Dear list members,
I have a problem using NFS.
Synopsis: Files that are hard linked on the server's root
NFS images for a Beowulf cluster become inaccesable
either imediately or after a few hours.
Boot example: From the server's logs (server is rg272) :-
July 30 10:05:45 rg272 kernel: nfsd Security: sbin/init bad
export.
July 30 10:05:46 rg272 kernel: nfsd Security: sbin/bash bad
export.
The client boot-up hangs at this point.
After time example: From the server's logs (as I remember it):-
July 30 11:34:02 rg272 kernel: nfsd Security:
lib/libc6.x.x.so bad export.
Client example: From the command line:-
cat /etc/inittab
cat: standard output: Stale NFS file handle
Other symptoms:
When booting the cluster each machine must be booted
on its own. Booting the machines in parallel fills the
servers logs with "bad export" messages.
Problem exists irrespective of NFS protocol
used (V2 or V3)
Cycling the power on the server clears the problem
temporarily but restarting the NFS service does not.
System Details:
Software:- Linux RedHat 7.3 on server and clients
(kernel 2.4.18)
Server:- Dell GX530 dual processor precision work
station.
Clients:- Dell GX260 single processor work station.
Progress in solving the problem to date:
By trawling the web I found one Gunnar Raetch who
experianced almost identical problems using kernel
2.4.14. He fixed the problem by applying the
following kernel patches:-
http://www.cse.unsw.edu.au/~neilb/patches/linux/2.4.14-pre4/
patch-D-NfsdOpsPrepare
patch-E-NfsdOps
patch-F-ReiserfsNfsdOps
I am unable patch my 2.4.18 kernel with these patches.
I also tried down grading to kernel 2.4.14, successfully
applied the patches & compiled, but the USB support
in 2.4.14 would not allow me use of my mouse and keyboard
(Oh Sh*t!).
I hope this provides sufficient info to resolve the problem.
Many thanks in advance for your anticipated kind help.
Best Regards,
Donald.
Reliance Gear Company Ltd.
Rowley Mills
Penistone Road
Lepton
Huddersfield
England
HD8 0LE
Tel: +44 1484 601000
fax: +44 1484 601001
http://www.reliance.co.uk
THE INFORMATION IN THIS COMMUNICATION IS CONFIDENTIAL AND INTENDED ONLY FOR
THE USE OF THE ADDRESSEE. DISCLOSURE, USE OR COPYING OF THIS MESSAGE OR
ITS CONTENTS OTHER THAN BY THE ADDRESSEE IS PROHIBITED. IF YOU RECEIVE THIS
COMMUNICATION IN ERROR WE WOULD BE GRATEFUL IF YOU WOULD NOTIFY US
IMMEDIATELY AND DELETE ANY MATERIAL RECEIVED
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
What options are you using to export the file systems from the server?
Are you using 'no_subtree_check' ?
Using no_subtree_check fixed a similar problem I had - see:
http://marc.theaimsgroup.com/?l=linux-nfs&m=101309591431249&w=2
James Pearson
Donald Burns wrote:
>
>
> Dear list members,
>
> I have a problem using NFS.
>
> Synopsis: Files that are hard linked on the server's root
> NFS images for a Beowulf cluster become inaccesable
> either imediately or after a few hours.
>
> Boot example: From the server's logs (server is rg272) :-
>
> July 30 10:05:45 rg272 kernel: nfsd Security: sbin/init bad
> export.
> July 30 10:05:46 rg272 kernel: nfsd Security: sbin/bash bad
> export.
>
> The client boot-up hangs at this point.
>
> After time example: From the server's logs (as I remember it):-
>
> July 30 11:34:02 rg272 kernel: nfsd Security:
> lib/libc6.x.x.so bad export.
>
> Client example: From the command line:-
>
> cat /etc/inittab
> cat: standard output: Stale NFS file handle
>
> Other symptoms:
>
> When booting the cluster each machine must be booted
> on its own. Booting the machines in parallel fills the
> servers logs with "bad export" messages.
>
> Problem exists irrespective of NFS protocol
> used (V2 or V3)
>
> Cycling the power on the server clears the problem
> temporarily but restarting the NFS service does not.
>
> System Details:
>
> Software:- Linux RedHat 7.3 on server and clients
> (kernel 2.4.18)
> Server:- Dell GX530 dual processor precision work
> station.
> Clients:- Dell GX260 single processor work station.
>
> Progress in solving the problem to date:
>
> By trawling the web I found one Gunnar Raetch who
> experianced almost identical problems using kernel
> 2.4.14. He fixed the problem by applying the
> following kernel patches:-
>
> http://www.cse.unsw.edu.au/~neilb/patches/linux/2.4.14-pre4/
>
> patch-D-NfsdOpsPrepare
> patch-E-NfsdOps
> patch-F-ReiserfsNfsdOps
>
> I am unable patch my 2.4.18 kernel with these patches.
> I also tried down grading to kernel 2.4.14, successfully
> applied the patches & compiled, but the USB support
> in 2.4.14 would not allow me use of my mouse and keyboard
> (Oh Sh*t!).
>
> I hope this provides sufficient info to resolve the problem.
>
> Many thanks in advance for your anticipated kind help.
>
> Best Regards,
>
> Donald.
>
> Reliance Gear Company Ltd.
> Rowley Mills
> Penistone Road
> Lepton
> Huddersfield
> England
> HD8 0LE
>
> Tel: +44 1484 601000
> fax: +44 1484 601001
>
> http://www.reliance.co.uk
>
> THE INFORMATION IN THIS COMMUNICATION IS CONFIDENTIAL AND INTENDED ONLY FOR
> THE USE OF THE ADDRESSEE. DISCLOSURE, USE OR COPYING OF THIS MESSAGE OR
> ITS CONTENTS OTHER THAN BY THE ADDRESSEE IS PROHIBITED. IF YOU RECEIVE THIS
> COMMUNICATION IN ERROR WE WOULD BE GRATEFUL IF YOU WOULD NOTIFY US
> IMMEDIATELY AND DELETE ANY MATERIAL RECEIVED
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> NFS maillist - [email protected]
> https://lists.sourceforge.net/lists/listinfo/nfs
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
On Thursday August 1, [email protected] wrote:
>
> Dear list members,
>
> I have a problem using NFS.
>
> Synopsis: Files that are hard linked on the server's root
> NFS images for a Beowulf cluster become inaccesable
> either imediately or after a few hours.
At a guess, I'd say "add the no_subtree_check option to your exports"
is the way to go.
NeilBrown
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
Dear All,
Many thanks to all those who responded so quickly -
one doesn't get support like this from Microsoft! My
coleagues were immensely impressed.
I added the option no_subtree_check to /etc/exports
and it has fixed all the problems described in my
original message.
Once again, I thank you.
Best Regards,
Donald.
Reliance Gear Company Ltd.
Rowley Mills
Penistone Road
Lepton
Huddersfield
England
HD8 0LE
Tel: +44 1484 601000
fax: +44 1484 601001
http://www.reliance.co.uk
THE INFORMATION IN THIS COMMUNICATION IS CONFIDENTIAL AND INTENDED ONLY FOR
THE USE OF THE ADDRESSEE. DISCLOSURE, USE OR COPYING OF THIS MESSAGE OR
ITS CONTENTS OTHER THAN BY THE ADDRESSEE IS PROHIBITED. IF YOU RECEIVE THIS
COMMUNICATION IN ERROR WE WOULD BE GRATEFUL IF YOU WOULD NOTIFY US
IMMEDIATELY AND DELETE ANY MATERIAL RECEIVED
NeilBrown
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs