|
Many IT organizations rely on Fibre Channel SAN architectures as a proven solution for Oracle deployments. Complex deployment and management processes have traditionally been viewed as an unavoidable trade-off associated with high performance SAN environments. It doesn’t have to be that way.
NetApp Fibre Channel SAN offerings deliver not just high performance but also a broad range of manageability features that lower the complexity associated with deploying and managing Oracle environments. This article explores NetApp engineering’s work in the FC SAN space, including:
- Developing unique management tools tailored to Oracle environments
- Enhancing FC SAN management and efficiency
- Delivering flexible and cost-effective DR solutions
- Ensuring full compatibility with SAN switches, HBAs and file systems
- Integrating with a broad suite of SRM tools
Developing Unique Management To ls Tailored to Oracle Environments
NetApp and Oracle have collaboratively developed and deployed solutions since 1995. In 2004 NetApp began a program of sustained investment in developing unique software tools designed specifically for Oracle environments, including SnapManager® for Oracle and SnapValidator™.
SnapManager for Oracle
Backup, restore (getting the files back from a backup) and recovery (getting a database back into a consistent state) are often complex tasks for Oracle Databases, especially in FC SAN environments where there’s a whole storage fabric to deal with. The goal of the SnapManager for Oracle (SMO) development team was to closely integrate the strengths of a NetApp storage system—including backup, restore, and cloning—with Oracle’s RMAN, RAC, and ASM (automatic storage management) products. This integration greatly simplifies the day-today management of test/dev clones, Snapshot™ copies, and backups.
SMO provides a variety of advantages discussed at length in a previous article (see http://www.netapp.com/go/ techontap/empower-dba.html), including three unique benefits no other SAN storage vendor can provide:
Ability to restore a single file within an ASM disk group. SMO understands the ASM file layout, making it possible to restore a single file from a Snapshot of an ASM disk group rather than the entire disk group (which may contain files from different databases).
- Ability to clone ASM disk groups and make them available to the same host. SMO understands the ASM label and integrates with ASM to create a unique new label for disk groups cloned with FlexClone®.
- Ability to create a consistent point-intime Snapshot copy across multiple storage systems. SMO can create consistent Snapshot copies of an ASM or Oracle instance, even when the instance spans multiple store systems, and even when the storage shares the underlying disks with other applications.
On rare occasions, faulty hardware can corrupt data can in flight between server and storage. Because Oracle keeps a checksum on each block, such corruption will be detected, but data corrupted during a write may not be read back and checked for many months. By then, the operational difficulties of recovering the data may be substantial.
NetApp SnapValidator for Oracle complies with the Oracle Hardware Assisted Resilient Data (HARD) initiative to verify the checksum and block offset every time a block is written. When a problem is detected, the write is failed, ensuring that corrupted blocks are never written to disk.
Enhancing FC SAN Management and Efficiency
NetApp offers all the standard features of mission-critical SAN storage, including high availability storage systems, non-disruptive upgrades, and a complete compatibility matrix. (More on compatibility later.) If you’re out of I/O capacity in one controller, for example, you can scale across controllers and still create consistent backups of your data. NetApp also offers a variety of unique features, including dual-parity RAID (see www.netapp.com/go/techontap/ matl/sample/0206tot_resiliency.html) (RAID-DP®), which provides the data reliability of mirroring (RAID1) at the price/performance of RAID5, and dynamic queue management (see www.netapp. com/go/techontap/matl/fc-sans.html), which shifts queue depth management from the host to the array target port.
NetApp also has in built protection against so called “lost writes.” Most vendors need to verify written data because they are unable to detect situations where the drive does not write the data but says it did. In those situations, most storage systems will see an old block with correct ECC and other checksums yet the RAID parity will be inconsistent. There is no way of recovering lost writes in these environments because it is impossible to tell which block is old. Because the NetApp WAFL file system writes new data to new places and uses a logical ID embedded in the checksum, NetApp offers the unique ability to identify lost writes.
|
