ipmi_watchdog.sh

#!/bin/bash

# Detect and install ipmitool (to be adjusted, this is run on an alpine system)
which ipmitool >/dev/null 2>&1 || (apk update && apk add ipmitool)

# IPMI hostname or IP address
IPMI_HOST=<<ipmi_ip_address>>

# IPMI username
IPMI_USER=<<username>>

# IPMI password
IPMI_PASS=<<password>>

# Check the system event log for errors
EVENT_LOG=$(ipmitool -I lanplus -H $IPMI_HOST -U $IPMI_USER -P $IPMI_PASS sel list)

# Check if the event log contains a specific error code
if echo "$EVENT_LOG" | grep -q "0x46"; then
  # Restart the machine
  ipmitool -I lanplus -H $IPMI_HOST -U $IPMI_USER -P $IPMI_PASS chassis power reset

  # Clear the event log
  ipmitool -I lanplus -H $IPMI_HOST -U $IPMI_USER -P $IPMI_PASS sel clear
fi