zerotier-docker/scripts/entrypoint-router.sh
Phill Kelley 4d0f21c52b
unexpected script termination - entrypoint-router.sh
Resolves issue raised in #15 where `entrypoint-router.sh` exits after
telling the pipe listener process to go away, with the result that the
temporary pipe file does not get cleaned up on a container restart.

> The temporary pipe file is not persisted so it will always get cleaned
 up when the container is terminated or recreated.

The pipe listener process exits automatically without needing any signal
from `entrypoint-router.sh` so the script lines doing that are removed.

Instead of creating the pipe file using `mktemp` with a random suffix,
the hard-coded name "/tmp/zerotier-ipc-log" will be used. The pipe file
is:

* still in `/tmp` so it is not persisted and will get cleaned up when
 the container is terminated.

* always initalised empty each time the script runs (important if the
 container restarts).

Fixes: #15

Signed-off-by: Phill Kelley <34226495+Paraphraser@users.noreply.github.com>
2022-08-03 12:45:22 +10:00

150 lines
4.9 KiB
Bash
Executable File

#!/usr/bin/env sh
set -Eeo pipefail
if [ "${1:0:1}" = '-' ]; then
set -- zerotier-one "$@"
fi
# useful paths
CONFIG_DIR="/var/lib/zerotier-one"
NETWORKS_DIR="${CONFIG_DIR}/networks.d"
# set up network auto-join if (a) the networks directory does not exist
# and (b) the ZEROTIER_ONE_NETWORK_IDS environment variable is non-null.
if [ ! -d "${NETWORKS_DIR}" -a -n "${ZEROTIER_ONE_NETWORK_IDS}" ] ; then
echo "Assuming container first run."
mkdir -p "${NETWORKS_DIR}"
for NETWORK_ID in ${ZEROTIER_ONE_NETWORK_IDS} ; do
echo "Configuring auto-join of network ID: ${NETWORK_ID}"
touch "${NETWORKS_DIR}/${NETWORK_ID}.conf"
echo "You will need to authorize this host at:"
echo " https://my.zerotier.com/network/${NETWORK_ID}"
done
fi
# make sure permissions are always as expected (self-repair)
PUID="${PUID:-"999"}"
PGID="${PGID:-"994"}"
if [ "$(id -u)" = '0' -a -d "${CONFIG_DIR}" ]; then
chown -Rc "${PUID}:${PGID}" "${CONFIG_DIR}"
fi
# is routing enabled?
if [ $(sysctl -n net.ipv4.ip_forward) -ne 1 ] ; then
# no! there is no point in setting up rules or termination handler
echo "$(date) - IPv4 forwarding not enabled - launching ZeroTier-One in non-routing mode"
# just exec the client (this script ends here)
exec "$@"
fi
echo "$(date) - launching ZeroTier-One in routing mode"
# use an appropriate default for a local physical interface
# (using eth0 maintains backwards compatibility)
PHY_IFACES="${ZEROTIER_ONE_LOCAL_PHYS:-"eth0"}"
# default to iptables (maintains backwards compatibility)
IPTABLES_CMD=iptables
# but support an override to use iptables-nft
[ "${ZEROTIER_ONE_USE_IPTABLES_NFT}" = "true" ] && IPTABLES_CMD=iptables-nft
# the default forwarding mode is inbound (backwards compatible)
GATEWAY_MODE="${ZEROTIER_ONE_GATEWAY_MODE:-"inbound"}"
# the wildcard for the local zerotier interface is
ZT_IFACE="zt+"
