ODROID H4+ Setup and Troubleshooting Guide

# Watch for hardware errors in real-time
sudo dmesg -wH --level=err,warn

# Monitor specifically for MCE (Machine Check Exceptions)
sudo dmesg -wH | grep -i "mce\|hardware error\|bert\|ras\|corrected"

# Check MCE logs (if rasdaemon is installed)
sudo ras-mc-ctl --errors
sudo ras-mc-ctl --summary

# In-system memory test (no reboot required, but less thorough)
sudo apt install memtester
# Test 8GB, 5 passes (adjust '8G' to match your available free RAM)
sudo memtester 8G 5

# Check for memory-related errors logged by the kernel
sudo dmesg | grep -i "memory\|dimm\|ecc"

# EDAC (Error Detection and Correction) monitoring
sudo modprobe edac_core
sudo edac-util -v

# Install tools
sudo apt install stress-ng lm-sensors htop

# Initialize sensors
sudo sensors-detect --auto

# Terminal 1: Monitor temperatures
watch -n 1 sensors

# Terminal 2: Run Stress Test (10 minutes)
# Tests CPU and Virtual Memory
sudo stress-ng --cpu 4 --vm 2 --vm-bytes 2G --timeout 600s --metrics-brief

while true; do
  clear
  date
  echo "=== CPU Temp & Freq ==="
  sensors | grep -E "Core|Package|temp"
  grep MHz /proc/cpuinfo | head -4
  echo "=== Memory ==="
  free -h
  echo "=== Load ==="
  uptime
  echo "=== Recent Errors ==="
  sudo dmesg -T | tail -5
  sleep 2
done

# CPU thermal stress (heats up CPU quickly)
sudo stress-ng --cpu 4 --cpu-method all --timeout 10m --metrics

# Combined stress (CPU + Memory)
sudo stress-ng --cpu 4 --vm 2 --vm-bytes 4G --matrix 2 --timeout 600s --verify --metrics-brief

# Long-run stability test (1 hour) logging to file
sudo stress-ng --cpu 4 --vm 2 --timeout 3600s --verify --log-file /tmp/stress.log &

#!/bin/bash
LOG_DIR="/tmp/stability_test_$(date +%Y%m%d_%H%M%S)"
mkdir -p "$LOG_DIR"

echo "Starting stability test - logs in $LOG_DIR"

# Capture initial state
sudo dmesg > "$LOG_DIR/dmesg_before.log"
sensors > "$LOG_DIR/sensors_before.log"
free -h > "$LOG_DIR/memory_before.log"

# Start monitoring background process
(while true; do
  echo "$(date +%s),$(sensors | grep 'Package id 0' | awk '{print $4}' | tr -d '+°C')"
  sleep 1
done) > "$LOG_DIR/temps.log" &
TEMP_PID=$!

sudo dmesg -wH > "$LOG_DIR/dmesg_live.log" &
DMESG_PID=$!

# Run stress test (10 mins)
echo "Running 10-minute stress test..."
sudo stress-ng --cpu 4 --vm 2 --vm-bytes 4G --timeout 600s --metrics-brief --verify 2>&1 | tee "$LOG_DIR/stress_output.log"

# Stop monitoring
kill $TEMP_PID $DMESG_PID

# Capture final state
sudo dmesg > "$LOG_DIR/dmesg_after.log"
sensors > "$LOG_DIR/sensors_after.log"
free -h > "$LOG_DIR/memory_after.log"

echo "Test complete. Check $LOG_DIR for results"
echo "Potential errors found:"
grep -i "error\|fail\|mce\|hardware" "$LOG_DIR/dmesg_after.log"