Files
packages/net/chrony/files/chrony.ntp-hotplug
Florian Wagner ce52e2ec21 chrony: make chrony-hotplug more reliable
At startup there's a race situation where "chronyc waitsync" (and thus
chrony-hotplug) will exit right away because it can't bind to loopback.

This change tries quite hard to make chrony-hotplug wait for loopback to
come up before running chronyc.

Fixes #28434.

Signed-off-by: Florian Wagner <florian@wagner-flo.de>
2026-03-15 20:40:39 +02:00

28 lines
778 B
Bash

#!/bin/sh
# Find interface corresponding to lo device
IFACE=$(ubus call network.interface dump | \
jsonfilter -e '@.interface[@.device="lo"].interface')
# During boot network.interface dump can be empty; try UCI instead
if [ -z "${IFACE}" ]; then
. /lib/functions.sh
handle_interface () {
local iface="$1"
local device
config_get device "${iface}" device
[ "${device}" = "lo" ] && IFACE=${iface}
}
config_load network
config_foreach handle_interface interface
fi
[ -z "${IFACE}" ] || \
ubus -t 30 wait_for network.interface.${IFACE}
# Wait for sync for up to 5 minutes and notify other services
/usr/bin/chronyc waitsync 300 1 0.0 1 || exit 0
ubus call hotplug.ntp call '{ "env": [ "ACTION=stratum" ] }'