pi-power-relay-moxad 2.0.2

power cycle device from a Raspberry Pi when network connectivity lost

pi-power-relay

power cycle a device from a Raspberry Pi when network connectivity lost

Description

If host(s) on an outside network are not reachable via ping, then reset a device, possibly a cable-modem, that is known to be flakey, by power cycling it via a power relay.

This is intended to be run on a Raspberry Pi, using an 'IoT Power Relay version 2' from Digital Loggers, to use GPIO to power-cycle a device. The power relay is plugged into a UPS, while the Raspberry Pi is plugged into the 'always ON' power port and the cable-modem is plugged into the 'normally ON' power port. The program, when detecting a loss of network connectivity, will drop power on the 'normally ON' port, and bring it back after a short period.

The power relay product used by this program can be seen at: https://dlidirect.com/products/iot-power-relay

This is typically run from the crontab every couple of minutes.

If run on something other than a Raspberry Pi, it will assume it is a development or test box and will NOT attempt to do the actual shutdown, but will still do everything else.

Example usages

% pi-power-relay --hosts 208.67.222.222,8.8.8.8 

This will ping the hosts 208.67.222.222 and 8.8.8.8. If either of them responds back correctly, then the program will exit with no action.

% pi-power-relay --maint 01:20-01:40

This will prevent a reset of the device if the network goes away from 1:20am to 1:40am during a known network maintenance period.

% pi-power-relay --device-name 'cable-modem' --logfile /var/log/power-relay

This will log any resets to the log file /var/log/power-relay and will use the device name of 'cable-modem' in the log entry.

% pi-power-relay --delay-exit 180

The reason for this option is that if you do not use the --quiet option, then an informational message is printed out before the program exits. But if you are relying on output from the crontab to then be sent to you on some other machine via e-mail, it will fail if you do not have a full blown mailer set up and the DNS lookup for the destination will fail because the device (cable-modem?) has not recovered yet. So, this delay will give the device time to recover, before you spit out the message and then exit. The default is no delay.

There is a help option. For eg:

% pi-power-relay --help

usage: pi-power-relay [options]*
    [-d|--debug]               debugging output
    [-e|--delay-exit num]      delay time before output and exit (0 secs)
    [-f|--force-reset]         reset now and quit, despite state or lock
    [-h|--help]                print this help info
    [-l|--logfile string]      log filename (none by default)
    [-m|--maint HH:MM-HH:MM]*  Maintenance time to NOT reset (none by default)
    [-p|--pin num ]            GPIO pin number (25)
    [-q|--quiet]               supress informational messages
    [-r|--reset-time num]      wait time between GPIO state change (15 secs)
    [-t|--tries num]           max number of ping attempts per host (3)
    [-w|--wait-time num]       reset wait time after previous reset (600 secs)
    [-x|--ping-timeout num]    wait time for ping to time out (2 secs)
    [-D|--device-name string]  name of thing being reset for log (device)
    [-H|--hosts string(s)]     comma-delimited hosts to ping (8.8.4.4,8.8.8.8)
    [-L|--lockfile string]     lock-file (/tmp/pi-power-relay--reset-time)
    [-V|--version]             print version of this program (2.0.2) 

Python 2

The code will work with both Python2 and Python3, but the packaging and distribution of this code is strongly tied to Python3. If you are using an ancient Python2 system, see the file README.install-python2 included in the downloadable source code.