OracleDBWorld

How to Migrate OCR , Vote File , ASM SPFILE and ASM Password file to new Diskgroup in Oracle 19c

Posted on by N Namrata

Similar to Database, storage also comes with support timelines, and this is one of the major reason storage team will ask you to migrate asm disks from old to new. Adding a disks and removing the old disks can be done with minor change in size for external redundancy diskgroups but not for high or normal redundancy diskgroups. Chances of getting the new disks same as old disks is kind of hard. Hence while migrating OCR,VOTING diskgroup we should be extra careful. Today, we will cover how to migrate ocr,vote file, ASM spfile and asm password file to new diskgroup

Please note all these commands were tested on Oracle 19c cluster

Create New diskgroup with Disks from new storage for OCR and Voting files

Check for new disks provided from new storage

Lets get started, once disks are provided to us, we need to check if those disks are visible on ASM. I will check it on both the nodes. 

SQL> select path from v$asm_disk where header_status='CANDIDATE' order by 1;

PATH
---------------------------------
/dev/ASM/OCR_VOTE_01
/dev/ASM/OCR_VOTE_02
/dev/ASM/OCR_VOTE_03
/dev/ASM/OCR_VOTE_04
/dev/ASM/OCR_VOTE_05

SQL>

Create Diskgroup with allocated disks from new Storage

Now I will create diskgroup named OCR_VOTE_NEW with new storage disks which are appearing as Candidate for us. Please note we are going with high redundancy, you can chose based on your company policy. 

export ORACLE_SID=+ASM1
sqlplus / as sysasm
SQL> create diskgroup OCR_VOTE_NEW high redundancy disk '/dev/ASM/OCR_VOTE_01','/dev/ASM/OCR_VOTE_02','/dev/ASM/OCR_VOTE_03','/dev/ASM/OCR_VOTE_04','/dev/ASM/OCR_VOTE_05'
attribute 'compatible.rdbms'='11.2.0.0', 'compatible.asm'='19.0.0.0';

Mount the newly created diskgroup on other nodes

export ORACLE_SID=+ASM2
sqlplus / as sysasm
SQL> select * from v$asm_diskgroup where state='DISMOUNTED';
SQL> alter diskgroup OCR_VOTE_NEW mount;

Final Check

You need to check the size of diskgroup, free_mb, total_mb and diskgroup state on both the nodes – 

set lines 3000 pages 300
select * from gv$asm_diskgroup;

Change ASM SPFILE and Password file location from current Diskgroup to New Diskgroup

Change of ASM SPFILE from current diskgroup to New Diskgroup

Lets get the details of spfile location – 

$ORACLE_HOME/bin/gpnptool get

Now lets create pfile from current spfile and create spfile from same pfile.

export ORACLE_SID=+ASM1
sqlplus / as sysasm
SQL> create pfile='/tmp/init_asm.ora' from spfile;

File created.

SQL>
SQL> create spfile='+OCR_VOTE_NEW' from pfile='/tmp/init_asm.ora';

File created.

SQL>
SQL> Disconnected from Oracle Database 19c Enterprise Edition Release
19.0.0.0.0 - Production
Version 19.14.0.0.0

Now lets crosscheck the location of spfile for ASM

$ORACLE_HOME/bin/gpnptool get

Move ASM Password file from current diskgroup to New diskgroup

Check the current location of password file in configuration of asm – 

grid@oracledbworld.com:~$ srvctl config asm
ASM home: <CRS home>
Password file: +OCR_VOTE/orapwASM
Backup of Password file: +MGMT/orapwASM_backup
ASM listener: LISTENER
ASM instance count: 2
Cluster ASM listener: ASMNET1LSNR_ASM
grid@oracledbworld.com:~$

Now let’s check password file location through asmcmd – 

grid@oracledbworld.com:~$ asmcmd pwget --asm
+OCR_VOTE/orapwASM

Lets move the password file using following command – 

grid@oracledbworld.com:~$ asmcmd pwmove --asm  -f +OCR_VOTE/orapwASM +OCR_VOTE_NEW/orapwASM
moving +OCR_VOTE/orapwASM -> +OCR_VOTE_NEW/orapwASM

If you don’t use -f option then you will get ASMCMD-8028 

grid@oracledbworld.com:~$ asmcmd pwmove --asm +OCR_VOTE/orapwASM +OCR_VOTE_NEW/orapwASM
ASMCMD-8028: Password file '+OCR_VOTE/orapwASM' is associated with 'asm'
already. Use the force option.

