Is there a bl_files corruption issue?

What has appeared to be a corruption issue with bl_files is actually a FUSE (Filesystem in userspace) issue unrelated to BetterLinux. The issue arises from the way FUSE shows its mountpoint. FUSE returns –EACCESS when root tries to stat() a FUSE mountpoint.

All other FUSE-based filesystems have the same problem. See GNOME desktop’s GVFS virtual filesystem, for example: https://bugs.launchpad.net/ubuntu/+source/gvfs/+bug/225361

This issue has also been discussed in the fuse-devel mailing list: http://thread.gmane.org/gmane.comp.file-systems.fuse.devel/7169/focus=7169

The Problem:

Because of the above-mentioned situation, some BetterLinux users have run into problems ranging from a broken quota system to receiving concerning, but harmless, filesystem errors (like the “permission denied” error when root tried to read /etc/betterlinux/bl_files during a backup).

The Solution:

In the case of GVFS, their workaround was to move its mountpoint into tmpfs. Backup tools ignore tmpfs, so this prevented what appeared to be corruption issues.

BetterLinux has implemented the same workaround. We moved our custom FUSE mount point from /etc/betterlinux/bl_files to tmpfs, or more specifically, to /dev/shm/bl_files. Although this prevents crashes, it does not prevent errors such as those returned by running readdir()/stat() on the mountpoint, or running a ‘df’ or listing the files.

Of course, this is not a perfect solution. To completely resolve this problem, one would have to patch FUSE and change its default behavior. For now, you can hide the error like this:

# mv /etc/betterlinux/bin/bl_cloak /etc/betterlinux/bin/bl_cloak.disabled

Add to Favourites

Print this Article