Setup n2n-supernode » Historie » Revision 2
Revision 1 (Jeremias Keihsler, 12.01.2017 09:56) → Revision 2/3 (Jeremias Keihsler, 12.01.2017 09:59)
h1. Install n2n-Supernode h2. Requirements To install n2n you will need the following: * a installed and supported operating system (e.g. CentOS 6.x) * root-access * EPEL repository * a fast internet connection h2. Preliminary note most of this is taken from * [[http://www.ntop.org/products/n2n/]] * [[https://github.com/lukablurr/n2n_v2_fork/blob/master/doc/multiple-supernodes.pdf]] * [[http://listgateway.unipi.it/mailman/listinfo/n2n]] * [[http://listgateway.unipi.it/pipermail/n2n/2009-October/000300.html]] h2. Install <pre><code class="bash"> yum install n2n </code></pre> h2. manual testing h3. starting Supernode <pre><code class="bash"> supernode -l 5422 </code></pre> h3. setting up firewall in order to access the supernode from one of the edge-clients the given UDP-port must be accessible from outside. Therefore we need to open the firewall. <pre><code class="bash"> system-config-firewall-tui </code></pre> h2. running n2n-Supernode as a service if everything is working as expected we maybe want n2n-Supernode to survive a system-reboot. The following should work without modification for RHEL/CentOS linux and similar: Create a folder /etc/n2n and create configuration files (as described below) using following file name format: <pre><code class="bash"> mkdir /etc/n2n </code></pre> <pre><code class="bash"> supernode-gw1.example.com supernode-gw2.example.com supernode-*.example.com </code></pre> The interesting contents of supernode-gw1.example.com would look similar to this: <pre><code class="bash"> vim /etc/n2n/supernode-gw1.example.com </code></pre> <pre> # == supernode config values parsed by /etc/rc.d/init.d/edged == # Read 'man supernode' for details about these values... # # Example port number to listen on... N2N_PORT="8765" # Verbose or not (-v logs to std out)... # N2N_OPTS="-v" N2N_OPTIONS="" # === end supernode-gw1.example.com config file ========== </pre> Read comments in the scripts below to make more sense out of how they work. Here are copies of some scripts that work when used with configuration files as described above: <pre><code class="bash"> vim /etc/init.d/supernoded </code></pre> <pre> #!/bin/sh # $Id: supernoded.init,v 1.9 2009/10/23 12:34:56 # n2n supernode startup script. # # chkconfig: 345 96 26 # processname: supernode # pidfile: /var/run/n2n-supernode.pid # # short-description: Provides client-to-client VPN access via Internet # description: This is startup script for n2n 'edge', # n2n edge + supernode provide internet connection for vpn peers # # SEE: # for more options n2n 'man edge' and 'man supernode' # # NOTE: # Option to enter 'debug' as second var: e.g 'supernode start debug' # Using grep -w switch searches for whole word reduces ambiguity # Comment in/out the 'echo' stuff depending if thigs work as expected # Config files located/named in: /etc/n2n/supernode-your.example.com # Assumes CentOS linux init scripts located in /etc/rc.d/init.d/ # # Created: FOOFORCE RB 2009-10-24 # License GPL: free to copy, improve and use. # http://fuji.unipi.it/pipermail/n2n/2009-October/000300.html # =============================================================== ========================================================= # ==== main init script variables =============================== ========================= # Make sure you include path where 'supernode' + sh + tools are installed.. # PATH="/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin" PATH="/sbin:/bin:/usr/sbin:/usr/bin" # The name + path to the program/daemon... NAME="supernode" DAEMON="/usr/sbin/${NAME}" # Modify $NAME to filter on 'grep -w [s]supernode' - faster than 'grep -v grep'... GREPNAME=$(echo "[s]${NAME#s}") # Enable some simple debug messages when running this script... if [[ "${2}" == "debug" ]]; then echo "DEBUG: ${GREPNAME}" fi PIDFILE="/var/run/supernode.pid" LOGFILE="/var/log/n2n-${NAME}.log" STAMP=$(/bin/date +"%d/%b/%Y %H:%M:%S") # Get the host/domain name to identify local configfile.. HOST=$(/bin/hostname) FQDN=$(/bin/hostname -f) # Initialise configfile variable to empty value... CONFIGDIR="/etc/n2n" # Initialise configfile variable to empty value... CONFIGFILE="" # Source function library. . /etc/rc.d/init.d/functions # Load values from configfile using our local domain name: # Example: # Our local FQDN = 'myhost.example.com' # Our configfile = 'supernode-myhost.example.com' # # if [[ "${HOST}" == "" ]]; then if [[ "${FQDN}" == "" ]]; then echo " ERROR: Missing ${FQDN} hostname in config file /etc/rc.d/init.d/${NAME}" exit 1 else CONFIGFILE="/etc/n2n/supernode-${FQDN}" fi if [[ -f "${CONFIGFILE}" ]]; then # Get values from local configuration file... if [[ "${2}" == "debug" ]]; then echo " Found config file: ${CONFIGFILE}" fi . ${CONFIGFILE} else echo "ERROR: Missing config file ${CONFIGFILE}" exit 1 fi # See if pid file exists