domotz-remote-pawn/src/external_processes/stallWatcher.js

Domotz Agent

/** This file is part of Domotz Agent.
 * Copyright (C) 2016  Domotz Ltd
 *
 * Domotz Agent is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * Domotz Agent is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with Domotz Agent.  If not, see <http://www.gnu.org/licenses/>.

 * Created by Iacopo Papalini <iacopo@domotz.com> on 07/12/16.
 *
 * This module spawns a process that checks periodically the responsiveness of the Domotz Agent and kills it
 * if is not responding over HTTP over socket requests in a sensible time
 */

module.exports.init = function (os, childProcess, fs) {
    var pollPeriod = parseInt(process.env.DOMOTZ_POLL_PERIOD || 60);
    var retryPeriod = parseInt(process.env.DOMOTZ_RETRY_PERIOD || 5);
    var retryNumber = parseInt(process.env.DOMOTZ_RETRY_NUMBER || 6);
    var gracePeriod = parseInt(process.env.DOMOTZ_GRACE_PERIOD || 120) * 1000;
    var tmpDir = os.tmpdir();
    var processName = 'domotz_stall_watcher';
    var fileName = tmpDir + '/' + processName + '.sh';
    var logFileName = process.env.DOMOTZ_LOG_DIR + '/' + processName + '.log';

    var contents = '#!/bin/sh\n' +
        '. /etc/domotz.env;\n' +
        'period=' + pollPeriod + '\n' +
        'failure=0\n' +
        'while true\ndo\n' +
        '\tsleep ${period}\n' +
        '\tdomotz_curl --fail --silent  -X POST -A domotz-zabbix-proxy -d \'{"action": "listener_status_check"}\' --unix-socket ' +
        process.env.DOMOTZ_HTTP_ON_UNIX_SOCKET_FILE + ' --max-time 2 http://localhost\n' +
        '\tif [ $? -ne 0 ]; then\n' +
        '\t\tfailure=$((failure+1))\n' +
        '\t\tif [ $failure -eq ' + retryNumber + ' ]; then\n' +
        '\t\t\techo "`date` Killing not responding listener process; pid: $(cat ${LISTENER_PID_FILE})" >> $1\n' +
        '\t\t\tkill -9 $(cat ${LISTENER_PID_FILE})\n' +
        '\t\t\texit 1\n' +
        '\t\telse\n' +
        '\t\t\techo "`date` Not responding listener process; pid: $(cat ${LISTENER_PID_FILE}) - tentative $failure" >> $1\n' +
        '\t\t\tif [ $failure -eq 1 ]; then\n' +
        '\t\t\t\tkill -1 $(cat ${LISTENER_PID_FILE})\n' +
        '\t\t\tfi\n' +
        '\t\t\tperiod=' + retryPeriod + '\n' +
        '\t\tfi\n' +
        '\telse\n' +
        '\t\tperiod=' + pollPeriod + '\n' +
        '\t\tfailure=0\n' +
        '\tfi\n' +
        'done\n';
    return {
        start: function () {
            "use strict";
            console.info('STALL WATCH - Start process. Grace Period: %d. ' +
                'Poll Period: %d. Retry Period: %d. #Retry: %d', gracePeriod,
                pollPeriod, retryPeriod, retryNumber);
            setTimeout(function () {
                fs.writeFileSync(fileName, contents, {mode: '0700'});
                var watcherProcess = childProcess.spawn(fileName, [logFileName], {detached: false});
                watcherProcess.on('exit', function (code, signal) {
                    console.info("STALL WATCH - Watcher process exited with %s", code || signal);
                    process.exit(1);
                });
            }, gracePeriod);
        }
    };
};