SUMMARY : Problems netbooting/jumpstarting a Netra t1 105

From: Jasper Wallace <jasper_at_ivision.co.uk>
Date: Tue Jun 11 2002 - 10:31:38 EDT
On Wed, 5 Jun 2002, Jasper Wallace wrote:

> Hi,
>
> I'm trying to jumpstart a Netra t1 105, but it hangs partway through the
> boot sequence.
>
> Booting with -v show that it hangs after:
>
> su2 at ebus1: offset 14,3803f8
> su2 is /pci@1f,0/pci@1,1/ebus@1/su@14,3803f8
> su3 at ebus1: offset 14,3602f8
> su3 is /pci@1f,0/pci@1,1/ebus@1/su@14,3602f8
>
> which is odd, beacause i think these two devices (serial port a and b) should
> be su0 and su1.
>
> This is using Solaris 8 02/02

It turns out that the jumpstart boot image (the directory structure that
gets nfs exported and mounted as / on the boxes being booted) is NOT read
only as i had always assumed it would be.

I compared boot messages from a working machine with the one i'm trying to
netboot and noticed that the su2 and su3 devices are at ebus1, where as
su0 and su1 are at ebus0:

Working:

PCI-device: ebus@1, ebus0
su0 at ebus0: offset 14,3803f8
su0 is /pci@1f,0/pci@1,1/ebus@1/su@14,3803f8
su1 at ebus0: offset 14,3602f8
su1 is /pci@1f,0/pci@1,1/ebus@1/su@14,3602f8

Not working:

PCI-device: ebus@1, ebus1
su2 at ebus1: offset 14,3803f8
su2 is /pci@1f,0/pci@1,1/ebus@1/su@14,3803f8
su3 at ebus1: offset 14,3602f8
su3 is /pci@1f,0/pci@1,1/ebus@1/su@14,3602f8

If these where different systems i'd guess it was somthing to do with prom
settings, but they are the same box.

I did some grepping around and in /etc there is a file called path_to_inst
which contains mappings between device paths in the prom and device
instances.

Beacause i had previously used this jumpstart setup to boot netra X1's and
sunfire V100's the mapping for ebus0 was somewhere @isa, which didn't help
netra t1's much...

This causes the ebus on the Netra t1 to be ebus1, not ebus0, so the Netra
didn't find su0 and su1 (which should be at ebus0), and so found them as su2
and su3.

Why it then hung i don't know, but running with a /etc/path_to_inst from
radicaly different hardware is obviously not healthy.

I commented out all the lines that refered to isa and the netra t1 boots
fine now.

So i guess the bug here is:

Booting either a Sunfire V100 or a Netra X1 changes the etc/path_to_inst
file in  the jumpstart boot image in such a way as to prevent Netra t1's
from net booting off of the same image.

I guess the last time we had this problem it was caused by either an Ultra 5
or an Ultra 2.

I don't know what will happen if we next try to net boot a V100 or X1, but
will try soon.

I may ammend our procedures here to delete, or empty (cat /dev/null >
etc/path_to_inst) the etc/path_to_inst file before each machine class is
jumpstarted.

Thanks to:

Pierre Zimmermann <Pierre.Zimmermann@tecnomen.fi>
and
David Foster <foster@dim.ucsd.edu>

for vairous sugesstions.

-- 
Internet Vision          Internet Consultancy           Tel: 020 7589 4500
60 Albert Court            & Web development            Fax: 020 7589 4522
Prince Consort Road                                   vision@ivision.co.uk
London SW7 2BE                                   http://www.ivision.co.uk/
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Received on Tue Jun 11 10:36:19 2002

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:42:46 EST