> > of the services has no effect what so ever. It simply fixes itself > > after some time. Umounting and mounting the fs immediately fixes it, > Any progress on this?
Hard to say, but you seem to be correct at least in some aspects: it
looks like the timestamps of the ticket caches have something to do
with it. When the bug occurs it seems as if it can be fixed by:
server ~> /etc/init.d/nfs-kernel-server restart
followed by
client ~> kinit
<type the password>
And after this, write-access is re-established. Thus far, the above has
never failed to fix the bug.
I'm sorry I forgot to report this earlier, but it has ceased to bother me
since it can now be pretty easily be worked around without umounting
anything. Telling users to do a "kinit" is not a big deal (since they
learn very quickly); I just have to keep restarting nfs-kernel-server
every once in a while.
-Juha
P.S. The server is now an up-to-date etch, so the kernel at that end has
been updated.
--
-----------------------------------------------
| Juha Jäykkä, [EMAIL PROTECTED] |
| Laboratory of Theoretical Physics |
| Department of Physics, University of Turku |
| home: http://www.utu.fi/~juolja/ |
-----------------------------------------------
signature.asc
Description: PGP signature
