Informix must rely on the OS to read a chunk from disk
in order to send it to the storage manager. The purpose
of this utility is to measure how fast the OS can read data
from a given chunk.
How To Use:
To use this uility you must first
compile the C code to an executable.
./diskreader : FILE_NAME BLOCK_SIZE OFFSET #_OF_BLOCKS(optional)
FILE_NAME: The complete pathname to the chunk/disk we want to read
BLOCK_SIZE: The blocksize (in Kbytes) used to read. I would suggest
OFFSET: The offset to start reading from (in Kbytes)
#_OF_BLOCKS: The number of blocks to read. If no values is specified, then
it will read until it encounters an End-Of-File marker.
$ ./diskreader /data6/tiller_test/data3 8 0 12500
Attempting to reading 12500 blocks of size 8 K from /data6/tiller_test/data3
Read 12500 blocks of 8K from /data6/tiller_test/data3. Total size read is: 100
It took 34 seconds. Rate : 2941.000000K per sec
If you are using cooked files then it can be very difficult to accurately
measure the throughput. The cooked file its self can be cached in memory
at some points and not cashed in others.
Not available yet. Please Email
firstname.lastname@example.org for anadvanced copy.
Compiling the Source:
You must compile the source code to an executable. On Solaris platforms
the complie line will look like the following:
cc diskreader.c -o diskreader