Dd command: Test hard disk I/O performance
I. dd replication and conversion
1. dd: convert and copy a file conversion and replication, bottom layer (Block) of the copy operation)
DESCRIPTION: Copy a file, converting and formatting according to the operands.
2. Usage: dd [OPTION] if =/PATH/FROM/SRC of =/PATH/TO/DEST
If =: the path is the source target path.
Of =: the location to be copied. The target path
[OPTION]
Bs = #: block size, the size of the replication unit. The default unit is byte;
Count = #: How many bs are copied;
3. Disk copy: dd if =/dev/sda of =/dev/sdb
4. MBR backup and damage
Backup MBR: dd if =/dev/sda of =/tmp/mbr. bak bs = 512 count = 1
Damage MBR Partition Table: dd if =/dev/zero of =/dev/sda bs = 512 count = 1
Destroy bootloader in MBR: dd if =/dev/zero of =/dev/sda bs = 256 count = 1
The first 446 bytes are the space occupied by the bootloader. Therefore, the entire damaged part will be damaged.
5. Two special devices:/dev/zero and/dev/null
/Dev/zero)
/Dev/null: data black hole; similar to recycle bin
Ii. dd hard disk I/O performance test
Several common DD commands: the difference lies in the way in which the cache is written in the memory.
Dd bs = 128 k count = 10 k if =/dev/zero of = test
Dd bs = 128 k count = 10 k if =/dev/zero of = test; sync
Dd bs = 128 k count = 10 k if =/dev/zero of = test conv = fdatasync
Dd bs = 128 k count = 10 k if =/dev/zero of = test oflag = dsync
WKioL1Z1KgGDda7VAABTcGLbp9w514.png
1. dd bs = 128 k count = 10 k if =/dev/zero of = test
(1) by default, the "sync" command is not included in the dd command. Before the dd command is completed, the system did not actually write the file to the disk.
(2) simply read data to the memory buffer (write cache [write cache])
(3) only the read speed is displayed, and the system does not actually write data to the disk until dd is complete.
Actual running result:
2. dd bs = 128 k count = 10 k if =/dev/zero of = test; sync
(1) only two independent commands are separated by semicolons (;).
(2) The dd command has displayed the "write speed" value on the screen, and cannot get the real write speed.
Actual running result:
3. dd bs = 128 k count = 10 k if =/dev/zero of = test conv = fdatasync
(1) After the dd command is executed, a "sync" operation will be executed.
(2) The time required to read the M data to the memory and write the data to the disk.
(3) comparison with actual conditions
Actual running result:
4. dd bs = 128 k count = 10 k if =/dev/zero of = test oflag = dsync
(1) dd performs synchronous write operations every time it is executed.
(2) write the 128 k to the disk after each k read, and then read the following k
(3) the slowest way, basically not using the write cache)
Actual running result:
The fourth method is the strictest. It can simulate database insert operations, so it is very slow.
It is also used to test the performance standard of the vps hard disk. Generally, if the test result is over 10 MB, it will not affect normal website construction. If it is over 50 MB, it will be very good.
In addition, the test damages the hard disk.
This article permanently updates the link address: