subject part
stay Oracle in , Replace OCR What are the steps for a disk group ？
Answer section
Core commands ：
1crsctl replace votedisk +newocr 2ocrconfig -add +newocr 3ocrconfig -delete +ocr 4more /etc/oracle/ocr.loc 5spget 6spcopy -u +CRSDG/xx/asmparameterfile/registry.xx +newocr/spfileASM.ora 7spset
The specific steps include ：
1. Add new storage
establish Normal Redundancy level OCR The volume group , You need at least 3 individual LUN, Every LUN The minimum capacity is 1GB, It needs to be configured on the storage side 3 individual LUN to RAC Each node in the cluster .
2. Multipath binding configuration
To configure /etc/multipatch.conf, Add new disk UUID, Suppose the path after the multi-path binding is /dev/mapper/newocr1、/dev/mapper/newocr2、/dev/mapper/newocr3, And reload Mutilpathd Services take effect .
3. ASMLib Configuration or udev binding
node 1 The operation is as follows ：
Will add new 3 individual LUN Add to ASMLIB in
1# /usr/sbin/oracleasm createdisk OCRNEW1 /dev/mapper/newocr1 2# /usr/sbin/oracleasm createdisk OCRNEW2 /dev/mapper/newocr2 3# /usr/sbin/oracleasm createdisk OCRNEW3 /dev/mapper/newocr3
node 2：
Scan the new 3 individual LUN：
1# /usr/sbin/oracleasm scandisks 2# /usr/sbin/oracleasm listdisks
4. Backup OCR
— Logical backup recovery
1ocrconfig -export /oracle/app/11.2.0/grid/cdata/ZFTPCCDB-crs/export_asm.bak
— Physical backup recovery
1ocrconfig -manualbackup 2ocrconfig -showbackup 3ocrcheck
5. newly build OCR The volume group
1set line 9999 2set pagesize 9999 3col path format a60 4SELECT a.group_number, disk_number,mount_status, a.name, path FROM v$asm_disk a order by a.disk_number; 5set line 999 6select name,state,free_mb,required_mirror_free_mb,usable_file_mb,a.group_number, disk_number,mount_status, path from v$asm_diskgroup a; 7select a.group_number,name,TYPE,state,TOTAL_MB,free_mb from v$asm_diskgroup a; 8select name,state,free_mb,required_mirror_free_mb,usable_file_mb,a.group_number from v$asm_diskgroup a;
node 1 newly build Normal Redundancy level ASM The volume group
1sqlplus / as sysasm 2create diskgroup newocr normal redundancy disk 3'/dev/oracleasm/disks/neworc1', 4'/dev/oracleasm/disks/neworc2', 5'/dev/oracleasm/disks/neworc3' 6attribute 'compatible.asm'='11.2','compatible.rdbms'='11.2';
node 1 perform mount Disk group ：
1ALTER DISKGROUP NEWOCR MOUNT; 2crsctl stat res -t
6. Replace VOTE
At one of the nodes ,root Execute under the user :
Replace VOTE
1crsctl replace votedisk +newocr
confirm
1crsctl query css votedisk
7. Replace OCR
stay grid or root Execute under the user ：
Node1 Execute the following command :
1ocrconfig -add +newocr 2ocrconfig -delete +ocr 3more /etc/oracle/ocr.loc
Node2 You don’t have to execute the following commands ：
1ocrconfig -repair -add +newocr 2ocrconfig -repair -delete +ocr
8. transfer ASM SPFILE
stay GRID Any node under the user executes
1$ asmcmd
View the current ASMSPFILE
1ASMCMD> spget
Copy SPFILE To the new volume group
1ASMCMD> spcopy -u +CRSDG/xx/asmparameterfile/registry.xx +newocr/spfileASM.ora
Check out the new SPFILE Absolute path
1ASMCMD> ls -l +newocr/spfileASM.ora 2+NEWOCR/XX/ASMPARAMETERFILE/REGISTRY.253.833623103 3ASMCMD>spset +NEWOCR/XX/ASMPARAMETERFILE/REGISTRY.253.833623103
Modify two nodes grid Under the user $ORACLE_HOME/dbs/init.ora The contents of the document , It can also be done without modification ：
1SPFILE='+newocr/spfileASM.ora'
9. Confirm old OCR Are there any other files in the pack
10. Delete old OCR The volume group
After restarting the cluster ：
node 2：
1su - grid 2sqlplus / as sysasm 3alter diskgroup OCR dismount; 4exit
node 1：
1su - grid 2sqlplus / as sysasm 3drop diskgroup ocr; 4drop diskgroup OVDISK INCLUDING CONTENTS; 5drop diskgroup OVDISK FORCE INCLUDING CONTENTS;
11. Delete disk
Any node ROOT User down operation ：
1# /usr/sbin/oracleasm dropdisk OCR
Delete multipath binding configuration , Delete the configuration /etc/multipatch.conf Central Plains OCR Contents of the disk .
Finally delete from storage to 2 Mapping of nodes .
12. Delete CRS Record
1crsctl delete res ora.OCR.dg