Table of Contents
Advertisement
Chapter 5 SCSI Interface
SCSI Message System
5
Bus Free
5
Bus Parity Errors
56
There are several situations when the device will go to BUS FREE
unexpectedly, as defined in the SCSI-2 standard (see sections 4.1.1
and 4.3):
• An internal hardware or firmware fault that makes it unsafe for
the device to continue operation without a full reset (similar to
a power-up reset).
• ATN asserted or bus parity error detected during non-tape
data transfers.
Bus parity errors (single bit errors) are very serious. They imply
the possibility of undetected double-bit errors on the bus. Most
likely these result in undetected data corruption. On properly
configured SCSI buses, parity errors are extremely rare. If any are
detected, they should be addressed quickly by improving the
configuration of the SCSI bus. A well-configured SCSI system in a
normal environment should be virtually free of bus parity errors.
Bus parity errors cause the library to either retry the operation, go
to Status phase, or go to BUS FREE and prepare sense data.
Retrying of parity errors during the data out phase is normally not
done, but can be enabled by changing the EnaParErrRetry
parameter in the Vendor Unique (VU) EEROM mode page. This
feature is disabled by default because of negative impacts on
device performance. Additionally, the data stream on writes
cannot be pipelined.
Quantum|ATL PowerStor L200 Series Library User's Guide
Advertisement
Table of Contents
