I can find reference on other tape types to how far before true
end-of-tape a physical early-warning marker appears, but not for DLT
IV. Does anyone know or have a pointer to an official reference?

Thanks,

- Gordon @ Internet Archive

[ Post a follow-up to this message ]

1,650 mm from the BOT hole, coming back on the last logical track
(logical EOT is actually at physical BOT due to serpentine recording
and an even number of logical tracks.)
It's in the ECMA standard.

Hope this helps,
Ralf-Peter

[ Post a follow-up to this message ]

Thanks, that helps. Looks like a 40GB tape is 557m long, so the
early-warning (EW) is somewhere around 110-120 MB from the end...

- Gordon @ IA

[ Post a follow-up to this message ]

In article <1108686977.145665.172710@f14g2000cwb.googlegroups.com>,
<gojomo@gmail.com> wrote:
>Thanks, that helps. Looks like a 40GB tape is 557m long, so the
>early-warning (EW) is somewhere around 110-120 MB from the end...
>
>- Gordon @ IA

But the data is recorded in serpentine fashion on many tracks
(208 for DLT8000) so isn't the early warning more like 0.6MB
from the end?

(40000 MB / 557 m ) * 1.65 m / 208

carl
--
carl lowenstein         marine physical lab     u.c. san diego
clowenst@ucsd.edu

[ Post a follow-up to this message ]

You have to base the calculation on 52 logical tracks (the DLT8000 has
a 4 track head)  which gives about 770 MB/logical track. The total tape
length is 557 m but only the area between BOT and EOT hole is used for
data. The BOT hole is 13.26 m from the beginning and the EOT hole is
6.096 m from the end of the tape, so you have a data area that is
nominally 537.6 m long.
770 * 1.65 / 537.6 = 2.36 MB. That's the latest that EW should be
asserted - given that the drive supports 16 MB data blocks the firmware
will have to set EW much earlier. I checked and there are some vague
hints that the FW calculates the EW position based on error rate. I
dimly remember that variability in tape length and error rates caused
issues when tapes were duplicated. One tape written to EW could not be
copied to another if that happened to be shorter or had a higher error
rate, so the firmware people based EW on worst case assumptions.

Ralf-Peter

[ Post a follow-up to this message ]