Forum Discussion

David_Howard's avatar
David_Howard
Icon for Nimbostratus rankNimbostratus
Jul 20, 2020

SNMP Monitor DCA Simple True/Green Red/false

I want to create a SNMP monitor that is basically just a true/false on the monitor check.

Following the article at

https://support.f5.com/csp/article/K14114

I created a SNMP DCA base

·        i3IcGenSwitchoverRole = 1.3.6.1.4.1.2793.2.1.5.0

·        i3IcGenSwitchoverRole_threshold =2

·        i3IcGenSwitchoverRole_coefficient = 1

 

The threshold is based on the chart below:

Type Access Description

Integer read-only This object displays the switchover role for this CIC server:

1 = NotApplicable because the Server is not in a Primary/Backup configuration.

2 = Primary

3 = Backup

4 = Failed

 

What I really want is for server to be green only if the i3IcGenSwitchoverRole is 2.  I don’t want a calculated value.  A calculated is wrong because the “Backup” is a higher number and is considered by the F5 to be the preferred. If I can do an SNMP true/false on a single value then the application can control its own primary / backup configuration.  Otherwise enabling and disabling of the servers has to be done manually.

BIG-IP
devops
monitor
snmp
  • David_Howard's avatar
    David_Howard
    Icon for Nimbostratus rankNimbostratus
    Jul 29, 2020

    I imported a script and did an external monitor check. I created an external monitor (https://devcentral.f5.com/s/articles/ltm-external-monitors-the-basics) with the variable PREFERRED_STATE=1 and a script. Here is the script:

    -----

    #!/bin/sh

    # remove IPv6/IPv4 compatibility prefix (LTM passes addresses in IPv6 format)

    # OID i3IcGenSwitchoverRole = 1.3.6.1.4.1.2793.2.1.5.0

    # Preferred States Numbers

    #1 = NotApplicable because the Server is not in a Primary/Backup configuration.

    #2 = Primary

    #3 = Backup

    #4 = Failed

     

    OID="1.3.6.1.4.1.2793.2.1.5.0"

    NODE_IP=`echo ${1} | sed 's/::ffff://'`

    NODE_PORT=${2}

     

    PIDFILE="/var/run/`basename ${0}`.${IP}_${PORT}.pid"

    # kill of the last instance of this monitor if hung and log current pid

    if [ -f $PIDFILE ]

    then

      echo "exceeded runtime needed to kill ${IP}:${PORT}" | logger -p local0.error

      kill -9 `cat $PIDFILE` > /dev/null 2>&1

    fi

     

    status=$(snmpwalk -v 2c -c Genesys $NODE_IP $OID| awk '{print $4}')

    if [ $status = $PREFERRED_STATE ]; then

     echo $status

    fi

    rm -f $PIDFILE

    exit