3 Configuring Storage for Oracle Grid Infrastructure and Oracle RAC

3.1.1 Overview of Oracle Clusterware and Oracle RAC Storage Options

There are two ways of storing Oracle Clusterware files:

• Oracle Automatic Storage Management (Oracle ASM): You can install Oracle Clusterware files (Oracle Cluster Registry and voting disks) in Oracle ASM disk groups.

  Oracle ASM is the required database storage option for Typical installations, and for Standard Edition Oracle RAC installations. It is an integrated, high-performance database file system and disk manager for Oracle Clusterware and Oracle Database files. It performs striping and mirroring of database files automatically.

  Only one Oracle ASM instance is permitted for each node regardless of the number of database instances on the node.

• A supported shared file system: Supported file systems include the following:

  • A supported cluster file system, such as a Sun QFS shared file system. Note that if you intend to use a cluster file system for your data files, then you should create partitions large enough for the database files when you create partitions for Oracle Clusterware.

    See Also:

    The Certify page on My Oracle Support for supported cluster file systems

  • Network File System (NFS): Note that if you intend to use NFS for your data files, then you should create partitions large enough for the database files when you create partitions for Oracle Grid Infrastructure. NFS mounts differ for software binaries, Oracle Clusterware files, and database files.

    Note:

    You can no longer use OUI to install Oracle Clusterware or Oracle Database files on block or raw devices.

    See Also:

    My Oracle Support for supported file systems and NFS or NAS filers

3.1.2 General Information About Oracle ACFS

Oracle Automatic Storage Management Cluster File System (Oracle ACFS) provides a general purpose file system. You can place Oracle Database binaries on this system, but you cannot place Oracle data files or Oracle Clusterware files on Oracle ACFS.

Note the following about Oracle ACFS:

• Oracle Restart does not support root-based Oracle Clusterware resources. For this reason, the following restrictions apply if you run Oracle ACFS on an Oracle Restart Configuration

  • You must manually load and unload Oracle ACFS drivers.

  • You must manually mount and unmount Oracle ACFS file systems, after the Oracle ASM instance is running

  • You can place Oracle ACFS database home file systems into the Oracle ACFS mount registry, along with other registered Oracle ACFS file systems.

• You cannot put Oracle Clusterware binaries and files on Oracle ACFS.

• Oracle ACFS provides a general purpose file system for other files.

3.1.3 General Storage Considerations for Oracle Grid Infrastructure and Oracle RAC

For all installations, you must choose the storage option to use for Oracle Grid Infrastructure (Oracle Clusterware and Oracle ASM), and Oracle RAC databases. To enable automated backups during the installation, you must also choose the storage option to use for recovery files (the Fast Recovery Area). You do not have to use the same storage option for each file type.

3.1.4 Guidelines for Using Oracle ASM Disk Groups for Storage

During Oracle Grid Infrastructure installation, you can create one disk group. After the Oracle Grid Infrastructure installation, you can create additional disk groups using ASMCA, SQL*Plus, or ASMCMD. Note that with Oracle Database 11g release 2 (11.2) and later releases, Oracle Database Configuration Assistant (DBCA) does not have the functionality to create disk groups for Oracle ASM.

If you install Oracle Database or Oracle RAC after you install Oracle Grid Infrastructure, then you can either use the same disk group for database files, OCR, and voting disk files, or you can use different disk groups. If you create multiple disk groups before installing Oracle RAC or before creating a database, then you can decide to do one of the following:

• Place the data files in the same disk group as the Oracle Clusterware files.

• Use the same Oracle ASM disk group for data files and recovery files.

• Use different disk groups for each file type.

If you create only one disk group for storage, then the OCR and voting disk files, database files, and recovery files are contained in the one disk group. If you create multiple disk groups for storage, then you can choose to place files in different disk groups.

3.1.5 Using Logical Volume Managers with Oracle Grid Infrastructure and Oracle RAC

Oracle Grid Infrastructure and Oracle RAC only support cluster-aware volume managers. This means, the volume manager that you want to use comes with a certain vendor cluster solution. To confirm that a volume manager you want to use is supported, look under the Certifications tab on My Oracle Support whether the associated cluster solution is certified for Oracle RAC. My Oracle Support is available at the following URL:

https://support.oracle.com

3.1.6 Supported Storage Options

The following table shows the storage options supported for storing Oracle Clusterware and Oracle RAC files.

