RH Switchover on OS Crash

G

The High-Availability RH architecture enables the system master board to perform a switchover to the backup host in the event of a system crash.

Under the Linux* operating system the RH Software patches the Linux kernel to perform a switchover whenever a kernel panic occurs. In addition, the host board can be forced to reboot under these circumstances by simply adding the string “panic=1” in the append statement found in the lilo.conf configuration file.

Under the VxWorks* operating system this same switchover/reboot functionality is attached to the NMI interrupt handler. To force a system switchover and reboot under VxWorks, you can configure the ZT 5524 watchdog timeout to force an NMI interrupt to be generated. This forces the modified NMI handler to be activated in the event that the ZT 5524 watchdog is not strobed in a suitable amount of time. This causes a switchover and reboot of the failed host to occur. See the Intel® NetStructure™ ZT 5524 System Master Processor Board Technical Product Specification for information about configuring the watchdog/NMI interrupt. For more information on obtaining documentation, see Section H.2, “User Documentation” on page 131.

High Availability Software for the Intel® NetStructureTM ZT 4901 Technical Product Specification

129

Page 129
Image 129
Intel ZT 4901 manual RH Switchover on OS Crash