmp/HC/dsk_3_4.dd
############################################
4) Compressing dd dumps in the next format:
(asm_dd_header_all_.tar)
############################################
/tmp/HC/dsk_1_0.dd
/tmp/HC/dsk_1_1.dd
/tmp/HC/dsk_2_0.dd
/tmp/HC/dsk_2_1.dd
/tmp/HC/dsk_2_2.dd
/tmp/HC/dsk_3_0.dd
/tmp/HC/dsk_3_1.dd
/tmp/HC/dsk_3_2.dd
/tmp/HC/dsk_3_3.dd
/tmp/HC/dsk_3_4.dd
./GetAsmDH.sh: line 81: compress: command not found
ls: /tmp/HC/*.Z: No such file or directory
[grid@vrh1 ~]$
4. 使用dd 命令 破坏ocr和votedisk所在diskgroup
[root@vrh1 ~]# dd if=/dev/zero of=/dev/asm-diskh bs=1024k count=1
1+0 records in
1+0 records out
1048576 bytes (1.0 MB) copied, 0.00423853 seconds, 247 MB/s
[root@vrh1 ~]# dd if=/dev/zero of=/dev/asm-diskg bs=1024k count=1
1+0 records in
1+0 records out
1048576 bytes (1.0 MB) copied, 0.0045179 seconds, 232 MB/s
[root@vrh1 ~]# dd if=/dev/zero of=/dev/asm-diski bs=1024k count=1
1+0 records in
1+0 records out
1048576 bytes (1.0 MB) copied, 0.00469976 seconds, 223 MB/s
[root@vrh1 ~]# dd if=/dev/zero of=/dev/asm-diskj bs=1024k count=1
1+0 records in
1+0 records out
1048576 bytes (1.0 MB) copied, 0.00344262 seconds, 305 MB/s
[root@vrh1 ~]# dd if=/dev/zero of=/dev/asm-diskk bs=1024k count=1
1+0 records in
1+0 records out
1048576 bytes (1.0 MB) copied, 0.0053518 seconds, 196 MB/s
5. 在一个节点上尝试重新启动HAS
[root@vrh1 ~]# crsctl start has
CRS-4123: Oracle High Availability Services has been started.
但是因为ocr和votedisk所在diskgroup丢失，所以CSS将无法正常启动，如以下日志所示：
alertvrh1.log [cssd(5162)]CRS-1714:Unable to discover any voting files, retrying discovery in 15 seconds; Details at (:CSSNM00070:) in /g01/11.2.0/grid/log/vrh1/cssd/ocssd.log 2012-08-09 03:35:41.207 [cssd(5162)]CRS-1714:Unable to discover any voting files, retrying discovery in 15 seconds; Details at (:CSSNM00070:) in /g01/11.2.0/grid/log/vrh1/cssd/ocssd.log 2012-08-09 03:35:56.240 [cssd(5162)]CRS-1714:Unable to discover any voting files, retrying discovery in 15 seconds; Details at (:CSSNM00070:) in /g01/11.2.0/grid/log/vrh1/cssd/ocssd.log 2012-08-09 03:36:11.284 [cssd(5162)]CRS-1714:Unable to discover any voting files, retrying discovery in 15 seconds; Details at (:CSSNM00070:) in /g01/11.2.0/grid/log/vrh1/cssd/ocssd.log 2012-08-09 03:36:26.305 [cssd(5162)]CRS-1714:Unable to discover any voting files, retrying discovery in 15 seconds; Details at (:CSSNM00070:) in /g01/11.2.0/grid/log/vrh1/cssd/ocssd.log 2012-08-09 03:36:41.328 ocssd.log 2012-08-09 03:40:26.662: [ CSSD][1078700352]clssnmReadDiscoveryProfile: voting file discovery string(/dev/asm*) 2012-08-09 03:40:26.662: [ CSSD][1078700352]clssnmvDDiscThread: using discovery string /dev/asm* for initial discovery 2012-08-09 03:40:26.662: [ SKGFD][1078700352]Discovery with str:/dev/asm*: 2012-08-09 03:40:26.662: [ SKGFD][1078700352]UFS discovery with :/dev/asm*: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diskf: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diskb: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diskj: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diskh: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diskc: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diskd: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diske: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diskg: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diski: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Fetching UFS disk :/dev/asm-diskk: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]OSS discovery with :/dev/asm*: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Handle 0xdf22a0 from lib :UFS:: for disk :/dev/asm-diskf: 2012-08-09 03:40:26.665: [ SKGFD][1078700352]Handle 0xf412a0 from lib :UFS:: for disk :/dev/asm-diskb: 2012-08-09 03:40:26.666: [ SKGFD][1078700352]Handle 0xf3a680 from lib :UFS:: for disk :/dev/asm-diskj: 2012-08-09 03:40:26.666: [ SKGFD][1078700352]Handle 0xf93da0 from lib :UFS:: for disk :/dev/asm-diskh: 2012-08-09 03:40:26.667: [ CSSD][1078700352]clssnmvDiskVerify: Successful discovery of 0 disks 2012-08-09 03:40:26.667: [ CSSD][1078700352]clssnmCompleteInitVFDiscovery: Completing initial voting file discovery 2012-08-09 03:40:26.667: [ CSSD][1078700352]clssnmvFindInitialConfigs: No voting files found 2012-08-09 03:40:26.667: [ CSSD][1078700352](:CSSNM00070:)clssnmCompleteInitVFDiscovery: Voting file not found. Retrying discovery in 15 seconds正式的恢复ocr和votedisk所在diskgroup的步骤如下： 1. 以-excl -nocrs 方式启动cluster，这将可以启动ASM实例 但不启动CRS
[root@vrh1 vrh1]# crsctl start crs -excl -nocrs
CRS-4123: Oracle High Availability Services has been started.
