On Sat, Aug 11, 2007 at 02:28:10PM +0100, Rory Beaton wrote:
> For the record, here's the output of amtapetype for the Hewlett Packard DAT160
> USB with hardware compression OFF.
>
> There are no visible jumpers or dip-switches to disable hardware compression
> like earlier DAT devices - inside the enclosure the drive itself is a fully
> shielded unit.
>
> Compression can be turned off using HP Library and Tape Tools for Linux but
> the
> setting appears to be volatile and although the LTT app runs in a console it
> seems a bit overkill for running what seems to be a bunch of compressed perl
> scripts. Anyone found a workaround for soft configuration?
If your linux supports the stinit facility, you could use it to
create multiple devices to access your drive, each with a different
set of options for block size, compression, etc. Then you could
chose which you wanted for amanda by device name.
Does your mtx support any compression on/off commands. Just an
alternative to LTT.
Do the LTT show any way to set the power-up state? That was
what some of those DIP switches did on the earlier HP DATs
>
> Can anyone suggest why the filemark is showing such a large value?
>
> grunt:/usr/sbin # amtapetype -o -e 80g -f /dev/nst0
> Writing 512 Mbyte compresseable data: 94 sec
> Writing 512 Mbyte uncompresseable data: 82 sec
> Estimated time to write 2 * 81920 Mbyte: 26240 sec = 7 h 17 min
> wrote 2490330 32Kb blocks in 95 files in 12562 seconds (short write)
> wrote 2477223 32Kb blocks in 189 files in 12784 seconds (short write)
> define tapetype unknown-tapetype {
> comment "just produced by tapetype prog (hardware compression off)"
> length 78236 mbytes
> filemark 4461 kbytes
> speed 6272 kps
> }
>
DAT tapetype reports have varied as to whether there was any
filemark at all. I don't know why. My own testing on two
HP DATs (dds2 & dds3) showed no filemarks.
You might do another tapetype run to check if it is consistant.
When a tape streamer reads an EOF from the application it will
stop streaming. Then when the next tape file is opened and
writing begins, it may backup the tape, get up to speed and
leave a small gap between the end of the last file and the
start of the next file. This "safety gap", if present, is
the filemark as reported by tapetype.
To put your 4.46MB gap between each tape file in perspective,
it is only a 0.75 second gap.
Put another way, 4.46MB is 0.006% of your measured capacity.
You would have to write 150 tape files (complete DLE's or
if spanning tapes, chunks) for the filemarks to even consume
1% of the tape.
jl
--
Jon H. LaBadie jon AT jgcomp DOT com
JG Computing
4455 Province Line Road (609) 252-0159
Princeton, NJ 08540-4322 (609) 683-7220 (fax)
|