Use the following guidelines when choosing storage options:

• You can choose any combination of the supported storage options for each file type provided that you satisfy all requirements listed for the chosen storage options.

• You can use Oracle ASM 11g release 2 (11.2) and later to store Oracle Clusterware files. You cannot use prior Oracle ASM releases to do this.

• If you do not have a storage option that provides external file redundancy, then you must configure at least three voting disk locations and at least three Oracle Cluster Registry locations to provide redundancy.

3.1.7 After You Have Selected Disk Storage Options

When you have determined your disk storage options, configure shared storage:

• To use a file system, refer to "Shared File System Storage Configuration".

• To use Oracle Automatic Storage Management, refer to "Oracle Automatic Storage Management Storage Configuration".

3.2.1 Requirements for Using a Shared File System

To use a shared file system for Oracle Clusterware, Oracle ASM, and Oracle RAC, the file system must comply with the following requirements:

• To use a cluster file system, it must be a supported cluster file system. Refer to My Oracle Support (https://support.oracle.com) for a list of supported cluster file systems.

• To use an NFS file system, it must be on a supported NAS device. Log in to the following URL, and click the Certification tab to find the most current information about supported NAS devices:

  https://support.oracle.com/

• If you choose to place your Oracle Cluster Registry (OCR) files on a shared file system, then Oracle recommends that one of the following is true:

  • The disks used for the file system are on a highly available storage device, (for example, a RAID device).

  • At least two file systems are mounted, and use the features of Oracle Clusterware 11g release 2 (11.2) to provide redundancy for the OCR.

• If you choose to place your database files on a shared file system, then one of the following should be true:

  • The disks used for the file system are on a highly available storage device, (for example, a RAID device).

  • The file systems consist of at least two independent file systems, with the database files on one file system, and the recovery files on a different file system.

• The user account with which you perform the installation (oracle or grid) must have write permissions to create the files in the path that you specify.

Use Table 3-2 and Table 3-3 to determine the minimum size for shared file systems:

In Table 3-2 and Table 3-3, the total required volume size is cumulative. For example, to store all Oracle Clusterware files on the shared file system with normal redundancy, you should have at least 2 GB of storage available over a minimum of three volumes (three separate volume locations for the OCR and two OCR mirrors, and one voting disk on each volume). You should have a minimum of three physical disks, each at least 500 MB, to ensure that voting disks and OCR files are on separate physical disks. If you add Oracle RAC using one volume for database files and one volume for recovery files, then you should have at least 3.5 GB available storage over two volumes, and at least 5.5 GB available total for all volumes.

3.2.2 Checking UDP Parameter Settings

The User Data Protocol (UDP) parameter settings define the amount of send and receive buffer space for sending and receiving datagrams over an IP network. These settings affect cluster interconnect transmissions. If the buffers set by these parameters are too small, then incoming UDP datagrams can be dropped due to insufficient space, which requires send-side retransmission. This can result in poor cluster performance.

On Oracle Solaris 10, the UDP parameters are udp_recv_hiwat and udp_xmit_hiwat. The default values for these parameters are 57344 bytes. Oracle recommends that you set these parameters to at least 65536 bytes.

To check current settings for udp_recv_hiwat and udp_xmit_hiwat, enter the following commands:

# ndd /dev/udp udp_xmit_hiwat
# ndd /dev/udp udp_recv_hiwat

To set the values of these parameters to 65536 bytes in current memory, enter the following commands:

# ndd -set /dev/udp udp_xmit_hiwat 65536
# ndd -set /dev/udp udp_recv_hiwat 65536

To set the UDP values for when the system restarts, the ndd commands have to be included in a system startup script. For example, The following script in /etc/rc2.d/S99ndd sets the parameters:

ndd -set /dev/udp udp_xmit_hiwat 65536 
ndd -set /dev/udp udp_recv_hiwat 65536

On Oracle Solaris 11, the UDP parameters are send_buf and recv_buf.

