[rescue] Testing ext. CD-ROM, how to boot?

Ahmed Ewing aewing at gmail.com
Wed Oct 10 21:18:43 CDT 2007

On 10/10/07, Scott Quinn <compoobah at valleyimplants.com> wrote:
> > (at the OK prompt)
> > boot /pci at 1f,0/pci at 1/scsi at 1,1/sd at 6,0:f
> Second SCSI chain, meaning bus 2?
> Wouldn't that be scsi at 2/sd@(n),0:f ?  either that or scsi at 1,2/sd at 6,0:f,
> can't recall right now, and my connected Sun is single-bus. I think you
> can drop the original /pci... stuff unless you have multiple similar
> devices.

The exact hardware device path depends wholly on the architecture of
the platform and has more to do with the way the device is presented
to the pci bus than with which "number" scsi chain it is in the system
as a whole. The syntax is not standardized across the Sun line and can
vary greatly from model to model. I've never seen a "scsi at n,2"
designation either--that would suggest the third bus on a single
physical chipset (scsi at n being the first, scsi at n,1 being the second).
Often but not always, scsi at n,1 represents the external port of a
chipset that handles an internal and external chain.

The easiest way to do it is to simply execute "probe-scsi-all" and
note the dev path under which the cdrom ultimately appears. It will be
<that path>/sd at 6,0:f assuming the scsi target ID of the cdrom is
jumper configured to be 6.

Another option: execute "show-disks", which will list all recognized
paths that could possibly contain a scsi disk. It will list each of
them, up to the "sd@". Choose one by number, and it will enter the
string into a buffer that you can invoke by hitting CTRL-Y. So you'd
then type:

ok   boot <CTRL-Y> (it pastes in the string up to sd@) 6,0:f

Lather, rinse, repeat until you get a dev path that works. Make sense?

Hope that helps,


More information about the rescue mailing list