Process Monitoring and Integrity

as valid recovery actions for cmd_hand. The default recovery action for cmd_hand process is 4 (failover and reboot) and that cannot be changed to anything else. A recovery action of 1 (No Action) is also not allowed because of the severity of the process.

In the event that cmd_hand process terminates unexpectedly, and the default recovery action kicks in, there is 2-3 minute delay before the CMM actually reboots. This is normal and expected because PMS makes multiple tries to failover, and times out because cmd_hand does not respond.

[PmsProcess053] UniqueID = 53 CommandLine = ./cmd_hand StartCommandLine = ./cmd_hand AdminState = 1 ProcessExistenceInterval = 2 ProcessRampUpTime = 10 ProcessSeverity = 3 RecoveryAction = 4 ProcessRestartEscalationAction = 2 ProcessRestartEscalationNumber = 5 ProcessRestartEscalationInterval = 300

6.9.3.8BPM

Note: PmsProc054 represents a crucial process of the CMM software stack. This process cannot be restarted properly if it terminates unexpectedly. Hence, none of the recovery actions that attempt to restart a process i.e., 2 (Restart) or 3 (Failover & Restart) are allowed as valid recovery actions for BPM. The default recovery action for BPM process is 4 (failover and reboot) which can only be changed to 1 (No Action).

[PmsProcess054] UniqueID = 54 CommandLine = ./BPM StartCommandLine = ./BPM AdminState = 1 ProcessExistenceInterval = 2 ProcessRampUpTime = 10 ProcessSeverity = 3 RecoveryAction = 4

62MPCMM0001 Chassis Management Module Software Technical Product Specification

Page 62
Image 62
Intel MPCMM0001 manual 3.8 BPM