matrix-ops/global/overlay/usr/local/bin/run-cosmos

#!/bin/bash
#
# Simplify running cosmos, with serialization if flock is available.
#

readonly PROGNAME=$(basename "$0")
readonly LOCKFILE_DIR=/tmp
readonly LOCK_FD=200
readonly FLEETLOCK_CONFIG=/etc/run-cosmos-fleetlock-conf
readonly FLEETLOCK_DISABLE_FILE=/etc/run-cosmos-fleetlock-disable
readonly FLEETLOCK_TOOL=/usr/local/bin/sunet-fleetlock
readonly HEALTHCHECK_TOOL=/usr/local/bin/sunet-machine-healthy
readonly HEALTHCHECK_DISABLE_FILE=/etc/run-cosmos-healthcheck-disable

lock() {
    local prefix=$1
    local fd=${2:-$LOCK_FD}
    local lock_file=$LOCKFILE_DIR/$prefix.lock

    # create lock file
    eval "exec $fd>$lock_file"

    # acquier the lock
    flock -n "$fd" \
        && return 0 \
        || return 1
}

eexit() {
    local error_str="$*"

    echo "$error_str"
    exit 1
}

fleetlock_lock() {
   if [ ! -f $FLEETLOCK_DISABLE_FILE ] && [ -f $FLEETLOCK_CONFIG ] && [ -x $FLEETLOCK_TOOL ]; then
      local fleetlock_group=""
      # shellcheck source=/dev/null
      . $FLEETLOCK_CONFIG || return 1
      if [ -z "$fleetlock_group" ]; then
          echo "Unable to set fleetlock_group"
	  return 1
      fi
      echo "Getting fleetlock lock"
      $FLEETLOCK_TOOL --lock-group "$fleetlock_group" --lock || return 1
   fi
   return 0
}

fleetlock_unlock() {
   if [ ! -f $FLEETLOCK_DISABLE_FILE ] && [ -f $FLEETLOCK_CONFIG ] && [ -x $FLEETLOCK_TOOL ]; then
      local fleetlock_group=""
      # shellcheck source=/dev/null
      . $FLEETLOCK_CONFIG || return 1
      if [ -z "$fleetlock_group" ]; then
          echo "Unable to set fleetlock_group"
	  return 1
      fi
      machine_is_healthy || return 1
      echo "Releasing fleetlock lock"
      $FLEETLOCK_TOOL --lock-group "$fleetlock_group" --unlock || return 1
   fi
   return 0
}

machine_is_healthy() {
   if [ ! -f $HEALTHCHECK_DISABLE_FILE ] && [ -x $HEALTHCHECK_TOOL ]; then
       echo "Running any health checks"
       $HEALTHCHECK_TOOL || return 1
   fi
   return 0
}

main () {
   lock "$PROGNAME" || eexit "Only one instance of $PROGNAME can run at one time."
   fleetlock_lock || eexit "Unable to acquire fleetlock lock."
   cosmos "$@" update
   cosmos "$@" apply
   fleetlock_unlock || eexit "Unable to release fleetlock lock."

   touch /var/run/last-cosmos-ok.stamp

   find /var/lib/puppet/reports/ -type f -mtime +10 -print0 | xargs -0 rm -f
}

main "$@"

if [ -f /cosmos-reboot ]; then
   rm -f /cosmos-reboot
   reboot
fi