Oracle Weblogic - Backup & Recovery Operations

Backup & Recovery Operations

Backing up once a month is too infrequent, whereas once an hour is too frequent, so what is right balance ? 

Difference between  restore and recover  :   Restore is a pure file system copy operation, where as recovery is restore plus some extra operations to get a specific point-in-time-recovery

• Review of Terms and Components

• Host
• Domain
• Administration Server
• Managed Server
• Application
• JMS Application
• SOA Application
• Instance
• Oracle HTTP Server
• Web Cache
• Database
• Metadata
• Node Manager
• Persistent Store
• Static artifacts -   These don't change very often, after patches, monthly or longer.
• Run-time artifacts -  These will change frequently, even multimes per second like logs.

• Homes :  Oracle, Middleware, WebLogic
• Oracle Home
• Middleware HOme
• Instance HOme

• Understanding Backup and Recovery

• Protects against failures of hardware
• Protects against failures of software
• Protects against failures of power
• Environmental disasters
• Accidental and malicious changes etc
• Guarantees a point of recovery
• Minimizes loss of business availability
• Insures an SLA

• Types of Backups
• Online
• Nondisruptive
• Possible inconsistent
• Can be tricky, especially for database
• Offline
• Requires all processes to be stopped
• Very easy
• Full
• Easier to recover
• Slower to create
• Incremental
• Harder to recover
• Faster to create

                          Note  :  In a hot backup/inconsistent backup the files are inconsistent, that is some 

                                       files may have different time stamps and need to be reconciled via a 

                                       transaction log.

• Backup Recommendations
• After the initial domain is created(offline)
• Scheduled backups ( online)
• After the component or the cluster changes(online)
• Before application deployment(online)
• Before patches or upgrades(offline)
• Database backups(online) for :
• LDAP
• Persistent stores
• SOA repository

• Limitations and Restrictions for Backing Up Data
• Performing a Full Offline Backup
• Shutdown All processes :
• Stop WebLogic via the Administration Console.
• Shut down the database
• Stop the Listener and the Node Managers
• Stop the Enterprise Manager and emAgent.
• Stop Web Cache and HTTP Server via OPMN.
• Perform the backup via OS tools:
• If using TAR, make sure that you keep permissions.
• If using ZIP, make sure that you include empty directories.
• Test the backup by performing recovery on another computer:
• Ideally, use an alternate computer in an alternate data center.
• Time the recovery for SLA input.
• Store the backup offsite

• Backing Up a Domain Configuration
• Enable autobackup of configuration 
• Admin Console-->settings of Domain -> Configuration-->General-->
• Configuration Archive Enabled  (  by default the checkbox disabled)
• Check new JAR files and Directories

• Backing Up an Instance Home
• Stop the Web tier(Oracle HTTP Server and Oracle Web Cache)
• opmnctl stopall
• opmnctl status
• Copy the Instance home
• As the superuser, change to the root directory
• Execute tar-zcvpf  techonline21.tar $ORACLE_INSTANCE
• Restart all services
• opmnctl startall
• opmnctl status

• Creating a Record of Installations
• 
  
• Directories to Restore
• Binaries
• Be mindful of preserving group ownership and permissions
• This should be read-only for most users
• Configurations
• If the last configuration caused the problem recover to a point in time prior to that.
• Logs 
• Not required for recovery
• Created if they do not exist
• Data
• Database restores data within tablespaces, not directories.
• RMAN restore brings data up to the last backup, then recover brings data up to a later point in time.

• Recovery After Disaster
• Possible causes of failure
• Data loss
• User error
• Malicious attack
• Corruption of data
• Media failure
• Application failure
• Recovery depends on the cause
• Repair
• Replace
• Relocate

• Recovery of Homes
• Recovery of Managed Server
• Recovery of the Administration Server Configuration
• Restarting an Administration Server on a New Computer
• Install Oracle WebLogic Server on a backup computer
• Copy the application files to a backup computer
• Copy the configuration files to a backup computer
• Restart the administration server on a new computer.

• Recovery of a cluster
• If you accidentally lost a member of a cluster or whole cluster, we can use several ways to recover it.  (  Below are most preferable way to recover)
• Undo the changes in the Change Center.
• Reenter the configuration changes that you made.
• User the configuration archive to go abck one or two versions.
• Recover the configuration
• Recover the domain
• Recover WebLogic
• Perform a full recovery
• 
  
• Restoring OPMN-Managed components to a New Computer