What You See Is What You Get Element

How to Use an Oracle Solaris 10 Zone Cluster to Easily Migrate Applications to Oracle Solaris 11 Environments

by Mahesh Subramanya

This article describes how to migrate a single instance of Oracle Database running on Oracle Solaris 10 with Oracle Solaris Cluster 3.3 into a clustered Oracle Solaris 11 environment without requiring any modification to the database, by using an Oracle Solaris 10 Zones cluster deployment.


Published February 2014


The objective of this article is to show how to migrate applications from an Oracle Solaris 10 clustered environment into an Oracle Solaris 11 environment with minimal effort. The following are the steps for this procedure:

Want to comment on this article? Post the link on Facebook's OTN Garage page.  Have a similar article to share? Bring it up on Facebook or Twitter and let's discuss.
  1. Preparing the source systems for migration
  2. Preparing the Oracle Solaris 10 Zones cluster on the target systems

    Note: The procedure in this article assumes that a two-node target cluster is already set up with the Oracle Solaris Cluster 4.1 software. See "How to Install and Configure a Two-Node Cluster" for instructions on how to configure a two-node cluster on Oracle Solaris 11.

  3. Installing the zone cluster
  4. Installing the cluster software in the zone cluster (optional)
  5. Re-creating the application setup (in this example, the database) on the target systems
  6. Verifying failover

About Oracle Solaris Zones Clusters

Oracle Solaris Zones are application containers that provide isolated and secure environments for applications. Applications that run inside an Oracle Solaris Zone are partitioned from other applications or processes that run in a different Oracle Solaris Zone. Oracle Solaris Zones differ from other virtual machine solutions because they share the Oracle Solaris kernel.

An Oracle Solaris Zones cluster is a virtual cluster that leverages the capabilities of Oracle Solaris Zones—such as security isolation, fault isolation, and flexible resource management—and extends them across a cluster of nodes, delivering high availability (HA) to the applications running inside the zones and enabling the consolidation of multiple cluster applications onto a single, global cluster.

Oracle Solaris 10 Zones use the solaris10 brand zone, which provides a complete runtime environment for Oracle Solaris 10 applications on SPARC and x86 machines running the Oracle Solaris 10 9/10 operating system or later releases. This enables you to run an Oracle Solaris 10 environment on Oracle Solaris 11. The solaris10 brand zone cluster is a key feature that was introduced in Oracle Solaris Cluster 4.1 software. It can be used to migrate Oracle Solaris 10 clustered applications that are deployed with Oracle Solaris Cluster 3.3 3/13 software to a solaris10 brand zone cluster on a cluster that runs Oracle Solaris Cluster 4.1 software. This is accomplished without any modification to the applications. Thus, this feature provides the ability to consolidate Oracle Solaris 10 applications on the latest Oracle hardware and leverage the powerful features introduced in Oracle Solaris 11 software.

The example in this article uses an Oracle Database 10g Enterprise Edition Release 10.2.0.5 deployment on a two-node cluster. The cluster runs Oracle Solaris Cluster 3.3 3/13 software, and the database is managed by the Oracle Solaris Cluster HA for Oracle (HA for Oracle) data service. The servers are connected to an Oracle ZFS Storage Appliance serving the Oracle Database data and binaries.

For further details on configuring the HA for Oracle data service, refer to the Oracle Solaris Cluster Data Service for Oracle Guide (for Oracle Solaris Cluster 3.3 or Oracle Solaris Cluster 4). For information about configuring Oracle ZFS Storage Appliance, refer to "Installing and Maintaining Oracle's Sun ZFS Storage Appliances as NAS Devices in an Oracle Solaris Cluster Environment."

The Source and Target Configurations

Figure 1 shows a view of the hardware and connectivity for the production source cluster, which is a two-node cluster called cluster-1 that has the following:

  • Two SPARC sun4u Sun Fire V240 servers (db-host-1 and db-host-2)
  • 10 gigabytes of main memory
  • Oracle Solaris 10 1/13
  • Oracle Solaris Cluster 3.3 3/13
  • HA for Oracle data service
  • Oracle Database 10g Release 2 (version 10.2.0.5)
  • Oracle's Sun ZFS Storage 7420 appliance
Figure 1

Figure 1. Source environment: hardware and connectivity.

Figure 2 shows a logical view of the source cluster, including the resource groups (rg) and resources (rs) for the HA for Oracle data service, as well as the dependencies between them.

Figure 2

Figure 2. Source environment: logical view of resources and resource groups.

Figure 3 shows a simplified view of the hardware and connectivity for the target cluster, new-phys-cluster, which is a two-node cluster with two of Oracle's Sun SPARC Enterprise T5120 servers (new-phys-host-1 and new-phys-host-2).

Figure 3

Figure 3. Target environment: hardware and connectivity.

Figure 4 shows a logical view of the target cluster running Oracle Solaris 11.1 and Oracle Solaris Cluster 4.1 software. This environment uses the solaris10 brand zone cluster to host the HA for Oracle Database 10g Release 2 configuration that is migrated from the source configuration shown in Figure 1.

