Configuring Platform Event Filters

This chapter includes the following sections:

Platform Event Filters

A platform event filter (PEF) can trigger an action and generate an alert when a critical hardware-related event occurs. For each PEF, you can choose the action to be taken (or take no action) when a platform event occurs. You can also choose to generate and send an alert when a platform event occurs. Alerts are sent as an SNMP trap, so you must configure an SNMP trap destination before the alerts can be sent.

You can globally enable or disable the generation of platform event alerts. When disabled, alerts are not sent even if PEFs are configured to send them.

Enabling Platform Event Alerts

Procedure
      Command or Action Purpose
    Step 1 Server# scope fault  

    Enters the fault command mode.

     
    Step 2 Server /fault # set platform-event-enabled yes  

    Enables platform event alerts.

     
    Step 3 Server /fault # commit  

    Commits the transaction to the system configuration.

     
    Step 4 Server /fault # show [detail]  

    (Optional) Displays the platform event alert configuration.

     

    The following example enables platform event alerts:

    Server# scope fault
Server /fault # set platform-event-enabled yes
Server /fault *# commit
Server /fault # show
SNMP Community String Platform Event Enabled  
--------------------- ----------------------- 
public                yes                    

Server /fault #

    Disabling Platform Event Alerts

    Procedure
        Command or Action Purpose
      Step 1 Server# scope fault  

      Enters the fault command mode.

       
      Step 2 Server /fault # set platform-event-enabled no  

      Disables platform event alerts.

       
      Step 3 Server /fault # commit  

      Commits the transaction to the system configuration.

       
      Step 4 Server /fault # show [detail]  

      (Optional) Displays the platform event alert configuration.

       

      The following example disables platform event alerts:

      Server# scope fault
Server /fault # set platform-event-enabled no
Server /fault *# commit
Server /fault # show
SNMP Community String Platform Event Enabled  
--------------------- ----------------------- 
public                no                     

Server /fault #

      Configuring Platform Event Filters

      You can configure actions and alerts for the following platform event filters:

      ID Platform Event Filter

      1

      Temperature Critical Assert Filter

      2

      Temperature Warning Assert Filter

      3

      Voltage Critical Assert Filter

      4

      Current Assert Filter

      5

      Fan Critical Assert Filter

      6

      Processor Assert Filter

      7

      Power Supply Critical Assert Filter

      8

      Power Supply Warning Assert Filter

      9

      Power Supply Redundancy Lost Filter

      10

      Discrete Power Supply Assert Filter

      11

      Memory Assert Filter

      12

      Drive Slot Assert Filter
      Procedure
          Command or Action Purpose
        Step 1 Server# scope fault  

        Enters the fault command mode.

         
        Step 2 Server /fault # scope pef id  

        Enters the platform event filter command mode for the specified event.

        See the Platform Event Filter table for event ID numbers.

         
        Step 3 Server /fault/pef # set action {none | reboot | power-cycle | power-off}  

        Selects the desired system action when this event occurs. The action can be one of the following:

        • none —No system action is taken.
        • reboot —The server is rebooted.
        • power-cycle —The server is power cycled.
        • power-off —The server is powered off.
         
        Step 4 Server /fault/pef # set send-alert {yes | no}  

        Enables or disables the sending of a platform event alert for this event.

        Note   

        For an alert to be sent, the filter trap settings must be configured properly and platform event alerts must be enabled.
         
        Step 5 Server /fault/pef # commit  

        Commits the transaction to the system configuration.

         

        This example configures the platform event alert for an event:

        Server# scope fault
Server /fault # scope pef 13
Server /fault/pef # set action reboot
Server /fault/pef *# set send-alert yes
Server /fault/pef *# commit
Server /fault/pef # show
Platform Event Filter Event                       Action      Send Alert         
--------------------- --------------------------- ----------- ------------------ 
13                    Memory Assert Filter        reboot      yes              

Server /fault/pef #
        What to Do Next

        If you configure any PEFs to send an alert, complete the following tasks:

        • Enable platform event alerts
        • Configure SNMP trap settings

        Configuring SNMP Trap Settings

        Before You Begin
        • You must log in with admin privileges to perform this task.
        • SNMP must be enabled and saved before trap settings can be configured.
        Procedure
            Command or Action Purpose
          Step 1 Server# scope snmp  

          Enters the SNMP command mode.

           
          Step 2 Server /snmp # set trap-community-str string  

          Enter the name of the SNMP community to which trap information should be sent.

           
          Step 3 Server /snmp # set trap-ver {1 | 2 | 3}  

          Specify the desired SNMP version of the trap message.

