SUMMARY: Strange msgs from RPC...

From: George R. Goffe (grgoffe@infosystest7.arc.nasa.gov)
Date: Fri Oct 04 1996 - 12:59:43 CDT


Hi there,

As usual the "group" has come through.

The problem turned out to be a wire which needed to re-seated. We
swapped wires at both ends on both interfaces and the problem went
away.

My thanks to those who replied (lots of thanks), specifically:

Casper Dik <casper@holland.Sun.COM>
Steve Phelps <steve@epic.co.uk>
jwf@ccuc.on.ca (John W. Funk)

Here's my origional post for those who missed it...

Hi there,

"We" are getting a lot of RPC msgs on various systems (strictly solaris
2.5.1) the msg indicates mostly getattr failures with a RPC timeout
indication. The NFS server is a Sparc 2000 with 2 network interfaces,
the "8" net and the "9" net. The "9" net seems to be functioning quite
well but the "8" net seems to have lots of trouble (ala RPC msgs).

"Is there a wizzard in the house?" Do "we" have a bad interface? I'm
somewhat in the dark when it comes to networking (unix). Can anyone
give me some pointers, hints, tips? Please?

This is what I was seeing...

"NFS getattr failed for server server_name. error 5 (RPC timed out)"
"NFS server not responding"

Their replies are below with the email cruft removed...

----- Begin Included Message -----

>From casper@holland.Sun.COM Thu Oct 3 00:23:44 1996

You're using soft mounts, you shouldn't.

When you start using hardmounts (your data has some value, I'd think)
you may start seeing "NFS server not responding still trying".

I'd check cabling and interface errors and wireload on the 8 net.

Casper

----- End Included Message -----

----- Begin Included Message -----

>From steve@epic.co.uk Thu Oct 3 02:56:06 1996

Could you give more details, specifically:

        The exact error message you're getting.
        The type of NFS client that you are using (PC-NFS, Solaris,
          etc..)
        Any reported problems on the client side (e.g. hangs when
          doing an ls?)

It might also be a good idea to run the following command on the
server (as root):

        snoop rpc nfs

----- End Included Message -----

----- Begin Included Message -----

>From jwf@ilap.com Thu Oct 3 05:35:58 1996

George,

I'm no wizard, but I manage a site with one Solaris 2.5.1 Ultra,
several Solaris 2.5 SS20's and SS10's, and several HP's running HP-UX
9.05.

The errors started to appear when the Ultra was added, and read
something like this.

"NFS getattr failed for server server_name. error 5 (RPC timed out)"
"NFS server not responding"

I just yesterday made a change, and am waitng to see if it solved the
problem, and am anxious to see what other recommendations you receive.

My attempted fix was to check all vfstab files for the mount option
"noac" and change it to "actimeo=0". Noac was recommended fix in
Solaris 2.4 to eliminate nfs truncating large files.

Good Luck,

jwf

John W. Funk, P.Eng
CAD/CAM & Unix Consultants Inc.
P.O. Box 61024, Maple Grove P.O.
Oakville, Ontario, L6J 7P5

voice: 905-829-5028
fax: 905-929-0517
email: jwf@ccuc.on.ca

----- End Included Message -----



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:11 CDT