To check current settings for send_buf and recv_buf, enter the following commands:

# ipadm show-prop -p send_buf udp
# ipadm show-prop -p recv_buf udp

To set the values of these parameters to 65536 bytes in current memory, enter the following commands:

# ipadm set-prop -p send_buf=65536 udp 
# ipadm set-prop -p recv_buf=65536 udp

3.2.3 Deciding to Use a Cluster File System for Oracle Clusterware Files

For new installations, Oracle recommends that you use Oracle Automatic Storage Management (Oracle ASM) to store voting disk and OCR files.

3.2.4 Deciding to Use Direct NFS Client for Data Files

Direct NFS Client is an alternative to using kernel-managed NFS. This section contains the following information about Direct NFS Client:

• About Direct NFS Client Storage

• Using the oranfstab File with Direct NFS Client

• Mounting NFS Storage Devices with Direct NFS Client

• Specifying Network Paths with the Oranfstab File

3.2.5 Deciding to Use NFS for Data Files

Network-attached storage (NAS) systems use NFS to access data. You can store data files on a supported NFS system.

NFS file systems must be mounted and available over NFS mounts before you start installation. Refer to your vendor documentation to complete NFS configuration and mounting.

Be aware that the performance of Oracle software and databases stored on NAS devices depends on the performance of the network connection between the Oracle server and the NAS device.

For this reason, Oracle recommends that you connect the server to the NAS device using a private dedicated network connection, which should be Gigabit Ethernet or better.

3.2.6 Configuring Storage NFS Mount and Buffer Size Parameters

If you are using NFS for the Grid home or Oracle RAC home, then you must set up the NFS mounts on the storage so that they allow root on the clients mounting to the storage to be considered root instead of being mapped to an anonymous user, and allow root on the client server to create files on the NFS filesystem that are owned by root.

On NFS, you can obtain root access for clients writing to the storage by enabling no_root_squash on the server side. For example, to set up Oracle Clusterware file storage in the path /vol/grid, with nodes node1, node 2, and node3 in the domain mycluster.example.com, add a line similar to the following to the /etc/exports file:

/vol/grid/ node1.mycluster.example.com(rw,no_root_squash)
node2.mycluster.example.com(rw,no_root_squash) node3.mycluster.example.com
(rw,no_root_squash) 

If the domain or DNS is secure so that no unauthorized system can obtain an IP address on it, then you can grant root access by domain, rather than specifying particular cluster member nodes:

For example:

/vol/grid/ *.mycluster.example.com(rw,no_root_squash)

Oracle recommends that you use a secure DNS or domain, and grant root access to cluster member nodes using the domain, as using this syntax allows you to add or remove nodes without the need to reconfigure the NFS server.

If you use Grid Naming Service (GNS), then the subdomain allocated for resolution by GNS within the cluster is a secure domain. Any server without a correctly signed Grid Plug and Play (GPnP) profile cannot join the cluster, so an unauthorized system cannot obtain or use names inside the GNS subdomain.

After changing /etc/exports, reload the file system mount using the following command:

# /usr/sbin/exportfs -avr

3.2.7 Checking NFS Mount and Buffer Size Parameters for Oracle Clusterware

On the cluster member nodes, you must set the values for the NFS buffer size parameters rsize and wsize to 32768.

The NFS client-side mount options for Oracle Grid Infrastructure binaries are:

rw,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,noac,vers=3,suid 0 0

If you have Oracle Grid Infrastructure binaries on an NFS mount, then you must include the suid option.

The NFS client-side mount options for Oracle Clusterware files (OCR and voting disk files) are:

rw,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,vers=3,noac,forcedirectio

Update the /etc/vfstab file on each node with an entry containing the NFS mount options for your platform. For example, if your platform is x86-64, and you are creating a mount point for Oracle Clusterware files, then update the /etc/vfstab files with an entry similar to the following:

nfs_server:/vol/grid - /u02/oracle/cwfiles nfs \ 
rw,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,vers=3,noac,forcedirectio

Note that mount point options are different for Oracle software binaries, Oracle Clusterware files (OCR and voting disks), and data files.

To create a mount point for binaries only, provide an entry similar to the following for a binaries mount point:

nfs_server:/vol/bin /u02/oracle/grid nfs \
rw,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,noac,vers=3,suid 0 0

3.2.8 Checking NFS Mount and Buffer Size Parameters for Oracle RAC

If you use NFS mounts, then you must mount NFS volumes used for storing database files with special mount options on each node that has an Oracle RAC instance. When mounting an NFS file system, Oracle recommends that you use the same mount point options that your NAS vendor used when certifying the device. Refer to your device documentation or contact your vendor for information about recommended mount-point options.

