Table 4–4 (Cont.) Error Types
Error Type
Definition
Double-Bit
memory
errors
Hardware reports a double-bit error (DBE) when the ECC checkers detect
this condition on a read from a main memory location. This read can occur
during a DMA or CPU cycle, with two possible error causes: a memory
failure or a programming error.
If system software attempts to access a location beyond the bounds of
physical memory, hardware will report a double-bit ECC error. This is a
programming error in the OpenVMS operating system and the EHS will
initiate a system crash. This will be seen as a FATMEMERR bugcheck.
If system software attempts to access a valid physical memory location
which does not respond, a DBE will be reported by the hardware. In
this case, the cause of the problem is failed memory. The CPU with this
memory failure is removed from the configuration.
If the system is operating in a non-Duplex mode, the OpenVMS operating
system is terminated by forcing an entry into the system console. In
Duplex, the failed CPU is removed and the system continues to operate in
Degraded Duplex mode.
DBEs due to memory failures are always treated as solid. The failed CPU
will not be reconfigured until the zone with the failure is removed and the
memory is repaired.
The FRU in most cases will be a pair of SIMMs on a memory mother board
(MMB). In all cases, FRU isolation is done at the time of the end action
when system registers are recovered from the failed CPU. In the case of
an end action timeout, the CPU module will be identified as the FRU.
(continued on next page)
4–6 Error Handling and Analysis
