Explanation of "fcd driver received notification of RISC Parity Error"
HP-UX 11i v2 - Explanation of "fcd driver received notification of RISC Parity Error" Message Reported by Fibre Channel fcd Driver
ISSUE:
A critical error for AB379A (4Gb Dual-port Fibre Channel Host Bus Adapter (HBA)) appears in syslog.log :
xxxx vmunix: 0/7/1/0: fcd driver received notification of RISC Parity Error.
xxxx vmunix: Driver and Firmware will be reset.
xxxx vmunix: 0/7/1/0: Fibre Channel Driver received a Fatal Error Notification.
xxxx vmunix: A possible cause is a Mailbox Command time out.
xxxx vmunix: Please see Preceding Error messages to isolate the cause.
xxxx vmunix: Fibre Channel Driver will now perform fatal error handling.
SOLUTION:
These messages indicated that the HBA reported fatal error and the fcd driver started recovery operation. The first event is the RISC processor on the HBA detected some fatal error on board, such as Parity error on on-board bus, or data RAM, firmware RAM. The RISC chip on HBA notified the event as external interrupt and suspended. The fcd driver received the interrupt. It decoded the status code of the event and reported the message 0/7/1/0: fcd driver received notification of RISC Parity Error . Because of the RISC processor on the HBA suspended, the yhr fcd driver could not receive any responses to commands via "Mailbox". ("Mailbox" is a set of control registers on the HBA. The fcd driver sends commands via Mailbox and waits for responses from the HBA.) Then it detected the condition as Fatal Error. Finally, the fcd driver started to recover the situation, which includes the reset of the HBA and creating firmware dump files.
So then the second event (Mailbox command time out) is considered as a subsidiary issue. The primary issue on the HBA is that RISC processor on the HBA detected some fatal error, such as Parity error on on-board bus, data RAM, firmware RAM.
Even if this HBA can be recovered by resetting, it is suggested to monitor this HBA and if similar events are reported, consider to have maintenance time to replace the HBA card.