Cream of the Crop 26/Cream of the Crop 26.iso/os2/gtak258.zip/DOC/tips.txt

Text File | 1994-09-03 | 18KB | 527 lines

Assorted mails and news on GTAK 2.12b.
Some of them are written in German.
    USE AT YOUR OWN RISK.

────────────────────────────────────────────────────────────────────────────────
────────────────────────────────────────────────────────────────────────────────

From:      Andreas Muschkat, 2:2426/3075.36
To:        Christian Hacker, 2:2447/102.31
Subject:   Re: Tar Backup
Date:      14.08.94 18:45

Hallo Christian Hacker,

Du meintest am 08.08.94 um 09:31 im Brett /PCPOOL/OS2.GER
zum Thema "Tar Backup":

CH>>> Nun kann es aber sein, dass auf Laufwerk F: und Laufwerk D:
CH>>> Dateien doppelt vorkommen. Diese koennen daher natuerlich
CH>>> nicht im gleichem Tar Archiv gesichert werden.
CH>>> Kann ich mit Tar unter OS2 mehrere Archive auf einem Band
CH>>> machen ?

MN>> Ich denke du bist hier der falschen Ueberzeugung das mehrere
MN>> tar-Aufrufe immer in das gleiche Archiv schreiben, was nicht der
MN>> Fall ist. Wenn du einfach alle PArtitionen hintereinander
MN>> sicherst, dann gibt es keine Probleme.

> Koentest Du mir vieleicht mal ein CMD-File zukommen lassen ? Evtl als PM.

Ich hatte Dir per NM schon ein "CMD-File" zugesandt, aber keine Antwort
erhalten.

Es wird zwar langsam Offtopic hier, aber da die mangelnde Antwort auch an
Problemen des Routings liegen koennte (sonst wuerdest Du ja hier nicht wieder
nachfragen), das ganze (hoffentlich letztes Mal) oeffentlich   :


       Archiv1                 Archiv2               Archiv3
|-----------------------|----------------------|----------------------|
Anfang                Pos1                   Pos2                   Ende


Legst Du ein frisches Band ein, so ist es leer. Hast Du kein leeres Band,
so musst Du

tape erase

eingeben, um es eines zu loeschen.
Nun sicherst Du z.B. drei Partitionen (unabhaengig davon, was drauf ist,
also ob da Files auf verschiedenen Partitionen gleich lauten) mit

c:
tar -cvp
d:
tar -cvp
e:
tar -cvp

Nun sieht Dein Band aus, wie in der "Zeichnung" oben.

Der Readpointer steht auf "Ende". Mit

tape rew

spulst Du das Band an den "Anfang" zurueck. Wenn Du nun

tar -xvp

eingibst, so wird das "Archiv1" vom Band auf das aktuelle Direktorie
entpackt und der Pointer steht auf "Pos1".

Wilst Du nun "Archiv2" nicht entpacken, dafuer aber "Archiv3", so musst Du

tape file           (um zu "Pos2" zu spulen)       und
tar -xvp

eingeben.

Ich hoffen, nun fallen Deine Schuppen von den Augen, wenn nicht, dann
experimentiere doch mal ein wenig mit den genannten Befehlen mit einem
leeren Band.

Tschaui, Andreas


--- CrossPoint v3.0
 * Origin: Neue Pointnummer : ----> (2:2426/3075.36)

────────────────────────────────────────────────────────────────────────────────
────────────────────────────────────────────────────────────────────────────────

From: [e-mail] (John P. Poet)
Newsgroups: comp.os.os2.apps
Subject: Re: How to turn of tape compression with GTAK???
Date: 12 May 1994 01:29:25 GMT
Organization: Engineering International Inc., Public Internet Access

In message <2qrku6$[e-mail]> - [e-mail] (Frank Paolino) writes:

>We have the EXABYTE TAPE 8505 and the GTAK SCSI tape package (two
>utils called TAR and TAPE).
>
>The 8505 defaults to write eith compression on.
>We want to exchange tape with another Exabyte drive that does not have compression.
>
>*** How can I turn off compression from the TAPE utility or from anohter piece of
>    software?

After running into this myself, I looked at the manual for an HP SCSI DAT
drive and determined that the following command would turn off the hardware
compression:

tape CMD \
cdb=15,10,0,0,14,0;wdata=0,0,10,0,0f,0e,40,80,0,0,0,20,0,0,0,0,0,0,0,0

Note that the \ is just means that it really all goes on one line.

To turn compression back on, use:

tape CMD \
cdb=15,10,0,0,14,0;wdata=0,0,10,0,0f,0e,c0,80,0,0,0,20,0,0,0,0,0,0,0,0

These work for me, but I can't guarentee that they will work for you.

