nbdkit-readahead-filter - prefetch data ahead of sequential reads
nbdkit --filter=readahead PLUGIN nbdkit --filter=readahead --filter=cache PLUGIN nbdkit --filter=readahead --filter=cow PLUGIN cow-on-cache=truenbdkit-readahead-filter is a filter that prefetches data when the client is reading.
When the client issues a read, this filter issues a parallel prefetch (.cache) for subsequent data. Plugins which support this command will prefetch the data, making sequential reads faster. For plugins which do not support this command, you can inject nbdkit-cache-filter(1) below (after) this filter, giving approximately the same effect. nbdkit-cow-filter(1) can be used instead of nbdkit-cache-filter, if you add the cow-on-cache=true option.
The filter uses a simple adaptive algorithm which accelerates sequential reads and requires no further configuration.
A similar filter is nbdkit-scan-filter(1) which reads ahead over the whole disk, useful if you know that the client will be reading sequentially across most or all of the disk.
In a number of significant cases this filter will do nothing. The filter will print a warning message if this happens.
For example nbdkit-curl-plugin(1) only supports serialize_requests, and so this filter cannot perform prefetches in parallel with the read requests.
We may be able to lift this restriction in future.
.cache (prefetch)Very many plugins do not have the concept of prefetching and/or do not implement the .cache callback, and so there is no way for this filter to issue prefetches.
You can usually get around this by adding --filter=cache after this filter as explained above. It may be necessary to limit the total size of the cache (see "CACHE MAXIMUM SIZE" in nbdkit-cache-filter(1)).
It may be the case that NBD clients are already issuing NBD_CMD_CACHE (NBD prefetch) commands. It may also be the case that your plugin is using local file functions where the kernel is doing readahead. In such cases this filter is not necessary and may be pessimal.
There are no parameters specific to nbdkit-readahead-filter. Any parameters are passed through to and processed by the underlying plugin in the normal way.
The filter.
Use nbdkit --dump-config to find the location of $filterdir.
nbdkit-readahead-filter first appeared in nbdkit 1.12.
nbdkit(1), nbdkit-cache-filter(1), nbdkit-cow-filter(1), nbdkit-file-plugin(1), nbdkit-retry-filter(1), nbdkit-scan-filter(1), nbdkit-torrent-plugin(1), nbdkit-vddk-plugin(1), nbdkit-filter(3), 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.