nfs client does not find nfs server

Hello,
both computer suse 10.1,
nfs server installed but client can not find the server.
What might be the reason?
Ping is working fine.

At the beginning I used used YAST to install the nfs server and
checked afterwards manually if everything is running and alright.
Thank you
Bastian

On Thu, 15 Mar 2007 09:03:16 -0700, bastian wrote:

> Hello,
> both computer suse 10.1,
> nfs server installed but client can not find the server. What might be the
> reason?
> Ping is working fine.

What exactly do you mean by "cannot find the server"? What is the error
message when you attempt the mount?

"Cannot find the server" would seem to mean a name resolution failure but
if this were so ping wouldn't work either.

Regards, Ian

>
> "Cannot find the server" would seem to mean a name resolution failure but
> if this were so ping wouldn't work either.

Ian,
Time out.
I have read in the meantime that something regarding the firewall is
not functioning the way it should.
Regards
Bastian
>
> Regards, Ian

On Thu, 15 Mar 2007 15:23:36 -0700, bastian wrote:

>
>> "Cannot find the server" would seem to mean a name resolution failure
>> but if this were so ping wouldn't work either.
>
> Ian,
> Time out.
> I have read in the meantime that something regarding the firewall is not
> functioning the way it should.

That's different. You can tell if it's a firewall problem by comparing the
output of:

rpcinfo -p on the server
rpcinfo -p <server> on the client

They should look the same.

Normally you wouldn't have a firewall between an NFS server and its
client(s). While NFS itself uses a predictable port number (2049), as
does RPC (111), some other components do not, so you don't know in advance
what ports to open. NFS isn't secure anyway (there are enhancements in V4
but it's still not that good). NFS is designed to be used within a LAN
type environment where the other machines are reasonably trustworthy and
no firewall is required between them.

I suggest that you open your firewalls on both client and
server so that all traffic between them is permitted. If this is not
acceptable, you should consider whether NFS is the correct solution to
your data sharing requirement.

Regards, Ian

Hello,
I see.
>
> rpcinfo -p on the server
> rpcinfo -p <server> on the client

They look the same.
I can find the server now using YAST after deactivating the firewalls
of both computers.

Nontheless I still have a problem:

server:/etc # mount -t nfs linuxserver.site:/ /linuxserver/
mount server reported tcp not available, falling back to udp
mount: RPC: Fehler des entfernten Systems - Keine Route zum
Zielrechner

Well last words are in German meaning:
mount: RPC: Mistake of the remote system - no route to the computer of
destination

server:/etc # rpcinfo -p
Program Vers Proto Port
100000 2 tcp 111 portmapper
100000 2 udp 111 portmapper
100003 2 udp 2049 nfs
100003 3 udp 2049 nfs
100003 4 udp 2049 nfs
100003 2 tcp 2049 nfs
100003 3 tcp 2049 nfs
100003 4 tcp 2049 nfs
100024 1 udp 32770 status
100021 1 udp 32770 nlockmgr
100021 3 udp 32770 nlockmgr
100021 4 udp 32770 nlockmgr
100024 1 tcp 45631 status
100021 1 tcp 45631 nlockmgr
100021 3 tcp 45631 nlockmgr
100021 4 tcp 45631 nlockmgr
100005 1 udp 675 mountd
100005 1 tcp 676 mountd
100005 2 udp 675 mountd
100005 2 tcp 676 mountd
100005 3 udp 675 mountd
100005 3 tcp 676 mountd

I have doubts about the nlockmgr and I wonder whether this is good or
bad?

This is my hosts

# mappings for the TCP/IP subsystem. It is mostly
# used at boot time, when no name servers are running.
# On small systems, this file can be used instead of a
# "named" name server.
# Syntax:
#
# IP-Address Full-Qualified-Hostname Short-Hostname
#

127.0.0.1 localhost

# special IPv6 addresses
::1 localhost ipv6-localhost ipv6-loopback

fe00::0 ipv6-localnet

ff00::0 ipv6-mcastprefix
ff02::1 ipv6-allnodes
ff02::2 ipv6-allrouters
ff02::3 ipv6-allhosts
127.0.0.2 server.site server
192.168.0.126 server.site server
192.168.0.134 linuxserver.site linuxserver

And hosts.allow is:

portmap: 192.168.0.134

Bastian

>
> They should look the same.
>
> Normally you wouldn't have a firewall between an NFS server and its
> client(s). While NFS itself uses a predictable port number (2049), as
> does RPC (111), some other components do not, so you don't know in advance
> what ports to open. NFS isn't secure anyway (there are enhancements in V4
> but it's still not that good). NFS is designed to be used within a LAN
> type environment where the other machines are reasonably trustworthy and
> no firewall is required between them.
>
> I suggest that you open your firewalls on both client and
> server so that all traffic between them is permitted. If this is not
> acceptable, you should consider whether NFS is the correct solution to
> your data sharing requirement.
>
> Regards, Ian

bastian <(E-Mail Removed)> wrote:

> mount: RPC: Mistake of the remote system - no route to the computer of
> destination

> 192.168.0.126 server.site server
> 192.168.0.134 linuxserver.site linuxserver

Confusing names, "linuxserver" is your NFS server, and "server" is your NFS
client. Is that right?

I would give the machines user friendly names, such as scoopy,shaggy,daphny,
etc, and maybe stick a sticker on the front showing the machine name.

> And hosts.allow is:
> portmap: 192.168.0.134

On the server, the address should be that of the client:

portmap: 192.168.0.126
lockd: 192.168.0.126
mountd: 192.168.0.126
rquotad: 192.168.0.126
statd: 192.168.0.126

On the server, cat /etc/exports

Post the output here

To share with any client on the local area network, this should read something
like:

/var/share/sharefolderfoo 192.168.0.0/255.255.255.0(rw)

On the client:

mount linuxserver://var/share/sharefolderfoo /var/share/sharefolderfoo

I hope this helps.

Mark.

--
Mark Hobley
393 Quinton Road West
QUINTON
Birmingham
B32 1QE

Telephone: (0121) 247 1596
International: 0044 121 247 1596

Email: markhobley at hotpop dot donottypethisbit com

http://markhobley.yi.org/

Hello,
many thanks
nfs is working fine.

I can only export directories. How can I export the whole harddisk
hda. My aim is to make a copy of hda with dd over the network. I just
tried to export /dev/. The root of the nfs client does not have the
necessary rights. So I have to change the rights. Well this is my
status quo at the moment.

>
> Confusing names, "linuxserver" is your NFS server, and "server" is your NFS
> client. Is that right?

yes correct.

> I would give the machines user friendly names, such as scoopy,shaggy,daphny,
> etc, and maybe stick a sticker on the front showing the machine name.

ok.

> On the server, cat /etc/exports

> Post the output here

/ 192.168.0.126(ro,root_squash,sync)

Well I have tried
/dev 192.168.0.126(ro,root_squash,sync)
/dev/hda 192.168.0.126(ro,root_squash,sync)

also.

>
> I hope this helps.

Definitely.

Bastian

>
> / 192.168.0.126(ro,no_root_squash,sync)
>
seems to be a solution.
Bastian

bastian <(E-Mail Removed)> wrote:
> / 192.168.0.126(ro,no_root_squash,sync)

> seems to be a solution.

Indeed, root is squashed (mapped to user nobody) by default.

Regards,

Mark.

--
Mark Hobley
393 Quinton Road West
QUINTON
Birmingham
B32 1QE

Telephone: (0121) 247 1596
International: 0044 121 247 1596

Email: markhobley at hotpop dot donottypethisbit com

http://markhobley.yi.org/

Mark,

> > seems to be a solution.
>
> Indeed, root is squashed (mapped to user nobody) by default.

unfortunately no_root_squashing is not enough.
I have no access onto /linuxserver/dev/hda.
Bastian