John Cui wrote: > So I want to confirm it is the desired result for NFSv4, is it? If so, > maybe it breaks the consistence of mount operation. >
If you can see c1, with the current set of shares, then it is a bug. If you can not see c1, then it is not a bug. The security model states that you can not see real fs state in the pseudo-fs. > Thanks, > > Tom Haynes wrote on 09/06/07 23:51: >> Okay, I don't need traces. You are right that NFSv4's pseudo-fs is >> letting you mount server:/a/b. >> >> The only mount call which will work from v2 or v3 is: >> >> # mount -o vers=3 ultralord:/a/b/c /mnt >> >> A way to check to see that a pseudo-fs is in play is to try the >> following: >> >> On the server: >> mkdir -p /a/b/c >> mkdir /a/b/c1 >> >> On the client: >> mount server:/a/b /mnt >> ls -la /mnt >> >> And see if you can see c1 or not. >> >
