Tuning the Storage Subsystem - Expert Oracle RAC Performance Diagnostics and Tuning - page 361

parity for every write operation. The parity disks avoid the cost of full duplication of the disk drives of RAID1. If a disk
fails, for example, due to a head crash, parity is used to reconstruct data without system loss using the information
stored on other disks in the array. Both data and parity are spread across all the disks in the array, thus reducing disk
bottleneck problems. Read performance is improved, but every write has to incur the additional overhead of reading
old parity, computing new parity, writing new parity, and then writing the actual data, with the last two operations
happening while two disk drives are locked.
Redolog Files
In the various discussions around datafiles, the operations performed on them are of a mixed-operations nature, as
there are writes and reads from the datafiles. Redolog files write often and read rarely (for example, in case a recovery
may be required). Since the type of operation is writing, no gains can be obtained by striping the storage array.
Figure 12-4 illustrates that redolog rows are written sequentially, one after the other. Due to the sequential nature of
writes, I/O is also sequential, so a RAID1 storage-array configuration would provide performance benefits for redolog
implementation.
RAID1
RAID1 is known as mirroring and is used when all the writes issued to a given disk are duplicated to another disk.
This provides a high-availability solution; if there is a failure of the first disk, the second disk or mirror can take over
without any data loss. Apart from providing redundancy for data on the disks, mirroring also helps reduce read
contention by directing reads to disks volumes that are less busy.
Similar to redolog files, backup sets generated by the recovery manager (RMAN) and archive logs generated
during redolog switch activity are all mostly write configurations and are good candidates to be configured on RAID1.
Testing to Determine Performance
Apart from tuning specific areas of the application for performance, it is important to look at the configuration of
the I/O subsystem and then tune the system for performance. For example, Figure 12-8 is an output from OEM that
illustrates high I/O activity against the storage array. As highlighted, the primary type of activity at the time of the
snapshot is direct reads, and because of direct-read waits, there are also high buffer-cache reads.