Fujitsu PRIMEQUEST 440, Руководство по проектированию систем

Страница: 198 / 280

6-20 C122-B001-02EN
CHAPTER 6 Redundant Configuration
6.6 Degradation Function
If a hardware abnormality is detected in the hardware power-on diagnosis or during
system operation, the degradation function stops the faulty hardware and has the
system inherit the operation with the remaining hardware.
Degradation reservation is the act of notifying a MMB about hardware that caused
more correctable errors than the threshold in a given period, and that is likely to fail.
The MMB automatically isolates such hardware at the next reboot.
Whether to use degradation varies depending on the location where the failure
occurred, the nature of the failure, the redundant configuration of a component where
a failure has occurred, and the mirror mode status. The following discussion includes
a case where the location of a failure can be identified and a case where it cannot be
identified.
Case where the location of a failure can be identified
If the location of a failure can be identified, degradation is performed as appropriate
for the failed location.
If an SB or an IO_Unit fails, and reserved SB set to be switched to in case of an SB
failure exists, and an IO_Unit to which redundant configuration is applied exists,
reserve degradation and stop the partition. Incorporate the reserved SB or the
IO_Unit into the partition at the next reboot.
Details on partitions are described in CHAPTER 8, "Partition."
Case where the location of a failure cannot be identified
If the location of a failure cannot be be identified, the partition is stopped.
If, after the partition is stopped, the user specifies reboot, reboot or the retry of reboot
is performed. If the system cannot be recovered after the specified times of retries has
been reached, the partition is stopped.
Remarks: It is very rare that the location of a failure cannot be identified. In most
cases, the location of a failure is identified and degradation is performed.