Hi there,
below quote is from steve adams suggesting to have db
block size and filesystem buffer size of same size to
avoid io issues in oracle.
Will this have any effect on postgresql. we use linux
7.3 with ext3 where fs buffersize is 4k however
postgres default blocksize is 8k. any performance gain
in having both sizes same in postgresql.
thanks,
Shankar
<quote from ixora.com.au>
"If the database block size is larger than the file
system buffer size, then all single block reads and
writes are split into a series of distinct physical
I/O operations - one for each file system buffer
addressed - and these I/O operations are performed one
at a time. For writes, a full rotational latency is
sustained between each pair of file system buffer
writes. For reads, all except the initial file system
buffer read can normally be satisfied cheaply from the
disk track buffers. However, the distinct contiguous
file system buffer reads trigger the file system read
ahead mechanism, even for random reads, causing
inappropriate read ahead."
</quote>
__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com