Process Monitoring and Integrity

6.7.12Excessive Failover/Reboots, Administrative Action

Prior to executing any failover/reboot the PMS will determine if the failover/reboot threshold has been exceeded. If it has, the PMS will be operationally disabled. When PMS is disabled, all process monitoring is halted. To re-enable the PMS, the operator must lock the global administrative state. The operator can then fix the problem and administratively unlock the global administrative state.

The following events are generated against the PMS Monitor (unique ID 1). The events for the process or processes that caused this condition to occur will also be present, but are not described in this table. They are defined in the scenarios provided above.

Table 17. Excessive Failover/Reboots, Administrative Action

Description

Event String

UID

Assert

Severity

 

 

 

 

 

PMS detects excessive failover/

Excessive reboots/failovers; all

1

Assert

Major

reboots

process monitoring disabled

 

 

 

 

 

 

 

 

Operator locks the global

None

-

N/A

N/A

administrative state

 

 

 

 

 

 

 

 

 

Operator takes actions to fix the

N/A

-

N/A

N/A

problem

 

 

 

 

 

 

 

 

 

Operator unlocks the global

 

1#a

 

 

administrative state causing

Monitoring initialized

De-assert

OK

monitoring to be resumed

 

 

 

 

 

 

 

 

 

a.The "Monitoring initialized" will be generated for the monitor (unique 1) as well as the individual processes that are admin- istratively unlocked.

6.8Process Integrity Executable (PIE)

The Process Integrity Executable (PIE) for the Chassis Management Module’s (CMM) Blade Proxy Manager (BPM) and Wrapper Processes is responsible for determining the health of the Wrapper Processes. Monitoring the integrity means not only monitoring the fact that the process is running but that it is functioning properly.

The PIE will monitor the BPM, CMM Wrapper Process (Wrapper Process number 255) and Chassis Wrapper Processes (23). It will also monitor the Wrapper Processes for intelligent (have a management controller) blades, power supplies, and fans. Wrapper Processes for non-intelligent devices will not be monitored.

PIE will monitor the BPM and Wrapper Processes. The Wrapper Processes have two categories for integrity monitoring. The first category contains the static processes. Static processes are processes that are always present while the CMM software is running. The CMM (255) and chassis (23)

Wrapper Processes are the static processes. The second category contains all the dynamic Wrapper Processes. Dynamic processes are ones that come and go as the configuration of the chassis changes (such as a blade insertion or removal). The fan, power supply, and blade Wrapper Processes belong to the dynamic category.

54MPCMM0001 Chassis Management Module Software Technical Product Specification

Page 54
Image 54
Intel MPCMM0001 manual Excessive Failover/Reboots, Administrative Action