John

____________________
John Patrick Poet   \
[e-mail]      \
Home: [phone]  \
Work: [phone]   \

────────────────────────────────────────────────────────────────────────────────
────────────────────────────────────────────────────────────────────────────────

From: [e-mail] (Kai Uwe Rommel)
Newsgroups: comp.os.os2.apps
Subject: Re: GTAK & RSH
Date: Fri, 12 Aug 1994 21:10:45 +0200
Organization: Private

[e-mail] (J POLTORAK) writes in article <[e-mail]>:
>
>Can someone show me how to pipe tar through rsh?
>
>I believe it can be done but I haven't figured it out...

[c:\] tar -cpv -f - | rsh unixhost -l user -b "dd of=/dev/tape-device bs=32k"

[c:\] rsh unixhost -l user -b "dd if=/dev/tapedevice bs=32k" | tar -xpv -f -

....

You need the latest CSD for TCP/IP to have a rsh.exe with a working -b
option (binary mode.

Kai Uwe Rommel

--
/* Kai Uwe Rommel                                      Muenchen, Germany *
 * [e-mail]                              CompuServe 100265,2651 *
 * [e-mail]                  Fax [phone] */

DOS ... is still a real mode only non-reentrant interrupt
handler, and always will be.                -Russell Williams

────────────────────────────────────────────────────────────────────────────────
────────────────────────────────────────────────────────────────────────────────

From:      "Andreas Kaiser" <[e-mail]>
Subject:   Re: Tape write protection error

> We have an HP C1533A external tape drive connected to an
> Adaptec SCSI expansion board. To date, we've had no luck
> writing to a DDS tape while connected to the Adaptec or
> the COMPAQ SCSI board (shipped with our server). While
> running Arcserv, a "tape write protection error" is
> reported using any tape (i.e.- DAT, DDS) even though we
> have the write protection tab out.

We had the very same problem recently. It looks like old DDS1 tapes
don't support the "media recognition system". Such tapes are treated
as read-only by default (perhaps interpreted as audio DATs). However
the device has a jumper to disable the media recognition system. If
disabled, the device can write to the old tapes, in DDS1 format.

────────────────────────────────────────────────────────────────────────────────
────────────────────────────────────────────────────────────────────────────────

From: [e-mail] (Gregory P. Smith)
Newsgroups: comp.os.os2.apps
Subject: Re: grr.... where is OS/2 tar (gtar)?????
Date: 18 Aug 1994 16:41:10 GMT
Organization: University of Colorado, Boulder

In article <32topr$[e-mail]>,
AB Wuysang <[e-mail]> wrote:
>
>I found that OS2 was not that good in multitasking my
>tape drive (Archive Viper 2150)...
...
>(virtually) this should work the same way, but instead it kept my tape
>write (short) - pause (short) - write (short) .., ie. sawing effect???
>since dd is not multi-threaded I can't use it in a way like:

This is because tar/gzip are all running at the regular priority.  I found
that with my drive if I use the sp (set priority) utility available on
hobbes (search for sp*.zip) and run tar/gzip as a time critical process
I get much better results without any sesawing.  What was happening before
was that tar wasn't keeping the data flowing fast enough for the tape
drive so it would pause quite often.
I have a 1542CF SCSI controller so the system load when I do this (I
haven't tried it with gzip turned on) is about 7% for a 9-10mb/min
backup.

--
Gregory P. Smith  --  [Team OS/2]
[e-mail], FidoNet: 1:104/332, RIME ->914
Happy OS/2 and Linux user.    (Finger for a PGP public key)

────────────────────────────────────────────────────────────────────────────────
────────────────────────────────────────────────────────────────────────────────

Path: spcvxb.spc.edu!terry
From: [e-mail] (Terry Kennedy, Operations Mgr.)
Newsgroups: comp.periphs.scsi
Subject: Everything you wanted to know about Wangtek 5150ES drives...
Date: 12 Jul 92 20:30:46 EDT

In article <[e-mail]>, [e-mail] (Dave Evans) writes:
> Including Terry Kennedys drive,  I believe there must be at least 4
> different versions of the Wangtek PCB.  I've seen at least 2 different
> EPROMS.   Does no one at Wangtek read this newsgroup who could offer
> some enligtenment ?   Our local distributor is quite helpfull, but
> their data sheets don't correspond with my drives, nor with Terry
> Kennedys'.

  First, the disclaimer: This is not an official representation of Wangtek
or of my employer. This is info I've discovered by reading publicly avail-
able reference material. When changing jumpers, always observe proper anti-
static precautions and be sure you have the current configuration written
down so you have a known starting point.

