Database Migration to ASM with short downtime

See how to migrate to ASM with the downtime it takes only to shutdown and restart the instance – in other words, the downtime will be in the range of only minutes on a production system. The example is done on a small Linux server using Oracle Database Enterprise Edition 11.2.0.2, but should work very similar on other platforms and with the 10g version as well.

SQL> select * from v$version;

BANNER
----------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production
PL/SQL Release 11.2.0.2.0 - Production
CORE    11.2.0.2.0      Production
TNS for Linux: Version 11.2.0.2.0 - Production
NLSRTL Version 11.2.0.2.0 - Production

SQL> select log_mode from v$database;

LOG_MODE
------------
ARCHIVELOG
SQL> select name from v$datafile union select name from v$tempfile;

NAME
---------------------------------
/home/oracle/prima/sysaux01.dbf
/home/oracle/prima/system01.dbf
/home/oracle/prima/temp01.dbt
/home/oracle/prima/undotbs01.dbf
/home/oracle/prima/users01.dbf

SQL> select name from v$controlfile;

NAME
--------------------------------
/home/oracle/prima/control01.ctl

SQL> select member from v$logfile;

MEMBER
-------------------------------
/home/oracle/prima/log_g1m1.rdo
/home/oracle/prima/log_g2m1.rdo

This is my standard demo Database. I have installed already Grid Infrastructure (Marketing name for the combination of Oracle Restart & ASM) for a standalone server. Also, I have already prepared 16 fake ‚Raw Devices‘, each of 250m in size. Yes, my system is tiny – it’s my notebook. Continuing to start the ASM instance and then create the two recommended diskgroups DATA and FRA. We have the option to use the comfortable GUI asmca (11g New Feature, it can also create  quorum failgroups meanwhile) or go with the command line:

[oracle@uhesse-pc ~]$ cat /u01/app/11.2.0/grid/dbs/init+ASM.ora
#init+ASM.ora

instance_type='asm'
asm_diskstring='/dev/raw/raw*'
remote_login_passwordfile='EXCLUSIVE'
diagnostic_dest='/u01/app/oracle/'
asm_diskgroups=data,fra

After connecting to the ASM instance and startup , which will produce an error message, because data and fra do not exist yet, we create them like this:

SQL> CREATE DISKGROUP data NORMAL REDUNDANCY
 FAILGROUP fg1 DISK
 '/dev/raw/raw1' NAME disk01,
 '/dev/raw/raw2' NAME disk02,
 '/dev/raw/raw3' NAME disk03,
 '/dev/raw/raw4' NAME disk04
 FAILGROUP fg2 DISK
 '/dev/raw/raw5' NAME disk05,
 '/dev/raw/raw6' NAME disk06,
 '/dev/raw/raw7' NAME disk07,
 '/dev/raw/raw8' NAME disk08;

Each file, placed on DATA will be mirrored across the fg1 and fg2 on the stripe layer. All drives in fg1 or in fg2 could fail without losing data.

SQL> CREATE DISKGROUP fra EXTERNAL REDUNDANCY
 DISK
 '/dev/raw/raw9'  NAME disk09,
 '/dev/raw/raw10' NAME disk10,
 '/dev/raw/raw11' NAME disk11,
 '/dev/raw/raw12' NAME disk12,
 '/dev/raw/raw13' NAME disk13,
 '/dev/raw/raw14' NAME disk14,
 '/dev/raw/raw15' NAME disk15,
 '/dev/raw/raw16' NAME disk16;

FRA has no redundancy – should one drive fail we would lose all data on FRA. Next is an Online Backup with Image Copies to the DATA diskgroup. No downtime involved.