Figure 4

Figure 4. Target environment: logical view.

In the logical view of the source cluster shown in Figure 2, the oracle-rg resource group contains the oracle-server-rs resource for the Oracle server instance; the logical host name resource (db-lh-rs) that manages logical host db-lh, which is used by clients to connect to the Oracle server; and the Oracle listener resource (oracle-listener-rs).

The scal-mnt-rg resource group contains the oracle-rs and oradata-rs scalable mount-point resources, which manage mount points /u01/app/oracle and /u02/oradata/, respectively.

These components are shown in the following status output.

db-host-1# clresourcegroup status

=== Cluster Resource Groups ===

Group Name    Node Name     Suspended  Status
----------    ---------    ---------   ------
oracle-rg     db-host-1     No         Online
              db-host-2     No         Offline

scal-mnt-rg   db-host-1     No         Online
              db-host-2     No         Online

db-host-1# clresource status

=== Cluster Resources ===

Resource Name       Node Name   State    Status Message
-------------       ---------   -----    --------------
oracle-listener-rs  db-host-1   Online   Online
                    db-host-2   Offline  Offline

db-lh-rs            db-host-1   Online   Online - LogicalHostname online.
                    db-host-2   Offline  Offline

oracle-server-rs    db-host-1   Online   Online
                    db-host-2   Offline  Offline

oracle-rs           db-host-1   Online   Online
                    db-host-2   Online   Online

oradata-rs          db-host-1   Online   Online
                    db-host-2   Online   Online

The oracle-listener-rs and oracle-server-rs resources have dependencies of type resource_dependencies_offline_restart on other resources. For more details on the types of dependencies that are available in Oracle Solaris Cluster, see the r_properties(5) man page.

db-host-1# clresource show -p Resource_dependencies_offline_restart \
oracle-listener-rs

=== Resources ===                              

Resource:                                    oracle-listener-rs
  Resource_dependencies_offline_restart:        oracle-rs

db-host-1# clresource show -p Resource_dependencies_offline_restart \
oracle-server-rs

=== Resources ===                              

Resource:                                    oracle-server-rs
  Resource_dependencies_offline_restart:        oracle-rs oradata-rs

The oracle-listener-rs and oracle-server-rs resources have the following extension properties. For more details, see "HA for Oracle Extension Properties" in the Oracle Solaris Cluster Data Service for Oracle Guide.

  • Extension properties for the oracle-listener-rs resource are as follows:

    Listener_name=LISTENER_DB1
    ORACLE_HOME=/u01/app/oracle/product/10.2.0/Db_1
    
  • Extension properties for the oracle-server-rs resource are as follows:

    ALERT_LOG_FILE=/u02/oradata/admin/testdb1/bdump/alert_testdb1.log
    ORACLE_HOME=/u01/app/oracle/product/10.2.0/Db_1
    CONNECT_STRING=hauser/hauser
    

Oracle Database software is installed on the NFS shares on the Sun ZFS Storage 7420 appliance. The following command shows information about the configuration of the Sun ZFS Storage 7420 appliance.

db-host-1# clnas show -v -d all

=== NAS Devices ===

Nas Device:                                  qualfugu
  Type:                                         sun_uss
  userid:                                       osc_agent
  Project:                                      qualfugu-1/local/oracle_db
    File System:                                  /export/oracle_db/oradata
    File System:                                  /export/oracle_db/oracle

File system /export/oracle_db/oracle, which is mounted on the cluster nodes at mount point /u01/app/oracle, is the installation path for the Oracle Database software. File system /export/oracle_db/oradata, which is mounted on the cluster nodes at mount point /u02/oradata/, is the path for database files. The following are the /etc/vfstab entries for these two files systems on both nodes of the cluster.

db-host-1# cat /etc/vfstab | grep oracle_db
qualfugu:/export/oracle_db/oracle - /u01/app/oracle nfs - no 
rw,suid,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,noac,vers=3
qualfugu:/export/oracle_db/oradata - /u02/oradata/ nfs - no 
rw,suid,bg,hard,forcedirectio,nointr,rsize=32768,wsize=32768,hard,noac,proto=tcp,vers=3

The public network is using domain name service (DNS), and DNS is running and configured as follows.

db-host-1# svcs dns/client
STATE          STIME    FMRI
online         Nov_04   svc:/network/dns/client:default

db-host-1# cat /etc/nsswitch.conf
#
# Copyright 2006 Sun Microsystems, Inc.  All rights reserved.
# Use is subject to license terms.
#

#
# /etc/nsswitch.dns:
#
# An example file that could be copied over to /etc/nsswitch.conf; it uses
# DNS for hosts lookups, otherwise it does not use any other naming service.
#
# "hosts:" and "services:" in this file are used only if the
# /etc/netconfig file has a "-" for nametoaddr_libs of "inet" transports.

