XENIX QIC Tape Questions

Alan Mintz alan at mq.com
Thu Dec 27 13:15:27 AEST 1990


In article <865 at anomaly.sbs.com>, mpd at anomaly.sbs.com (Michael P. Deignan) writes:
> 
> I have a few questions on XENIX tape drives which maybe someone can address.
> Using an Archive 60mb Internal system, in the /dev directory, we have the
> following devices defined:
> 
> crw-rw-rw-   2 root     superuser 25, 32 Dec  2 15:28 erct0
> crw-rw-rw-   2 root     superuser 25,  8 Dec  2 15:28 nrct0
> crw-rw-rw-   2 root     superuser 25, 24 Dec  2 15:28 nrct2
> crw-rw-rw-   2 root     superuser 25,  0 Dec 24 14:17 rct0
> crw-rw-rw-   2 root     superuser 25, 16 Dec  2 15:28 rct2
> 
> Now:
> 
> 	1. What is the difference between RCT0 and RCT2?

rct0 and rct1 refer to QIC24 tapes. rct2 and rct3 refer to QIC11 (whatever
that is - anybody ?)

> 	2. "ERCT0" is a device for a tape system w/ error correction, correct?

Yes.

> 	3. What are "NRCT#" and why use one of them over the RCT#
> 	   counterpart?

The nrct0 and nrct2 devices do not forceably rewind the tape before (after?)
each operation. This allows you to write multiple file system backups on
one tape by using the nrct0 device and the "tape" command to write file marks
between each backup (see a recent article in either c.u.x.sco or c.u.sysv386
from paulz at sco.com describing the exact procedure).

ECC allows a higher recoverable error rate on the tape than the standard
encoding scheme (2/64 blocks can be bad, yet recoverable, according to TFM).

And a question...

If you look at the minors for these devices, it seems that the 8s bit
is set for no rewind, the 16s bit is set for QIC11 (as opposed to QIC24), and
the 32s bit is set for ECC. This says that the nrct0 and nrct2 devices do
NOT (as they are defined) use ECC. Would creating a /dev/enrct0 device
with a minor of 40 create a no-rewind ECC device ?
-- 
< Alan H. Mintz             | Voice +1 714 980 1034 >
< Micro-Quick Systems, Inc. | FAX   +1 714 944 3995 >
< 10384 Hillside Road       | ...!uunet!ahmcs!alan  >
< Alta Loma, CA  91701 USA  | alan at mq.com           >



More information about the Comp.unix.xenix.sco mailing list