  Ok. Here's the complete scoop on Wangtek 5150ES drives:

  The current part number for a "generic" 5150ES is:

  33685-201 (black faceplate)
  33685-202 (beige faceplate)

  These are referred to as the "ACA version" of the drive.

  There are _many_ other part numbers for 5150ES drives. If you have one that
isn't one of the above, it doesn't mean you have an old or an out of rev drive,
it just means its a special version created for a distributor or OEM, or with
different default jumper settings.

  You can order the Wangtek 5150ES OEM Manual from Wangtek. It is part number
63045-001 Revision D. Wangtek can be reached at:

  WANGTEK Incorporated
  41 Moreland Road
  Simi Valley, CA 93065
  [phone] [voice]
  [phone] [FAX]
  [phone] [BBS]

  WANGTEK-Europe
  Unit 1A, Apollo House
  Calleva Industrial Park
  Aldermaston, Reading
  RG7 4QW England
  (44) 734-811463 [voice]
  (44) 734-816076 [FAX]
  851-848135 [telex]

  There are 5 possible logic boards. Here are the jumper options for each:

  Logic assembly #33678
  ---------------------

  (J10)
  0 - SCSI unit LSB
  1 - SCSI unit
  2 - SCSI unit MSB
  K - not documented

  J32 - Diagnostic test connector, default is not installed

  E1, F1 - SCSI termination power. E1 in = power from drive and to cable,
       E1 out - power from cable. F1 = terminator power fuse, 1.5A FB.
       Default is IN.

  E2 - Chassis ground. E2 in jumpers logic to chassis ground. E2 out isolates
       through a .33 uFD capacitor. Default is IN.

  E5 - Master oscillator enable. Test only. Must be IN.

  E20 - Factory test. Must be OUT.

  RP1, RP2, RP3 - SIP terminators. Default is IN, remove for no termination.

  Logic assembly #30559
  ---------------------

  HDR1 - Factory testing. Setting depends on drive. Don't touch.

  HDR2 - Factory testing. Defaults are pins 15-16, 17-18, 19-20. Don't touch.

  HDR3 pin 1 - A-B enables buffered mode. B-C disables. Can be overridden by
               SCSI Mode Select.

  HDR3 pin 2, 3 - Default data format. Set to B-C for a 5150ES.

  HDR3 pin 4 - parity enable. A-B enables, B-C disables.

  (J10)
  0 - SCSI unit LSB
  1 - SCSI unit
  2 - SCSI unit MSB
  K - not documented

  E1 - SCSI termination power. E1 in = power from drive and to cable,
       E1 out - power from cable.

  E2 - Chassis ground. E2 in jumpers logic to chassis ground. E2 out isolates
       through a .33 uFD capacitor. Default is IN.

  E3 - Master oscillator enable. Test only. Must be IN.

  E4 - Write test mode. Test only. Must be OUT.

  E5 - Write oscillator enable. Test only. Must be IN.

  E6 - Disable HDR2. Test only. Must be IN.

  E7 - Microcontroller clock select. In for a 5150ES.

  E8 - Write precomp select. Set on a per-drive basis. Don't touch.

  E9 - RAM size. Don't touch.

  E10 - Erase frequency. Don't touch.

  RP2, RP3 - DIP and SIP terminators. Default is IN, remove for no termination.

  Logic assembly #30600
  ---------------------

  HDR1 - Factory testing. Setting depends on drive. Don't touch.

  HDR2 - Write precomp select. Set on a per-drive basis. Don't touch.

  HDR3 pin 1, 2, 3 - SCSI device address. 1 is LSB, 3 is MSB. A-B=1, B-C=0

  HDR3 pin 4 - Parity enable. IA-B is enabled.

  HDR3 pin 5, 6 - Default data format. B-C for a 5150ES.

  HDR3 pin 7 - Buffered mode select. A-B is enabled.

  HDR3 pin 8 - Reserved. Must be OUT.

  HDR4 - Write frequency select. Don't touch.

  E1 - SCSI termination power. E1 in = power from drive and to cable,
       E1 out - power from cable.

  E2 - Chassis ground. E2 in jumpers logic to chassis ground. E2 out isolates
       through a .33 uFD capacitor. Default is IN.

  E3 - Hard/soft reset. IN enables hard reset.

  E4 - Write precomp select. Don't touch.

  E5 - Clock speed. Don't touch.

  E6 - Tape hole test. Don't touch.

  Logic assembly #30552
  ---------------------

  HDR1 - Factory testing. Setting depends on drive. Don't touch.

  HDR2 - Write precomp select. Set on a per-drive basis. Don't touch.

  HDR3 pin 1, 2, 3 - SCSI device address. 1 is LSB, 3 is MSB. [Note - HDR3
             pins 1-3 are duplicated at another location on the board]

  HDR3 pin 4 - Parity enable. IN is enabled.