RMAN> backup as copy database format '+DATA';

Starting backup at 01-DEC-10
using target database control file instead of recovery catalog
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=19 device type=DISK
channel ORA_DISK_1: starting datafile copy
input datafile file number=00001 name=/home/oracle/prima/system01.dbf
output file name=+DATA/prima/datafile/system.256.736599607 tag=TAG20101201T110002 RECID=1 STAMP=736599624
channel ORA_DISK_1: datafile copy complete, elapsed time: 00:00:25
channel ORA_DISK_1: starting datafile copy
input datafile file number=00002 name=/home/oracle/prima/sysaux01.dbf
output file name=+DATA/prima/datafile/sysaux.257.736599629 tag=TAG20101201T110002 RECID=2 STAMP=736599634
channel ORA_DISK_1: datafile copy complete, elapsed time: 00:00:07
channel ORA_DISK_1: starting datafile copy
input datafile file number=00003 name=/home/oracle/prima/undotbs01.dbf
output file name=+DATA/prima/datafile/undotbs1.258.736599641 tag=TAG20101201T110002 RECID=3 STAMP=736599646
channel ORA_DISK_1: datafile copy complete, elapsed time: 00:00:15
channel ORA_DISK_1: starting datafile copy
input datafile file number=00004 name=/home/oracle/prima/users01.dbf
output file name=+DATA/prima/datafile/users.259.736599655 tag=TAG20101201T110002 RECID=4 STAMP=736599655
channel ORA_DISK_1: datafile copy complete, elapsed time: 00:00:01
channel ORA_DISK_1: starting datafile copy
copying current control file
output file name=+DATA/prima/controlfile/backup.260.736599657 tag=TAG20101201T110002 RECID=5 STAMP=736599662
channel ORA_DISK_1: datafile copy complete, elapsed time: 00:00:07
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
including current SPFILE in backup set
channel ORA_DISK_1: starting piece 1 at 01-DEC-10
channel ORA_DISK_1: finished piece 1 at 01-DEC-10
piece handle=+DATA/prima/backupset/2010_12_01/nnsnf0_tag20101201t110002_0.261.736599665 tag=TAG20101201T110002 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:07
Finished backup at 01-DEC-10

Now we change some dynamic parameters to point to the new Database resp. Recovery Area:

SQL> alter system set db_recovery_file_dest_size=1800m;

System altered.

SQL> alter system set db_recovery_file_dest='+FRA';

System altered.

SQL> alter system set db_create_file_dest='+DATA';

System altered.

Archivelogs will now be created into FRA. We also put our spfile there. The ‚from memory‘ clause is an 11g New Feature.

SQL> create spfile='+DATA/spfileprima.ora' from memory;
 
 File created.

We remove our spfile from $ORACLE_HOME/dbs and replace it with a pointer to the new spfile:

SQL> host cat /u01/app/oracle/product/11.2.0/dbhome_1/dbs/initprima.ora
spfile='+DATA/spfileprima.ora'

Attention: Now we need to shutdown and restart the productive instance and have a short downtime:

SQL> startup force nomount

ORACLE instance started.

Total System Global Area  417546240 bytes
Fixed Size                  2227072 bytes
Variable Size             352322688 bytes
Database Buffers           54525952 bytes
Redo Buffers                8470528 bytes

The controlfiles should also be (mirrored) on ASM. Therefore:

SQL> alter system set control_files='+DATA','+FRA' scope=spfile;

System altered.

Restart to make the modified CONTROL_FILES parameter active:

SQL> startup force nomount
ORACLE instance started.

Total System Global Area  417546240 bytes
Fixed Size                  2227072 bytes
Variable Size             352322688 bytes
Database Buffers           54525952 bytes
Redo Buffers                8470528 bytes

We only need to restore controlfiles to the new location, switch to the new datafiles on DATA and recover the latest changes that where done since the online backup:

[oracle@uhesse-pc ~]$ rman target /

Recovery Manager: Release 11.2.0.2.0 - Production on Wed Dec 1 11:13:33 2010

Copyright (c) 1982, 2009, Oracle and/or its affiliates.  All rights reserved.

connected to target database: PRIMA (not mounted)

RMAN> restore controlfile from '/home/oracle/prima/control01.ctl';

Starting restore at 01-DEC-10
using target database control file instead of recovery catalog
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=99 device type=DISK

channel ORA_DISK_1: copied control file copy
output file name=+DATA/prima/controlfile/current.263.736600443
output file name=+FRA/prima/controlfile/current.256.736600443
Finished restore at 01-DEC-10