# function to add and remove the requisite rules
# - $1 is either "A" (add) or "D" (delete)
# - $2 is comment
update_iptables() {
case "${GATEWAY_MODE}" in
"inbound" )
echo "$2 ${IPTABLES_CMD} rules for inbound traffic (ZeroTier to local interfaces ${PHY_IFACES})"
for PHY_IFACE in ${PHY_IFACES} ; do
${IPTABLES_CMD} -t nat -${1} POSTROUTING -o ${PHY_IFACE} -j MASQUERADE
${IPTABLES_CMD} -${1} FORWARD -i ${PHY_IFACE} -o ${ZT_IFACE} -m state --state RELATED,ESTABLISHED -j ACCEPT
${IPTABLES_CMD} -${1} FORWARD -i ${ZT_IFACE} -o ${PHY_IFACE} -j ACCEPT
${IPTABLES_CMD} -${1} FORWARD -i ${PHY_IFACE} -o ${ZT_IFACE} -j DROP
done
;;
"outbound" )
echo "$2 ${IPTABLES_CMD} rules for outbound traffic (local interfaces ${PHY_IFACES} to ZeroTier)"
${IPTABLES_CMD} -t nat -${1} POSTROUTING -o ${ZT_IFACE} -j MASQUERADE
for PHY_IFACE in ${PHY_IFACES} ; do
${IPTABLES_CMD} -${1} FORWARD -i ${ZT_IFACE} -o ${PHY_IFACE} -m state --state RELATED,ESTABLISHED -j ACCEPT
${IPTABLES_CMD} -${1} FORWARD -i ${PHY_IFACE} -o ${ZT_IFACE} -j ACCEPT
${IPTABLES_CMD} -${1} FORWARD -i ${ZT_IFACE} -o ${PHY_IFACE} -j DROP
done
;;
"both" )
echo "$2 ${IPTABLES_CMD} rules for bi-directional traffic (local interfaces ${PHY_IFACES} to/from ZeroTier)"
${IPTABLES_CMD} -t nat -${1} POSTROUTING -o ${ZT_IFACE} -j MASQUERADE
for PHY_IFACE in ${PHY_IFACES} ; do
${IPTABLES_CMD} -t nat -${1} POSTROUTING -o ${PHY_IFACE} -j MASQUERADE
${IPTABLES_CMD} -${1} FORWARD -i ${ZT_IFACE} -o ${PHY_IFACE} -j ACCEPT
${IPTABLES_CMD} -${1} FORWARD -i ${PHY_IFACE} -o ${ZT_IFACE} -j ACCEPT
done
;;
* )
echo "Warning: ZEROTIER_ONE_GATEWAY_MODE=${GATEWAY_MODE} is not supported - ignored"
;;
esac
}
# add rules to set up NAT-routing
update_iptables "A" "adding"
# define where the ZeroTier daemon will write its output (if any)
TAIL_PIPE="/tmp/zerotier-ipc-log"
cat /dev/null >"${TAIL_PIPE}"
# start listening and echoing anything that appears there into this process
tail -f "${TAIL_PIPE}" &
# make a note of the process ID for tail
TAIL_PIPE_PID=${!}
# start the ZeroTier daemon in detached state
nohup "$@" </dev/null >"${TAIL_PIPE}" 2>&1 &
# make a note of the process ID
ZEROTIER_DAEMON_PID=${!}
# report
echo "$(date) - ZeroTier daemon is running as process ${ZEROTIER_DAEMON_PID}"
# function to handle cleanup
termination_handler() {
echo "$(date) - terminating ZeroTier-One"
# remove rules
update_iptables "D" "removing"
# relay the termination message to the daemon
# (the pipe listener is cleaned up automatically)
if [ -d "/proc/${ZEROTIER_DAEMON_PID}" ] ; then
kill -TERM ${ZEROTIER_DAEMON_PID}
wait ${ZEROTIER_DAEMON_PID}
fi
}
# set up termination handler (usually catches TERM)
trap termination_handler INT TERM HUP
# suspend this script while the zerotier daemon is running
wait ${ZEROTIER_DAEMON_PID}
# would not usually expect to arrive here inside a Docker container but
# it can happen if the user does a "sudo killall zerotier-one" rather
# that use Docker commands
echo "$(date) - the ZeroTier daemon has quit unexpectedly - cleaning up"
# run the termination handler
termination_handler