Exabyte 8500C

Exabyte 8500C

Post by Eckhard Rueggebe » Sat, 24 Apr 1993 18:37:43



I have an IPX with an Exabyte 8550C attached. When I boot 4.1.2, I can write to
the tape drive with no problem (after editing /usr/sys/scsi/targets/st_conf.c),
even in compression mode. mt -f /dev/rst9 stat says "Exabyte 8500C ...."
But when I boot Solaris 2.1, no cable, no disk, no tape changed, I can't do any-
thing useful with the tape. First, it is not reckognized correctly, mt says
only "SCSI tape drive:" , and sundiag finds a QIC tape. Second, when I try to
access it with tar, dump or even cat, I get numerous SCSI error messages :

ikarus:/home/ikarus/eckhard/rlogins 17 #        data transfer overrun
esp:            State=DATA Last State=DATA_DONE
esp:            Latched stat=0x10<XZERO> intr=0x10<BUS> fifo 0x80
esp:            last msg out: IDENTIFY; last msg in: COMMAND COMPLETE
esp:            DMA csr=0x94400000
esp:            addr=fff20290 dmacnt=8000 last=fff1da90 last_cnt=2800
esp:            Cmd dump for Target 5 Lun 0:
esp:            cdblen=6, cdb=[ 0xa 0x1 0x0 0x0 0x14 0x0 ]
esp:            pkt_state=0xf<XFER,CMD,SEL,ARB> pkt_flags=0x0 pkt_statistics=0x0
esp:            cmd_flags=0x10c22 cmd_timeout=120

        Reverting to slow SCSI cable mode

        SCSI transport failed: reason 'data_ovr': giving up


        Error for command 'write file mark'     Error Level: Fatal
        Block 0, Absolute Block: 0
        Sense Key: Unit Attention
        Vendor 'EXABYTE': ASC = 0x29 (reset), ASCQ = 0x0

Does somebody know why and what to change ?
Thanks a lot

Eckhard R"uggeberg

 
 
 

Exabyte 8500C

Post by M.A.Sc.. » Sat, 24 Apr 1993 19:37:53


You might try checking the SCSI terminator. We had similar problems:
system ok under 4.1.2, but failed under 2.1; fixed by adding an active
terminator instead of relying on the passive one within the exabyte.

--
--
Regards.   Mike Scott. BNR Europe Ltd, London Road, Harlow, Essex  CM17 9NA, UK

phone  +44-279-403133 (direct line) or  +44-279-429531 xtn 3133

 
 
 

Exabyte 8500C

Post by Christian Ra » Sat, 24 Apr 1993 22:46:06



|> I have an IPX with an Exabyte 8550C attached. When I boot 4.1.2, I can write to
|> the tape drive with no problem (after editing /usr/sys/scsi/targets/st_conf.c),
|> even in compression mode. mt -f /dev/rst9 stat says "Exabyte 8500C ...."
|> But when I boot Solaris 2.1, no cable, no disk, no tape changed, I can't do any-
|> thing useful with the tape. First, it is not reckognized correctly, mt says
|> only "SCSI tape drive:" , and sundiag finds a QIC tape. Second, when I try to
|> access it with tar, dump or even cat, I get numerous SCSI error messages :
[...]
|> Does somebody know why and what to change ?
|> Thanks a lot
|>
|> Eckhard R"uggeberg

The following program should solve your problem:
---------------------------------------------------------------------------------
/* Solaris 2.1 patch for Exabyte 8500C tape drives */

/* Compile this program with
        cc -o pexa pexa.c
   Run it on the workstation on which you want the patch installed.
   The program should produce the following output:
        00006d91: EXB-8500 -> EXB8500C
        0000733c: EXB-8500 -> EXB8500C
   Perform the following commands:
        mv /kernel/drv/st /kernel/drv/st.old
        mv /tmp/st.patch /kernel/drv/st
   Reboot your machine with option -v
   If the patch works, the Exabyte should identify with
        sdx: <EXABYTE EXB8500C 8mm Helial Scan Drive>
   or something like that.
*/

#include <stdio.h>
#include <string.h>

#define BUFSIZE 100000

char *changefrom = "EXB-8500";
char *changeto =   "EXB8500C";

int eerror (e, s)
int e;
char *s;
{
  fprintf (stderr, "%s\n", s);
  exit (e);

