Db file sequential read
WilhelmOct 27 2007 — edited Oct 29 2007Due to high Disk i/o we took an AWR report.
The first of Top 5 Events was db file sequential read. So i looked up 'db file sequential read' in Oracle Documentation. It says
'A sequential read is a single-block read'.Single block I/Os are usually the result of using indexes. Rarely, full table scan calls could get truncated to a single block call due to extent boundaries, or buffers already present in the buffer cache. These waits would also show up as 'db file sequential read'.
From Tablespace I/O stats and File I/O stats of the AWR i came to know that the most number of Physical Reads are from datafiles of one particular Tablespace. So i assumed that the single block reads(which caused 'db file sequential read') are happening in this tablespace.
The current block size of this tablespace is 8K. I am thinking of creating a tablespace with non-standard block size of 16k or 32k so that more rows will fetched single block reads and moving all segments into this tablespace.
But i read in the below mentioned link that this might be a bad idea eventhough the author hasn't mentioned any particular reason:
"This feature should not be used as a performance feature and tempt the dba's to create larger block size tablespaces so that it reads more number of rows in one I/O"
http://viveklsharma.blogspot.com/2007/05/effect-of-non-standard-block-sizes-on.html
Is the creation of this large non-standard block size tablespace a bad idea?