Now check the configuration to confirm the same 

grid@oracledbworld.com:~$ srvctl config asm
ASM home: <CRS home>
Password file: +OCR_VOTE_NEW/orapwASM
Backup of Password file: +MGMT/orapwASM_backup
ASM listener: LISTENER
ASM instance count: 2
Cluster ASM listener: ASMNET1LSNR_ASM

Change OCR and Vote disks from Current diskgroup to New diskgroup

Check the current OCR diskgroup and integrity using root user

root@oracledbworld.com:~#
root@oracledbworld.com:~#
/GRIDHOME/oracle/app/product/grid/19.3.0/bin/ocrcheck
Status of Oracle Cluster Registry is as follows :
          Version                  :          4
          Total space (kbytes)     :     491684
          Used space (kbytes)      :      84956
          Available space (kbytes) :     406728
          ID                       : 1317199346
          Device/File Name         :   +REDO_01
                                     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

root@oracledbworld.com:~#

Validate the current vote disk status

root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/crsctl query css votedisk
##  STATE    File Universal Id                File Name Disk group
--  -----    -----------------                --------- ---------
  1. ONLINE   410c026e40234f78bf9fbb0da838be6f (/dev/ASM/REDO_DUB_01)
[REDO_01]
Located 1 voting disk(s).
root@oracledbworld.com:~#

 Backup of current OCR Voting disk 

Following command will take backup in same diskgroup/backuploc which is configured earlier

root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/ocrconfig -manualbackup

or

You can take backup at specific location – 

root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/ocrconfig -export <location>

Main Activity

Open CRS alert.log in one session and in another session you can execute change of diskgroup commands –

Check cluster status – 

crsctl stat res -t -init
crsctl stat res -t
crsctl check crs
crsctl check cluster -all

Lets move OCR and Vote disk from current diskgroup to New diskgroup – 

root@oracledbworld.com:~# /GRIDHOME/oracle/app/product/grid/19.3.0/bin/ocrconfig -add +OCR_VOTE_NEW
root@oracledbworld.com:~# /GRIDHOME/oracle/app/product/grid/19.3.0/bin/ocrconfig -delete +REDO_01
root@oracledbworld.com:~# /GRIDHOME/oracle/app/product/grid/19.3.0/bin/crsctl replace votedisk
+OCR_VOTE_NEW
Successful addition of voting disk 55c21209de7d4f73bf1d0a1b2bbb0b48.
Successful addition of voting disk a9de904b4d414feebf4784e1a19ae1ac.
Successful addition of voting disk 7e519858f0944ffbbf4003dc798811b8.
Successful addition of voting disk bc3a763700a24fecbf340e8dc10b5771.
Successful addition of voting disk 40ec69d53a254feebf2739743c03415d.
Successful deletion of voting disk 410c026e40234f78bf9fbb0da838be6f.
Successfully replaced voting disk group with +OCR_VOTE_NEW.
CRS-4266: Voting file(s) successfully replaced