          Note   

          SNMPv3 traps will be delivered only to locations where the SNMPv3 user and key values are configured correctly.
           
          Step 4 Server /snmp # set inform-type {trap | inform}  

          Specifies whether SNMP notification messages are sent as simple traps or as inform requests requiring acknowledgment by the receiver.

           
          Step 5 Server /snmp # scope trap-destination number  

          Enters the SNMP trap destination command mode for the specified destination. Four SNMP trap destinations are available. The destination number is an integer between 1 and 4.

           
          Step 6 Server /snmp/trap-destination # set enabled {yes | no}  

          Enables or disables the SNMP trap destination.

           
          Step 7 Server /snmp/trap-destination # set addr ip-address  

          Specifies the destination IP address to which SNMP trap information is sent.

           
          Step 8 Server /snmp/trap-destination # commit  

          Commits the transaction to the system configuration.

           

          This example configures general SNMP trap settings and trap destination number 1 and commits the transaction:

          Server# scope snmp
Server /snmp # set trap-community-str public
Server /snmp # set trap-ver 3
Server /snmp # set inform-type inform
Server /snmp *# scope trap-destination 1
Server /snmp/trap-destination *# set enabled yes
Server /snmp/trap-destination *# set addr 192.0.20.41
Server /snmp/trap-destination *# commit
Server /snmp/trap-destination # show
Trap Destination IP Address       Enabled  
---------------- ---------------- -------- 
1                192.0.20.41      yes      

Server /snmp/trap-destination #

          Sending a Test SNMP Trap Message

          Before You Begin

          You must log in with admin privileges to perform this task.

          Procedure
              Command or Action Purpose
            Step 1 Server# scope snmp  

            Enters the SNMP command mode.

             
            Step 2 Server /snmp # scope trap-destination number  

            Enters the SNMP trap destination command mode for the specified destination. Four SNMP trap destinations are available. The destination number is an integer between 1 and 4.

             
            Step 3 Server /snmp/trap-destination # sendSNMPtrap  

            Sends an SNMPv1 test trap to the configured SNMP trap destination.

            Note   

            The trap must be configured and enabled in order to send a test message.

             

            This example sends a test message to SNMP trap destination 1:

            Server# scope snmp
Server /snmp # scope trap-destination 1
Server /snmp/trap-destination # sendSNMPtrap
SNMP Test Trap sent to Destination:1
Server /snmp/trap-destination #

            Interpreting Platform Event Traps

            A CIMC platform event alert sent as an SNMP trap contains an enterprise object identifier (OID) in the form 1.3.6.1.4.1.3183.1.1.0.event. The first ten fields of the OID represent the following information: iso(1).org(3).dod(6).internet(1).private(4).enterprises(1).wired_for_management(3183).PET(1).version(1).version(0), indicating an IPMI platform event trap (PET) version 1.0 message. The last field is an event number, indicating the specific condition or alert being notified.

            Platform Event Trap Descriptions

            The following table provides a description of the event being notified in a platform event trap message, based on the event number in the trap OID.

            Event Number Platform Event Description
            0 Test Trap
            131330 Under Voltage
            131337 Voltage Critical
            196871 Current Warning
            262402 Fan Critical
            459776 Processor related (IOH-Thermalert/Caterr sensor) predictive failure deasserted
            459777 Processor related (IOH-Thermalert/Caterr sensor) predictive failure asserted
            460032 Power Warning
            460033 Power Warning
            524533 Power Supply Critical
            524551 Power Supply Warning
            525313 Discrete Power Supply Warning
            527105 Power Supply Redundancy Lost
            527106 Power Supply Redundancy Restored
            552704 Power Supply Inserted
            552705 PSU Failure
            552707 Power Supply AC Lost
            65799 Temperature Warning
            65801 Temperature Critical
            786433 Memory Warning
            786439 DDR3_INFO sensor LED - RED bit asserted (Probable ECC error on a DIMM)
            818945 Memory Warning
            818951 DDR3_INFO sensor LED - RED bit asserted (Probable ECC error on a DIMM)
            851968 Related to HDD sensor
            851972 Related to HDD sensor
            854016 HDD Absent
            854017 HDD Present
            880384 HDD Present, no fault indicated
            880385 HDD Fault
            880512 HDD Not Present
            880513 HDD is deasserted but not in a fault state
            884480 Drive Present
            884481 Drive Slot Warning
            884485 Drive in Critical Array
            884488 Drive Rebuild/Remap Aborted
            884489 Drive Slot Warning