Thanks to Alan _at_ Digital who correctly diagnosed my "problem":
If the 18 is a hex SCSI status and I read the SCSI-2 spec
correctly, it means "Reservation conflict". Since this
is a cluster, it would see that the other member of the
cluster is up and is using those disks. To prevent the
other node from accessing them, it uses the SCSI Reserve
command. So, the booting node simply sees the devices
reserved and reports that. One might wish that the console
would spends a few more bytes to translate the SCSI status
to the corresponding message, but all seems normal.
It seems that if we shutdown the other node, the error disappears. So
we're not worrying about it anymore.
Pete
Original Question:
Has anyone seen this? Can you point me in the direction of a solution?
When booting our DEC Alpha 1000A 5/333 system, during the Initialize
phase..
---
Testing the Disks (read only)
bad SCSI status (18) received from pkc0.0.0.2003.0
file open failed for dkc1.0.0.2003.0
Testing the Network
---
I'm no firmware guru, but I believe "pkc" is our KZPSA that attaches to
our RAID Array 450 w/HSZ50.
There are six "units" defined in this cabinet: D1 D2 D3 D4 D5 D6
The SCSI bus is differential, with another KZPSA on another system on
the
same bus. (This is a cluster).
Thanks in advance,
Pete
Received on Mon Feb 01 1999 - 19:47:49 NZDT