Check the cluster registry integrity and Vote disk status to revalidate – 

root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/ocrcheck
Status of Oracle Cluster Registry is as follows :
          Version                  :          4
          Total space (kbytes)     :     491684
          Used space (kbytes)      :      84956
          Available space (kbytes) :     406728
          ID                       : 1317199346
          Device/File Name         : +OCR_VOTE_NEW
                                     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

root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/crsctl query css votedisk
##  STATE    File Universal Id                File Name Disk group
--  -----    -----------------                --------- ---------
  1. ONLINE   55c21209de7d4f73bf1d0a1b2bbb0b48 (/dev/ASM/OCR_VOTE_01)
[OCR_VOTE_NEW]
  2. ONLINE   a9de904b4d414feebf4784e1a19ae1ac (/dev/ASM/OCR_VOTE_02)
[OCR_VOTE_NEW]
  3. ONLINE   7e519858f0944ffbbf4003dc798811b8 (/dev/ASM/OCR_VOTE_03)
[OCR_VOTE_NEW]
  4. ONLINE   bc3a763700a24fecbf340e8dc10b5771 (/dev/ASM/OCR_VOTE_04)
[OCR_VOTE_NEW]
  5. ONLINE   40ec69d53a254feebf2739743c03415d (/dev/ASM/OCR_VOTE_05)
[OCR_VOTE_NEW]
Located 5 voting disk(s).
root@oracledbworld.com:~#

Keep on monitoring the CRS Alert log to ensure no errors

Cluster bounce is required

Until above steps everything is online activity

To start the ASM with new spfile and to ensure everything is working fine post our activity we will take node by node bounce of cluster services.

Please check with application team for Cluster bounce and schedule it in lean period as fluctuation can observed during cluster bounce.

Prechecks for Node 1

crsctl stat res -t -init
crsctl stat res -t
crsctl check crs
crsctl check cluster -all

Stop the database and cluster services on first node

From database owner – 

srvctl config database
srvctl status database -d <db_name> -v
srvctl stop instance -d <db_name> -i <first_instance>
srvctl status database -d <db_name> -v

From root user –

ps -ef | grep tns
ps -ef | grep pmon
ps -ef | grep d.bin
root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/crsctl stop crs -f

Monitor the crs alert log, And check if cluster services are down on node 1 and start the cluster services

ps -ef | grep d.bin
root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/crsctl start crs -wait

Post Check for node 1

crsctl stat res -t -init
crsctl stat res -t
crsctl check crs
crsctl check cluster -all

Once cluster services are started and all the services are online, start the database services from database owner 

srvctl config database
srvctl status database -d <db_name> -v
srvctl start instance -d <db_name> -i <first_instance>
srvctl status database -d <db_name> -v

Please monitor database alert log during execution of above commands.

Prechecks for Node 2

crsctl stat res -t -init
crsctl stat res -t
crsctl check crs
crsctl check cluster -all

Stop the database and cluster services on second node

From database owner – 

srvctl config database
srvctl status database -d <db_name> -v
srvctl stop instance -d <db_name> -i <second_instance>
srvctl status database -d <db_name> -v

From root user –

ps -ef | grep tns
ps -ef | grep pmon
ps -ef | grep d.bin
root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/crsctl stop crs -f

Monitor the crs alert log, And check if cluster services are down on node 2 and start the cluster services

ps -ef | grep d.bin
root@oracledbworld.com:~#/GRIDHOME/oracle/app/product/grid/19.3.0/bin/crsctl start crs -wait

Post Check for node 2

crsctl stat res -t -init
crsctl stat res -t
crsctl check crs
crsctl check cluster -all

Once cluster services are started and all the services are online, start the instance services from database owner 

srvctl config database
srvctl status database -d <db_name> -v
srvctl start instance -d <db_name> -i <second_instance>
srvctl status database -d <db_name> -v

Change of Backup location if it is pointing to old diskgroup which you want do drop 

ocrconfig -backuploc <new location>

Final Check

Once everything looks clean, dismount the old diskgroup from both the nodes – 

export ORACLE_SID=+ASM1
sqlplus / as sysasm
alter diskgroup redo01 dismount;

export ORACLE_SID=+ASM2
sqlplus / as sysasm
alter diskgroup redo01 dismount;

Wait for one day, now you can drop the diskgroup and Release the storage to system administrator and storage team.

Reference

Move OCR , Vote File , ASM SPILE to new Diskgroup (Doc ID 1638177.1)

Leave a Reply

Your email address will not be published. Required fields are marked *