Perform bechmarks on the filesystem: Difference between revisions
Created page with "To test the speed of the filesystem (not the physical disk!) you can use the following script: <pre> #!/bin/bash # USAGE: # ./speed_test.sh /path/to/my/file /path/to/destinat..." |
No edit summary |
||
| Line 1: | Line 1: | ||
To test the speed of the filesystem (not the physical disk!) you can use the following script: | *To test the speed of the filesystem (not the physical disk!) you can use the following script: | ||
<pre> | <pre> | ||
#!/bin/bash | #!/bin/bash | ||
| Line 25: | Line 25: | ||
</pre> | </pre> | ||
*To make a big file, use | |||
cat /dev/zero > zerofile | cat /dev/zero > zerofile | ||
and abort it with CTRL-C after a few seconds. | and abort it with CTRL-C after a few seconds. | ||
| Line 33: | Line 33: | ||
to create a 500MB file. | to create a 500MB file. | ||
*Start the script like this: | |||
Start the script like this: | |||
./speed_test.sh /mnt/sda1/zerofile /mnt/sdb1/zerofilecopy 3 | ./speed_test.sh /mnt/sda1/zerofile /mnt/sdb1/zerofilecopy 3 | ||
Revision as of 19:49, 31 January 2012
- To test the speed of the filesystem (not the physical disk!) you can use the following script:
#!/bin/bash
# USAGE:
# ./speed_test.sh /path/to/my/file /path/to/destination number_of_tests
NUM_TESTs=$3
SUM=0
for i in $( seq 1 $NUM_TESTs ); do
REC=`dd if=$1 of=$2 2> some_random_file_ ; cat some_random_file_ | cut -d " " -f8 | tail -1`
SUM=`echo $SUM + $REC | bc`
done
RESULT=`echo $SUM / $NUM_TESTs | bc | awk '{ str1=str1 $0 }END{ print str1 }'`
echo $RESULT MB/s
#clean up
rm some_random_file_
rm $2
- To make a big file, use
cat /dev/zero > zerofile
and abort it with CTRL-C after a few seconds.
Alternatively, you can use something like
dd if=/dev/zero of=file.out bs=1MB count=500
to create a 500MB file.
- Start the script like this:
./speed_test.sh /mnt/sda1/zerofile /mnt/sdb1/zerofilecopy 3