# DNS service expects that an instance of svc:/network/dns/client be
# enabled and online.

passwd:     files
group:      files

# You must also set up the /etc/resolv.conf file for DNS name
# server lookup.  See resolv.conf(4).
#hosts:      files dns
hosts:      cluster files dns

# Note that IPv4 addresses are searched for in all of the ipnodes databases
# before searching the hosts databases.
#ipnodes:   files dns
ipnodes:   files dns [TRYAGAIN=0]
networks:   files
protocols:  files
rpc:        files
ethers:     files
#netmasks:   files
netmasks:   cluster files
bootparams: files
publickey:  files
# At present there isn't a 'files' backend for netgroup;  the system will
#   figure it out pretty quickly, and won't use netgroups at all.
netgroup:   files
automount:  files
aliases:    files
services:   files
printers:	user files

auth_attr:  files
prof_attr:  files
project:    files

tnrhtp:     files
tnrhdb:     files
bash-3.2# cat /etc/resolv.conf
domain mydomain.com
nameserver 13.35.29.41
nameserver 19.13.8.13
nameserver 13.35.24.52
search mydomain.com

The following is the identity of the oracle user.

db-host-1# id -a oracle
uid=602(oracle) gid=4051(oinstall) groups=4052(dba)

The oracle user is configured with the following profile:

db-host-1# su - oracle
Oracle Corporation      SunOS 5.10      Generic Patch   January 2005
db-host-1$ cat .profile
export ORACLE_BASE=/u01/app/oracle
export ORACLE_HOME=/u01/app/oracle/product/10.2.0/Db_1
export ORACLE_SID=testdb1
export PATH=$PATH:$ORACLE_HOME/bin
export TERM=vt100

Preparing the Source Systems for Migration

Before you begin the migration, perform the following tasks.

  1. Copy the contents of /var/opt/oracle/ to NFS location /net/qualfugu/archive/optbkp/.

    db-host-1# cp -rf /var/opt/oracle/ /net/qualfugu/archive/optbkp/
    
  2. Unconfigure the Sun ZFS Storage 7420 appliance from the old cluster.

    db-host-1# clresourcegroup offline oracle-rg scal-mnt-rg
    db-host-1# clresource disable -g oracle-rg,scal-mnt-rg +
    db-host-1# clresourcegroup unmanage oracle-rg scal-mnt-rg
    db-host-1# clnas remove-dir -d qualfugu-1/local/oracle_db
    db-host-1# clnas remove qualfugu
    
  3. (Optional) Delete both resource groups.

    db-host-1# clresourcegroup delete oracle-rg scal-mnt-rg
    
  4. Shut down the cluster.

    db-host-1# cluster shutdown -y -g0
    

Note: If this cluster is reinstated, the NAS shares from the Sun ZFS Storage 7420 appliance can be added back by using the clnas add command, and the resource groups can be brought online by using the clresourcegroup online -emM command (if they were not deleted).

Preparing the Oracle Solaris 10 Zones Cluster on the Target Systems

The target two-node cluster new-phys-cluster includes the following (as shown in Figure 3 and Figure 4): two Sun SPARC Enterprise T5120 systems (new-phys-host-1 and new-phys-host-2) running Oracle Solaris 11.1 and Oracle Solaris Cluster 4.1.

We will create a new solaris10 brand zone cluster, cluster-1, on the target systems. The zone cluster will host the Oracle Database service that was configured in the source cluster running in an Oracle Solaris 10 environment.

Note: The procedure below assumes that the two-node target cluster is already set up with the Oracle Solaris Cluster 4.1 software. See "How to Install and Configure a Two-Node Cluster" for instructions on how to configure a two-node cluster on Oracle Solaris 11.

  1. Ensure that the cluster nodes are online.

    new-phys-host-1:~# clnode status
    
    === Cluster Nodes ===
    
    --- Node Status ---
    
    Node Name                              Status
    ---------                              ------
    new-phys-host-2                        Online
    new-phys-host-1                        Online
    
  2. Install the solaris10 brand zone's support package on both nodes.

    new-phys-host-1:~#  pkg publisher
    PUBLISHER                           TYPE     STATUS   URI
    solaris                             origin   online   <solaris-repository>
    ha-cluster                          origin   online   <ha-cluster-repository>
    
    new-phys-host-1:~# pkg install pkg:/system/zones/brand/brand-solaris10
               Packages to install:   1
           Create boot environment:  No
    Create backup boot environment: Yes
    
    DOWNLOAD                        PKGS       FILES    XFER (MB)
    Completed                                  1/1       44/44      0.4/0.4
    
    PHASE                                        ACTIONS
    Install Phase                                  74/74
    
    PHASE                                         	 	ITEMS
    Package State Update Phase                       	1/1
    Image State Update Phase                         	2/2
    
    new-phys-host-2:~# pkg publisher
    PUBLISHER                           TYPE     STATUS   URI
    solaris                             origin   online   <solaris-repository>
    ha-cluster                          origin   online   <ha-cluster-repository>
    
    new-phys-host-2:~# pkg install pkg:/system/zones/brand/brand-solaris10
               Packages to install:   1
           Create boot environment:  No
    Create backup boot environment: Yes
    
