Update OS procedure / Java updates

Different OS updates:

recommended procedure (also it works without stopping the services, but is recommended to stop them first).

  • Stop services with ver_stop command on the server (bash /etc/veridiumid/scripts/veridium_services.sh stop)

  • Do os update or other operation

  • Reboot servers – the services will start automatically

  • Run check_services to see that all services are up and running (green), before going to next server.

    • Normally the services will go up in 3,4 minutes after a reboot.

You need to look specifically for elasticsearch – cluster in status green, cassandra – all nodes are in UN status, zookeeper – all nodes are up in the cluster – one leader and 2 followers.

Java update:

  • Stop services with ver_stop command on the server (bash /etc/veridiumid/scripts/veridium_services.sh stop)

  • do java update

  • start services with ver_start command on the server (bash /etc/veridiumid/scripts/veridium_services.sh start)

  • Run check_services to see that all services are up and running (green), before going to next server.

    • Normally the services will go up in 3,4 minutes after a reboot.

Individual Veridium Processes used by Java

for pid in $(pgrep -f java); do
    exe=$(readlink -f /proc/$pid/exe)

    echo "===================================================="
    echo "PID      : $pid"
    echo "JAVA     : $exe"
    echo "USER     : $(ps -o user= -p "$pid")"
    echo "STARTED  : $(ps -o lstart= -p "$pid")"
    echo "CPU%     : $(ps -o %cpu= -p "$pid")"
    echo "MEM%     : $(ps -o %mem= -p "$pid")"
    echo "RSS(MB)  : $(awk '/VmRSS/ {printf "%.1f", $2/1024}' /proc/$pid/status 2>/dev/null)"
    echo "THREADS  : $(awk '/Threads/ {print $2}' /proc/$pid/status 2>/dev/null)"
    echo "UPTIME   : $(ps -o etime= -p "$pid")"

    echo "COMMAND  :"
    tr '\0' ' ' < /proc/$pid/cmdline
    echo

    echo "JAVA VERSION:"
    "$exe" -version 2>&1 | head -n 2

    echo
done

Last updated: