nbdkit-cow-filter - nbdkit copy-on-write (COW) filter
nbdkit --filter=cow plugin [plugin-args...] [cow-block-size=N] [cow-on-cache=false|true] [cow-on-read=false|true|/PATH]
nbdkit-cow-filter is a filter that makes a temporary writable copy on top of a plugin. It can also be used to enable writes for plugins which are read-only.
The underlying plugin is opened read-only. This filter does not pass any writes or write-like operations (like trim and zero) through to the underlying plugin.
Note that anything written is thrown away as soon as nbdkit exits. If you want to save changes, either copy out the whole disk using a tool like nbdcopy(1), or use the method described in "NOTES" below to create a diff.
Limitations of the filter include:
The underlying file/device must not be resized.
The underlying plugin must behave “normally”, meaning that it must serve the same data to each client.
Set the block size used by the filter. This has to be a power of two and the minimum block size is 4K. The maximum block size depends on the plugin, but a block size larger than a few megabytes is not usually a good idea.
The default is 64K.
Do not save data from cache (prefetch) requests in the overlay. This leaves the overlay as small as possible. This is the default.
When the client issues a cache (prefetch) request, preemptively save the data from the plugin into the overlay.
Do not save data from read requests in the overlay. This leaves the overlay as small as possible. This is the default.
When the client issues a read request, copy the data into the overlay so that the same data can be served more quickly later.
When /PATH (which must be an absolute path) exists, this behaves like
cow-on-read=true, and when it does not exist like
cow-on-read=false. This allows you to control the
cow-on-read behaviour while nbdkit is running.
Serve the file disk.img, allowing writes, but do not save any changes into the file.
nbdkit-xz-filter(1) only supports read access, but you can provide temporary write access by using the command above. Because xz decompression is slow, using
cow-on-read=true causes reads to be cached as well as writes, improving performance at the expense of using more temporary space. Note that writes are thrown away when nbdkit exits and do not get saved into the file.
Although nbdkit-cow-filter itself cannot save the differences, it is possible to do this using an obscure feature of qemu-img(1). nbdkit must remain continuously running during the whole operation, otherwise all changes will be lost.
nbdkit --filter=cow file disk.img
and then connect with a client and make whatever changes you need. At the end, disconnect the client.
qemu-img commands to construct a qcow2 file containing the differences:
qemu-img create -F raw -b nbd:localhost -f qcow2 diff.qcow2 qemu-img rebase -F raw -b disk.img -f qcow2 diff.qcow2
diff.qcow2 now contains the differences between the base (disk.img) and the changes stored in nbdkit-cow-filter.
nbdkit can now be killed.
All connections to the nbdkit instance see the same view of the disk. This is different from nbd-server(1) -c option where each connection sees its own copy-on-write overlay and simply disconnecting the client throws that away. It also allows us to create diffs as above.
The copy-on-write changes are stored in a temporary file located in /var/tmp by default. You can override this location by setting the
TMPDIR environment variable before starting nbdkit.
nbdkit --dump-config to find the location of
nbdkit-cow-filter first appeared in nbdkit 1.2.
nbdkit(1), nbdkit-file-plugin(1), nbdkit-cache-filter(1), nbdkit-cacheextents-filter(1), nbdkit-xz-filter(1), nbdkit-filter(3), nbdcopy(1), qemu-img(1).
Richard W.M. Jones
Copyright Red Hat
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
Neither the name of Red Hat nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY RED HAT AND CONTRIBUTORS ''AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL RED HAT OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.