Quote:}

int bstrcmp (s1, s2, len)
char *s1, *s2;
int len;
{ int res, i;

  res = 1;
  i = 0;

  while (res && i < len)
    if (s1[i] == s2[i]) i++;
    else res = 0;

  return res;

Quote:}

void bstrcpy (s1, s2, len)
char *s1, *s2;
int len;
{ int i;

  for (i = 0; i < len; i++)
    s1[i] = s2[i];

Quote:}

main ()
{ FILE *fp, *ofp;
  int loc = 28041;
  int rd, wr;
  char buf[BUFSIZE], *ptr;

  fp = fopen ("/kernel/drv/st", "r");
  if (fp == NULL) eerror (2, "cannot access input file");

  ofp = fopen ("/tmp/st.patch", "w");
  if (ofp == NULL) eerror (2, "cannot access output file");

  rd = fread (buf, 1, BUFSIZE, fp);

  for (ptr = buf; ptr < buf + rd; ptr++)
    if (bstrcmp(changefrom, ptr, strlen(changefrom))) {
      bstrcpy (ptr, changeto, strlen(changeto));
      printf ("%08x: %s -> %s\n", ptr - buf, changefrom, changeto);
    }

  wr = fwrite (buf, 1, rd, ofp);
  if (wr != rd) eerror (3, "wrong number of bytes written to output file");

  fclose (fp); fclose (ofp);

  exit (0);

Quote:}

-----------------------------------------------------------------------------

I hope that SUN will support the EXABYTE 8500C by default in the following
Solaris releases...

--
Christian Rank
Lehrstuhl fuer Wirtschaftsinformatik * Universitaet Passau *
Innstr. 29 * D-8390 Passau

 
 
 

Exabyte 8500C

Post by ]ke Sandgre » Thu, 29 Apr 1993 00:16:26


In comp.unix.solaris you write:

Quote:>I have an IPX with an Exabyte 8550C attached. When I boot 4.1.2, I can write to
>the tape drive with no problem (after editing /usr/sys/scsi/targets/st_conf.c),
>even in compression mode. mt -f /dev/rst9 stat says "Exabyte 8500C ...."
>But when I boot Solaris 2.1, no cable, no disk, no tape changed, I can't do any-
>thing useful with the tape. First, it is not reckognized correctly, mt says
>only "SCSI tape drive:" , and sundiag finds a QIC tape. Second, when I try to
>access it with tar, dump or even cat, I get numerous SCSI error messages :

Solaris 2.1 doesn't know about 8500C's you need to patch /kernel/drv/st
to make it recognize the 8500C. It's a bit tricky though since that file is
a binary. Use emacs and the /usr/include/sys/scsi/targets/stdef.h
struct st_drivetype and the values from /usr/sys/scsi/targets/st_conf.c
for the 8500C.
What you need to do is change the 8200 entry to an 8500C entry.
When done it works like a charm.

If you give up, mail me and i'll send you my patched st.
--
Ake Sandgren, Inst. of Info. Proc, Umea University, S-90187 Umea, Sweden

 
 
 

1. Exabyte 8500C on SS1000

An exabyte 8500C is connected to our Sparc Server 1000 (Solaris 2.3).
Sometimes, when reading a file with a large amount of data, a timeout occurs
and an EOT is found. The message in /var/adm/messages is :


     Connected command timeout for Target 4.0

     SCSI transport failed: reason 'timeout': giving u

Any ideas ?

Thanks in advance.

------------------------------------------------------------
Anne Thomas Doare                     IFREMER/Centre de Brest          
                                            DITI/RIC
tel:98-22-45-68                             B.P 70
fax:98-22-45-46                         29280  PLOUZANE
                                            FRANCE


------------------------------------------------------------

2. FreeBSD and Novell Netware

3. Exabyte 8500C woes in SOlaris 2.3

4. Mandrake 7.0 & KDE

5. Here is a fix for Exabyte 8500c problem

6. "Top" for Solaris

7. Exabyte 8500C, Solaris 2.6 and dd

8. killall -CONT & killall -STOP

9. Exabyte 8500C and Solaris 2.x

10. exabyte 8500c woes - FIXED!

11. Exabyte 8500c unusable with Solaris 2.3

12. Exabyte 8500c errors

13. Exabyte 8500C on Solaris 2.2?