Hi
(This is reposted as the previous message didn't seem to arrive. Sorry
if it is there twice).
We have a fileserver here which gives us problems: when a lot of people
are using a particular filesystem from that server (which is a partition
on that server), the client machines start to unmount that filesystem
and it becomes impossible to remount. The error message when trying to
remount is "bad block or too many filesystems mounted".
The server is a Redhat 6.0 machine and the rpm for nfsd, mountd and
quotad daemons is knfsd-1.4.6-2.
Is there a limit to the number of client machines the server can supply
with its filesystems? And how is that limit reached? All the machines
mount the filesystems happily, but it is only when people are logged in
and actively using a particular filesystem that the mounts start
dropping. So it seems it is not the number of machines that mount which
matters but the degree of active use of a filesystem. If this is the
case, can something be done about to stop filesystems unmounting?
The second problem, possibly related to the first, is that the
rpc.rquotad daemon keeps stopping. There is also an often repeated
message in /var/log/messages:
Apr 11 12:03:16 kidna kernel: nfsd Security: /// bad export.
I checked the /etc/exports file - everything is normal. I also checked
the /etc/hosts file where it maps the IP addresses in the export files
to computer names - that file is correct and there are no erroneous
entries in there.
Does anyone have any idea how the three problems may be related - the
stopping of the rpc.quotad daemon, the message "bad export" and the
dropping of mounts?
Any help will be most appreciated.
Thanks
Hugo
Dr Hugo Bouckaert - Systems Administrator, Computer Science UWA
Tel: +(61 8) 9380 2878 / Fax: +(61 8) 9380 1089
Email: [EMAIL PROTECTED] / Web: http://www.cs.uwa.edu.au/~hugo
--
To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
as the Subject.
