Question: What is VERITAS® Database Edition™ for Oracle®?
Answer: VERITAS Database Edition for Oracle is the optimal foundation for managing Oracle database
environments. The Database Edition offers key storage virtualization capabilities – the ability to
manage logical pools of storage rather than physical storage devices – in addition to providing raw
device performance with the manageability of the VERITAS File System, online administration of
storage, and the flexibility of storage hardware independence. VERITAS Database Edition for Oracle
provides a management platform for Oracle that helps IT organizations manage larger and more
complex environments with existing resources.
In addition, the high availability (HA) version, VERITAS Database Edition/HA for Oracle, for multi-
system configurations, is designed to meet demanding 7x24 availability requirements. The HA version
includes VERITAS Cluster Server™ (VCS), the VCS Agent for Oracle and the QuickLog agent. These
components automatically detect system or Oracle database failure and manage recovery (either by
restarting Oracle databases on the same system or by failing over to a designated takeover system).
Question: What problems do VERITAS Database Edition for Oracle address?
Answer: VERITAS Database Edition for Oracle addresses two critical issues in today’s enterprise computing
environments -- the rising cost of managing data and storage and, at the same time, static or shrinking
IT resources used to manage those environments. As businesses increasingly depend on databases, the
amount of data and the number of users accessing the databases continue to grow. This growth raises
corporate requirements for data and system availability and performance. The Database Edition is
designed to reduce storage costs by focusing on improving system performance and availability, as
well as enabling system and database administrators to become more efficient.
Question: What are the new features in Database Edition 3.0 for Oracle?
Answer: Database Edition 3.0 for Oracle offers support for Oracle9i and Oracle Disk Manager, integration with
VERITAS FlashSnap, and new improvements to the VxDBA utility.
Database Edition for Oracle is the first solution to implement the Oracle9i Oracle Disk Manager
(ODM). The ODM interface provides streamlined I/O and file management for Oracle servers. ODM
increases I/O performance, uses system resources more efficiently and improves file management.
While running with ODM, Database Edition for Oracle write performance equals that of raw disk
partitions. For read-intensive environments, such as reporting, performance greater than raw can be
experienced. In conjunction with ODM, Database Edition for Oracle provides Oracle database
administrators with a new level of manageability of Oracle9i environments.
The Database Edition also integrates with VERITAS FlashSnap, which enhances the ability of
administrators to create point-in-time copies of data with minimal impact to applications or users. This
relieves them from having to worry about issues like shrinking maintenance windows and performance
impacts of taking consistent copies of data for mission critical processes like backups, decision
support, queries and reporting.
Enhancements to the VxDBA utility include the addition of a Java-based graphical user interface and a
Storage Checkpoint scheduler that allows DBAs to automate the creation of Storage Checkpoints.
Explanation of Key Features
Question: How does VERITAS Quick I/O improve performance and reliability?
Answer: VERITAS Quick I/O database accelerator enables Oracle databases built on VERITAS File System to
achieve at least raw disk performance. Quick I/O presents a VERITAS File System file to Oracle as a
raw character device. This bypasses the UNIX system buffer cache, eliminating double-buffering and
ensuring data reliability by writing data safely from the Oracle’s own buffer cache, System Global
Area (SGA), to the disk. Without double-buffering, Oracle also eliminates redundant buffer copying,
and thus excess data movement on every read and write.
Quick I/O attains its performance gains through shared update locks and kernel-supported
asynchronous I/O. Quick I/O supports shared update locks for database files at the VERITAS File
System level. Traditionally, the exclusive write lock at the UNIX file system level caused write
transactions to be serialized, and it slowed down database performance, especially with many
simultaneous users. Quick I/O eliminates this bottleneck and allows Oracle to process parallel update
requests, relying on Oracle to manage locking. Quick I/O further improves database performance by
taking advantage of asynchronous I/O, which has previously been available only to raw devices. For
example, on the Solaris® 2.x operating system, Oracle uses asynchronous writes for parallel database
writes and asynchronous reads for database recovery. Solaris’ enhanced kernelized asynchronous I/O
(KAIO) is performed on Quick I/O database files.
