From: Desrosiers, Jean (Jean.Desrosiers@VISAER.com)
Date: Wed Oct 02 2002 - 09:57:16 EDT
Thanks for all the responses. ( included below original query ) They were
prompt and right on!
The hung step during boot was bypassed by doing a "ctrl \" ( control
backslash ).
Once boot completed the /etc/rc.config.d/nfsconf file was edited to make
NFS_CLIENT=0 and NFS_SERVER=0.
Reboot was successful.
Once I can gain access to the machine myself I'll be ( completing ) going
through more of the recommendations I received below.
I'll update the summary with significant info then.
Read on!
ORIGINAL QUERY
==============
Hi all
I've got a problem with a machine loaned to a client. ( HPUX 10.20 ) It
booted fine after shipment and I coached the network guy through set_parms
to change the ip address, default gateway, and subnet mask so the machine
could be connected to their network. Now the machine hangs during boot on
the NFS startup with the error "NFS Server [pid 776 @/net] not responding.
Still trying".
I am advised the machine can be pinged but remains hung there. Telnet isn't
working yet.
I was hoping it would "give up" but it hasn't yet, at least 30 mins elapsed.
My plan was to disable startup of NFS services as they won't be needed.
After looking at the archives I noticed a past issue where an NFS server on
the network was incorrectly configured and caused a similar problem.
Does this ring any bells?
Does anyone have a plausible explanation for this and (hopefully) a
method/process to overcome it?
=====================================================
RESPONSES ( in order received )
===========
On HP's you should be able to do a ctrl-\ (Pressing Control key along with
backward slash) to skip a startup process. As far as the NFS Server not
starting up, once you get in to the system try running
/sbin/init.d/nfs.server stop, start and see if it still breaks. Also, check
in /var/adm/messages to see if there is any other specific error messages
for nfs.server.
Srinath Rajagopalan
----------------------------------------------
Default gateway is invalid or not responding.
Don't ask why, it just does.
Try setting to 0.0.0.0 if you have no valid router.
Mike.
----------------------------------------------------------
It will not give up. Reboot into single user mode. Manually mount /usr and
vi /etc/rc.config.d/nfsconf and change the NFS_CLIENT=1 to NFS_CLIENT=0 also
change NFS_SERVER=1 to NFS_SERVER=0 and then boot up as usual. Figure out
your problem with NFS, check out your /etc/rc.config.d/netconf file
thoroughly for any funny looking IP addresses. Maybe something did not go
right with the set_params. Try this and when you are satisfied and ready to
try it again change the NFS_CLIENT and NFS_SERVER fields back to 1 in the
/etc/rc.config.d/nfsconf. (That is if both of them were set to 1. It is
possible that only one was set to 1.)
Hope this helps!
Bryan Douglas Quinn
UNIX Support Specialist
--------------------------------------------------------
It sounds like your system is trying to make an NFS mount.
You should be able to issue a break (break key) to interrupt
the services startup and log in. Then determining what mount
is being attempted (probably via automount) and cleaning up
the configuration should be possible.
Brian Brocklesby
Alcatel USA
----------------------------------------------------------
/net relates to automounter. I would boot the system to single user mode.
Edit the:
/etc/rc.config.d/nfsconf file and ensure you have:
AUTOMOUNT=0
and:
AUTOFS=0
Reboot - you should come up ok and still be able to do NFS mounts if
required. Hope this helps.
Regards,
Andy Cranston.
-----------------------------------------------------------
I don 't remember in 10.20, but in 11.00, the default installation
sets a "/net -hosts -soft" configuration line in /etc/auto_master
that causes that problem.
2 options:
1) unplug the network cable. I 'm not sure but it can make it
give an error and pass.
2) reboot and go into maintenance mode. There you can disable
NFS startup
Hope it helps, .
Martín Arpón
Sistemas Operativos
---------------------------------------------------------
A VERY common problem. You need to reboot the
machine into single user mode and disable all
NFS services in the /etc/rc.config.d/nfsconf
file. Turn off both NFS client and server!
