> Suppose I have an NFS mount on /foo and another on top of that, say > /foo/bar. Now, when the NFS server dies, I can't umount /foo because > of the /foo/bar mount. But I also cannot umount /foo/bar because > that apparently tries to access /foo/bar (the bar directory in the > /foo NFS mount) and that failes.
> Is that intentional? No, I would guess. > Any way to get out short of rebooting? Well, "wait for the NFS server to come back" is the obvious one, but given what you said later that's not an option in your case. Other than that? Possibly, possibly not. I once ran into trouble unmounting something and added an option to prevent umount(8) from trying to be smart about anything, simply passing the path to unmount(2). That helped in my case, but it was long enough ago I can't recall how close it was to what you describe. Maybe write a tiny program to just call unmount(2) with "/foo/bar" and maybe MNT_FORCE? (I'm assuming the client system is NetBSD.) Might not work, but it sounds to me as though you don't have much to lose by trying it. /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mo...@rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
