Use /usr/bin/time -h dd if = /dev / zero of = sometestfile bs = 1024 count = 10:
10240 bytes transferred in 0.000221 secs (46331902 bytes/sec)
The reverse is also true:
10240 bytes transferred in 0.000107 secs (95656287 bytes/sec)
Converted to 46 MB/s write and 95 MB/s read speed. This seems a bit slow. Or is this completely normal?
Thank you
Tobias Timpe
< p>I run this command on my home ZFS file server (Core i3-4130T with mirror WDC WD80):
# /usr/bin/time -h dd if=/dev/zero of=sometestfile bs=1024 count=10000000
10240000000 bytes transferred in 55 secs (186900359 bytes/sec)
It’s not good, but it’s for actual writing Into a better test of performance. An interesting result is that I reached 187MB/s, and my disk rated performance is 178MB/s. This is a bit strange b/c typical benchmark is a small part of the rated performance. More than it can pass ZFS disk compression to explain.
My disk is in a ZFS mirror (2x read performance is more useful for me than storage efficiency) but assuming your disk is also 5400 RPM, you should expect comparable Write performance.
To understand why even using a larger count of dd is not a good test, please check:
# /usr /bin/time -h dd if=/dev/zero of=sometestfile bs=1M count=1000000
1048576000000 bytes transferred in 223.56 secs (4690377542 bytes/sec)
Just increase the blocksize By 1M, my disk performance has now exceeded the level of the chart. Until you realize that reading from /dev/zero means that the data source has no entropy and is highly compressible. dd is not the best tool for testing disk performance. /p>
There are other tools in the FreeBSD ports tree (I used to Used bonnie & iozone) can tell you more about disk performance. If you are really curious, then go check it out. Otherwise, the better performance test for the main file server is between the two disks Copy the contents of DVD or Blu-ray and the time of copying.
My home storage server, currently consists of 6 x 3 TB Seagate drives, 12 GB DDR3 RAM and Core i3 540 , The performance is very poor when reading or writing the raidz2 I set.
Use /usr/bin/time -h dd if = /dev / zero of = sometestfile bs = 1024 count = 10:
10240 bytes transferred in 0.000221 secs (46331902 bytes/sec)
The reverse is also true:
10240 bytes transferred in 0.000107 secs (95656287 bytes/sec)
Converted to 46 MB/s write and 95 MB/s read speed. This seems a bit slow. Or is this completely normal ?
Thank you
Tobias Timpe
Using the commands shown above, you will not test the disk write performance. You are testing various factors, but mainly controller and disk latency. To test disk read performance, you need to jump through the hoops to exclude disk caching from the test, which is very important for ZFS because it means Disable ARC cache. It’s much easier to test write performance: write larger files, preferably files larger than the system’s built-in RAM.
I’m working on my home ZFS file server (with Run this command on Core i3-4130T with mirrored WDC WD80:
# /usr/bin/time -h dd if=/dev/zero of=sometestfile bs= 1024 count=10000000
10240000000 bytes transferred in 55 secs (186900359 bytes/sec)
It is not good, but it is a better test of actual write performance. An interesting result is I reached 187MB/s and my disk rated performance is 178MB/s. This is a bit strange b/c typical benchmark is a small part of the rated performance. Exceeding it can be explained by ZFS’s disk compression.
< p>My disk is in a ZFS mirror (2x read performance is more useful for me than storage efficiency) but assuming your disk is also 5400 RPM, you should expect comparable write performance.
To understand why even using a larger count of dd is not a good test, please check:
# /usr/bin/time -h dd if=/dev/ zero of=sometestfile bs=1M count=1000000
1048576000000 bytes transferred in 223.56 secs (4690377542 bytes/sec)
Just increase the blocksize to 1M, and my disk performance has now exceeded the chart Until you realize that reading from /dev/zero means that the data source has no entropy and is highly compressible. dd is not the best tool for testing disk performance.
There are also in the FreeBSD ports tree Other tools (I have used bonnie & iozone before) can tell you more about disk performance. If you are really curious, then go check it out. Otherwise, perform The better performance test is to copy the contents of DVD or Blu-ray between two disks and the copy time.