search for: fuamode

Two hopefully useful additions to the fua filter. The second one is kind of like cache=unsafe in qemu, in that it exchanges correctness for speed. Useful for data which is easily recreated in the event of a crash or for people who like living on the edge and have good backups. Rich.

...to test client or +server fallbacks, and for evaluating timing differences between proper +use of FUA compared to a full flush. =head1 PARAMETERS @@ -53,6 +55,15 @@ returns C<NBDKIT_FUA_NONE>. Pass through FUA and flush requests unchanged. Turns the filter into a no-op. +=item B<fuamode=discard> + +The filter will discard FUA and flush requests. + +B<This mode is unsafe>: If the NBD disk contains a filesystem then you +will likely lose data in the event of a crash. It should only be used +for ephemeral data which you can easily recreate, such as caches, +builds, test dat...

...ating timing differences between proper > +use of FUA compared to a full flush. > > =head1 PARAMETERS > > @@ -53,6 +55,15 @@ returns C<NBDKIT_FUA_NONE>. > Pass through FUA and flush requests unchanged. Turns the filter into > a no-op. > > +=item B<fuamode=discard> > + > +The filter will discard FUA and flush requests. > + > +B<This mode is unsafe>: If the NBD disk contains a filesystem then you > +will likely lose data in the event of a crash. It should only be used > +for ephemeral data which you can easily recreate, suc...

...nbdkit will be emulating FUA with a flush, by comparing the log filter output on top of different fua filter modes: - nbdkit --filter=blocksize --filter=log --filter=fua file file=disk.img \ + nbdkit --filter=blocksize --filter=log --filter=fua file disk.img \ maxlen=4k logfile=fua_emulated fuamode=emulate - nbdkit --filter=blocksize --filter=log --filter=fua file file=disk.img \ + nbdkit --filter=blocksize --filter=log --filter=fua file disk.img \ maxlen=4k logfile=fua_native fuamode=native Serve the file F<disk.img> in write-through mode, where all writes from the client are i...

...nbdkit will be emulating FUA with a flush, by comparing the log filter output on top of different fua filter modes: - nbdkit --filter=blocksize --filter=log --filter=fua file file=disk.img \ + nbdkit --filter=blocksize --filter=log --filter=fua file disk.img \ maxlen=4k logfile=fua_emulated fuamode=emulate - nbdkit --filter=blocksize --filter=log --filter=fua file file=disk.img \ + nbdkit --filter=blocksize --filter=log --filter=fua file disk.img \ maxlen=4k logfile=fua_native fuamode=native Serve the file F<disk.img> in write-through mode, where all writes from the client are i...

Remove the need to use file= (and in future other) parameters for many plugins. eg. Using the file plugin becomes: nbdkit file disk.img Rich.

...ts/test-fua.sh index f95aa18..88c14b5 100755 --- a/tests/test-fua.sh +++ b/tests/test-fua.sh @@ -31,6 +31,7 @@ # OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF # SUCH DAMAGE. +source functions.sh set -e set -x @@ -83,34 +84,22 @@ trap cleanup INT QUIT TERM EXIT ERR # 2: fuamode=emulate: log shows that blocksize optimizes fua to flush # 3: fuamode=native: log shows that blocksize preserves fua # 4: fuamode=force: log shows that fua is always enabled -nbdkit -P fua1.pid -U fua1.sock --filter=log --filter=fua \ - file logfile=fua1.log fua.img -nbdkit -P fua2.pid -U f...

v1 was here: https://www.redhat.com/archives/libguestfs/2018-September/msg00024.html v2: - As discussed in the patch review, tighten up the characters permitted in keys. - Update documentation to note that relative paths can be made safe by prefixing with ./ and absolute paths do not need any extra steps. - I pushed patch 1/6 from the v1 series since it was just a trivial

v1: https://www.redhat.com/archives/libguestfs/2018-September/msg00024.html v2: https://www.redhat.com/archives/libguestfs/2018-September/msg00034.html v3: - Fixed is_config_key zero length test. - Fixed is_config_key so it uses strspn and is not O(n^2). - Changed >= 1.7 to >= 1.8 in the documentation. Rich.