    DOWNLOAD                          PKGS       FILES    XFER (MB)
    Completed                                  1/1       44/44      0.4/0.4
    
    PHASE                                        ACTIONS
    Install Phase                                  74/74
    
    PHASE                                          	   	ITEMS
    Package State Update Phase                       	1/1
    Image State Update Phase                         	2/2
    
  3. On each node, verify the status of the IPMP group for the public network, sc_ipmp0. This interface will be used to configure the node-scope network resource for the zone cluster.

    new-phys-host-1:~# ipmpstat -g
    GROUP       GROUPNAME   STATE     FDT       INTERFACES
    sc_ipmp0     sc_ipmp0    ok       --            net0
    new-phys-host-2:~# ipmpstat -g
    GROUP       GROUPNAME   STATE     FDT       INTERFACES
    sc_ipmp0     sc_ipmp0    ok       --            net0
    
  4. Create the following cluster-1.config zone configuration file so the solaris10 brand zone cluster (cluster-1) can be created in subsequent steps:

    This configuration uses the same host names for the cluster nodes and for the logical hosts as were used in the source setup shown in Figure 1.

    new-phys-host-1:~# cat /cluster-1.config
    create -b
    set zonepath=/zones/cluster-1
    set brand=solaris10
    set autoboot=true
    set limitpriv=default,proc_priocntl,proc_clock_highres
    set enable_priv_net=true
    set ip-type=shared
    add net
    set address=db-lh
    set physical=auto
    end
    add capped-memory
    set physical=10G
    set swap=20G
    set locked=10G
    end 
    add dedicated-cpu
    set ncpus=1-2
    set importance=2
    end
    add attr
    set name=cluster
    set type=boolean
    set value=true
    end
    add node
    set physical-host=new-phys-host-1
    set hostname=db-host-1
    add net
    set address=db-host-1/24
    set physical=sc_ipmp0
    end
    end
    add node
    set physical-host=new-phys-host-2
    set hostname=db-host-2
    add net
    set address=db-host-2/24
    set physical=sc_ipmp0
    end
    end
    add sysid
    set root_password=ZiitH.NOLOrRg
    set name_service="DNS{domain_name=mydomain.com name_server=13.35.24.52,13.35.29.41,
    19.13.8.13 search=mydomain.com}"
    set nfs4_domain=dynamic
    set security_policy=NONE
    set system_locale=C
    set terminal=vt100
    set timezone=US/Pacific
    end
    

    Note: Oracle Solaris Cluster 4.1 software supports only the shared-ip type of solaris10 brand zone cluster. Oracle Solaris Cluster 4.1 SRU 3 supports the exclusive-ip type.

  5. Configure the new zone cluster.

    The name chosen for the new zone cluster is functional. Since the old cluster setup is being preserved, the same name is used for the new zone cluster. (Use the file cluster-1.config defined in the previous step.)

    new-phys-host-1:~# clzonecluster configure -f /cluster-1.config cluster-1
    

    If you encounter any errors during the zone cluster configuration or verification, refer to "Creating Zone Clusters."

  6. Verify the zone cluster configuration.

    new-phys-host-1:~# clzonecluster verify cluster-1
    
    new-phys-host-1:~# clzonecluster status cluster-1
    
    === Zone Clusters ===
    
    --- Zone Cluster Status ---
    
    Name      Brand       Node Name        Zone Host Name  Status      Zone Status
    ----      -----       ---------         --------------   ------    -----------
    cluster-1 solaris10   new-phys-host-1  db-host-1       Offline     Configured
                          new-phys-host-2  db-host-2       Offline     Configured
    

Installing the Zone Cluster

The following archive types are supported as the source archive for installing a solaris10 brand zone cluster:

  • A native brand zone on an Oracle Solaris 10 system
  • A cluster brand zone on an Oracle Solaris 10 cluster that has the proper patch level
  • An Oracle Solaris 10 physical system
  • An Oracle Solaris 10 physical cluster node
  • A solaris10 zone archive (derived from an installed solaris10 brand zone)
  • An Oracle VM Template for Oracle Solaris 10 Zones (which can be downloaded here; this requires some additional steps to be performed to extract the archive)

Although it is possible to use one of the original Oracle Solaris 10 physical cluster nodes from cluster-1 to create an archive and use it for the zone cluster installation, that option would involve making sure the original cluster meets the patch level requirements to ensure the zone cluster installation is successful.

Instead, this document will use a known solaris10 zone archive (by installing the archive from an Oracle VM Template), to ensure the results described in this article can be reproduced by readers on their own systems. This approach includes extra steps for creating a dummy zone, obtaining the archive out of the zone installed from the Oracle VM Template, and deleting the dummy zone that was used to obtain the archive.