Update the /etc/vfstab file on each node with an entry similar to the following:

nfs_server:/vol/DATA/oradata  /u02/oradata     nfs\   
rw,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,noac,forcedirectio, vers=3,suid 0 0

The mandatory mount options comprise the minimum set of mount options that you must use while mounting the NFS volumes. These mount options are essential to protect the integrity of the data and to prevent any database corruption. Failure to use these mount options may result in the generation of file access errors. Refer to your operating system or NAS device documentation for more information about the specific options supported on your platform.

https://support.oracle.com

3.2.9 Enabling Direct NFS Client Oracle Disk Manager Control of NFS

Complete the following procedure to enable Direct NFS Client:

1. Create an oranfstab file with the following attributes for each NFS server to be accessed using Direct NFS Client:

   • Server: The NFS server name.

   • Local: Up to four paths on the database host, specified by IP address or by name, as displayed using the ifconfig command run on the database host.

   • Path: Up to four network paths to the NFS server, specified either by IP address, or by name, as displayed using the ifconfig command on the NFS server.

   • Export: The exported path from the NFS server.

   • Mount: The corresponding local mount point for the exported volume.

   • Mnt_timeout: Specifies (in seconds) the time Direct NFS Client should wait for a successful mount before timing out. This parameter is optional. The default timeout is 10 minutes (600).

   • Dontroute: Specifies that outgoing messages should not be routed by the operating system, but instead sent using the IP address to which they are bound.

   The examples that follow show three possible NFS server entries in oranfstab. A single oranfstab can have multiple NFS server entries.

   Example 3-1 Using Local and Path NFS Server Entries

   The following example uses both local and path. Since they are in different subnets, we do not have to specify dontroute.

   server: MyDataServer1
   local: 192.0.2.0
   path: 192.0.2.1
   local: 192.0.100.0
   path: 192.0.100.1
   export: /vol/oradata1 mount: /mnt/oradata1
   

   Example 3-2 Using Local and Path in the Same Subnet, with dontroute

   The following example shows local and path in the same subnet. dontroute is specified in this case:

   server: MyDataServer2
   local: 192.0.2.0
   path: 192.0.2.128
   local: 192.0.2.1
   path: 192.0.2.129
   dontroute
   export: /vol/oradata2 mount: /mnt/oradata2
   

   Example 3-3 Using Names in Place of IP Addresses, with Multiple Exports

   server: MyDataServer3
   local: LocalPath1
   path: NfsPath1
   local: LocalPath2
   path: NfsPath2
   local: LocalPath3
   path: NfsPath3
   local: LocalPath4
   path: NfsPath4
   dontroute
   export: /vol/oradata3 mount: /mnt/oradata3
   export: /vol/oradata4 mount: /mnt/oradata4
   export: /vol/oradata5 mount: /mnt/oradata5
   export: /vol/oradata6 mount: /mnt/oradata6
   

2. By default, Direct NFS Client is installed in a disabled state. To enable Direct NFS Client, complete the following steps on each node. If you use a shared Grid home for the cluster, then complete the following steps in the shared Grid home:

   1. Log in as the Oracle Grid Infrastructure installation owner.

   2. Change directory to Grid_home/rdbms/lib.

   3. Enter the following commands:

      $ make -f ins_rdbms.mk dnfs_on
      

3.2.10 Creating Directories for Oracle Clusterware Files on Shared File Systems

Use the following instructions to create directories for Oracle Clusterware files. You can also configure shared file systems for the Oracle Database and recovery files.

To create directories for the Oracle Clusterware files on separate file systems from the Oracle base directory, follow these steps:

1. If necessary, configure the shared file systems to use and mount them on each node.

   Note:

   The mount point that you use for the file system must be identical on each node. Ensure that the file systems are configured to mount automatically when a node restarts.

2. Use the df command to determine the free disk space on each mounted file system.

3. From the display, identify the file systems to use. Choose a file system with a minimum of 600 MB of free disk space (one OCR and one voting disk, with external redundancy).

   If you are using the same file system for multiple file types, then add the disk space requirements for each type to determine the total disk space requirement.

