Pre-Summary

From: Jim Harm <harm1_at_llnl.gov>
Date: Wed, 04 Mar 1998 15:28:26 -0800

My original message:
>Greetings;
> I have been plagued with this warning message on machineA:
>"NFS2 server machineA not responding still trying
>NFS2 server machineA ok"
>
>The real mystery is
> we run NFS3
> this machine is not an NFS server
> It is cerainly not mounting anything NFS from itself
>
>It is a alpha 8400 running 4.0d.
>
>Anyone else have this problem and maybe some suggestions?
>I'd ignore it, but it seems we have extreeeeemely sluggish performance
>at about the same time; like the machine seems to just stall
>for 15 to 45 minutes or more.
>

Most responses have been like these:

1.
>Greetings Bill:

> I have been plagued with this warning message on machineA:
> "NFS2 server machineA not responding still trying
> NFS2 server machineA ok"

Did you get any responses to this post? I am very curious because we
get the same problem from time to time.

-C

--
2.
This may happen if you shutdown the machine(A) which has mounted nfs
directories on other machine(B) without umounting on B. (B has mounted a
directory on A) .
3.
Hi Jim,
I have the same problem with one of our machines, and I would be interested
in any replies you get. If you could forward them to me, or post a summary
would be greatly appreciated.
TIA
Glenn.
4.
Jim,
yesterday i have overlooked your question, but today my collegue
(Harald Baumgartner, who also answered you) comes up with it again...
We have been plagued with this message too (for years, but it has become
worse with time, so now we believed it to be solved - for us...)
There are many possibilities for this message and for us the solution
has been fragmentation of a heavily used (homedirectories!), nfs-exported
disk. But of course we are using NFS.
So my first item for you would be to check within /etc/rc.config
whether you are sure to have no typos nfs3/nfs2? Then, do you export 
filesystem(s) to oldfashioned ultrix-boxes? Those don't have nfs3 at all
(AFAIK) and would automatically use nfs2.
These messages can be found in the files /var/adm/syslog.dated/*/kern.log
on your nfs-clients.
Check whether it's really not your nfs-server who is not responding...
(I would doubt such an error...)
Then check all the (net)devices between server and clients to work okay,
to be setup consistently.
Then ask you dec-support for the sys_check utility (as my collegue suggested).
If all these bring up no answers, you may have to look into the problem
more thoroughly - and don't hesitate to ask me again!
With regards,
Bernt Christandl
5.
one suggestion is, to defragment the advfs-disks. It will help a lot.
Other possibilities are many. Try sys_check from DEC-Support.
			Regards, H. Baumgartner
6.
Just to let you know that you ate not alone, one of our workstations also
reported the same error messaage. I believe the problem was due to an
incorrect entry in the 'Local Domain' as used by BIND.
Regards,
Rich
7.
May be your server runs automount?
Messages on that kind appear on some Unixes , I do not remember which..
8.
> What do you get for output when you issue a 'mount' command without
> any
> arguments? I haven't tracked it all the way down yet, but I had a file
> system mount as nfsv2, and my suspicion is that the nfs (v3) mount
> request
> happened at boot time, some time later the cpu that serves the
> filesystem
> in question boots and starts serving the filesystem, but due to the
> wait
> time of the initial request, the mount timesout and tries to mount as
> nfs v2.
9.
Try forcing the use of NFS V2 over UDP.
10.
DEC support said to change the kernel parameter ubc-nfsloopback=1
on a machine that happens to be on the same subnet which does nfs mount
from itself to itself.
Conclusion:
	We have triple checked the configuration of DNS, NFS, and auto-mounter.
It occurs on machines that are configured very differently and are on completely
unconnected subnets and domains.
We are currently analyzing a dump with DEC.
I will not further summarize unless we find a real difinitive answer.
Sorry I couldn't help the others in the same boat.
Thank you for your ideas; we have checked alot based on those suggestions
and therefore have a more narrow search.
"Eliminate that which cannot be and what remains must be."
	with appologies to Sir A. C. Doyle.
}}}===============>>
James E. Harm (Jim); jharm_at_llnl.gov
(510) 422-4018 L-073 Pg: 423-7705x57152
fax: (510) 423-6374 UNIX Sys.Admin...-.-

Received on Thu Mar 05 1998 - 00:28:27 NZDT