[comp.sys.att] "Shoe-shining" with 60 meg Wangtek tape on Sys V 386 3.2.1

wtm@uhura.neoucom.EDU (Bill Mayhew) (10/18/90)

I've been trying to figure out a way to prevent shoe-shining of the
tape (repeated short back-and-forth movement) on a 60 meg
Sperry/CMS Wangtek tape drive on my Sys V 386 3.2.1 system. I did

find /usr/local -name "*" -print | cpio -oc -C1024 > /dev/rmt/c0s0

The system has 8 megs of RAM, and there weren't many processes
running, so cpio should have gotten the whole 1 meg buffer (I
think??).  It took about 1/2 an hour to back up about 5 megs of
data from /usr/local.  Yuck!  Running the same hardware under the
DOS partiton with the CMS QIC streaming software backed up up 30 megs
in just under 6 minutes with the tape running continuously.

The tape control board and drive have the exact same part numbers
as the respective AT&T labeled Mountain tape drive parts.  The only
diff is the copyright sticker on the ROM on the controller board.

Any ideas?

==Bill==
-- 
Bill Mayhew      NEOUCOM Computer Services Department
Rootstown, OH  44272-9995  USA    phone: 216-325-2511
wtm@uhura.neoucom.edu   ....!uunet!aablue!neoucom!wtm
via internet: (140.220.001.001)

wsinpdb@wsinis02.info.win.tue.nl (Paul de Bra) (10/19/90)

In article <1990Oct18.152955.14275@uhura.neoucom.EDU> wtm@uhura.neoucom.EDU (Bill Mayhew) writes:
>I've been trying to figure out a way to prevent shoe-shining of the
>tape (repeated short back-and-forth movement) on a 60 meg
>Sperry/CMS Wangtek tape drive on my Sys V 386 3.2.1 system. I did
>
>find /usr/local -name "*" -print | cpio -oc -C1024 > /dev/rmt/c0s0

According to the manual the -Cn option gives you a buffer of
n BYTES, not n BLOCKS. There is a bug in the 3.2.1 cpio however
which makes cpio give you a buffer of 10xn BYTES. So you really have
asked for a 10k buffer.

You can safely use cpio -oc -C100000 to get a 1000000 byte buffer.
I always use -C131072 or -C262144 to get a multiple of 1024 bytes
after the 10x bug.

Paul.
(debra@research.att.com)

ram@attcan.UUCP (Richard Meesters) (10/19/90)

You might want to try upping the value of the -C option from 1024 to say 
maybe 64k or so.  Your mileage will vary and it takes a little time to find
the correct block size to make your tape drive stream.  Try different values.


Regards,
Rich.

tneff@bfmny0.BFM.COM (Tom Neff) (10/22/90)

You can also pipe everything through a big 'dd' before it hits the tape

	find blah blah | cpio blah blah | dd bs=2048k of=/dev/rmt/c0s0

This double buffers 2MB at a time.  Watch that tape spin!

-- 
"DO NOT, repeat, DO NOT blow the hatch!"  /)\   Tom Neff
"Roger....hatch blown!"                   \(/   tneff@bfmny0.BFM.COM