  1. Download the Oracle VM Template to either of the two Sun SPARC Enterprise T5120 servers in the target cluster.

    See the Oracle VM Template for Oracle Solaris 10 Zones README (which is embedded in the template) for more details.

    Also, ensure you have the following components, which are used in this procedure:

    • /net/qualfugu/archive/: A directory on the Sun ZFS Storage 7420 appliance that contains the Oracle VM Template and is also used to store the dummy zone archive.
    • /net/qualfugu/osc-dir: A DVD or DVD image path for the Oracle Solaris Cluster 3.3 3/13 software, which is available here, and the relevant patches, which are available on My Oracle Support.
    • /net/qualfugu/nas/: A directory that contains the NAS client package (SUNWsczfsnfs), which is required to configure the Sun ZFS Storage 7420 appliance for the Oracle Solaris Cluster configuration. (See Step 5 of "How to Install a Sun ZFS Storage Appliance in a Cluster" for download instructions.)
  2. Install the dummy zone to obtain the archive, as shown below, where -a 10.134.90.201 is the IP address and /zones/cluster-1 is the root path for the dummy zone.

    new-phys-host-1:~# zfs create -o mountpoint=/zones rpool/zones
    new-phys-host-1:~# cd /net/qualfugu/archive
    new-phys-host-1:/net/qualfugu/archive#./solaris-10u11-sparc.bin \
    -a 10.134.90.201 -i net0 -p /zones/cluster-1 -z s10zone
    
           This is an Oracle VM Template for Oracle Solaris Zones.
     Copyright 2011, Oracle and/or its affiliates. All rights reserved.
    
     This software and related documentation are provided under a license agreement 
    containing  restrictions  on use  and  disclosure and are protected  by  intellectual  
    property  laws.   Except  as  expressly permitted in  your license agreement  or allowed 
    by law, you may not use,  copy,   reproduce,   translate,  broadcast,  modify,  license, 
    transmit,  distribute,  exhibit,  perform,  publish,  or display any part,   in  any  
    form,   or  by  any  means.   Reverse  engineering, disassembly,  or decompilation of this
    software,  unless required by law for interoperability, is prohibited.
    
    Checking disk-space for extraction
      Ok
    
    Extracting in /net/qualfugu/archive/bootimage.wuaWaV ...
    100% [===============================>]
    
    Checking data integrity
      Ok
    
    Checking platform compatibility
    
          The host  and  the image  do not have  the same Solaris release:
            host  Solaris release:   5.11
            image Solaris release:   5.10
    
          Will create a Solaris 10 branded Zone.
    
    IMAGE:      /net/qualfugu/archive/solaris-10u11-sparc.bin
    ZONE:       s10zone
    ZONEPATH:   /zones/cluster-1
    INTERFACE:  net0
    VNIC:       vnicZBI43632
    MAC ADDR:   2:8:20:92:88:96
    IP ADDR:    10.134.90.201
    NETMASK:    255.0.0.0
    DEFROUTER:  # # This file is deprecated. Default routes will be created for any router # 
    addresses specified here, but they will not change when the underlying # network 
    configuration profile (NCP) changes. For NCP-specific static # routes, the '-p'
    option of the route(1M) command should be used. # # See netcfg(1M) for information about 
    network configuration profiles.
    TIMEZONE:   US/Pacific
    
    Checking disk-space for installation
      Ok
    
    Installing in /cpool/s10zone ...
      100% [==========================>]
    
  3. Create the archive from the installed zone and copy it to a location, such as NFS, where it is accessible to the cluster nodes. (We will use the /net/qualfugu/archive/ directory as the location where we will copy the archive.)

    new-phys-host-1:~# cd /zones
    new-phys-host-1:~# find cluster-1 -print | cpio -oP@/ | gzip > \ 
    /net/qualfugu/archive/disk-image.cpio.gz
    
  4. Destroy the dummy zone.

    new-phys-host-1:~# zoneadm -z s10zone uninstall
    Are you sure you want to uninstall zone s10zone (y/[n])? y
    new-phys-host-1:~# zonecfg -z s10zone delete
    Are you sure you want to uninstall zone s10zone (y/[n])? y
    
  5. Install the zone cluster by using the obtained zone image.

    new-phys-host-2:~# zfs create -o mountpoint=/zones rpool/zones
    new-phys-host-1:~# clzonecluster install \
    -a /net/qualfugu/archive/disk-image cluster-1
    Waiting for zone install commands to complete on all the nodes of the zone cluster "cluster-1"...
    
  6. Log in to the console of zone cluster-1 on all nodes of the zone cluster.

    new-phys-host-1:~# zlogin -C cluster-1
    new-phys-host-2:~# zlogin -C cluster-1
    
  7. Boot the zone cluster into the offline/running mode.

    new-phys-host-1:~# clzonecluster boot -o cluster-1
    Waiting for zone reboot commands to complete on all the nodes of the zone cluster "cluster-1"...
    