Edit all /etc/rc.config.d/nfs* files.
Best regards,
Bill Hassell
---------------------------------------------------
We've seen this before when moving servers between networks.
When /etc/fstab contains an NFS mount point or automounter is
running the system will try to go off and mount stuff but not find
the source server and appear to hang.
I recommend doing a "ctrl + \" (hold down ctrl, press \) at the
console to interrupt NFS and finish the boot cycle, then turning
off NFS in /etc/rc.config.d/nfsconf and reboot again to validate.
Good luck!
john.adams
---------------------------------------------------
Does the systems have an "/etc/nsswitch.conf" file? I am making an
assumption that the subnet, IP, duplex, and speed are all correct and
verified.
It could be that the systems is hanging because it cannot resolve its own
address. I believe the machine is defaulted to searching DNS.
To get around the boot hanging, have the customer boot to single user mode.
>From there they should be able to use one the nsswitch.hp_default file (they
will to make sure it resolves to "files" for the the "hosts" entry. Once
the nsswitch.conf file modification are complete, then they can "init" to
their default run level.
The nsswitch.conf file would need to first look in files, than in DNS.
Myron Schneider
-------------------------------------------------------
You can comment the file system that is mounting from the NFS Server in
your "/etc/fstab" or you can set the auto mounter. This allows the system
not to mount
the file system until the file system is accessed. This will prevent the
system from waiting on NFS server to come up.
James
--------------------------------------------------------
Solution is very simple. Boot into single user mode and fix
/etc/rc.config.d/netconf.
Some entries related with network are not corrects. Disabling NFS will not
solve
your problem at all.
Hope it helps
---------------------------------------------
Check your default gateway setting.....
To Truc
----------------------------------------------
There are a couple of things that you might want to check.
1) the NIS domain name may still be set to yours not the one where the
server is on loan to
2) automounter is attempting to mount NFS directories that you had set up at
your site which do not exist at the loan site.
I ran into a similar situation and it turned out to be a combination of both
of these items.
I was able to break out of the hang and kill the automounter process which
was the actual process the was hanging the machine.
Randy Linton
------------------------------------------------------
probably means you have an NFS mounted file system in your /etc/fstab,
and due to the IP address change the NFS server is either not reachable or
hasn't got an export FOR this IP address.
And the NFS mount is not "soft" (which is somewhat dangerous, but lets it
"give up" after a certain timeout) so the NFS client is keeping on Retrying
to mount that file system (and so never gets any further THAN starting the
NFS
client in the startup).
Either modify the fstab, to temporarily comment OUT all NFS mounts, or
change
/etc/rc.config.d/nfsconf, to make the machine NOT a NFS client.
Both will have to be done in single user mode, I'm afraid, as all
"multi-user
levels" will already start the NFS client (the default for that is runlevel
2).
> Does anyone have an plausible explanation for this and (hopefully) a
> method/process to overcome it?
Not a method which will work over the network, I'm afraid.
You can ENable NFS over the network and then start it, but you cannot get
the
machine started with a NFS server that isn't reachable, other than in single
user level (how to start a machine in single user level has been discussed
in this list a lot of times and is somewhat model dependant, but IT also can
only be done sitting at the physical console OF that machine).
--
Eef Hartman
--------------------------------------------
That's all folks.......hope this helps others.
J.Desrosiers
Operational Oracle DBA
VISaer Inc.
jean.desrosiers@visaer.com
--
---> Please post QUESTIONS and SUMMARIES only!! <---
To subscribe/unsubscribe to this list, contact majordomo@dutchworks.nl
Name: hpux-admin@dutchworks.nl Owner: owner-hpux-admin@dutchworks.nl
Archives: ftp.dutchworks.nl:/pub/digests/hpux-admin (FTP, browse only)
http://www.dutchworks.nl/htbin/hpsysadmin (Web, browse & search)
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 11:02:20 EDT