...mp;& kill $pid3 - test "$pid4" && kill $pid4 - # For easier debugging, dump the final log files before removing them. echo "Log 1 file contents:" cat fua1.log || : echo "Log 2 file contents:" @@ -78,34 +72,18 @@ cleanup_fn cleanup # 2: fuamode=emulate: log shows that blocksize optimizes fua to flush # 3: fuamode=native: log shows that blocksize preserves fua # 4: fuamode=force: log shows that fua is always enabled -nbdkit -P fua1.pid -U fua1.sock --filter=log --filter=fua \ - file logfile=fua1.log fua.img -nbdkit -P fua2.pid -U f...

Combine much common code into tests/functions.sh. Patch 1: Preparation for patch 3. Patch 2: Fix a long-standing bug in how man pages links are generated. Patch 3: Common code for iterating a test function over every plugin. Patch 4: Common code for starting nbdkit in a test and waiting for the PID file to appear. This is the largest and most complex of the patches but is

After more than a month since v2 [1], I've finally got my FUA support series polished. This is all of my outstanding patches, even though some of them were originally posted in separate threads from the original FUA post [2], [3] [1] https://www.redhat.com/archives/libguestfs/2018-January/msg00113.html [2] https://www.redhat.com/archives/libguestfs/2018-January/msg00219.html [3]

...hich is deprecated and will be removed in nbdkit 1.26. New nbdkit-tls-fallback-filter(1) provides alternate content when a client connects without using TLS to a server that permits but does not require TLS (Eric Blake). nbdkit-fua-filter(1) has new option "fuamode=pass" to pass through FUA and flush requests unchanged (disabling the filter). A new, unsafe "fuamode=discard" mode was added which discards FUA and flush requests. Language bindings New nbdkit-cc-plugin(3) lets you write small nbdkit plugins as inline...

v1 was here: https://www.redhat.com/archives/libguestfs/2018-September/msg00057.html v2: - Fix tab vs spaces in configure.ac. - To generate list of plugins, use printf instead of xargs. - Use 'source ./functions.sh' instead of 'source functions'. - functions.sh: Consistent quoting in foreach_plugin function. - functions.sh: Change the contract of start_nbdkit so it

Context: https://bugzilla.redhat.com/show_bug.cgi?id=1837809#c28 https://rwmj.wordpress.com/2020/03/21/new-nbdkit-remote-tmpfs-tmpdisk-plugin/ http://libguestfs.org/nbdkit-tmpdisk-plugin.1.html https://github.com/libguestfs/nbdkit/blob/0632acc76bfeb7d70d3eefa42fc842ce6b7be4f8/plugins/tmpdisk/tmpdisk.c#L182 I did a bit of testing to try to see if this is really feasible, and yes I think it is.

.../* rate, 0.0 = never, 1.0 = always */ diff --git a/filters/fua/fua.c b/filters/fua/fua.c index 120ff0a..9d0e561 100644 --- a/filters/fua/fua.c +++ b/filters/fua/fua.c @@ -41,8 +41,6 @@ #include <nbdkit-filter.h> -#define THREAD_MODEL NBDKIT_THREAD_MODEL_PARALLEL - static enum FuaMode { NONE, EMULATE, diff --git a/filters/log/log.c b/filters/log/log.c index 133e352..9c0f35c 100644 --- a/filters/log/log.c +++ b/filters/log/log.c @@ -47,8 +47,6 @@ #include "cleanup.h" -#define THREAD_MODEL NBDKIT_THREAD_MODEL_PARALLEL - static uint64_t connections; static char...

...uot; && kill $pid3 @@ -73,10 +70,8 @@ cleanup () echo "Log 4 file contents:" cat fua4.log || : rm -f $files - - exit $status } -trap cleanup INT QUIT TERM EXIT ERR +cleanup_fn cleanup # Run four parallel nbdkit; to compare the logs and see what changes. # 1: fuamode=none (default): client should send flush instead diff --git a/tests/test-log.sh b/tests/test-log.sh index f811de4..5ad5f22 100755 --- a/tests/test-log.sh +++ b/tests/test-log.sh @@ -31,6 +31,7 @@ # OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF # SUCH DAMAGE. +source ./funct...

Being able to programmatically force nbdkit to be less parallel can be useful during testing. I was less sure about patch 3, but if you like it, I'm inclined to instead squash it into patch 1. This patch is written to apply after my NBD_CMD_CACHE work (since I touched the nocache filter); but can be rearranged if we think this series should go in first while that one undergoes any adjustments