  8. After the zone boots, make sure system configuration was completed.

    If the system configuration was not completed, complete any pending system configuration.

    new-phys-host-1:~# clzonecluster status cluster-1
    
    === Zone Clusters ===
    
    --- Zone Cluster Status ---
    
    Name        Brand       Node Name           Zone Host Name   Status    Zone Status
    ----        -----       ---------           --------------   ------    -----------
    cluster-1   solaris10   new-phys-host-1     db-host-1        Offline   Running
                            new-phys-host-2     db-host-2        Offline   Running
    

Installing the Cluster Software in the Zone Cluster (Optional)

Note: You can skip this step if the archive contains cluster software, for example, if the archive is from an Oracle Solaris 10 physical cluster node or a cluster brand zone on an Oracle Solaris 10 system.

  1. Install the Oracle Solaris Cluster 3.3 software from the DVD or the downloaded DVD image from the global zone.

    new-phys-host-1:~# clzonecluster install-cluster \
    -d /net/qualfugu/osc-dir/ \
    -p patchdir=/net/qualfugu/osc-dir,patchlistfile=plist-sparc \
    -s all cluster-1
    Preparing installation. Do not interrupt ...
    Installing the packages for zone cluster "cluster-1" ...
    

    Where:

    • -d specifies the location of the cluster software DVD image.
    • -p patchdir specifies the location of the patches to install along with the cluster software. The location must be accessible to all nodes of the cluster.
    • patchlistfile specifies the file that contains the list of patches to install along with cluster software inside the zone cluster. The location must be accessible to all nodes of the cluster. In the following example, the patch list plist-sparc is as follows:

      new-phys-host-1:~# cat /net/qualfugu/osc-dir/plist-sparc
      145333-15
    • -s specifies which agent packages to install along with core cluster software. In this example, all is specified to install all the agent packages.
  2. Reboot the zone cluster to boot the zone into online/running mode.

    new-phys-host-1:~# clzonecluster reboot cluster-1
    
  3. Verify that the zone cluster is now in online/running mode.

    new-phys-host-1:~# clzonecluster status cluster-1
    
    === Zone Clusters ===
    
    --- Zone Cluster Status ---
    
    Name        Brand       Node Name         Zone Host Name    Status  Zone Status
    ----        -----       ---------         --------------    ------  -----------
    cluster-1   solaris10   new-phys-host-1   db-host-1         Online  Running
                            new-phys-host-2   db-host-2         Online  Running
    

Re-creating the Application Setup on the Target Systems

  1. Configure the Sun ZFS Storage 7420 appliance for the zone cluster that was used in the source cluster setup.

    On db-host-1, install the required NAS software package in the zones of the zone cluster.

    new-phys-host-1:~# zlogin cluster-1
    [Connected to zone 'cluster-1' pts/2]
    Last login: Mon Nov  5 21:20:31 on pts/2
    Oracle Corporation      SunOS 5.10      Generic Patch   January 2005
    db-host-1# cd /net/qualfugu/nas/
    db-host-1# pkgadd -d . SUNWsczfsnfs
    
  2. Repeat Step 1 on db-host-2.
  3. Add the Sun ZFS Storage 7420 appliance to the cluster configuration.

    db-host-1# clnas add -t sun_uss -p userid=osc_agent qualfugu
    Enter password: <password you set when you configured the appliance>
    db-host-1# clnas find-dir qualfugu
    === NAS Devices ===
    
    Nas Device:                                   qualfugu
      Type:                                            sun_uss
      Unconfigured Project:                 qualfugu-1/local/oracle_db
    
    db-host-1# /usr/cluster/bin/clnas add-dir \
    -d qualfugu-1/local/oracle_db qualfugu
    db-host-1# /usr/cluster/bin/clnas show -v -d qualfugu
    
    === NAS Devices ===
    
    Nas Device:                                     qualfugu
      Type:                                            sun_uss
      userid:                                          osc_agent
      Project:                                         qualfugu-1/local/oracle_db
        File System:                                      /export/oracle_db/oradata
        File System:                                      /export/oracle_db/oracle
    
  4. Create mount points on both zone cluster nodes and add /etc/vfstab entries.

    db-host-1# mkdir -p /u01/app/oracle /u02/oradata/
    db-host-1# echo "qualfugu:/export/oracle_db/oracle - /u01/app/oracle nfs - no 
    rw,suid,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,noac,vers=3" >> /etc/vfstab
    db-host-1# echo "qualfugu:/export/oracle_db/oradata /u02/oradata/ -  nfs - no 
    rw,suid,bg,hard,forcedirectio,nointr,rsize=32768,wsize=32768,hard,noac,proto=tcp,vers=3" >> /etc/vfstab
    
    db-host-2# mkdir -p /u01/app/oracle /u02/oradata/
    db-host-2# echo "qualfugu:/export/oracle_db/oracle - /u01/app/oracle nfs - no 
    rw,suid,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,noac,vers=3" >> /etc/vfstab
    db-host-2# echo "qualfugu:/export/oracle_db/oradata /u02/oradata/ -  nfs - no 
    rw,suid,bg,hard,forcedirectio,nointr,rsize=32768,wsize=32768,hard,noac,proto=tcp,vers=3" >> /etc/vfstab
    