and get pid number... if [[ -f "${PIDFILE}" ]]; then CURPID=$(cat ${PIDFILE}) fi # Quit if we can't find application... test -f ${DAEMON} || exit 0 set -e # Set any value to enable this script... # N2N_RUN=1 # # test -n "${N2N_RUN}" || exit 0 # Optionaly change into n2n config directory... if [[ -d "${CONFIGDIR}" ]]; then cd ${CONFIGDIR} fi # ============================================================== ======================================================== # Should not need to change much below here... # ==== see how we were called + act accordingly =============== ========= if [ "${N2N_PORT}" == "" ]; then N2N_PORT="1304" echo "No port specified in ${CONFIGFILE} - using default port: ${N2N_PORT}" fi case "${1}" in start) # echo " ${NAME} ${1} process is ${0} on pid: ${$})..." echo "${STAMP} start n2n /etc/rc.d/init.d/${NAME}" >> ${LOGFILE} if [[ -n "${CURPID}" ]]; then echo " A PID:${CURPID} already exists for ${0}:" echo " To re-start with current config settings, run '$0 stop' then '$0 start' ..." else if [ -d /var/lock/subsys ] ; then touch /var/lock/subsys/${NAME}; fi # ==================================================== ============================================== # All variables read from ${CONFIGFILE}... # If variable value not empty then add command-line switch to variable... if [[ -n "${N2N_PORT}" ]]; then N2N_PORT="-l ${N2N_PORT}"; fi # next line for debug: if [[ "${2}" == "debug" ]]; then echo " DEBUG: ${DAEMON} ${N2N_PORT} ${N2N_OPTIONS}" fi # Execute the command + switches + values sources from config file... echo -n $"Starting $NAME: " ${DAEMON} ${N2N_PORT} ${N2N_OPTIONS} >> ${LOGFILE} & RETVAL=$? [ $RETVAL -eq 0 ] && echo_success [ $RETVAL -ne 0 ] && echo_failure echo # ================================================= =========================================== if [[ "${2}" == "debug" ]]; then echo " NOTICE: Creating new ${NAME} pidfile - deleting old ${CURPID}..." fi rm -f ${PIDFILE} touch ${PIDFILE} for PID in $(ps ax | grep -w "${GREPNAME}" | grep -v "(restart|stop|start)" | awk '{ print $1 }'); do if [[ "${2}" == "debug" ]]; then echo " Saving into pidfile - new pid: ${PID}" fi echo ${PID} >> ${PIDFILE} done fi ;; stop) # echo " DEBUG: Searching for at least one live process id..." PID=$(ps ax | grep -w "${GREPNAME}" | grep -v "(restart|stop|start)" | awk '{ print $1 }') if [[ "${2}" == "debug" ]]; then echo " This ${NAME} ${1} init query ${0} is running on pid: ${$}) for ${PID}..." fi # echo " ${NAME} ${1} process is ${0} on pid: ${$})..." echo "${STAMP} stop /etc/rc.d/init.d/${NAME}" >> ${LOGFILE} # echo " NOTICE: Kill name:${NAME} pid:${pid} before stopping..." # kill -9 $(cat /var/run/${NAME}_pid) && rm /var/run/${NAME}_pid # If a pid value found in file or in process id, then kill it... if [[ -n "${PID}" ]]; then echo -n $"Stopping ${NAME}: " # kill -9 $(cat /var/run/${NAME}.pid) && rm /var/run/${NAME}.pid kill ${PID} RETVAL=$? sleep 4s rm -f ${PIDFILE} if [ $RETVAL -eq 0 ] ; then echo_success echo else echo_failure echo fi else # Else assume stale pid... if [[ "${2}" == "debug" ]]; then echo " NOTICE: No current active process ${PID}..." fi if [[ -f "${PIDFILE}" ]]; then rm -f "${PIDFILE}" fi fi # Indiscriminate pid killer... # for PID in $(ps ax | grep -w ${NAME} | grep -v stop | awk '{ print $1 }'); do # echo " Killing currently active ${NAME} pid(s): ${PID}" # # kill -HUP $PID # kill ${PID} # done if [ -f /var/lock/subsys/${NAME} ]; then if [[ "${2}" == "debug" ]]; then echo " Removing stale lock..." fi rm -f /var/lock/subsys/${NAME} fi if [ -f "${PIDFILE}" ]; then echo " Removing stale pidfile..." rm -f ${PIDFILE} fi ;; restart) echo "* Re-start ${NAME}..." ${0} stop ${2} # Not a race... sleep 4s ${0} start ${2} ;; status) # status ${NAME} if [[ "${2}" == "debug" ]]; then echo " This init script ${NAME} ${1} query ${0} is running on pid: ${$})..." fi PID=$(ps ax | grep -w "${GREPNAME}" | grep -v status | awk '{ print $1 }') # If at least one active session then find all... if [[ -n "${PID}" ]]; then for PID in $(ps ax | grep -w "${GREPNAME}" | grep -v status | awk '{ print $1 }'); do if [[ -n "${PID}" ]]; then echo "${NAME} (pid ${PID}) is running..." else echo "${NAME} is stopped" fi done else echo "${NAME} is stopped" fi ;; *) echo " Usage: ${0} {start|stop|status|restart|any debug}" >&2 exit 1 ;; esac exit 0 # === end supernoded ====================================== ================================ </pre> make init-script executeable <pre><code class="bash"> chmod +x /etc/init.d/supernoded </code></pre> Using CentOS tools to activate the systemV init scripts: <pre><code class="bash"> chkconfig --add supernoded chkconfig supernoded on </code></pre> Command line syntax: <pre><code class="bash"> service supernoded [start|stop|restart|any debug] </code></pre> For example: <pre><code class="bash"> service supernoded restart </code></pre> … OR… <pre><code class="bash"> service supernoded restart debug </code></pre> provides some extra clues.