RMAN> alter database mount;

database mounted
released channel: ORA_DISK_1

RMAN> switch database to copy;

datafile 1 switched to datafile copy "+DATA/prima/datafile/system.256.736599607"
datafile 2 switched to datafile copy "+DATA/prima/datafile/sysaux.257.736599629"
datafile 3 switched to datafile copy "+DATA/prima/datafile/undotbs1.258.736599641"
datafile 4 switched to datafile copy "+DATA/prima/datafile/users.259.736599655"

RMAN> recover database;

Starting recover at 01-DEC-10
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=99 device type=DISK

starting media recovery
media recovery complete, elapsed time: 00:00:01

Finished recover at 01-DEC-10

RMAN> alter database open;

Downtime is over! End users can connect while we do some additional work:

SQL> select name from v$datafile union select name from v$tempfile;

NAME
--------------------------------------------------------------------------------
+DATA/prima/datafile/sysaux.257.736599629
+DATA/prima/datafile/system.256.736599607
+DATA/prima/datafile/undotbs1.258.736599641
+DATA/prima/datafile/users.259.736599655
/home/oracle/prima/temp01.dbt

The tempfile was not touched by RMAN during the backup or switch to copy. We need to do that manually:

SQL> alter database  tempfile '/home/oracle/prima/temp01.dbt' drop;

Database altered.

SQL> alter tablespace temp add tempfile size 50m;

Tablespace altered.

SQL> select name from v$datafile union select name from v$tempfile;

NAME
--------------------------------------------------------------------------------
+DATA/prima/datafile/sysaux.257.736599629
+DATA/prima/datafile/system.256.736599607
+DATA/prima/datafile/undotbs1.258.736599641
+DATA/prima/datafile/users.259.736599655
+DATA/prima/tempfile/temp.264.736600915

Everything nice. But our Online Logs are still on the filesystem:

SQL> select member from v$logfile;

MEMBER
-----------------------------------
/home/oracle/prima/log_g1m1.rdo
/home/oracle/prima/log_g2m1.rdo

This can also be fixed online:

SQL> alter database add logfile size 20m;
 
 Database altered.
 
 SQL> alter database add logfile size 20m;
 
 Database altered.

Gave us two new groups mirrored across DATA and FRA. Now we drop the old groups:

SQL> alter system switch logfile;
 
 System altered.
 
 SQL> alter system switch logfile;
 
 System altered.
 
 SQL> alter system checkpoint;
 
 System altered.
 
 SQL> alter database drop logfile group 1;
 
 Database altered.
 
 SQL> alter database drop logfile group 2;
 
 Database altered.
 
 SQL> select member from v$logfile;
 
 MEMBER
 --------------------------------------------------------------------------------
 +DATA/prima/onlinelog/group_3.264.736612757
 +FRA/prima/onlinelog/group_3.257.736612759
 +DATA/prima/onlinelog/group_4.265.736612765
 +FRA/prima/onlinelog/group_4.258.736612769

That was it. We may now do a backup of the database to FRA:

RMAN> backup database;

Starting backup at 01-DEC-10
using target database control file instead of recovery catalog
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=105 device type=DISK
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
input datafile file number=00001 name=+DATA/prima/datafile/system.256.736608851
input datafile file number=00002 name=+DATA/prima/datafile/sysaux.257.736608873
input datafile file number=00003 name=+DATA/prima/datafile/undotbs1.258.736608885
input datafile file number=00004 name=+DATA/prima/datafile/users.259.736608899
channel ORA_DISK_1: starting piece 1 at 01-DEC-10
channel ORA_DISK_1: finished piece 1 at 01-DEC-10
piece handle=+FRA/prima/backupset/2010_12_01/nnndf0_tag20101201t145715_0.259.736613837 tag=TAG20101201T145715 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:16
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
including current control file in backup set
including current SPFILE in backup set
channel ORA_DISK_1: starting piece 1 at 01-DEC-10
channel ORA_DISK_1: finished piece 1 at 01-DEC-10
piece handle=+FRA/prima/backupset/2010_12_01/ncsnf0_tag20101201t145715_0.260.736613853 tag=TAG20101201T145715 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:01
Finished backup at 01-DEC-10

Conclusion: It is relatively simple and only needs a short downtime to migrate your Database to ASM, mainly the process is an RMAN Online Backup with Image Copies.

Addendum: See here for a recent example by Pete Sharman about using Enterprise Manager 12c to do an ASM Migration.