4. Note the names of the mount point directories for the file systems that you identified.

5. If the user performing installation (typically, grid or oracle) has permissions to create directories on the storage location where you plan to install Oracle Clusterware files, then OUI creates the Oracle Clusterware file directory.

   If the user performing installation does not have write access, then you must create these directories manually using commands similar to the following to create the recommended subdirectories in each of the mount point directories and set the appropriate owner, group, and permissions on the directory. For example, where the user is oracle, and the Oracle Clusterware file storage area is cluster:

   # mkdir /mount_point/cluster
   # chown oracle:oinstall /mount_point/cluster
   # chmod 775 /mount_point/cluster
   

   Note:

   After installation, directories in the installation path for the Oracle Cluster Registry (OCR) files should be owned by root, and not writable by any account other than root.

When you have completed creating a subdirectory in the mount point directory, and set the appropriate owner, group, and permissions, you have completed NFS configuration for Oracle Grid Infrastructure.

3.2.11 Creating Directories for Oracle Database Files on Shared File Systems

Use the following instructions to create directories for shared file systems for Oracle Database and recovery files (for example, for an Oracle RAC database).

1. If necessary, configure the shared file systems and mount them on each node.

   Note:

   The mount point that you use for the file system must be identical on each node. Ensure that the file systems are configured to mount automatically when a node restarts.

2. Use the df -h command to determine the free disk space on each mounted file system.

3. From the display, identify the file systems:

   File Type	File System Requirements
   Database files	Choose either: A single file system with at least 1.5 GB of free disk space. Two or more file systems with at least 1.5 GB of free disk space in total.
   Recovery files	Choose a file system with at least 2 GB of free disk space.

   
   

   If you are using the same file system for multiple file types, then add the disk space requirements for each type to determine the total disk space requirement.

4. Note the names of the mount point directories for the file systems that you identified.

5. If the user performing installation (typically, oracle) has permissions to create directories on the disks where you plan to install Oracle Database, then DBCA creates the Oracle Database file directory, and the Recovery file directory.

   If the user performing installation does not have write access, then you must create these directories manually using commands similar to the following to create the recommended subdirectories in each of the mount point directories and set the appropriate owner, group, and permissions on them:

   • Database file directory:

     # mkdir /mount_point/oradata
     # chown oracle:oinstall /mount_point/oradata
     # chmod 775 /mount_point/oradata
     

   • Recovery file directory (Fast Recovery Area):

     # mkdir /mount_point/fast_recovery_area
     # chown oracle:oinstall /mount_point/fast_recovery_area
     # chmod 775 /mount_point/fast_recovery_area
     

By making members of the oinstall group owners of these directories, this permits them to be read by multiple Oracle homes, including those with different OSDBA groups.

When you have completed creating subdirectories in each of the mount point directories, and set the appropriate owner, group, and permissions, you have completed NFS configuration for Oracle Database shared storage.

3.2.12 Disabling Direct NFS Client Oracle Disk Management Control of NFS

Complete the following steps to disable Direct NFS Client:

1. Log in as the Oracle Grid Infrastructure installation owner, and disable Direct NFS Client using the following commands, where Grid_home is the path to the Oracle Grid Infrastructure home:

   $ cd Grid_home/rdbms/lib
   $ make -f ins_rdbms.mk dnfs_off
   

   Enter these commands on each node in the cluster, or on the shared Grid home if you are using a shared home for the Oracle Grid Infrastructure installation.

2. Remove the oranfstab file.

3.3.1 Configuring Storage for Oracle Automatic Storage Management

This section describes how to configure storage for use with Oracle Automatic Storage Management (Oracle ASM).

• Identifying Storage Requirements for Oracle ASM

• Creating Files on a NAS Device for Use with Oracle ASM

• Using an Existing Oracle ASM Disk Group

• Configuring Disk Devices for Oracle ASM

3.3.2 Using Disk Groups with Oracle Database Files on Oracle ASM

Review the following sections to configure Oracle Automatic Storage Management (Oracle ASM) storage for Oracle Clusterware and Oracle Database Files:

• Identifying and Using Existing Oracle Database Disk Groups on Oracle ASM

• Creating Disk Groups for Oracle Database Data Files

3.3.3 Configuring Oracle Automatic Storage Management Cluster File System

Oracle ACFS is installed as part of an Oracle Grid Infrastructure installation (Oracle Clusterware and Oracle Automatic Storage Management) for 11g release 2 (11.2). You can configure Oracle ACFS for a database home, or use ASMCA to configure ACFS as a general purpose file system.

