Hi,
I am trying to use a Quantum DLT8000 drive on an ES40 with KZPCM-DA
controllers. The drives are LVD but the HP Surestore 2/20, that the drive
is in, correctly reports that Single ended signalling is being used. I am
getting the Compaq LVD adapter to run them on.
SCU and hwmgr both show the device as identifying itself as
"QUANTUM" "DLT8000".
If I use
ddr_config -s tape COMPAQ DLT8000
then the output is perfect as you would expect and the densities all line
up with those specified in the man page for tz.
ddr_config -s tape QUANTUM DLT8000
does not match the device and uses defaults, setting the device up as type
rdat rather than tk and all the density codes as 00 (default), no refernce
to compression etc.
I have looked at
1) The man page for tz
2) The man page for ddr.dbase
3) The man page ddr_config
4) The manuals for Quantum DLT4000, Quantum DLT7000, and DLT8000 and it is
clear where the density codes for the "COMPAQ" branded devices come from.
5) ddr.dbase
the following summarises what I found out
1) ddr.dbase does not contain entries for the Quantum DLT4000, DLT7000 or
DLT8000
2) the /dev/ntape/tape_d1 uses maximum density - not because the explicit
density code
for the maximum density is set but because the density code is 00 -
i.e. default
3) I cannot see any standard usage of the _d3 - _d7 in terms of
intermediate densities that
the drive supports
4) If I copy the ddr.dbase entries for Compaq DLT4000, DEC TZ89, and Compaq
DLT8000
to create QUANTUM DLT4000, QUANTUM DLT7000 and QUANTUM DLT8000
respectively and compile it then ddr_config -s shows what you would
expect by
reading the tz, ddr.dbase, and the Quantum manuals.
What I am having trouble with is:
1) When is ddr.db used ? At device creation time? I would expect that
there would be a link between the hwmgr's scsi did and the ddr.db info for
any recognised device. Is this how it works and/or how does it work?
2) I am getting lots of SCSI events when I try to use the device either
using default parameters or those generated by modifying ddr.dbase. I don't
know how to interpret the errors coming out of evmshow. Should I be using
something else to figure out what is happening?
3) Is the LVD HBA likely to help? My expectation is no, but I will be able
to run longer cables and the Data Xfer rate will be better.
TIA,
Regards,
Greg PAlmer
Technical Leader, Central Systems Services,
Queensland University of Technology
Received on Tue Apr 10 2001 - 18:25:32 NZST