Oracle8 Server Unleashed:Backup and Recovery

Listing 23.2. Import recovery scenario.


1: IMP system/manager FULL=Y INCTYPE=SYSTEM FILE=I10.exp

2: IMP system/manager FULL=Y INCTYPE=RESTORE FILE=C7.exp

3: IMP system/manager FULL=Y INCTYPE=RESTORE FILE=I8.exp

4: IMP system/manager FULL=Y INCTYPE=RESTORE FILE=I9.exp

5: IMP system/manager FULL=Y INCTYPE=RESTORE FILE=I10.exp

High Availability Options

High availability is supported by many hardware features and by options in the Oracle software.

Hardware features include: UPS (uninterruptable power supply), multiple processors, and RAID technology disk drives. Some vendors, such as Tandem, specialize in redundant fault-tolerant systems that automatically switch all applications to a duplicate system where processing continues without interruption. RAID, or redundant array of inexpensive devices, has performance options (hardware data striping) and data- protection features (hardware disk mirroring with or without parity checking).

This section will concentrate on the features and benefits of RAID technology in an Oracle environment, and various Oracle options that increase availability and limit downtime.

RAID Drives

RAID is a bank of disk drives the computer will view as one physical drive. The Oracle environment will view a single RAID device as a single disk device as well.

There are five levels of RAID. RAID level 0 stripes data files across the available disk drives without any mirroring or parity checking. RAID level 1 identically mirrors disk drives. RAID levels 2, 3, and 4 are not used and do not apply to business systems. RAID level 5 is a combination of RAID level 0 and RAID level 1. RAID level 5 stripes data files, mirrors the striped disk segments across many available drives, and provides parity to ensure the accuracy of the reads and writes. Both RAID level 1 and RAID level 5 support a disk drive going bad with no impact on the application. Both RAID level 1 and RAID level 5 will support the bad disk drive being “hot swapped,” that is, replaced with another physical drive, and the RAID environment will rebuild the information (striped data or a mirror) on the replacement drive.

TIP:
RAID level 1 provides excellent fault tolerance for data availability.

NOTE:
For performance reasons, RAID level 5 should not be used for Oracle structures that require performance such as the redo logs, TEMPORARY tablespaces, and ROLLBACK tablespaces.

Backups Versus RAID

RAID devices should not take the place of a regular backup schedule. Once again, business needs for data availability will dictate which scenario is best for the application. RAID technology brings a high degree of inexpensive fault tolerance to the computing environment. It certainly is NOT a replacement for system and/or database level backups.

NOTE:
If the RAID controller or the RAID disk management software (anything other than the physical drives themselves) develops problems, the recoverability of information stored on the RAID device is unpredictable.

Standby Database

A standby database is a duplicate database environment, on duplicate equipment (with some exceptions), with identical levels of operating systems and identical levels of Oracle software. This standby database is in a constant state of recovery; as archive log files are closed on the primary database, they are transferred and applied to this standby database. This method incurs absolutely no overhead on the primary database and the only disadvantage is that the standby database is only as current as the last log file applied. A standby database is easy to use to replace the primary database so the amount of application downtime is minimal. These archive log files are easy to size and tune.

TIP:
Mirror the archive log files to help prevent any data loss at all.

TIP:
There is no limit to the number of standby database environments.

Standby databases are easy to create and maintain (see Figure 23.3). Standby databases require the exact same level of Oracle software and hardware operating system release levels as the database they back up. Standby databases also require the exact same computer processor type. It is recommended, but not required, that the same disk configuration exist. The Oracle database files are backed up utilizing either the cold or hot methods (as previously described) and moved to the standby system. This SQL statement will create the standby database’s control file:

Figure 23.3. Standby database configuration.


alter database create standby controlfile as <filename>

Move this file to the standby database. Use the following SQL statement to archive the current redo log and then manually copy this archive log file to the standby database: alter system archive log current.

TIP:
I recommend a cold backup for the data file collection and transfer.

Maintenance of the standby database is quite easy. Start the standby database with the startup nomount option and issue the following command:


alter database mount standby database

The standby database is not open, so users cannot access this standby environment. As archive log files become available on the primary system, move them to the standby system and use the following command to apply the logged changes to the standby database: alter database recover logfile <filename>.

NOTE:
There are options on some Oracle processes that do not create logged information, such as any “direct path load” or DDL with the UNRECOVERABLE option set. These operations will have to be applied to the standby environment as well.

Table of Contents

Используются технологии uCoz