Re: Where is the problem: Tape Drive? Cartridge(s)? Cable? SAS Controller?

["Thomas Schmitt" <scdbackup@gmx.net>]

Hi,

> Nov 13 08:02:29 BigMutt kernel: [34669.493879] st 0:0:0:0: [st0] Error e0000
> (driver bt 0x0, host bt 0xe).
> Nov 13 08:02:31 BigMutt kernel: [34671.743714] st 0:0:0:0: [st0] Error 10000
> (driver bt 0x0, host bt 0x1).

This looks electrical, not mechanical or magnetical.

I am sure that above messages are not the complaint of healthy drive firmware
about a bad medium or a drive problem. It probably comes from line 391 of
  https://github.com/torvalds/linux/blob/master/drivers/scsi/st.c
It seems to get emitted when no SCSI error condition was sent back by the
drive, but the transaction failed nevertheless.

Speculating more, assumed that "driver_bt" is sg_io_hdr_t.driver_status
and "host_bt" is .host_status of the low level SCSI driver, then the numbers
could be decoded by the comments in kernel header file
  https://github.com/torvalds/linux/blob/master/include/scsi/scsi.h
Line 133 ff. lists the "Host byte codes".
Yours would be first

  #define DID_TRANSPORT_DISRUPTED 0x0e /* Transport error disrupted execution
                                        * and the driver blocked the port to
                                        * recover the link. Transport class will
                                        * retry or fail IO */

and at the second occasion

  #define DID_NO_CONNECT 0x01	/* Couldn't connect before timeout period */

The other number "driver bt 0x0" would be

  #define DRIVER_OK 0x00	/* Driver status                   */

in contrast to the possible values DRIVER_BUSY to DRIVER_SENSE.


So i agree to Dan Ritter's opinion, that the problem is in plugs, cable,
tape drive electronics, or computer electronics.


Have a nice day :)

Thomas