Database Reference
In-Depth Information
■
in a database operation full-scan operations are treated as a direct-path read. a direct-path read is a physical
i/o from a data file that bypasses the buffer cache and reads the data block directly into process-private memory.
however, in environments that support asynchronous i/o, oracle can submit i/o requests and continue processing. oracle
can then later pick up the results of the i/o request and wait on “direct-path read” until the required i/o completes.
Note
Random Data Access
Random data access is typically observed in an OLTP environment or an environment that has high transaction-based
workloads. In such environments, a smaller volume of data is retrieved from different locations of the I/O system,
causing frequent movement of the disk headers. Appropriate tuning or optimization of objects using data partitioning,
as well as creating appropriate data-retrieval indexes, could minimize contention in such environments. Oracle
reports resource contention during such operations using a “
db file sequential read
” wait event.
I/Os are the size of the database block, e.g., 8 KB
•
•
Synchronous single reads (corresponds to single block or singleton select operations)
•
Asynchronous non-contiguous writes, i.e., scattered in buffer and on disk (corresponds to
DBWR writes)
•
Typically noticed in OLTP databases
Variables Affecting I/O Performance
The basic variables that affect I/O performance are the I/O size, stripe size, stripe width, concurrency, and alignment.
•
I/O size
—
the unit size in bytes of the single I/O operation; typically dictated by the block size
and or the
MULTI_BLOCK_READ_COUNT
parameter value
Stripe size
•
—
the size of an allocation unit on each disk device in an array unit. Each disk
device will have the same stripe size. In an ASM environment, apart from the strip size defined
at the LUN level, the graininess of the file (coarse grain or fine grain)
•
Stripe width
—
the product of the stripe size and number of disks in that stripe array
•
Concurrency
—
the number of distinct concurrent I/O operations against the disk array
■
aSm supports two types of file-extent distributions: coarse and fine. in a coarse-grain distribution, each file
extent is mapped to a single allocation unit. in the case of fine-grain distribution, each grain is interleaved 128K across
groups of 8 aus. Fine distribution breaks up large-sized i/o operations into multiple 128K i/o operations that can execute
in parallel, which benefits sequential i/os. Coarse- and fine-grain attributes are pre-defined as part of system templates
for all system-related files; e.g., redo and archive log files are defined as fine grain, whereas datafiles are coarse.
Note
Search WWH ::
Custom Search