Cached Quick I/O, a feature of Quick I/O, enables systems with large available memory to experience
faster than raw performance by running Oracle on VERITAS File System. This performance
improvement is due to database blocks being cached in the system buffer cache, reducing the number
of physical I/Os and improving read performance. Although Cached Quick I/O buffers reads, it still
performs direct writes so as not to jeopardize data integrity.
Question: What is the Oracle Disk Manager?
Answer: Oracle Disk Manager (ODM) is a new Application Programming Interface (API) co-developed by
Oracle and VERITAS to deliver optimal I/O performance and manageability for Oracle servers. ODM
is one of the new Orcacle9i features.
Question: What are the main features of ODM?
Answer: ODM improves performance and manageability through the support for file system I/O, simplified
system calls within Oracle, and improved file management file identification.
Support for File System I/O
Historically, database administrators chose between file system files and raw partition storage by
selecting the “lesser of two evils.” File system files are easier to manage, but raw partitions provide
better I/O performance. Using ODM, database I/O types are supported on both file system files and
raw partitions. In addition, ODM supports mixed file types, both raw and file system files, which
benefits existing databases and supports migration from raw partitions to file system files.
Simplified System Calls
Oracle instances routinely service a large number of concurrent requests involving table and index
scans, single block and scattered reads, and temporary segment reads. To support this wide variety of
functions, the Oracle server uses a collection of system calls such as read(), write(), lio_listio(), kaio(),
pread(), pwrite(), aio_read(), and aio_write(). Ideally, DBAs should not have to be concerned with
these low-level I/O details. ODM was designed to address these challenges. ODM reduces
complexity for database administrators by allowing all datafile I/O types on both file systems and raw
partitions, using a single system call: odm_io(). The odm_io() call supports all Oracle file I/O types,
including sequential reads, sequential writes, direct reads, direct writes, scattered reads, Database
Writer batch writes, Log Writer Redo Log buffer flushing and archiving of Redo Logs by the Archiver
process. Using ODM, all of these I/O types are enhanced on both file system files and raw partitions.
Improved File Management and File Identification
When the Oracle server is in the process of initializing a new datafile, any number of things can go
wrong. The DBA must clean up from the point of failure, which may entail removing a file from a file
system that was too small for the data being added. On large systems with many existing datafiles,
these cleanup operations can be dangerous. ODM reduces the failures that can occur when adding a
datafile to a database by not committing the file to the database until after it has been initialized.
Contiguous file system space allocation is the other essential benefit of the ODM file creation
functionality. In traditional file system files, the disk space allocated to a file is not contiguous. The
potential degradation of table scan throughput for noncontiguous datafiles is one of the factors that
drives DBAs to choose raw partitions for datafile storage. Datafiles created with ODM consist of
contiguous file system blocks.
Oracle9i also offers a feature called Oracle Managed Files (OMF). OMF manages datafile attributes,
such as file names, file locations, storage attributes, and whether the file is used in a database. In
addition, OMF files are created with the AUTOEXTEND capability by default. Using this feature
significantly reduces the capacity planning associated with maintaining existing applications and
deploying new databases. In the past, DBAs have been cautious about using AUTOEXTEND, due to
concerns over file system fragmentation. ODM eliminates this concern through the introduction of the
odm_resize() routine, which allocates contiguous file system extents to files. Eliminating potential
fragmentation assures that table and index scan throughput does not degrade as the tablespace grows.
Lastly, ODM streamlines file identification by producing shared file descriptors. Traditionally, the
Oracle server obtains a file descriptor from the operating system and then uses that file descriptor to
perform file operations such as reading and writing. File descriptors require an allocation structure in
the OS kernel and also incur kernel overhead. ODM file identifiers are shareable – once a file has been
identified in ODM, subsequent processes using the shared identifier do not incur any kernel overhead.
This removes the limit on the number of files that can be opened.
Question: When will ODM be available?
Answer: Now. ODM is available through the Oracle9i database that is currently shipping. Currently,
VERITAS Database Edition for Oracle is the only file system with support for the ODM API!
Question: Does ODM replace Quick I/O?
