Fujitsu PRIMEQUEST 440, System Design Manual

Page: 138 / 280

4-32 C122-B001-02EN
CHAPTER 4 Hardware System Management
4.3.4 Hardware error monitoring
PSA monitors the partition for errors output by PCI cards and SCSI device drivers,
and periodically monitors predictive signs detected by the S.M.A.R.T. (Self-
Monitoring Analysis and Reporting Technology) function of the hard disk. When an
error is detected, error analysis is performed to identify the unit. The analysis result is
recorded as logging information and reported to the MMB and the management
software at an upper layer.
As shown in Figure 4.24 , the PSA also monitors various logs on errors the OS and
hardware detect in the CPU, memory, chip sets, and inter-chip busses.
The PSA performs an error analysis to identify a component or unit that had more
correctable errors than defined in the threshold for a given period. The PSA then
records this information as log information and reports the error information to the
MMB and upper-level system management software. The PSA notifies the MMB for
the purpose of having the MMB automatically disconnect the faulty component or
unit at the next reboot. (This is because frequent reoccurrence of an error, even if the
error is a correctable error, is assumed to be a sign of an impending failure.) When
the PSA reports the error information to the MMB, the MMB disconnects the above
faulty unit or component. This is called degradation reservation.