Kill geth when it is stuck

geth-restart.sh

# This bash script check if your geth is stuck in the same block number every X seconds 

FIRST_TIME=true
NB_SECONDS_CHECK=300 #5 minutes

while true
do
    JSON_RPC=$(curl -H "Content-Type: application/json" -X POST --data '{"jsonrpc":"2.0","method":"eth_blockNumber","params":[],"id":83}' http://0.0.0.0:8080)
    NEW_BLOCK_NB=$(echo $JSON_RPC | python -c 'import json,sys;obj=json.load(sys.stdin);print int(obj["result"], 0)')

    if [ "$FIRST_TIME" = true ] ; then
        # First call - Just memorize the block number to compare it in the next call 
        echo "Getting block number [$NEW_BLOCK_NB]"
        FIRST_TIME=false
        LAST_BLOCK_NUMBER=$NEW_BLOCK_NB
    else
        echo "Checking if geth is stuck.... [last_block: $LAST_BLOCK_NUMBER] - [new_block: $NEW_BLOCK_NB]"
        if [ "$NEW_BLOCK_NB" -eq "$LAST_BLOCK_NUMBER" ] ; then
            # let's kill geth. You need to have something (like supervisord) to handle geth restart
            echo "geth stuck with the same block... Kill it'..."
            kill -INT $(pgrep geth)
        else
            # All is fine - Let's update the last block to compare it again
            LAST_BLOCK_NUMBER=$NEW_BLOCK_NB
        fi
    fi                                                  
    sleep $NB_SECONDS_CHECK
done