Hi Netters:
This is an attempt to summarize the 20 or so responses I received
from my post about a week ago concerning the capacity of the 8mm
tapes, both 2.3 and 5.0 GB.
In general:
Most managers mentioned the inter record gap takes up space. Others
noted that each file has a marker which takes up quit a bit of space.
The number of files actually written was stated as being a variable
in the capacity of the tape as well.
The particular command that is used to write to the tape (tar, dump,
etc) and its parameters were keys as to the amount of data able
to be written. Especially block size. Too small a block sizes
wastes large amounts of tape.
The quality of the tape also should be mentioned. The error rate
can cause the capacity to decrease, since the drive seems to do
erasing and rewriting of the tape as it encounters bad spots. The
age and number of times the tape has been used is a factor.
Some said you lose about 10% of tape capacity in overhead, such as
the amount of tape needed for each EOF (2MB). I'm not real clear
on what was meant by overhead other than the EOF.
Some specific information was given:
From the EXB-8500 Product Specification:
physical blocks (1 KB)
Cartridge length 8200-mode 8500-mode
P5-90 112m 2413984 4827968
P6-120 106m 2292768 4585536
One manager mentioned that I could be running up against the
2GB limitation in the SunOS 4.x software. He said the standard
kernel returns an error if an attempt is made to write more than
2GB to any file or device because the file offset overflows.
And I received some exact commands that work for various managers
at their sites, such as
/etc/dump bdsf 112 54000 14400 /dev/nrst9 # 8200
/etc/dump bdsf 126 54000 6000 /dev/nrst0 # 8500
/usr/etc/dump 0ubsdf 126 6000 54000 /dev/[tape_device_name] /[filesystem_name]
(for both 2.2 and 5.0GB tapes)
/etc/dump 0ufbsd /dev/nrst9 112 10380 410000 /dev/rid000a
(for an EXB-8500)
/etc/dump 0ufbsd /dev/nrst1 100 6000 54000 /dev/rsd2a
(for an EXB-8200)
One manager described a test to see how much you can really fit
onto a tape:
dd if=/dev/zero of=/dev/nrst8 bs=1024k
which when done, would display the count in MB the size of the
single file the tape could contain. (I did not have time yet
to try this...I wanted to get this summary out as soon as I
could).
There were some other comments that might be of use but were just
too numerous to mention:
Bottom Line:
It looks like there are no EXACT numbers for the capacities of
both the 2GB and 5GB 8mm tapes. The actual capacity is determined
by many factors, such as blocksize used, error rates, is the tape
new or used, the format of the command (dump,tar), the blocking
factor, the parameters used in the command, and the angle of the
earth with respect to the sun...:>) With all these different
variables, the actual capacity is anyone's guess (or as EP Tinnel
put it "your milage may vary!"; thanks EP!)
MANY, MANY thanks to all those who took the time and effort to
send information to me. I am absolutely amazed at the amount of
knowledge and good people behind that knowledge on this newsgroup.
It is a valuable resource. Keep it up.
For those of you that would like the EXACT responses that I received,
please contact me via email. I'll be glad to oblige.
The helpful contributors were:
silver!jennine@uunet.UU.NET (Jennine Townsend)
Perry_Hutchison.Portland@xerox.com
tim@canon.co.uk (Tim F O'Donoghue)
Geert Jan de Groot <geertj@ica.philips.nl>
klarsen@afit.af.mil (Kristin L. Larsen)
walfish@jpmorgan.com (Joshua Walfish
dan@bellcore.com (Daniel Strick)
Patrick O'Callaghan <poc@usb.ve>
Tommy Reingold <boole.att.com!tommy>
don@mars.dgrc.doc.ca (Donald McLachlan)
Ray Brownrigg ray@isor.vuw.ac.nz
alida!swj@uunet.UU.NET (Stephen W. Jay)
Stephen W. Jay, swj@alida.com
David Wiseman <magi@csd.uwo.ca>
Bill Morrow, morrow@cns.ucalgary.ca
pluta@cgc9.eda.mke.ab.com (Rick Pluta)
ckelly@math.nwu.edu (Christopher Kelly)
ept@eptsun1.ctd.ornl.gov (E P Tinnel)
trdlnk!mike@uunet.UU.NET (Michael Sullivan)
(I hope I didn't miss anyone....)
Dan Zambon
dzambon@afit.af.mil
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:07:33 CDT