$plugindir/nbdkit-sparse-random-plugin.so
The plugin.
Use "nbdkit --dump-config" to find the location of $plugindir.
nbdkit-sparse-random-plugin1
Make sparse random disks
Usage
nbdkit sparse-random
[size=]
SIZE
[seed=SEED]
[percent=N]
[runlength=N]
[random-content=true]
Overview
"nbdkit-sparse-random-plugin" is a plugin for nbdkit(1) which makes disks containing sparse blocks of random data. These disks have a similar shape to virtual machine disks, and this plugin can be used to benchmark disk copying tools like nbdcopy(1). To get a non-sparse random disk, see nbdkit-random-plugin(1). For a plugin containing a predictable test pattern see nbdkit-pattern-plugin(1).
The size of the virtual disk must be specified using the "size" parameter. If you specify the "seed" parameter then you will get the same random data over multiple runs with the same seed.
The "percent" parameter controls the percentage of the disk which contains random data versus sparse empty space.
The plugin does not generate random data spread evenly over the disk. Instead to make it look more like a real virtual machine disk, it tries to create runs of data and runs of empty space. The "runlength" parameter controls the average length of each run of random data.
The data in each block normally consists of the same random non-zero byte repeated over the whole block. If you want fully random content within each block use "random-content=true". This is not the default because earlier testing of this plugin showed that a great deal of time was spent generating random content. The random content is generated using a method which is not cryptographically secure.
Writes and testing copying
Writes to the disk verify that the data written is the same as the data read (if not, returning EIO). Thus when testing copies you can use a single instance of this plugin for both read and write:
nbdkit sparse-random size=1T --run 'nbdcopy "$uri" "$uri"'"qemu-img convert" could be used in place of nbdcopy. See also nbdkit-checkwrite-filter(1).
Parameters
percent=N-
Specify the approximate percentage of the disk which contains random data versus sparse empty space. The default is 10 (10%).
"percent=0"will create a completely empty disk and"percent=100"will create a completely full disk.
random-content=true-
By default a single random non-zero byte is repeated over the whole block, which is fast to generate and check. If you want blocks where each byte is random, use this setting.
runlength=N-
Specify the average length of runs of random data. This is expressed in bytes and the usual modifiers can be used. The default is 16M, meaning that (on average) data runs will be 16 megabytes in length.
seed=SEED-
Specify the random seed to get repeatable data over multiple runs.
If not specified then a random seed is chosen.
- [
size=]SIZE -
Specify the virtual size of the disk image.
This parameter is required.
"size="prefix may be omitted in most cases. See "Magic parameters" in nbdkit(1).
Files
Version
"nbdkit-sparse-random-plugin" first appeared in nbdkit 1.24.
Releases
3 releases since Debian 11 (Bullseye); the text changed 2 times.