oracle 11gr2 針對ocr/vote asm diskgroup損壞的處理方法

The following procedure describe the steps to recreate the diskgroup where OCR and Voting disks are stored and restore the OCR data from a backup.

1. List the backups of the OCR

# $GI/bin/ocrconfig -showbackup

2. Validate the integrity of the backup

# $GI/bin/ocrdump -backupfile backup_file_name

3. Stop clusterware in all the nodes

# $GI/bin/crsctl stop crs [-f]

4. Before 11.2.0.2: In one node, start crs in exclusive mode

          # $GI/bin/crsctl start crs –excl

and then apply step 5

   Since 11.2.0.2: In one node, start crs in exclusive mode without crsd

         # $CRS_HOME/bin/crsctl start crs -excl -nocrs

         and then continue at step 6 (skip step 5).

IMPORTANT:
A new option '-nocrs' has been introduced with  11.2.0.2, which prevents the start of the ora.crsd resource.
It is vital that this option is specified, otherwise the failure to start the ora.crsd resource will tear down ora.cluster_interconnect.haip, which in turn will cause ASM to crash.

5. Stop crsd

# $GI/bin/crsctl stop resource ora.crsd -init

6. Create Diskgroup

sql>create diskgroup <name normal redundancy disk '<path to disks' FORCE ATTRIBUTE 'compatible.asm'='11.2.0.2.0','compatible.rdbms'='11.2.0.2.0','au_size'='4M','cell.smart_scan_capable'='TRUE';

7. Restore OCR from one of the backups

# $GI/bin/ocrconfig -restore file_name

8. Validate OCR integrity

# $GI/bin/ocrcheck

9. Validate the list of voting disks. If the diskgroup was rebuilt,it will report empty

#$GI/bin/crsctl start resource ora.crsd -init
# $GI/bin/crsctl query css votedisk

10. Add the voting disk:

# $GI/bin/crsctl replace votedisk <DISKGROUP NAME

11. Restart the CRS in normal mode

# $GI/bin/crsctl stop crs
# $GI/bin/crsctl start crs

12. Validate Voting disks and OCR

# $GRI/bin/crsctl query css votedisk
# $GI/bin/ocrcheck

13. Start cluster:

#$GRI/bin/crsctl start cluster