Board index Linux Networking

NFS - mount: RPC: Timed out - how to increase timeout

NFS - mount: RPC: Timed out - how to increase timeout

Post by martin.wool.. » Thu, 15 Dec 2005 19:18:00



Normally our nfs mounts just work, but occaisonally they fail due to
network traffic issues.   When they do fail I need to mount them
manually.

I would like to increase the timeout values so that the nfs mount will
try basically forever until the mount occurs, which it will given
enough time.  I have tried mount commands with "timeo=100,retrans=100",
which I think should fail after 1,000 seconds.  However this is clearly
not the case as the RPC Timed out message appears after 20 seconds,
which is exactly the same time as without the timeo and retrans options
(eg mount 172.24.0.19:/home  -t nfs /home/MAUD) .  Even ridiculously
high values eg  "timeo=100000000,retrans=100000000" do not increase the
timeout.

How can I force an nfs mount over a busy network?  I don't care if the
network gets slower, but our nfs mounts must work.
--
Regards
Martin Woolley
ICT Support
Handsworth Grammar School
Isis Astarte Diana Hecate Demeter Kali Inanna

 
 
 
Top

NFS - mount: RPC: Timed out - how to increase timeout

Post by Geronimo W. Christ Es » Fri, 16 Dec 2005 08:38:27



> Normally our nfs mounts just work, but occaisonally they fail due to
> network traffic issues.   When they do fail I need to mount them
> manually.

Does NFS over TCP make any difference ? (mount ... -otcp)

I've found that it greatly improves things over heavily congested
networks. Some older versions of Linux have a tendancy to lock up if
there are a lot of dropped packets, TCP helps deal with that.

 
 
 
Top

NFS - mount: RPC: Timed out - how to increase timeout

Post by martin.wool.. » Sun, 18 Dec 2005 00:02:25



Quote:

> Does NFS over TCP make any difference ? (mount ... -otcp)

> I've found that it greatly improves things over heavily congested
> networks. Some older versions of Linux have a tendancy to lock up if
> there are a lot of dropped packets, TCP helps deal with that.

Thanks for the reply.  Using tcp makes do difference; according to the
man page for nfs "The tcp and namlen options are implemented but are
not  currently  supported by the Linux kernel."  Mind you the man page
dates from 1993 so it is likely that tcp are now supported.  Today the
mount went just fine, but yesterday it refused to work at all.
Ethereal shows no difference in the amount of network traffic on either
day.
--
Regards
Martin Woolley
ICT Support - Handsworth Grammar School
Isis Astarte Diana Hecate Demeter Kali Inanna
 
 
 
Top

NFS - mount: RPC: Timed out - how to increase timeout

Post by Geronimo W. Christ Es » Sun, 18 Dec 2005 09:19:45



> Thanks for the reply.  Using tcp makes do difference; according to the
> man page for nfs "The tcp and namlen options are implemented but are
> not  currently  supported by the Linux kernel."  Mind you the man page
> dates from 1993 so it is likely that tcp are now supported.

The man page is wrong. NFS over TCP was standard in the Linux kernel
years ago. Long before 2.4.20. Did you try it ?
 
 
 
Top

1. * rwall and shutdown: rpc time-outs *

hello -

i was hoping someone could assist with a problem i having using shutdown under
solaris 2.5.1 -

my situation is that when i issue a shutdown, it attemps to rwall: to anyone
that mounted an NFS partition, (i assume because we are using the SUN automounter
facility.  Consequently, i get errors of the nature:

rwall: Can't send to [username]
RCP: Rpcbind failure - RPC: Timed out

I examined the various solutions, and the easiest appeared to be modifying
/etc/shutdown.  I was wondering if anyone could please tell me what lines
need to be edited in the notify() function.  

Currently i have added a   " - a" after the rwall  b/c all our users are
logged in via Xterminals.  however, i have not done anything else.  

Could you kindly advise which lines i should take out for remote broadcasts,
and which i should retain?

Here is the notify portion of my /etc/shutdown:

notify() {
        /usr/sbin/wall -a <<-!
        $1
        !
        if [ -x /usr/sbin/showmount -a -x /usr/sbin/rwall ]
        then
                remotes=`/usr/sbin/showmount`
                if [ "X${remotes}" != "X" ]
                then
                        /usr/sbin/rwall ${remotes} <<-!
                        $1
                        !
                fi
        fi

Any feedback would be most appreciated.

Sincerely,

-kari

2. IOCC Performance/Programming Questions

3. RPC time outs at boot time

4. Problems booting 2.2.2 with loadlin

5. nfs mount gives mount:RPC:Timed out error

6. UNIX: create: can't allocate buffer

7. NFS RPC Timeout RH9 to RH8 mount

8. Netscape Mail server sendmail error

9. NFS mounting problem (RPC timeout)

10. RPC timeout on NFS mount

11. Netscape time-outs

12. Q: HELP: Cirrus PCMCIA IRQ Device time-outs...

13. Idle time-outs


All times are UTC
Board index