Amanda-Users

Re: more with Amanda and Sun L9

2007-09-20 06:03:15
Subject: Re: more with Amanda and Sun L9
From: Craig Dewick <cdewick AT lios.apana.org DOT au>
Date: Thu, 20 Sep 2007 20:00:37 +1000 (EST)
Hi Brian,

I'm sorry, missed this thread until this morning and don't know
your config or the actual problem.

That's ok.

I have an L9 on solaris and might be able to help.

See that you've configured sgen, the L9 uses different scsi id
for the robot and the tape, these have to be reflected in the
amanda.conf. This is somewhat unlike the C2 which used a single
scsi id and different LUNs.

Ok I've confirmed it's using different SCSI ID's for the robotics car and the DLT drive using the front panel LCD controls. The unit has quite an old firmware version though - 2.06.

Don't know which glue script you are using, I'm running the mtx
version, chg-zd-mtx, however I found that the version of mtx
I was running had a problem with the script itself. I have to dig
through my notes to find both the version and the problem, it was
something I think as simple as the shell the script invoked.

Yes I'm using chg-zd-mtx as well. I have wondered of the version of mtx is not compatible. I currently have v1.3.11 of mtx from Eric Lee Green installed on the current host system (Sun Ultra 2 with a 370-1704 HVD Sbus SCSI card).

Something that bothers me is that the 'sgen' driver won't pickup the robotics card. /kernel/drv/sgen.conf contains:

device-type-config-list="changer", "sequential";
inquiry-config-list="HP", "QUANTUM";

and all the:

name="sgen" class="scsi" target=0 lun=0;
name="sgen" class="scsi" target=1 lun=0;
name="sgen" class="scsi" target=2 lun=0;

lines are uncommented. Reading I've done on the sgen driver suggests that the conf file shouldn't need any 'parent="xxx"' additions to the individual target config lines.

The 'st' driver finds the actual DLT drive without a problem and that shows in the boot-time messages.

Craig.




<Prev in Thread] Current Thread [Next in Thread>