  5. Create the oracle user on both nodes of the zone cluster. Make sure that the identity is the same as that of the source setup.

    db-host-1# id -a oracle
    uid=602(oracle) gid=4051(oinstall) groups=4052(dba)
    
    db-host-2# id -a oracle
    uid=602(oracle) gid=4051(oinstall) groups=4052(dba)
    
  6. Mount the file systems and verify that the file systems are accessible by the oracle user on both nodes of zone cluster.

    Mount the file system on both nodes of the zone cluster using the following command:

    # mount /u01/app/oracle
    

    Verify the owner, group, and mode of $ORACLE_HOME/bin/oracle using the following command:

    # ls -l $ORACLE_HOME/bin/oracle
    

    Confirm that the owner, group, and mode are as follows:

    • Owner: oracle
    • Group: dba
    • Mode: -rwsr-s-x
  7. Create the required file system resource group and resources.

    db-host-1# clresourcegroup create -S scal-mnt-rg
    
    db-host-1# clresource create -g scal-mnt-rg \
    -t SUNW.ScalMountPoint \
    -p MountPointDir=/u02/oradata/ \
    -p FileSystemType =nas \
    -p TargetFileSystem=qualfugu:/export/oracle_db/oradata \
    oradata-rs
    db-host-1# clresource create -g scal-mnt-rg \
    -t SUNW.ScalMountPoint \
    -p MountPointDir=/u01/app/oracle \
    -p FileSystemType=nas \
    -p TargetFileSystem=qualfugu:/export/oracle_db/oracle \
    oracle-rs
    
  8. Bring online the file system resources.

    db-host-1# clresourcegroup online -eM scal-mnt-rg
    
  9. Verify that the file system is mounted properly on both nodes of the zone cluster.

    db-host-1# clresource status -g scal-mnt-rg
    
    === Cluster Resources ===
    
    oracle-rs        		db-host-1	Online    Online
                         	db-host-2	Online    Online
    
    oradata-rs           	db-host-1	Online    Online
    	             	db-host-2	Online    Online
    
    db-host-1# mount -p |grep qualfugu
    qualfugu:/export/oracle_db/oracle - /u01/app/oracle nfs - no 
    rw,suid,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,noac,vers=3,xattr,
    zone=cluster-1,sharezone=10
    qualfugu:/export/oracle_db/oradata /u02/oradata/ -  nfs - no 
    rw,suid,bg,hard,forcedirectio,nointr,rsize=32768,wsize=32768,hard,noac,proto=tcp,
    vers=3,xattr,zone=cluster-1,sharezone=10
    
    db-host-2# mount -p |grep qualfugu
    qualfugu:/export/oracle_db/oracle - /u01/app/oracle nfs - no 
    rw,suid,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,noac,vers=3,xattr,
    zone=cluster-1,sharezone=10
    qualfugu:/export/oracle_db/oradata /u02/oradata/ -  nfs - no 
    rw,suid,bg,hard,forcedirectio,nointr,rsize=32768,wsize=32768,hard,noac,proto=tcp,
    vers=3,xattr,zone=cluster-1,sharezone=10
    

    Note: Often the database has a specific SRM project (entries in /etc/projects) to be used for the database to obtain the correct level of system resources. For such configurations, these project specifications must be reproduced in the solaris10 brand zone cluster as well.

  10. Restore the contents of /var/opt/oracle on all nodes of zone cluster.

    db-host-1# mkdir -p /var/opt/oracle/
    db-host-2# mkdir -p /var/opt/oracle/
    db-host-1# cp -rf /net/qualfugu/archive/optbkp/ /var/opt/oracle/
    db-host-2# cp -rf /net/qualfugu/archive/optbkp/ /var/opt/oracle/
    
  11. Create the resource groups and resources required for HA for Oracle.

    db-host-1# clresourcetype register SUNW.oracle_server
    db-host-1# clresourcetype register SUNW.oracle_listener
    
    db-host-1# clreslogicalhostname create -g oracle-rg \
    -h db-lh db-lh-rs
    db-host-1# clresource create -g oracle-rg \
    -t oracle_listener \
    -p ORACLE_HOME=/u01/app/oracle/product/10.2.0/Db_1 \
    -p Listener_name=LISTENER_DB1 \
    -p Resource_dependencies_offline_restart=oracle-rs \
    oracle-listener-rs
    
    db-host-1# clresource create -g oracle-rg \
    -t oracle_server -p ORACLE_SID=testdb1 \
    -p ALERT_LOG_FILE=/u02/oradata/admin/testdb1/bdump/alert_testdb1.log \
    -p ORACLE_HOME=/u01/app/oracle/product/10.2.0/Db_1 \
    -p CONNECT_STRING=hauser/hauser \
    -p Resource_dependencies_offline_restart=oradata-rs,oracle-rs \
    oracle-server-rs
    
