Hello,
I'm having problem replacing a SCSI disk on my sun SparcServer 1000E.
I have an old (NARROW) micropolis 2.1 Gb drive that is giving me
errors in the /var/adm/messages. I wanted to replace it with a new
barracuda 9Gb drive that I configured to use the same SCSI ID as the
old disk.
I removed the mount entry for the drive to be rewplaced from
/etc/vfstab, shut the server down, turned off all peripherals and
proceeded to replace the micropolis drive with the new drive, this
required a change of cables (See figure below)
Old setup:
Sun N---------C
-C Micropolis drive
|
-C Other drives
-C
|_...T
New setup:
Sun N---------W
-W Barracuda
|
-C Other drives
-C
|_...T
- = Cable
W = Wide SCSI connector
N = Narrow SCSI connector (Not centronics type)
C = Narrow, "centronics" type connector.
The lenght of the SCSI chain did not increase, actually it probably
decreased. The barracuda is NOT TERMINATED, I removed the appropriate
jumper (I tested it on an NT machine and it was seen as ID 4 correctly
and was functionnal)
What happens when I boot is that no devices are recognized anymore on
the SCSI chain where a drive was replaced.
What might be the problem here ?
a) Terminator issues (The barracuda had a terminator jumper that I
removed, but there are two other jumpers that say:
- Drive supplies Bus termination power
- Drive supplies own termination power (Default)
- Bus supplies drive termination power
b) N,W,C connector mixing causing problems ? We have a similar setup
on another chain and it runs fine
c) Should I touch /reconfigure even if I'm not adding a new drive but
REPLACING an existing drive with a different drive ? I don't thinks so
because at the prom prompt, the probe-scsi-all gave errors on that
particular bus as well.
d) Cable problem ? On of the cables is new
Any suggestions, hints would be appreciated.
Thanks
NOTE: To reply via e-mail, remove the 'anti-spam-' from my e-mail adress