On Wed, Aug 20, 2008 at 04:31:24PM -0700, Andrew Lentvorski wrote:
One of the goals of ZFS was to be "fsck-less". You don't have to fsck to
get a usable file system, but it seems that you have to run some vacuuming
periodically to keep the performance up.
XFS is "fsck-less", in that their fsck implementation is, literally:
int main(int argc, char**argv)
{
printf("%s: XFS file system.\n", argv[0]);
exit(0);
}
I suspect even ZFS is subject to catastrophic corruption, such as
garbage being written to unusual places on the disk, although, they
might just recover that by having less data available on the next use
of the filesystem.
XFS does include a utility xfs_repair to try and help recover this
kinds of situations. With my vmware-scribble-on-image events that
have happened, I haven't been very happy with the results.
They do also include an xfs_check you can run, but only on unmounted
filesystems. It's more of a sanity check, since pretty much anything
it prints indicates a bug in the XFS code, even after a bad powerdown.
David
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
