We are trying to migrate from 4100 series hardware to ES40 series
hardware. The 4100 series server is directly attached to a storage
array via a KZPBA-CB UltraSCSI Storage Adapter Module. The boot device
for the 4100 series server and all its data is located on this storage
array. The operating system for the 4100 series server is Tru64 UNIX
5.1 patch kit-003.
Our migration plan was to simply to detach the 4100 series server from
the storage array and then to attach the ES40 series server to that
storage array using the same SCSI cable previously attached to 4100
series server. We also figured on rebuilding the kernel because the
operating system on the boot device was previously generated for 4100
series type hardware. Anyway that was our plan, it sounded really easy
to do so we set about to try it.
One of the good things about Tru64 UNIX 5.1 is you can pull the cards
out of the expansion slots mix them up, put them back into the computer
in any order, and it doesn't matter becuase TRU64 UNIX 5.1 remembers the
configuration of the cards via the world-wide-id or mac-type-address on
the I/O cards and not by the ordinal postion of the cards within the
slots on the chassis. That is the good news about TRU64 UNIX 5.1 and
unfortunately for us; if one of our theories holds true, this is also
the bad news too. I think they call it persistence of memory accept for
our
expirement we required less remembering and more persistence. Needless
to say we were not successful with our first expirement. We couldn't
make the operating system on the boot device forget about its previous
life as a 4100 series type hardware and we couldn't get the ES40 to boot
from
that device even when using the genvmunix kernel located on that device.
Anyways I didn't mention the ES40 came with its own raid-controller and
local drives with Tru64 UNIX 5.1 pre-installed on those local drives.
We were able to boot the ES40 from those local drives and see all the
devices associated with the SCSI cable attached to the storage array,
but try as we may we couldn't boot from the boot device on the storage
array even when we trying to use a generic vmunix kernel located on
that device. This behavior was totally unexpected.
We think the hardware manger database is crux of our problem. Another
theory suggests it may be related to the local raid-controller on the
ES40. Heheh, I think we are stuck. If anyone has any suggestions or
alternate routes on how to reincarnate a 4100 series server as an ES40
server we would appreciate your comments.
Anyway we hope for good karma this Saturday.
Sincerely
Kevin Criss
--
-------------------------------------------------------------------------------
IC 5-14-1.5 Indiana's open door to state government law.
http://www.ai.org/legislative/ic/code/title5/ar14/ch1.5.html
-------------------------------------------------------------------------------
Received on Tue Nov 13 2001 - 20:10:22 NZDT