  HDR3 pin 5, 6, 7, 8 - Default data format. 5,5 B-C, 7-8 A-B for a 5150ES.

  HDR4 - Write frequency select. Don't touch.

  E1 - SCSI termination power. E1 in = power from drive and to cable,
       E1 out - power from cable.

  E2 - Chassis ground. E2 in jumpers logic to chassis ground. E2 out isolates
       through a .33 uFD capacitor. Default is IN.

  E3 - Hard/soft reset. IN enables hard reset.

  E4 - Write precomp select. Don't touch.

  E5 - Clock speed. Don't touch.

  E6 - Tape hole test. Don't touch.

  Logic assembly #30427
  ---------------------

  HDR1 - Factory testing. Setting depends on drive. Don't touch.

  HDR2 - Write precomp select. Set on a per-drive basis. Don't touch.

  HDR3 pin 1, 2, 3 - SCSI device address. 1 is LSB, 3 is MSB. A-B=1, B-C=0

  HDR3 pin 4 - Parity enable. IA-B is enabled.

  HDR3 pin 5, 6, 7, 8 - Default data format. 5,5 B-C, 7-8 A-B for a 5150ES.

  E1, E3 - Factory test. Must be IN.

  E2 - SCSI termination power. E2 in = power from drive and to cable,
       E2 out - power from cable.

  E4 - Chassis ground. E4 in jumpers logic to chassis ground. E4 out isolates
       through a .33 uFD capacitor. Default is IN.

  Firmware - There are many flavors of firmware. I have seen the following
parts:

  24115-xxx
  24144-xxx
  21158-xxx

  the -xxx suffix changes as the firmware is updated. According to the folks
I spoke to at Wangtek, the standard firmware is the 21158. The latest version
as of this writing is 21158-007. All of these will work with the Adaptec and
GTAK.

  The firmware options (as returned by a SCSI Identify) are on the end of the
product string, which is "WANGTEK 5150ES SCSI ES41C560 AFD QFA STD" for the
21158-007 firmware. The 3-letter codes have the following meaning:

  AFD - Automatic Format Detection - the drive will recognize the format (such
    as QIC-24, QIC-120, or QIC-150) that the tape was written in.

  QFA - Quick File Access - the ability to rapidly locate a tape block, and
    to implement the "position to block" and "report block" SCSI commands.
    This is compatible with the Tandberg implementation.

  STD - Standard feature set.

  I hope you find this information useful.

────────────────────────────────────────────────────────────────────────────────
────────────────────────────────────────────────────────────────────────────────

From: [e-mail] (Andreas Kaiser)
Newsgroups: comp.os.os2.apps
Subject: Re: GTAK problem
Date: Sat, 03 Sep 94 22:30:23 GMT

[e-mail] writes in article <3483ho$[e-mail]>:

>         tar -cp -pp C:/.
> I wonder if just
>         tar -cp -pp C:/
> (without the ".") would be better (or equivalent)?

Equivalent.

> Also, I am not sure what "quoting" means.  Would that be
>         tar -cp -pp "C:/*"
> or
>         tar -cp -pp 'C:/*'

Well, my knowledge on this subject is not necessarily much better than
yours, since I did not write this part of GNU-TAR (I just added the
initial wildcard match mimicing Unix behaviour) and I rarely use
wildcards with TAR, except to save a finite set of known files. So when
I encounter such questions, all I do is trial-and-error, perhaps eased
by some looks into the source code.

As a result of such a trial-and-error survey made right now, I
discovered that TAR does not match quoted wildcards at all when
archiving (I did not knew this before). Quoted wildcards are matched
when diff'ing or reading the archive (restore, diff, list).

Examples of properly quoted wildcards (use double quotes):
    tar xpv "c:/os2/*.add"
    tar xpv c:/os2/"*.add"

The general rules are:

- When archiving, use wildcards to save a finite set of files only. To
save whole filesystems or directories, do not use wildcards, just
specify the root- or subdirectory. And remember that the only directory
which ends in a "/" is the root directory, all others should be
specified without the trailing "/".

- To restore (or diff, list) wildcard-selectable files, use quoted
wildcards only. Unquoted wildcards try to match a set of files on your
disk before restoring, so you may get much less restored as expected if
the pattern matches something on your disk.

- To restore a whole archive, do not specify a filename or pattern
argument.

- To restore a directory (or a filesystem, if you used -P when
archiving) of the archive, do not use wildcards. Just specify the
directory (or filesystem).

        Gruss, Andreas

-------------------- internet: [e-mail]
-- Andreas Kaiser -- fidonet:  2:246/8506.9
--------------------

────────────────────────────────────────────────────────────────────────────────
────────────────────────────────────────────────────────────────────────────────