I solved this problem by removing the 8mm tape drive from the end of the SCSI
bus.
The configuration file and terminations were correct.
The console parameter FAST_SCSI_B was set to TRUE.
In using the utility 'scu' I was able to see...
the disk drives reported: ANSI Version: SCSI-2 Compliant
while the tape drive reported: ANSI Version: SCSI-1 Compliant
It appears 'fsck' on reboot was having a problem with the tape on the bus.
After the system came-up alright (fsck ran fine during reboot) I was able to
put the tape drive back on the bus and run 'fsck -o' without problems.
Thanks to those who responded.
Mike Vihel
New Mexico State University
Origional posting:
Here's a little problem with disk installation:
We bought 3 identical Seagate 3.5 inch 2.1Gb drives. I installed two of them in
the Alpha - one as SCSI ID 1 on bus 0 (internal rz1) and the other as
scsi id 0 on bus 1 (external rz8). I removed rz9. rz3 (system disk) and
rz10 (users disk) are still installed.
The active termination jumper was removed from both rz1 and rz8.
I used the following procedure to install rz1 and rz8:
cp /genvmunix /vmunix
shutdown now
power off
power on
reboot
/sbin/bcheckrc
cd /usr/sys/conf
cp NMSNPA NMSNPA.old
edt NMSNPA : add devices rz1 and rz8
/usr/sbin/doconfig -c NMSNPA
mv /usr/sys/NMSNPA/vmunix /vmunix
cd /dev
./MAKEDEV rz1 rz8
shutdown now
power off
power on
reboot
disklabel -rw rz1 scratch1
disklabel -rw rz8 scratch2
At this point the console locked up. I hit the halt button, rebooted and tried
labeling rz8 again, which again failed and locked up the console. After another
reboot I finished the installation of rz1 by doing:
newfs /dev/rz1a
newfs /dev/rz1g
mkdir /tmp2
mkdir /scr1
mount -t ufs /dev/rz1g /scr1
cd /etc
edt fstab: added rz1, removed rz9
After a third reboot I was able to label rz8, do newfs, change fstab
successfully, but the whole system locked up again within 10 minutes. The system
would not reboot because it failed the fsck. I had to remove the scsi bus from
rz8 to get the system to reboot.
Today I deleted all files referring to rz8 and rz9 from /dev. Changed the scsi
id of the external disk to 1 (rz9) just in case scsi id 0 is reserved and to
make the settings identical to the internal disk, which has no problem. I
repeated all of the above with the same behavior.
I then exchanged the disk for an identical one which had been installed in
nmsnp4 without incident. Same results.
I then exchanged the power and scsi cables between rz9 and rz10 (which works
flawlessly) and was able to label the disk, but could not reboot.
On the rare occasions when I am able to label the disk, the system fails to
reboot. The error message given is
/sbin/ufs_fsck -p
/dev/rrz3a: 1002 files, 47390 used, 15841 free (57 frags...)
Can'tcam_logger: CAM_ERROR packet
cam_logger: bus 1 target 1 lun 0
dme_tcds_resume
Invalid DME DAT element.
cam_logger: CAM_ERROR packet
cam_logger: bus 1 target 1 lun 0
sim_err_sm
Target went to data in phase
It appears to be happy until it tries to fsck the bus 1 target 1 (rz9) disk.
So I'm stumped. The internal disk with exactly the same setup works fine. Either
of the two other disks fail. I haven't tried swapping the rz1 disk to the rz9
place.
Perhaps you have some ideas and could come over sometime to look at the problem?
Received on Mon Feb 27 1995 - 12:59:26 NZDT