Ubuntu Manpage: nbdkit-scan-filter - scan disk prefetching data ahead of sequential reads

Provided by: nbdkit_1.36.3-1ubuntu10_amd64

NAME

       nbdkit-scan-filter - scan disk prefetching data ahead of sequential reads

SYNOPSIS

        nbdkit --filter=scan PLUGIN [scan-ahead=false] [scan-clock=false]
                                    [scan-forever=true] [scan-size=]NN

        nbdkit --filter=scan --filter=cache PLUGIN

        nbdkit --filter=scan --filter=cow PLUGIN cow-on-cache=true

DESCRIPTION

"nbdkit-scan-filter" is a filter that scans the disk prefetching data. It is sometimes useful if you
expect that the client will read the disk sequentially.

The basic operation of the filter is that when a client connects, the filter will start issuing ".cache"
(prefetch) requests to the plugin across the whole disk. Plugins which support this command will
prefetch the data, making subsequent 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.

Various parameters can be used to tune scanning, although the defaults should be suitable in most cases.

A similar filter is nbdkit-readahead-filter(1).

Limitations
In a number of significant cases this filter will do nothing. The filter will print a warning message if
this happens.

Thread model must be parallel*
For example nbdkit-curl-plugin(1) only supports "serialize_requests", and so this filter cannot
perform prefetches in parallel with the read requests.

Only scans while clients are connected*
The current filter only scans while there is at least one client connected.

Only scans the default export*
The current filter only scans the default export and ignores all clients connecting to the non-
default export name.

*We may be able to lift these restrictions in future.

Underlying filters or plugin must support ".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.

Prefetching the whole disk may load it all into cache
In particular if you use this filter together with nbdkit-cache-filter(1) or nbdkit-cow-filter(1),
they will cache the whole content of the plugin into a temporary file. This may be many gigabytes of
data, consuming all space in /var/tmp. Of course this is the whole point of using this filter, but
you should be aware of it.

If using the cache filter, the total size of the cache can be limited (see "CACHE MAXIMUM SIZE" in
nbdkit-cache-filter(1)).

PARAMETERS

scan-ahead=false
By default the filter tries to stay ahead of incoming read requests. That is to say, it starts
prefetching at the beginning of the disk and continues incrementally, but if the client issues a read
beyond the current prefetch point then the filter skips forward and begins prefetching after the
read.

However if you set this parameter to false, then this behaviour is disabled. The filter simply
prefetches sequentially regardless of client requests.

scan-clock=false
By default, if all clients disconnect and then another client connects, prefetching resumes at the
same place in the disk. (Like stopping and starting a clock.)

If you set this parameter to false, then the filter starts prefetching from the beginning of the disk
again.

scan-forever=true
By default the filter scans over the disk once and then stops.

If you set this parameter to true, then after the disk has been prefetched completely, the filter
goes back to the beginning and starts over, repeating this for as long as nbdkit is running and there
are clients connected.

scan-size=NN
This parameter controls the prefetch block size. The default is "2M". This must be a power of 2 and
most plugins will have their own limits on the amount of data they can prefetch in a single request.

FILES

       $filterdir/nbdkit-scan-filter.so
           The filter.

           Use "nbdkit --dump-config" to find the location of $filterdir.

VERSION

       "nbdkit-scan-filter" first appeared in nbdkit 1.32.

AUTHORS

       Richard W.M. Jones

COPYRIGHT

       Copyright Red Hat

LICENSE

       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.

nbdkit-1.36.3                                      2024-03-31                              nbdkit-scan-filter(1)