Hi, I can confirm this is not a configuration issue, but a regression. I keep nfs-kernel-server version 1.2.0-4.1 pinned on my server, since that's a known working version. The current version in unstable (1.2.1-3) is broken.
The behaviour seems to be very strange; at first mount from the client, the mount is using wrong option, but at subsequent mounts, it's using the right ones. Maybe it's a bug in mountd, in the way it parses /var/lib/nfs/etab? Replacing just the mountd binary with the old one fixes the issue. Removing completely the crossmnt option makes the server behave again normally. It seems to me that the crossmnt option doesn't work reliably even with single-host exports? iustin -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org