CRS-2672: Attempting to start 'ora.mdnsd' on 'vrh1'
CRS-2676: Start of 'ora.mdnsd' on 'vrh1' succeeded
CRS-2672: Attempting to start 'ora.gpnpd' on 'vrh1'
CRS-2676: Start of 'ora.gpnpd' on 'vrh1' succeeded
CRS-2672: Attempting to start 'ora.cssdmonitor' on 'vrh1'
CRS-2672: Attempting to start 'ora.gipcd' on 'vrh1'
CRS-2676: Start of 'ora.cssdmonitor' on 'vrh1' succeeded
CRS-2676: Start of 'ora.gipcd' on 'vrh1' succeeded
CRS-2672: Attempting to start 'ora.cssd' on 'vrh1'
CRS-2672: Attempting to start 'ora.diskmon' on 'vrh1'
CRS-2676: Start of 'ora.diskmon' on 'vrh1' succeeded
CRS-2676: Start of 'ora.cssd' on 'vrh1' succeeded
CRS-2679: Attempting to clean 'ora.cluster_interconnect.haip' on 'vrh1'
CRS-2672: Attempting to start 'ora.ctssd' on 'vrh1'
CRS-2681: Clean of 'ora.cluster_interconnect.haip' on 'vrh1' succeeded
CRS-2672: Attempting to start 'ora.cluster_interconnect.haip' on 'vrh1'
CRS-2676: Start of 'ora.ctssd' on 'vrh1' succeeded
CRS-2676: Start of 'ora.cluster_interconnect.haip' on 'vrh1' succeeded
CRS-2672: Attempting to start 'ora.asm' on 'vrh1'
CRS-2676: Start of 'ora.asm' on 'vrh1' succeeded
2.重建原ocr和votedisk所在diskgroup,注意compatible.asm必须是11.2:
[root@vrh1 vrh1]# su - grid [grid@vrh1 ~]$ sqlplus / as sysasm SQL*Plus: Release 11.2.0.3.0 Production on Thu Aug 9 04:16:58 2012 Copyright (c) 1982, 2011, Oracle. All rights reserved. Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production With the Real Application Clusters and Automatic Storage Management options SQL> create diskgroup systemdg high redundancy disk '/dev/asm-diskh','/dev/asm-diskg','/dev/asm-diski','/dev/asm-diskj','/dev/asm-diskk' ATTRIBUTE 'compatible.rdbms' = '11.2', 'compatible.asm' = '11.2';3.从ocr backup中恢复ocr并做ocrcheck检验:
[root@vrh1 ~]# ocrconfig -restore /g01/11.2.0/grid/cdata/vrh-cluster/backup00.ocr
[root@vrh1 ~]# ocrcheck
Status of Oracle Cluster Registry is as follows :
Version : 3
Total space (kbytes) : 262120
Used space (kbytes) : 3180
Available space (kbytes) : 258940
ID : 1238458014
Device/File Name : +systemdg
Device/File integrity check succeeded
Device/File not configured
Device/File not configured
Device/File not configured
Device/File not configured
Cluster registry integrity check succeeded
Logical corruption check succeeded
4. 准备恢复votedisk ，可能会遇到下面的错误:
[grid@vrh1 ~]$ crsctl replace votedisk +SYSTEMDG CRS-4602: Failed 27 to add voting file 2e4e0fe285924f86bf5473d00dcc0388. CRS-4602: Failed 27 to add voting file 4fa54bb0cc5c4fafbf1a9be5479bf389. CRS-4602: Failed 27 to add voting file a109ead9ea4e4f28bfe233188623616a. CRS-4602: Failed 27 to add voting file 042c9fbd71b54f5abfcd3ab3408f3cf3. CRS-4602: Failed 27 to add voting file 7b5a8cd24f954fafbf835ad78615763f. Failed to replace voting disk group with +SYSTEMDG. CRS-4000: Command Replace failed, or completed with errors.需要重新配置一下ASM的参数，并重启ASM:
SQL> alter system set asm_diskstring='/dev/asm*';
System altered.
SQL> create spfile from memory;
File created.
SQL> startup force mount;
ORA-32004: obsolete or deprecated parameter(s) specified for ASM instance
ASM instance started
Total System Global Area 283930624 bytes
Fixed Size 2227664 bytes
Variable Size 256537136 bytes
ASM Cache 25165824 bytes
ASM diskgroups mounted
SQL> show parameter spfile
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
spfile string /g01/11.2.0/grid/dbs/spfile+AS
M1.ora
[grid@vrh1 trace]$ crsctl replace votedisk +SYSTEMDG
CRS-4256: Updating the profile
Successful addition of voting disk 85edc0e82d274f78bfc58cdc73b8c68a.
Successful addition of voting disk 201ffffc8ba44faabfe2efec2aa75840.
Successful addition of voting disk 6f2a25c589964faabf6980f7c5f621ce.
Successful addition of voting disk 93eb315648454f25bf3717df1a2c73d5.
Successful addition of voting disk 3737240678964f88bfbfbd31d8b3829f.
Successfully replaced voting disk group with +SYSTEMDG.
CRS-4256: Updating the profile
CRS-4266: Voting file(s) successfully replaced
5. 重启has服务，检验cluster是否正常：
1 ONLINE INTERMEDIATE vrh1 FAILED OVER
本文转自maclean_007 51CTO博客，原文链接:
http://blog.51cto.com/maclean/1278445