Answer: No. You can think of ODM as the latest generation of Quick I/O and ODM is only supported on
Oracle9i. Database Edition 3.0 for Oracle supports Oracle9i and ODM. DBE 3.0 will continue to
support Quick I/O for all Oracle users who are running Oracle 7, 8, 8i and 9i in order to ease
migration. Keep in mind that Oracle users do not migrate quickly to new major releases such as 9i.
Question: What are the differences between ODM and Quick I/O?
Answer: ODM is fully automatic and does not require any setup intervention by the DBA. While Quick I/O is
also easy to implement, it does require minimal setup by the DBA and it uses file system resources.
Comparing performance of the two technologies, they are equal. Keep in mind, customers are not
comparing ODM vs. Quick I/O. They are comparing them separately to RAW partitions and to
conventional file systems. They both perform equal to RAW and they both improve conventional file
system performance by up to six times in I/O performance.
The bottom line is that Oracle customers who are moving to Oracle9i should understand that when
they run Oracle9i with DBE Edition 3.0 (or later) they will receive optimal I/O performance
automatically with Oracle ODM and the VERITAS File System. VERITAS is the first file system
vendor to support the new ODM API for 9i. If customers are running Oracle 7, 8, or 8i, they can
continue to use Quick I/O for proven I/O performance equal to RAW.
Question: What are Storage Checkpoints?
Answer: The VERITAS Database Edition Storage Checkpoint facility is a disk and I/O efficient snapshot of file
systems or the database. A Storage Checkpoint provides a consistent view of a file system at the point
in time when the Storage Checkpoint was taken. Instead of making a physically separate copy of the
file system, a Storage Checkpoint identifies and maintains only the changed file system blocks, saving
disk space and significantly reducing I/O overhead. By keeping track of changed blocks, the Storage
Checkpoint facility enables Block Level Incremental (BLI) Backup (via VERITAS NetBackup for
Oracle Advanced BLI Agent), Storage Rollback and the ability to set up a clone Oracle database.
Like VERITAS File System snapshots, the Storage Checkpoint feature is a file system image at a
particular point in time. However, there are three key differences between these file system image
features. First, Storage Checkpoints are persistent while VERITAS File System snapshots are
transient. For example, if the system crashes, Storage Checkpoints will be available upon reboot while
snapshots are lost. Second, file system snapshot only allows for backup of entire datafiles, while
Storage Checkpoints enable just the changed blocks to be backed up because unlike snapshot file
systems, Storage Checkpoints are aware of one another. Third, a snapshot file system requires
allocating a separate storage pool for saving before-images, whereas a Storage Checkpoint uses the
free space pool from the primary file system. Therefore, storage requirements for Storage Checkpoints
can be easily monitored and managed.
Question: What is Block Level Incremental Backup (BLI Backup)?
Answer: BLI Backup is an incremental-backup method that only backs up changed data blocks, virtually
eliminating the database backup window. Having the ability to just back up changed blocks is
important. Unlike traditional file level incremental backup in which an entire multi-gigabyte datafile
must be backed up if only a single 2K block is changed, BLI Backup enables just the modified 2K
block to be backed up.
BLI Backup offers several key benefits. First, since only changed blocks are backed up, tremendous
time and media savings are possible. Second, with a very short backup window, both system and
network overhead is greatly reduced, thereby enhancing performance. Third, more frequent backups
are possible, so each backup contains a more up-to-date image of data for improved recoverability.
Fourth, BLI Backup enables faster database recovery from hot backup than traditional techniques.
BLI Backup works through integration of VERITAS Database Edition with VERITAS NetBackup™
for Oracle Advanced BLI Agent. DBAs can set up a NetBackup class and schedule a BLI Backup just
as they are able to do for traditional backups.
Question: What is Storage Rollback and how does it differ from Block-Level Incremental Backup
restore, especially since they both use Storage Checkpoints?
Answer: Storage Rollback is a quick restore from an on-disk backup image that helps the DBA recover from
logical database errors such as a dropped tablespace or datafile without having to restore from
secondary media. Using VxDBA, the DBA can rollback a datafile, tablespace or the entire database to
a point-in-time (a particular Storage Checkpoint) and apply Oracle redo logs up to the point of failure
and repair the logical error with a minimum of application downtime.
