ado@uunet.uu.net (Arthur David Olson) (02/14/90)
Before: We've got an all-Sun Ethernet, with a 3/280 acting as a server
for a diskless 3/60 (among other machines). All is well.
After: We attach an SGI 4D/70G to the Ethernet. We can no longer boot
the 3/60.
Diagnosis: The 4D/70G's "bootparamd" daemon is responding to the 3/60's
requests for boot parameters; since we'd failed to tell the SGI about boot
parameters for Sun machines, the 3/60 gets a useless answer and can't
boot.
Cure: We make the information in the 3/280's "/etc/bootparams" file
available to the 4D/70G. (Another tack would be to comment out the
"bootparamd" line in the 4D/70G's "inetd.conf" file; I decided against
this to avoid changing a vendor-supplied file.)
The Question: Who's to blame--
* Me, for not having taken the "cure" step in the first place?
* The 4D/70G, for responding to boot parameter requests of
machines it knows nothing about?
* Or the 3/60, for paying attention to useless responses?
Arthur David Olson ado@alw.nih.gov ADO is a trademark of Ampex.