SUMMARY : mount nfs failed at boot

Date: Thu Feb 18 1999 - 19:31:41 CST


it has been about two weeks since I send my question, yet I
could not get the server down for troubleshoot. So, I guess I should
summarise first.

Many thanks to :
Michael Little
Bismark Espinoza
Kevin Sheehan
Rick Fincher
Markus Roellig
Nickolai Zeldovich
Phil Scarr
Stephen Harris
Jeff Wimmer

for their solution. I'm still working on it. ( the machine is quiet
critical, so I have to be very careful).
Again, thank you very much.


qqqqqqqqqqqqqq ORIGINAL QUESTION qqqqqqqqqqqqqqqqq

Date sent: Wed, 10 Feb 1999 12:43:41 +0800
Subject: mount nfs failed at boot


I have nfs mount failed at boot but success if done manually with
"mount -a".

earth is sunos 4.1.4
neptune is sunos 4.1.4
jupiter is solaris2.3

at boot time earth cannot automatically mount neptune and
jupiter's exported file system.

Could someone please point me for possible mistake.



content of /etc/fstab at earth
------------------------------------------ /var/local nfs rw,soft,bg 0 0 /usr/local nfs rw,soft,bg 0 0

content of /etc/dfs/dfstab at jupiter
share -F nfs -o /usr/local

content of /etc/exports at neptune
/var -

rrrrrrrrrrrrrrrrrrrr RESPONSES RECEIVED rrrrrrrrrrrrrrrrrrrr

-----------"Michael Little" <>------------


And idea. BTW, mounting rw filesystems as soft? You don't mind
corrupted data on those filesystems then? Out of curiousity, in
/etc/hosts file, does it have entries for neptune or
I am
wondering if it is not doing a correct DNS lookup and resolution on
FQDN until it finishes booting. Try changing on earth:

content of /etc/fstab at earth
neptune:/usr/share /var/local nfs rw,intr,bg 0 0 # This
mounts them as hard but interuptable
jupiter:/usr/local /usr/local nfs rw,intr,bg 0 0

And on the others:

content of /etc/dfs/dfstab at jupiter
share -F nfs -o /usr/local

content of /etc/exports at neptune
/var -

And see what happens.

Michael Little
Solid Systems, Inc.
(713) 934-9494 x131
(800) 324-0500 x131
(281) 289-9737 pg
------------Bismark Espinoza <> ---------

Remove the "-" in neptune's /etc/exports file.

In jupiter /etc/nsswitch.conf, a line should read:

hosts:files nis dns

Put the earth's and neptune's IP address in
jupiter's /etc/hosts.

--------------- Rick Fincher <> --------------


If earth doesn't have any NFS volumes exported at boot time NFS
won't start and
you'll have to include a line in your startup files to manually run the

Hope this helps,


----------- (Kevin Sheehan/Uniq
{Consulting Poster Child}) --------------------------------------------------------

[ Regarding "mount nfs failed at boot",
writes on Feb 10: ]

> Could someone please point me for possible mistake.

Maybe DNS or routing or something isn't working right yet...

> content of /etc/fstab at earth
> ------------------------------------------
> /var/local nfs rw,soft,bg 0 0
> /usr/local nfs rw,soft,bg 0 0

... but rw,soft is a disaster looking for a place to happen.

                l & h,

-----------Markus Roellig <>--------------

I had the same problem. Our machines ( EARTH,
SATURn,NEPTUN :)) ) with
Sun OS 4.1.x and Solaris 2.5/6 respectively. The solution could be
a missing
entry in the file /etc/hosts on your earth machine, which is under
NIS only
consulted at boot time. If the entry is missing the both machines
are unknown at
the moment, the system tries to mount, afterwards the NIS server
supports the
proper IP addresses.


Markus Roellig
Institut fuer theoretische Physik/Astrophysik
J.W.Goethe Universitaet, Frankfurt, Germany

----------------Nickolai Zeldovich <>----------------

hmm.. is the network fully configured at the time when it does
mounting from rc*?

-- [ Nickolai Zeldovich // ]

------------Phil Scarr <>-------------------

Try forcing everything to NFS Version 2 by adding vers=2 to the


---------------Stephen Harris <>-------------------

Are you access names via NIS? Maybe NIS isn't running fully at
that stage...? If the server can not resolve "" early
then mistakes may happen. Maybe try putting the names into the
/etc/hosts file?


Looks like a timing problem. You need to enable retries and set
them to
something like 100 at first. After it works, decrease the number
until you
can still mount the drive. This will reduce overhead network traffic.
mount for the exact format, it goes after the rw,soft,bg line.

Jeff Wimmer
Associate Network Engineer
Litton Data Systems
6608 Sun Scope Drive
Ocean Springs, MS 39564
voice: 228.872.7376

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:13:15 CDT