BLI Backup is true backup to secondary media for recovery from media or server failure. Therefore,
secondary media is needed for a restore. To restore using BLI Backup technology, a full backup must
be restored and then each of the BLI Backups, resulting in a complete image of the database.
Even though Storage Rollback is a form of on-disk recovery, backup is still necessary for media
recovery as Storage Rollback can only be used to recover from logical errors.
Benefits of VERITAS Database Edition for Oracle
Question: What are the performance benefits of the Database Edition for Oracle?
Answer: The key performance benefit is raw disk performance for Oracle databases built on VERITAS File
System with VERITAS Quick I/O or ODM for Oracle9i environments. In addition, the Database
Edition offers various storage layouts that can be reconfigured online to optimize database application
performance. Using striping (RAID-0), for example, DBAs can reduce I/O bottlenecks and increase
throughput for heavily used tables.
The Database Edition addresses performance issues by:
• Delivering raw disk database performance through file systems and greater than raw performance
for systems with large available memory (Cached Quick I/O)
• Improving throughput and bandwidth for transaction-intensive workloads, using disk striping to
provide parallel data transfer to multiple disks, and disk mirroring (RAID-1) to provide I/O
balancing for reads as well as data redundancy
• Enabling unsurpassed performance for large file I/O applications such as multimedia and imaging
Question: What are the availability benefits of the Database Edition for Oracle?
Answer: The availability goal of the VERITAS Database Edition architecture is to provide continual access to
data through “online everything” – allowing administrators to reconfigure layouts and capacity without
interrupting user access to data. This can greatly reduce or eliminate planned downtime. Database
Edition can also minimize unplanned downtime due to application or system failures through the use
of the VERITAS Cluster Server.
Database Edition’s availability benefits to Oracle environments result in more uptime for mission-
critical applications, increasing the overall application return-on-investment to IT organizations.
The Database Edition addresses data availability requirements by:
• Providing online database layout reconfiguration options including resizing file systems, and
adding, detaching, and reattaching mirrors
• Achieving fast recovery of file systems in seconds following a system failure or reboot, thus
minimizing database downtime
• Off-loading production servers with tasks such as backup, decision support and testing with
• Enabling backup at the block level to minimize or eliminate the database backup window
• Providing Storage Checkpoint and Storage Rollback to allow fast, online recoverability form
logical database errors
• Allowing redundant storage configurations (RAID-1, RAID 0+1, RAID 1+0, and RAID-5) to
provide fault-tolerance and improved data availability and reliability for Oracle databases
• Offering a cost-effective, fault-tolerant solution for read-intensive workloads, such as data
warehousing applications, through RAID-5, which provides data redundancy while requiring less
disk space than mirroring
Furthermore, the HA version enhances availability of Oracle databases and applications for clustered
configurations by recovering from any single point of failure without operator intervention. The HA
version addresses data availability requirements by:
• Monitoring and restarting the Oracle database on the same system after detecting Oracle server
failure or failing over to a designated take-over server
• Automatically detecting and recovering from a system failure by failing over to a designated
takeover server system and restarting the Oracle database and other monitored services
• Keeping the Oracle database online during system maintenance through manual failover
Question: What are the manageability benefits of the Database Edition for Oracle?
Answer: The manageability goal is to reduce complexity and time required to perform routine administrative
tasks, and to make administrators more efficient. With the Database Edition, DBAs can enjoy the
flexibility of file systems such as familiar file system naming conventions and utilities to manage
Oracle datafiles. Oracle DBAs and system administrators can also develop common strategies for
backing up and restoring databases as well as non-database files.
The Database Edition addresses manageability issues by:
• Allowing DBAs to use familiar file system naming conventions and utilities to manage Oracle
• Simplifying management of large database configurations with large file systems whose sizes are
not limited to underlying device sizes, eliminating the need to manage multiple file systems per
• Providing online database storage configuration (including defragmentation, file system and
volume resizing, and migration of data from one physical drive to another)
• Offering VxDBA utility to manage Storage Checkpoints and Storage Rollback, as well as simplify
database management through an easy-to-use, Java-based graphical user interface
• Allowing quick, reliable deployment of a high availability database configuration through simple
installation and configuration of VCS