To configure Oracle ACFS for an Oracle Database home for an Oracle RAC database:

1. Install Oracle Grid Infrastructure for a cluster (Oracle Clusterware and Oracle Automatic Storage Management)

2. Change directory to the Oracle Grid Infrastructure home. For example:

   $ cd /u01/app/11.2.0/grid
   

3. Ensure that the Oracle Grid Infrastructure installation owner has read and write permissions on the storage mount point you want to use. For example, if you want to use the mount point /u02/acfsmounts/:

   $ ls -l /u02/acfsmounts
   

4. Start Oracle ASM Configuration Assistant as the grid installation owner. For example:

   ./asmca
   

5. The Configure ASM: ASM Disk Groups page shows you the Oracle ASM disk group you created during installation. Click the ASM Cluster File Systems tab.

6. On the ASM Cluster File Systems page, right-click the Data disk, then select Create ACFS for Database Home.

7. In the Create ACFS Hosted Database Home window, enter the following information:

   • Database Home ADVM Volume Device Name: Enter the name of the database home. The name must be unique in your enterprise. For example: dbase_01

   • Database Home Mount Point: Enter the directory path for the mount point. For example: /u02/acfsmounts/dbase_01

     Make a note of this mount point for future reference.

   • Database Home Size (GB): Enter in gigabytes the size you want the database home to be.

   • Database Home Owner Name: Enter the name of the Oracle Database installation owner you plan to use to install the database. For example: oracle1

   • Database Home Owner Group: Enter the OSDBA group whose members you plan to provide when you install the database. Members of this group are given operating system authentication for the SYSDBA privileges on the database. For example: dba1

   • Click OK when you have completed your entries.

8. Run the script generated by Oracle ASM Configuration Assistant as a privileged user (root). On an Oracle Clusterware environment, the script registers the ACFS as a resource managed by Oracle Clusterware. Registering ACFS as a resource helps Oracle Clusterware to mount the ACFS automatically in proper order when ACFS is used for an Oracle RAC database Home.

9. During Oracle RAC installation, ensure that you or the DBA who installs Oracle RAC selects for the Oracle home the mount point you provided in the Database Home Mountpoint field (in the preceding example, /u02/acfsmounts/dbase_01).

3.3.4 Upgrading Existing Oracle ASM Instances

If you have an Oracle ASM installation from a prior release installed on your server, or in an existing Oracle Clusterware installation, then you can use Oracle Automatic Storage Management Configuration Assistant (ASMCA, located in the path Grid_home/bin) to upgrade the existing Oracle ASM instance to 11g release 2 (11.2), and subsequently configure failure groups, Oracle ASM volumes and Oracle Automatic Storage Management Cluster File System (Oracle ACFS).

During installation, if you are upgrading from an Oracle ASM release prior to 11.2, and you chose to use Oracle ASM and ASMCA detects that there is a prior Oracle ASM version installed in another Oracle ASM home, then after installing the Oracle ASM 11g release 2 (11.2) binaries, you can start ASMCA to upgrade the existing Oracle ASM instance. You can then choose to configure an Oracle ACFS deployment by creating Oracle ASM volumes and using the upgraded Oracle ASM to create the Oracle ACFS.

If you are upgrading from Oracle ASM 11g release 2 (11.2.0.1) or later, then Oracle ASM is always upgraded with Oracle Grid Infrastructure as part of the rolling upgrade, and ASMCA is started by the root scripts during upgrade. ASMCA cannot perform a separate upgrade of Oracle ASM from release 11.2.0.1 to 11.2.0.2.

On an existing Oracle Clusterware or Oracle RAC installation, if the prior version of Oracle ASM instances on all nodes is 11g release 1, then you are provided with the option to perform a rolling upgrade of Oracle ASM instances. If the prior version of Oracle ASM instances on an Oracle RAC installation are from a release prior to 11g release 1, then rolling upgrades cannot be performed. Oracle ASM on all nodes will be upgraded to 11g release 2 (11.2).