Very Computer

Hello Solaris Community.

I'm wondering if anyone else has seen this behaviour.

With Solaris 8, we've had a process in place for a while now which has
worked well.  We've simply been creating compressed configuration
tarfiles, parking them on an nfs share, then booting the Sun client
from cd/dvd rom drive, using the appropriate obp command line:

boot cdrom - install nowin dhcp \
nfs://<nfsserver>/<path_to_compressed_config>/<config_tarball>

The machine would then go through the process of starting the kernel,
grabbing an IP address, grabbing the compressed config file, then
kicking off the automated install.  No need to mess with creating boot
servers, etc.

With the inception of Solaris 10, however, this seems to have become
broken.  The system will start the kernel, obtain an IP address (with
success), then promptly bust into an un-breakable loop of death.
Here's the output:

.... [snip]

Configuring devices.
Trying DHCP on hme0
Success
Using DHCP for network configuration information.

<unknown>: warning: dhcpagent: daemonize: timed out awaiting adoption.

DHCP: using local media

/sbin/install-discovery: /usr/bin/loadkeys: not found

[ network/physical:default failed (see 'svcs -x' for details) ]

Feb 24 06:38:17 svc.startd[7]: svc:/network/loopback:default: Method
"/lib/svc/method/net-loopback" failed with exit status 1.

[ network/loopback:default failed (see 'svcs -x' for details) ]

Feb 24 06:38:17 svc.startd[7]: svc:/system/filesystem/usr:default:
Method "/lib/svc/method/fs-usr" failed with exit status 1.

Feb 24 06:38:18 svc.startd[7]: svc:/system/filesystem/usr:default:
Method "/lib/svc/method/fs-usr" failed with exit status 1.

Feb 24 06:38:18 svc.startd[7]: svc:/system/filesystem/usr:default:
Method "/lib/svc/method/fs-usr" failed with exit status 1.

[ system/filesystem/usr:default failed (see 'svcs -x' for details) ]

Requesting System Maintenance Mode

(See /lib/svc/share/README for more information.)

Console login serice(s) cannot run

svc.startd: Could not exec() sulogin: No such file or directory

Requesting System Maintenance Mode

(See /lib/svc/share/README for more information.)

....... endless loop.

Something's obviously been modified for 10 -- has Sun removed the
option for using compressed tar configuration files or something?

Thanks.

Not a method I have ever used, but surely flash archives would be more
appropriate

Quote:

> The machine would then go through the process of starting the kernel,
> grabbing an IP address, grabbing the compressed config file, then
> kicking off the automated install.  No need to mess with creating boot
> servers, etc.

As you have an NFS server, why not just use that for rthe boot server ?
You would then not have to walk/drive/run to the server to be built ot put
a CD/DVD in and just do a boot net<:dhcp> - install

Quote:

> With the inception of Solaris 10, however, this seems to have become
> broken.  The system will start the kernel, obtain an IP address (with
> success), then promptly bust into an un-breakable loop of death.
> Here's the output:

Browsing through http://docs.sun.com, there appears to be no method for
compressed tar install with nfs on Sol10, but there is for http, I am sure
more delving will find the answer to why it is removed (either security or
no one was using it :-) ).

So, maybe setup as http ? or configure JumpStart, Sun supply JET
(JumpStart Enterprise Toolkit) free on http://sun.com/downloads and there
is usefull setup  info at

http://www.sun.com/bigadmin/content/jet/
and Resources at http://jet.maui.co.uk

HTH

Quote:

> .... [snip]

> Configuring devices.
> Trying DHCP on hme0
> Success
> Using DHCP for network configuration information.

> <unknown>: warning: dhcpagent: daemonize: timed out awaiting adoption.

> DHCP: using local media

> /sbin/install-discovery: /usr/bin/loadkeys: not found

> [ network/physical:default failed (see 'svcs -x' for details) ]

> Feb 24 06:38:17 svc.startd[7]: svc:/network/loopback:default: Method
> "/lib/svc/method/net-loopback" failed with exit status 1.

> [ network/loopback:default failed (see 'svcs -x' for details) ]

> Feb 24 06:38:17 svc.startd[7]: svc:/system/filesystem/usr:default:
> Method "/lib/svc/method/fs-usr" failed with exit status 1.

> Feb 24 06:38:18 svc.startd[7]: svc:/system/filesystem/usr:default:
> Method "/lib/svc/method/fs-usr" failed with exit status 1.

> Feb 24 06:38:18 svc.startd[7]: svc:/system/filesystem/usr:default:
> Method "/lib/svc/method/fs-usr" failed with exit status 1.

> [ system/filesystem/usr:default failed (see 'svcs -x' for details) ]

> Requesting System Maintenance Mode

> (See /lib/svc/share/README for more information.)

> Console login serice(s) cannot run

> svc.startd: Could not exec() sulogin: No such file or directory

> Requesting System Maintenance Mode

> (See /lib/svc/share/README for more information.)

> ....... endless loop.

> Something's obviously been modified for 10 -- has Sun removed the
> option for using compressed tar configuration files or something?

> Thanks.

--
        Bruce

"The internet is a huge and diverse community and  
not every one is friendly"                            
        http://www.ytc1.co.uk                    

I don't think this is the solution, but check.  Ensure you jumpstart
shares are shared out exactly as they are on the original operational
server.

I too would not have attempted just a tar bar.  I would have only
tarred up some of the profiles and config start scripts, but
re-installed the software images from scratch.

Good luck.