Re: /home: Stale NFS file handle
* Dave Sherohman (esper@sherohman.org) spake thusly:
> On Mon, Feb 11, 2002 at 10:10:46AM -0600, Dimitri Maziuk wrote:
> > Short answer: no.
>
> Ah, well. I've already rebooted them anyhow, though, so it doesn't
> really matter.
>
> > Next time unmount on the clients before doing something
> > like that.
>
> I just realized... The stale handles are because of the new drives
> (and resulting new inodes), not because of the server being offline
> for an extended time, aren't they?
At a guess, yes (IANA NFS guru).
...I wasn't too worried about it
> since NFS usually recovers when the server comes back up, but hadn't
> thought about things getting shifted around on the disk.
"Usually" is the keyword here. IME GNOME + NFS-mounted /home is
a very bad combination: even when NFS recovers fine, apps barf
about "no working directory" etc.
> Unmounting all the clients isn't too bad when there are only 8 of
> them, but I would expect it to be a bit of a PITA if you've got 50 or
> 100...
Having to reboot 50 or 100 machines because you can't unmount those
stale NFS handles may be a worse PITA, esp. when some of them run
obscure services that aren't documented anywhere and/or have no startup
scripts in /etc/init.d. OTGH, getting lusers to log off before mucking
around the NFS server is usually enough, unmounting everything is
a bit of an overkill...
Dima
--
Q276304 - Error Message: Your Password Must Be at Least 18770 Characters
and Cannot Repeat Any of Your Previous 30689 Passwords -- RISKS 21.37
Reply to: