NIS, NFS & Solaris 2.5.1 : Need Help

NIS, NFS & Solaris 2.5.1 : Need Help

Post by Guillaume Pio » Sat, 18 Oct 1997 04:00:00



Hi,

We are running Solaris 2.5.1 on a network with NIS setup on it.
We are accessing the HOME directory of the users via NFS. For one of our
Clients, when a user with a home directory NFS mounted from the server
logs in, the HOME directory is not accessible (ksh reports "bad
directory"). Yet the NFS mounted file system is mounted on the client,
according to mount.

on the client Box Solaris 2.5.1 with NIS.

/var/adm/messages :

Oct 16 12:03:46 refmark-poller unix: NOTICE: authdes_create: unable to
get client netname
Oct 16 12:03:47 refmark-poller unix: WARNING: authget: authdes_create
failed: Invalid argument
Oct 16 12:03:47 refmark-poller unix: WARNING: clget: authget failed
(scanning chtable): Invalid argument

On the NIS Server & NFS Server Solaris 2.5.1.

/var/adm/messages :

Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: t_bind to wrong
address
Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: Cannot establish
NFS service over /dev/udp: transport setup problem.
Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: t_bind to wrong
address
Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: Cannot establish
NFS service over /dev/tcp: transport setup problem.
Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: Could not start
NFS service for any protocol. Exiting.
Oct  6 17:29:51 refmark-q-db sendmail[23297]: RAA22588: SYSERR(oracle):
Cannot exec /usr/lib/mail.local: Not enough space
Oct  6 17:29:51 refmark-q-db sendmail[22991]: NOQUEUE: SYSERR(oracle):
dowork: cannot fork: Not enough space

If someone know a solution please send me a mail


Thank in advance

regards

        Guillaume PIOT

 
 
 

NIS, NFS & Solaris 2.5.1 : Need Help

Post by Chip Bennet » Mon, 20 Oct 1997 04:00:00



> Hi,

> We are running Solaris 2.5.1 on a network with NIS setup on it.
> We are accessing the HOME directory of the users via NFS. For one of
> our
> Clients, when a user with a home directory NFS mounted from the server

> logs in, the HOME directory is not accessible (ksh reports "bad
> directory"). Yet the NFS mounted file system is mounted on the client,

> according to mount.

> on the client Box Solaris 2.5.1 with NIS.

> /var/adm/messages :

> Oct 16 12:03:46 refmark-poller unix: NOTICE: authdes_create: unable to

> get client netname
> Oct 16 12:03:47 refmark-poller unix: WARNING: authget: authdes_create
> failed: Invalid argument
> Oct 16 12:03:47 refmark-poller unix: WARNING: clget: authget failed
> (scanning chtable): Invalid argument

> On the NIS Server & NFS Server Solaris 2.5.1.

> /var/adm/messages :

> Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: t_bind to
> wrong
> address
> Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: Cannot
> establish
> NFS service over /dev/udp: transport setup problem.
> Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: t_bind to
> wrong
> address
> Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: Cannot
> establish
> NFS service over /dev/tcp: transport setup problem.
> Oct  6 17:33:18 refmark-console /usr/lib/nfs/nfsd[757]: Could not
> start
> NFS service for any protocol. Exiting.
> Oct  6 17:29:51 refmark-q-db sendmail[23297]: RAA22588:
> SYSERR(oracle):
> Cannot exec /usr/lib/mail.local: Not enough space
> Oct  6 17:29:51 refmark-q-db sendmail[22991]: NOQUEUE: SYSERR(oracle):

> dowork: cannot fork: Not enough space

> If someone know a solution please send me a mail


> Thank in advance

> regards

>         Guillaume PIOT

   There's nothing in the NFS errors that I recognize, but I _did_
notice the "Not enough space" messages from sendmail.  Are you running
out of virtual memory on the server?

Also, "nfsd" is multi-threaded and does not fork.  I see where it says
it is "exiting".  Does this happen every time you attempt to mount?  The
nfsd messages appear on the 6th, but the client messages are marked the
16th.  If you do not reboot, or at least restart nfsd, you are being
left without an nfs daemon.

If the server ran out of virtual memory on the 6th, and you haven't
rebooted it, or restarted nfsd, maybe the real problem for the client is
that the servers NFS deamon isn't around anymore.

The nfsd messages give the impression that nfsd is exiting as soon as
the server is booted.  It appears to be attempting NFS version 2 setup
(UDP) and then NFS version 3 setup (TCP).  When it can't setup either
protocol, it exits (at least that's the way it looks).  If nfsd really
is exiting on boot up, do you have enough VM to even boot properly?
(BTW: Why are the date-time stamps in /var/adm/message in the wrong
order?)  Could the network interface be configured wrong?

After you're confident that you are not presently out of VM on the
server, and if you are restarting nfsd and it's actually exiting each
time you attempt to contact it from the client (rather than at boot
time, despite the way it looks),  I would try "truss"ing nfsd, and
"snoop"ing the network interface at the same time and see if I could
gain a little more info on what's happening.

Chip Bennett

 
 
 

1. NIS & NIS+ Need help!

                        Hello,
I would like install NIS or NIS+ in my network.
I have 20 PC with FreeBSD last version and what is the process step by
step
to create a server and clients.

                                    Thank you,
                                                    Antoine NIVARD.


2. Intellipoint mouse on SuSE

3. Questions & Help needed with NIS<-->NIS+

4. USB Under RH9 (2.4.20-8) locks hard

5. Secure NFS under Solaris 2.5/2.5.1 without NIS/NIS+ ?

6. errors compiling zImage in p-10-ALPHA

7. Netware NFS <-> (Sun) NFS & NIS

8. Prompt

9. NFS/NIS Help Needed

10. Need help with NFS and NIS

11. Reasonable nis security between Solaris & Linux (was Re: Is nis (yp) a security worry?

12. Solaris 2.x NFS and NIS server with Solaris 1.x clients???

13. NIS & NYS & NIS+