    db-host-1# clresourcegroup online -eM oracle-rg
    
    db-host-1# clresource status -g oracle-rg
    
    === Cluster Resources ===
    
    Resource Name          Node Name    State      Status Message
    -------------          ---------    -----      --------------
    oracle-server-rs       db-host-2    Offline    Offline
                           db-host-1    Online     Online
    
    oracle-listener-rs     db-host-2    Offline    Offline
                           db-host-1    Online     Online
    
    db-lh-rs               db-host-2    Offline    Offline - LogicalHostname offline.
                           db-host-1    Online     Online - LogicalHostname online.
    

Verifying Failover

  1. Perform failover verification by halting one zone where the resources are online.

    db-host-1# halt -q
    halt: can't turn off auditd
    
    [Connection to zone 'cluster-1' pts/2 closed]
    
  2. Verify that the status of the zone cluster on node new-phys-host-1 is offline/installed.

    new-phys-host-1:~# clzonecluster status cluster-1
    
    === Zone Clusters ===
    
    --- Zone Cluster Status ---
    
    Name        Brand       Node Name         Zone Host Name  Status  Zone Status
    ----        -----       ---------         --------------  ------  -----------
    cluster-1   solaris10   new-phys-host-1   db-host-1       Offline Installed
                         	new-phys-host-2   db-host-2       Online  Running
    
  3. Log in to the other node of the zone cluster and verify the status of the resource.

    new-phys-host-2:~# zlogin cluster-1
    [Connected to zone 'cluster-1' pts/2]
    Last login: Mon Nov  5 21:30:31 on pts/2
    Oracle Corporation      SunOS 5.10      Generic Patch   January 2005
    
    db-host-2# clresource status -g oracle-rg
    
    === Cluster Resources ===
    
    Resource Name          Node Name      State      Status Message
    -------------          ---------      -----      --------------
    oracle-server-rs       db-host-2      Online     Online
                           db-host-1      Offline    Offline
    
    oracle-listener-rs     db-host-2      Online     Online
                           db-host-1      Offline    Offline
    
    db-lh                  db-host-2      Online     Online - LogicalHostname online
                           db-host-1      Offline    Offline - LogicalHostname offline
    
  4. Boot the zone cluster node that was halted in Step 1 and switch back the resource group:

    Note that the following output is same as that obtained from the source setup shown in Figure 1.

    new-phys-host-1:~# clzonecluster boot -n new-phys-host-1 cluster-1
    
    new-phys-host-1:~# zlogin cluster-1
    [Connected to zone 'cluster-1' pts/2]
    Last login: Mon Nov  5 21:30:31 on pts/2
    Oracle Corporation      SunOS 5.10      Generic Patch   January 2005
    
    db-host-1# clresourcegroup switch -n db-host-1 oracle-rg
    db-host-1# clresourcegroup status
    
    === Cluster Resource Groups ===
    
    Group Name    Node Name     Suspended  Status
    ----------    ---------    ---------   ------
    oracle-rg     db-host-1    No          Online
                  db-host-2    No          Offline
    
    scal-mnt-rg   db-host-1    No          Online
                  db-host-2    No          Online
    
    db-host-1# clresource status
    
    === Cluster Resources ===
    
    Resource Name          Node Name    State      Status Message
    -------------          ---------    -----      --------------
    oracle-server-rs       db-host-1    Online     Online
                           db-host-2    Offline    Offline
                           
    
    oracle-listener-rs     db-host-1   Online      Online
                           db-host-2   Offline     Offline
                           
    
    db-lh-rs               db-host-1   Online      Online - LogicalHostname online.
                           db-host-2   Offline     Offline - LogicalHostname offline.
                           
    
    oracle-rs              db-host-1   Online      Online
                           db-host-2   Online      Online
    
    oradata-rs             db-host-1   Online      Online
                           db-host-2   Online      Online
    

Migration to the solaris10 brand zone cluster is now complete.

Conclusion

This article described how to migrate an Oracle database running on Oracle Solaris 10 in an Oracle Solaris Cluster environment to Oracle Solaris 11 without upgrading the database software by using Oracle Solaris 10 Zones and Oracle Solaris zone cluster features.

This migration is an example that can be extended to other application environments and shows how Oracle Solaris can help preserve infrastructure investments and lower migration costs.

See Also

For more information on Oracle Solaris Cluster, see the following resources.

And here are some Oracle Solaris resources:

About the Author

Mahesh has been a software engineer in the Oracle Solaris Cluster High Availability Infrastructure group for the past three years. Previously, Mahesh worked in the Oracle Solaris Revenue Product Engineering group, and prior to that he worked at Nexus Info as a software engineer.

Revision 1.0, 02/14/2014

Follow us:
Blog | Facebook | Twitter | YouTube