Restart Geth from another process

Question

I want to make a script to restart my server and it's prerequisites, which one of them is geth. I don't see a command line to restart it. I want something like

geth --restart --xxxx --yyyyy

Does this exist?

In caso no, is it safe to just kill the process and run it again?

Answer 1

This is what I use in my Linux environment. I save the following in /home/user/bin/runGeth

#!/bin/sh

# Graceful exit, like pressing Control-C on a program
killall -q --signal SIGINT geth
sleep 10

# Hard kill, only to stop a process that refuses to terminate
killall -q geth

# Clear IPC as this can sometimes cause problems
rm -f /home/user/.ethereum/geth.ipc

DATE=`date +%Y%m%d_%H%M%S`
mv /home/user/ethlogs/geth.log /home/user/logarchive/geth.log_$DATE

# Message <= 32 bytes
MESSAGE="BokkyPooBah wuz here!"

# Use 6 for full details
VERBOSITY=3

geth --support-dao-fork --rpc --rpcaddr "192.168.7.123" --rpcport 8545 --extradata "$MESSAGE" --verbosity $VERBOSITY 2>> /home/user/ethlogs/geth.log &

Then chmod 700 /home/user/bin/runGeth.

And in /etc/rc.local, I add:

sudo -u user /home/user/bin/runGeth

I also restart geth periodically as I have had several instances where it locks up or dies. I add the following to crontab -e

# m h  dom mon dow   command
10 1,7,13,19 * * * sudo -u user /home/user/bin/runGeth 

Q: In case no, is it safe to just kill the process and run it again?

The graceful exit for geth is necessary. I leave some time for geth to gracefully shut down. I then try a hard kill as geth has in the past refused to gracefully exit.

I used only the hard kill before and the blockchain data got corrupted.

I also use similar scripts to send me periodic emails that report on the coinbase balance, ethminer hashrate, GPU temperature, ETH prices and other statistics. And a watchdog script to shutdown ethminer processes if the GPUs reach a specified temperature.

Answer 2

I use a couple of scripts for managing my geth instances

For starting geth (here i provide as parameter the verbosity but you can provide more if you want)

verbosity=3
while [ "$1" != "" ]; do
    case $1 in
    -v | --verbosity )  shift
                        verbosity=$1
                        ;;
    esac
    shift
done
if test -z `ps -ef | grep 'SCREEN -dmS geth' | grep -v grep | awk '{print $2}'`
then
echo "Starting Geth..."
screen -dmS geth /usr/bin/geth --networkid "189" --identity "firstMiner" --nodiscover --rpc --rpcapi "admin,db,eth,debug,miner,net,shh,txpool,personal,web3" --rpcaddr "0.0.0.0" --rpcport "8545" --rpccorsdomain "*"  --maxpeers "25" --mine --etherbase "3a4ac889777bce619ad88eef2ad7b9cace99273e" --gasprice "0" --targetgaslimit "9999999" --verbosity "$verbosity"
echo "Geth started with verbosity $verbosity"
else
echo "Geth is alreay running"
fi

For stopping geth

if test -z `ps -ef | grep 'SCREEN -dmS geth' | grep -v grep | awk '{print $2}'`
then
echo "Geth is already Stopped"
else
echo "Stopping Geth..."
ps -ef | grep 'SCREEN -dmS geth' | grep -v grep | awk '{print $2}' | xargs kill
echo "Geth stopped"
fi

For checking the status of geth

if test -z `ps -ef | grep 'SCREEN -dmS geth' | grep -v grep | awk '{print $2}'`
then
echo "Geth is stopped"
else
echo "Geth is running"
fi

For attaching to geth console (I provide a relative path to the ipc here but you can put an absolute path)

if test -z `ps -ef | grep 'SCREEN -dmS geth' | grep -v grep | awk '{print $2}'`
then
echo "Geth is not running"
else
geth attach ipc:../.ethereum/geth.ipc
fi

To restart you could put the geth stop and geth start scripts together in another script.