Solaris 9 Jumpstart Failure

Question

Greetings.
I am in the process of changing a system over to Solaris 9 (9/04) from Solaris 8 (we cannot move to Solaris 10 due to ClearCase incompatiblities).
We use flash archives in our jumpstart process. The master system is created using a very Spartin profile (SUNWCreq with a number of other required packages). There are also a number of additional tweaks made to the master system to stop unrequired services, deamons, etc. The only additional patches installed are the Java cluster patches.
I have been able to successfully jumpstart the jumpstart server host from CD. However any attempts to jumpstart other clients using the jumpstart server have failed. I suspect that it is related to the inability to copy the sysidcfg file during the jumpstart process.
The address for the jumpstart server is 10.1.1.1; the hostname is n1; the MAC address is 0:3:ba:35:80:88
The address for the jumpstart client is 10.1.1.34; the hostname is n34; the MAC address is 0:3:ba:14:c6.
On the jumpstart server some of the relevant files are included below.
/etc/bootparams:
n34 root=n1:/jumpstart/OS/Solaris_9_2004-09/Solaris_9/Tools/Boot install=n1:/jumpstart/OS/Solaris_9_2004-09 boottype=:in sysid_config=n1:/jumpstart/Sysidcfg install_config=n1:/jumpstart rootopts=:rsize=32768
/etc/hosts:
# Internet host table
127.0.0.1 localhost
10.1.1.1 n1 oam1a loghost
10.1.1.2 n2 db1a
10.1.1.34 n34
/etc/ethers:
0:3:ba:14:c6:cd n34
/etc/dfs/dfstab:
share -F nfs -o ro,anon=0 /jumpstart
share -F nfs -o ro,anon=0 /jumpstart/OS/Solaris_9_2004-0�9
/tftpboot directory:
lrwxrwxrwx 1 root root 26 Aug 8 10:36 0A010122 -> inetboot.SUN4U.Solaris_9-1
lrwxrwxrwx 1 root root 26 Aug 8 10:36 0A010122.SUN4U -> inetboot.SUN4U.Solaris_9-1
-rwxr-xr-x 1 root root 152376 Aug 8 10:36 inetboot.SUN4U.Solaris_9-1
-rw-r--r-- 1 root root 313 Aug 8 10:36 rm.10.1.1.34
ls -l /jumpstart/Sysidcfg/sysidcfg:
-rw-r--r-- 1 root root 375 Aug 4 17:12 /jumpstart/Sysidcfg/sysidcfg
/jumpstart/Sysidcfg/sysidcfg:
system_locale=en_AU
timezone=Australia/NSW
name_service=none
root_password=<removed for this post>
terminal=xterm
network_interface=primary { protocol_ipv6=no netmask=255.255.240.0 default_route
=10.1.0.1 }
timeserver=localhost
timeserver=47.153.235.110
Once the install is started on the client the following output is generated (note the sysidcfg copy failure):
ok boot net - install
Res
LOM event: +21h32m35s host reset
etting ...
�
Netra 120 (UltraSPARC-IIe 648MHz), No Keyboard
OpenBoot 4.0, 1024 MB memory installed, Serial #51693261.
Ethernet address 0:3:ba:14:c6:cd, Host ID: 8314c6cd.
Executing last command: boot net - install
Boot device: /pci@1f,0/pci@1,1/network@c,1 File and args: - install
SunOS Release 5.9 Version Generic_117171-07 64-bit
Copyright 1983-2003 Sun Microsystems, Inc. All rights reserved.
Use is subject to license terms.
whoami: no domain name
Configuring /dev and /devices
Using RPC Bootparams for network configuration information.
Skipping interface eri1
Configured interface eri0
Searching for configuration file(s)...
cp: cannot create /etc/sysidcfg: Permission denied
chmod: WARNING: can't change /etc/sysidcfg
Using sysid configuration file 10.1.1.1:/jumpstart/Sysidcfg/sysidcfg
Search complete.
WARNING: IP: Hardware address '00:03:ba:35:80:88' trying to be our address 010.001.001.001!
WARNING: IP: Hardware address '00:03:ba:35:80:88' trying to be our address 010.001.001.001!
The IP address conflict is with the jumpstart server. The address for the jumpstart server is 10.1.1.1. The warning message is generated for a number of minutes after it starts. I figure that it is caused by the inability to copy the sysidcfg file.
This same system can be installed successfully using a Solaris 8 jumpstart configuration.
Note that the client system is currently installed with Solaris 8.
I did use the command "boot net -v - install" from the OK prompt, but no additional information was provided relating to when the sysidcfg file could not be copied.
If anyone has any ideas about what could be causing this problem or has any information about additional debugging which could be used to figure out this issue, I would greatly appreciate your thoughts.
Thanks in advance.
Cheers,
Jason.

Answer 1

Ideas.. Hmm, none which seems that correct, but you could try some things.
If it gets the wrong IP that could explain why it fails to copy the sysidcfg file.
First you could try and do a snoop on the ethernet address;
snoop ether 0:3:ba:14:c6:cd
(you could also try the -v flag to increase the verbosity).
The things you should look for are arp/rarp requests, the jumpstart client will use arp/rarp to determine its IP address, snoop will show you which server that responds, and what address it gets, futher more its a good idea to verify that the response to the bootparams requests comes from the correct server.
Of course there might be other oddnesses as well.
Was the data you provided extracts? If it was you should check the /etc/ethers and /etc/bootparams for duplicates entries, so there are no other occourances of the clients ehternet address in /etc/ethers, and no bogus entries starting with * or the same hostname in /etc/bootparams.
If you added the client manually you could always try and use the Tools/rm_install_client and Tools/add_install_client scripts to add it again, these script sometimes detect problems with the configuration.
Lastly you didn't show us your /etc/nsswitch.conf file, but i assume that it has "files" first for the ethers, hosts and bootparams entries?
Good luck
//Magnus