It was discovered `rclone mount` (but not `rclone cmount`) cached
directories after rename which it shouldn't have done.
This caused IO errors when trying to access files in renamed
directories on bucket based file systems.
This turned out to be the kernel caching the directories as basil/fuse
sets their expiry time to 60s for some reason.
This fix invalidates the relevant kernel cache entries in the for the
directories which fixes the problem.
Fixes: #4977
See: https://forum.rclone.org/t/after-a-directory-renmane-using-mv-files-are-not-visible-any-longer/22797
It turns out that NFS calls mknod in FUSE even though we have create
defined. This was causing EIO errors when creating files.
This patch fixes it by implementing mknod. The way it is implemented
means that to write to an NFS file system you'll need --vfs-cache-mode
writes.
This limit was previously 4k set in 59026c4761 however leaf
names above 1k now produce an IO error.
WinFSP seems to have its own method for dropping too long file names
above 255 long.
Before this change, the current working directory could disappear
according to the Linux kernel.
This was caused by mount returning different nodes with the same
information in.
This change uses vfs.Node.SetSys to cache the information so we always
return the same node.
The tests are now run for the mount commands and for the plain VFS.
This makes the tests much easier to debug when running with a VFS than
through a mount.
This problem was introduced in "mount: allow files of unkown size to
be read properly" 0baafb158f by failure to check that the
DirEntry was nil or not.
Before this change, files of unknown size (eg Google Docs) would
appear in file listings with 0 size and would only allow 0 bytes to be
read.
This change sets the direct_io flag in the FUSE return which bypasses
the cache for these files. This means that they can be read properly.
This is compatible with some, but not all applications.
Before this change rclone marked files opened for write without VFS
cache with the non seekable flag.
This caused problems with rclone mount layerd with mergerfs.
This change removes the hint and lets rclone do all the checking for
seekability.
OSX FUSE only supports 32 bit number of blocks which means that block
counts have been wrapping. This causes f_bavail to be 0 which in turn
causes problems with programs like borg backup.
Fixes#2356
By default the timeout is 60s which isn't long enough for long
transactions. The symptoms are rclone just quitting for no reason.
Supplying the --daemon-timeout flag fixes this causing the kernel to
wait longer for rclone.
When running `rclone mount`, there were 2 signal handlers for `os.Interrupt`.
Those handlers would run concurrently and in some cases cause either unmount or `atexit.Run()` being skipped.
In addition `atexit.Run()` will get called in `resolveExitCode` to ensure cleanup on errors.
Before this change we would unconditionally set the OSXFUSE options
noappledouble and noapplexattr.
However the noapplexattr options caused problems with copies in the
Finder.
Now the default for noapplexattr is false so we don't add the option
by default and the user can override the defaults using the
--noappledouble and --noapplexattr flags.
Before this change rclone would set the volume name from the
remote:path normally. However this has `:` and `/` in which make it
difficult to use in macOS.
Now rclone will remove the special characters and replace them with
spaces. It also allows the volume name to be set with the --volname
flag.
By default bazil fuse will return ENOTSUPP for these. However if we
return ENOSYS then OSXFUSE (at least) will never call them again
saving round trips though fuse.
