 |
 |
 |
 |
>I was interested in your comments. Like you, we also FACSnet our data to a
>DEC DECstation (unix machine we use as a server). At the moment we then
>archive on high capacity tapes (2.2 Gbyte). As one always worries about the
>longevity of tapes, I make multiple copies. I am currently interested in
>storing flow data on optical disks but can't decide whether to use WORM or
>re-writable. I am interested in your experiences with either type; which
>brands are you using etc?
We use a DEC MicroVAX II running VMS 5.4. The WORM drive is a Maxtor RXT
800S, and we use a U.S. Design 1108 Qbus to SCSI adapter. Since this is hardly a
DEC peripheral, the WORM uses VOS 2 software to interface with VMS. This means
that most VMS commands don't work with the WORM: instead of MOUNTing a WORM you
have to OPMOUNT it, instead of COPYing to the WORM you have to OPSAVE to it, but
it's okay to COPY from the WORM to the hard disk. This situation makes direct
network access to the WORM drive impossible. <strike one - not truly compatible
with VMS> WORM directories are stored on the hard disk until the WORM is full;
at this time an OPFREEZE command is issued, and hopefully it will take. The
original drive that we bought was a 5.25 inch full-height device. It had the bad
habit of not being able to complete every OPFREEZE. This meant that the
directory had to stay on the hard disk (taking up 2.5 to 4 Meg for each SIDE of
the WORM platter), and you hoped nothing happened to your hard disk (without the
directory, the WORM is TOTALLY unreadable: goodbye 400 Meg of data!) The
replacement drives have both been 5.25 inch HALF-height devices, and the current
one has been very reliable. <strike two - not as 'fail-safe' as a WORM should
be> The Maxtor optical platters store about 800 Meg (400 Meg per side) and cost
$102 (U.S.) in March of 1991 (= roughly 13 cents per megabyte). Unfortunately,
they are not 'standard' platters, so they are only available through U.S.
Design/Maxtor. <foul ball? - proprietary/non-standard media>
But most of the time the darn thing works. I believe that other
companies have released better units since the time when we bought our WORM in
May of 1989. Re-writable (Magneto-Optical, or MO) drives were not available (?)
during our evaluation/purchase (I was not here at the time, so I am not sure).
We have one MO on a Northgate 486 PC. Unfortunately, DOS doesn't like
our system configuration (a bitter fight between the network driver and the MO's
SCSI driver). Maybe MAK can tell you about the manufacturer and his experiences
with this system ...
An aside: one PI was worried about old data (7-8 years old) stored on
9-track tape. After several hours (spread out over several days), I copied the
data from approximately 100 tapes to one SIDE of a WORM platter. Of the 100 or
so tapes, only 2 couldn't be recovered, and they were some of the last tapes to
be archived. Since that time (almost a year ago), the data has NOT been
accessed. Was it worth it? No.
Another aside: we have data on 8-inch floppies (a couple hundred) from
an MDADS, but we no longer have an MDADS. Prior to its removal, I was encouraged
to transfer the data from the 8-inch floppies to our MicroVAX II. Unfortunately,
we didn't have a network card in the MDADS, and its RS-232 port couldn't
reliably transfer all the data. Should we go to another site where they have an
MDADS on the internet and transfer the data now, before MDADS become obsolete?
No, because the data is becoming obsolete (in most cases) along with the
hardware it ran on. Just one request was made for this data over the last 3
years (or more).
/\/\/\_ Eric Van Buren vanburen%flovax.dnet@rocdec.roc.wayne.edu
\ \ \-' Immunology/Microbiology 313 577 1009 voice
\_^_/ Wayne State University 313 577 1155 fax
|
|
|
|